[{"title":"A comparative analysis of deep learning and deep transfer learning approaches for identification of rice varieties","authors":["Komal Sharma","Ganesh Kumar Sethi","Rajesh Kumar Bawa"],"abstract":"Rice is an essential staple food for human nutrition. Rice varieties worldwide have been planted, imported, and exported. During production and trading, different types of rice can be mixed. Due to rice impurities, rice importers and exporters may lose trust in each other, requiring the development of a rice variety identification system. India is a significant player in the global rice market, and this extensive study delves into the importance of rice there. The study uses state-of-the-art deep learning and TL classifiers to tackle the problems of rice variety detection. An enormous dataset consisting of more than 600,000 rice photographs divided into 22 different classes is presented in the study to improve classification accuracy. With a training accuracy of 96% and a testing accuracy of 80.5%, ResNet50 stands well among other deep learning models compared by the authors. These models include CNN, Deep CNN, AlexNet2, Xception, Inception V3, DenseNet121, and ResNet50. Finding the best classifiers to identify varieties accurately is crucial, and this work highlights their possible uses in rice seed production. This paper lays the groundwork for future research on image-based rice categorization by suggesting areas for development and investigating ensemble strategies to improve performance.","doi":"10.1007\/s11042-024-19126-7","cleaned_title":"a comparative analysis of deep learning and deep transfer learning approaches for identification of rice varieties","cleaned_abstract":"rice is an essential staple food for human nutrition. rice varieties worldwide have been planted, imported, and exported. during production and trading, different types of rice can be mixed. due to rice impurities, rice importers and exporters may lose trust in each other, requiring the development of a rice variety identification system. india is a significant player in the global rice market, and this extensive study delves into the importance of rice there. the study uses state-of-the-art deep learning and tl classifiers to tackle the problems of rice variety detection. an enormous dataset consisting of more than 600,000 rice photographs divided into 22 different classes is presented in the study to improve classification accuracy. with a training accuracy of 96% and a testing accuracy of 80.5%, resnet50 stands well among other deep learning models compared by the authors. these models include cnn, deep cnn, alexnet2, xception, inception v3, densenet121, and resnet50. finding the best classifiers to identify varieties accurately is crucial, and this work highlights their possible uses in rice seed production. this paper lays the groundwork for future research on image-based rice categorization by suggesting areas for development and investigating ensemble strategies to improve performance.","key_phrases":["rice","an essential staple food","human nutrition","rice varieties","production","trading","different types","rice","rice impurities","rice importers","exporters","trust","the development","a rice variety identification system","india","a significant player","the global rice market","this extensive study","the importance","rice","the study","the-art","tl classifiers","the problems","rice variety detection","an enormous dataset","more than 600,000 rice photographs","22 different classes","the study","classification accuracy","a training accuracy","96%","a testing accuracy","80.5%","resnet50","other deep learning models","the authors","these models","cnn","deep cnn","alexnet2","xception","inception v3","densenet121","resnet50","the best classifiers","varieties","this work","their possible uses","rice seed production","this paper","the groundwork","future research","image-based rice categorization","areas","development","ensemble strategies","performance","rice","india","more than 600,000","22","96%","80.5%","resnet50","cnn","cnn","v3","resnet50"]},{"title":"Interpretable deep learning methods for multiview learning","authors":["Hengkang Wang","Han Lu","Ju Sun","Sandra E. Safo"],"abstract":"BackgroundTechnological advances have enabled the generation of unique and complementary types of data or views (e.g. genomics, proteomics, metabolomics) and opened up a new era in multiview learning research with the potential to lead to new biomedical discoveries.ResultsWe propose iDeepViewLearn (Interpretable Deep Learning Method for Multiview Learning) to learn nonlinear relationships in data from multiple views while achieving feature selection. iDeepViewLearn combines deep learning flexibility with the statistical benefits of data and knowledge-driven feature selection, giving interpretable results. Deep neural networks are used to learn view-independent low-dimensional embedding through an optimization problem that minimizes the difference between observed and reconstructed data, while imposing a regularization penalty on the reconstructed data. The normalized Laplacian of a graph is used to model bilateral relationships between variables in each view, therefore, encouraging selection of related variables. iDeepViewLearn is tested on simulated and three real-world data for classification, clustering, and reconstruction tasks. For the classification tasks, iDeepViewLearn had competitive classification results with state-of-the-art methods in various settings. For the clustering task, we detected molecular clusters that differed in their 10-year survival rates for breast cancer. For the reconstruction task, we were able to reconstruct handwritten images using a few pixels while achieving competitive classification accuracy. The results of our real data application and simulations with small to moderate sample sizes suggest that iDeepViewLearn may be a useful method for small-sample-size problems compared to other deep learning methods for multiview learning.ConclusioniDeepViewLearn is an innovative deep learning model capable of capturing nonlinear relationships between data from multiple views while achieving feature selection. It is fully open source and is freely available at https:\/\/github.com\/lasandrall\/iDeepViewLearn.","doi":"10.1186\/s12859-024-05679-9","cleaned_title":"interpretable deep learning methods for multiview learning","cleaned_abstract":"backgroundtechnological advances have enabled the generation of unique and complementary types of data or views (e.g. genomics, proteomics, metabolomics) and opened up a new era in multiview learning research with the potential to lead to new biomedical discoveries.resultswe propose ideepviewlearn (interpretable deep learning method for multiview learning) to learn nonlinear relationships in data from multiple views while achieving feature selection. ideepviewlearn combines deep learning flexibility with the statistical benefits of data and knowledge-driven feature selection, giving interpretable results. deep neural networks are used to learn view-independent low-dimensional embedding through an optimization problem that minimizes the difference between observed and reconstructed data, while imposing a regularization penalty on the reconstructed data. the normalized laplacian of a graph is used to model bilateral relationships between variables in each view, therefore, encouraging selection of related variables. ideepviewlearn is tested on simulated and three real-world data for classification, clustering, and reconstruction tasks. for the classification tasks, ideepviewlearn had competitive classification results with state-of-the-art methods in various settings. for the clustering task, we detected molecular clusters that differed in their 10-year survival rates for breast cancer. for the reconstruction task, we were able to reconstruct handwritten images using a few pixels while achieving competitive classification accuracy. the results of our real data application and simulations with small to moderate sample sizes suggest that ideepviewlearn may be a useful method for small-sample-size problems compared to other deep learning methods for multiview learning.conclusionideepviewlearn is an innovative deep learning model capable of capturing nonlinear relationships between data from multiple views while achieving feature selection. it is fully open source and is freely available at https:\/\/github.com\/lasandrall\/ideepviewlearn.","key_phrases":["backgroundtechnological advances","the generation","unique and complementary types","data","views","e.g. genomics","proteomics","metabolomics","a new era","multiview learning research","the potential","new biomedical discoveries.resultswe propose ideepviewlearn","(interpretable deep learning method","multiview learning","nonlinear relationships","data","multiple views","feature selection","ideepviewlearn","deep learning flexibility","the statistical benefits","data","knowledge-driven feature selection","interpretable results","deep neural networks","an optimization problem","that","the difference","observed and reconstructed data","a regularization penalty","the reconstructed data","the normalized laplacian","a graph","bilateral relationships","variables","each view","selection","related variables","ideepviewlearn","simulated and three real-world data","classification","clustering","reconstruction tasks","the classification tasks","ideepviewlearn","competitive classification results","the-art","various settings","the clustering task","we","molecular clusters","that","their 10-year survival rates","breast cancer","the reconstruction task","we","handwritten images","a few pixels","competitive classification accuracy","the results","our real data application","simulations","small to moderate sample sizes","that ideepviewlearn","a useful method","small-sample-size problems","other deep learning methods","multiview learning.conclusionideepviewlearn","an innovative deep learning model","nonlinear relationships","data","multiple views","feature selection","it","fully open source","https:\/\/github.com\/lasandrall\/ideepviewlearn","ideepviewlearn","three","10-year","https:\/\/github.com\/lasandrall\/ideepviewlearn"]},{"title":"Deep-learning-based image reconstruction with limited data: generating synthetic raw data using deep learning","authors":["Frank Zijlstra","Peter Thomas While"],"abstract":"ObjectDeep learning has shown great promise for fast reconstruction of accelerated MRI acquisitions by learning from large amounts of raw data. However, raw data is not always available in sufficient quantities. This study investigates synthetic data generation to complement small datasets and improve reconstruction quality.Materials and methodsAn adversarial auto-encoder was trained to generate phase and coil sensitivity maps from magnitude images, which were combined into synthetic raw data.On a fourfold accelerated MR reconstruction task, deep-learning-based reconstruction networks were trained with varying amounts of training data (20 to 160 scans). Test set performance was compared between baseline experiments and experiments that incorporated synthetic training data.ResultsTraining with synthetic raw data showed decreasing reconstruction errors with increasing amounts of training data, but importantly this was magnitude-only data, rather than real raw data. For small training sets, training with synthetic data decreased the mean absolute error (MAE) by up to 7.5%, whereas for larger training sets the MAE increased by up to 2.6%.DiscussionSynthetic raw data generation improved reconstruction quality in scenarios with limited training data. A major advantage of synthetic data generation is that it allows for the reuse of magnitude-only datasets, which are more readily available than raw datasets.","doi":"10.1007\/s10334-024-01193-4","cleaned_title":"deep-learning-based image reconstruction with limited data: generating synthetic raw data using deep learning","cleaned_abstract":"objectdeep learning has shown great promise for fast reconstruction of accelerated mri acquisitions by learning from large amounts of raw data. however, raw data is not always available in sufficient quantities. this study investigates synthetic data generation to complement small datasets and improve reconstruction quality.materials and methodsan adversarial auto-encoder was trained to generate phase and coil sensitivity maps from magnitude images, which were combined into synthetic raw data.on a fourfold accelerated mr reconstruction task, deep-learning-based reconstruction networks were trained with varying amounts of training data (20 to 160 scans). test set performance was compared between baseline experiments and experiments that incorporated synthetic training data.resultstraining with synthetic raw data showed decreasing reconstruction errors with increasing amounts of training data, but importantly this was magnitude-only data, rather than real raw data. for small training sets, training with synthetic data decreased the mean absolute error (mae) by up to 7.5%, whereas for larger training sets the mae increased by up to 2.6%.discussionsynthetic raw data generation improved reconstruction quality in scenarios with limited training data. a major advantage of synthetic data generation is that it allows for the reuse of magnitude-only datasets, which are more readily available than raw datasets.","key_phrases":["objectdeep learning","great promise","fast reconstruction","accelerated mri acquisitions","large amounts","raw data","raw data","sufficient quantities","this study","synthetic data generation","small datasets","reconstruction quality.materials","methodsan adversarial auto-encoder","phase and coil sensitivity maps","magnitude images","which","mr reconstruction task","deep-learning-based reconstruction networks","varying amounts","training data","20 to 160 scans","test set performance","baseline experiments","experiments","that","synthetic raw data","reconstruction errors","increasing amounts","training data","this","magnitude-only data","real raw data","small training sets","synthetic data","the mean absolute error","mae","up to 7.5%","larger training sets","the mae","2.6%.discussionsynthetic raw data generation","improved reconstruction quality","scenarios","limited training data","a major advantage","synthetic data generation","it","the reuse","magnitude-only datasets","which","raw datasets","20","160","up to 7.5%"]},{"title":"Topic- and learning-related predictors of deep-level learning strategies","authors":["Eve Kikas","Gintautas Silinskas","Eliis H\u00e4rma"],"abstract":"The aim of this study was to examine which topic- and learning-related knowledge and motivational beliefs predict the use of specific deep-level learning strategies during an independent learning task. Participants included 335 Estonian fourth- and sixth-grade students who were asked to read about light processes and seasonal changes. The study was completed electronically. Topic-related knowledge was assessed via an open question about seasonal changes, and learning-related knowledge was assessed via scenario-based tasks. Expectancies, interest, and utility values related to learning astronomy and using deep-level learning strategies were assessed via questions based on the Situated Expectancy-Value Theory. Deep-level learning strategies (using drawings in addition to reading and self-testing) were assessed while completing the reading task. Among topic-related variables, prior knowledge and utility value\u2014but not interest or expectancy in learning astronomy\u2014were related to using deep-level learning strategies. Among learning-related variables, interest and utility value of effective learning\u2014but not metacognitive knowledge of learning strategies or expectancy in using deep-level learning strategies\u2014were related to using deep-level learning strategies. This study confirms that it is not enough to examine students\u2019 knowledge and skills in using learning strategies with general or hypothetical questions, instead, it is of crucial importance to study students in real learning situations.","doi":"10.1007\/s10212-023-00766-6","cleaned_title":"topic- and learning-related predictors of deep-level learning strategies","cleaned_abstract":"the aim of this study was to examine which topic- and learning-related knowledge and motivational beliefs predict the use of specific deep-level learning strategies during an independent learning task. participants included 335 estonian fourth- and sixth-grade students who were asked to read about light processes and seasonal changes. the study was completed electronically. topic-related knowledge was assessed via an open question about seasonal changes, and learning-related knowledge was assessed via scenario-based tasks. expectancies, interest, and utility values related to learning astronomy and using deep-level learning strategies were assessed via questions based on the situated expectancy-value theory. deep-level learning strategies (using drawings in addition to reading and self-testing) were assessed while completing the reading task. among topic-related variables, prior knowledge and utility value\u2014but not interest or expectancy in learning astronomy\u2014were related to using deep-level learning strategies. among learning-related variables, interest and utility value of effective learning\u2014but not metacognitive knowledge of learning strategies or expectancy in using deep-level learning strategies\u2014were related to using deep-level learning strategies. this study confirms that it is not enough to examine students\u2019 knowledge and skills in using learning strategies with general or hypothetical questions, instead, it is of crucial importance to study students in real learning situations.","key_phrases":["the aim","this study","which","the use","specific deep-level learning strategies","an independent learning task","participants","335 estonian fourth- and sixth-grade students","who","light processes","seasonal changes","the study","topic-related knowledge","an open question","seasonal changes","learning-related knowledge","scenario-based tasks","expectancies","interest","utility values","astronomy","deep-level learning strategies","questions","the situated expectancy-value theory","deep-level learning strategies","drawings","addition","reading","self-testing","the reading task","topic-related variables","prior knowledge","utility value","not interest","expectancy","astronomy","deep-level learning strategies","learning-related variables","interest","utility value","effective learning","not metacognitive knowledge","strategies","expectancy","deep-level learning strategies","deep-level learning strategies","this study","it","students\u2019 knowledge","skills","learning strategies","general or hypothetical questions","it","crucial importance","students","real learning situations","335","sixth"]},{"title":"Urban traffic signal control optimization through Deep Q Learning and double Deep Q Learning: a novel approach for efficient traffic management","authors":["Qazi Umer Jamil","Karam Dad Kallu","Muhammad Jawad Khan","Muhammad Safdar","Amad Zafar","Muhammad Umair Ali"],"abstract":"Traffic congestion remains a persistent challenge in urban areas, necessitating efficient traffic control strategies. This research explores the application of advanced reinforcement learning techniques, specifically Deep Q-Learning (DQN) and Double Deep Q-Learning (DDQN), to address this issue at a four-way traffic intersection. The RL agents are trained using a reward function based on minimizing waiting times, enabling them to learn effective traffic signal control policies. The study focuses on comparing the performance of a simple non-reinforcement learning (Non RL) agent, a Deep Q-Network (DQN) agent, and an improved Double Deep Q-Learning (DDQN) agent in different traffic scenarios. The Non RL agent, which follows a fixed order of traffic phases, demonstrates limitations in both low and high traffic situations, leading to inefficiencies and imbalanced queue lengths. On the other hand, the DQN agent exhibits promising results in low traffic conditions but struggles in high traffic due to its greedy behavior. The DDQN agent, with an extended green light base time, outperforms both the Non RL agent and the original DQN agent in high traffic scenarios, making it more suitable for real-world traffic conditions. However, it shows some inefficiencies in low traffic scenarios. Future research is recommended to address multi-agent deep reinforcement learning challenges, incorporate attention mechanisms and hierarchical reinforcement learning, explore graph theory applications, and develop efficient communication protocols among agents to further enhance traffic control solutions.","doi":"10.1007\/s11042-024-20060-x","cleaned_title":"urban traffic signal control optimization through deep q learning and double deep q learning: a novel approach for efficient traffic management","cleaned_abstract":"traffic congestion remains a persistent challenge in urban areas, necessitating efficient traffic control strategies. this research explores the application of advanced reinforcement learning techniques, specifically deep q-learning (dqn) and double deep q-learning (ddqn), to address this issue at a four-way traffic intersection. the rl agents are trained using a reward function based on minimizing waiting times, enabling them to learn effective traffic signal control policies. the study focuses on comparing the performance of a simple non-reinforcement learning (non rl) agent, a deep q-network (dqn) agent, and an improved double deep q-learning (ddqn) agent in different traffic scenarios. the non rl agent, which follows a fixed order of traffic phases, demonstrates limitations in both low and high traffic situations, leading to inefficiencies and imbalanced queue lengths. on the other hand, the dqn agent exhibits promising results in low traffic conditions but struggles in high traffic due to its greedy behavior. the ddqn agent, with an extended green light base time, outperforms both the non rl agent and the original dqn agent in high traffic scenarios, making it more suitable for real-world traffic conditions. however, it shows some inefficiencies in low traffic scenarios. future research is recommended to address multi-agent deep reinforcement learning challenges, incorporate attention mechanisms and hierarchical reinforcement learning, explore graph theory applications, and develop efficient communication protocols among agents to further enhance traffic control solutions.","key_phrases":["traffic congestion","a persistent challenge","urban areas","efficient traffic control strategies","this research","the application","advanced reinforcement learning techniques","specifically deep q-learning (dqn","double deep q-learning","(ddqn","this issue","a four-way traffic intersection","the rl agents","a reward function","waiting times","them","effective traffic signal control policies","the study","the performance","a simple non-reinforcement learning","(non rl) agent","a deep q-network (dqn) agent","an improved double deep q-learning (ddqn) agent","different traffic scenarios","the non rl agent","which","a fixed order","traffic phases","limitations","both low and high traffic situations","inefficiencies","queue lengths","the other hand","the dqn agent","promising results","low traffic conditions","struggles","high traffic","its greedy behavior","the ddqn agent","an extended green light base time","both the non rl agent","the original dqn agent","high traffic scenarios","it","real-world traffic conditions","it","some inefficiencies","low traffic scenarios","future research","multi-agent deep reinforcement learning challenges","attention mechanisms","hierarchical reinforcement learning","graph theory applications","efficient communication protocols","agents","traffic control solutions","four"]},{"title":"Research trends in deep learning and machine learning for cloud computing security","authors":["Yehia Ibrahim Alzoubi","Alok Mishra","Ahmet Ercan Topcu"],"abstract":"Deep learning and machine learning show effectiveness in identifying and addressing cloud security threats. Despite the large number of articles published in this field, there remains a dearth of comprehensive reviews that synthesize the techniques, trends, and challenges of using deep learning and machine learning for cloud computing security. Accordingly, this paper aims to provide the most updated statistics on the development and research in cloud computing security utilizing deep learning and machine learning. Up to the middle of December 2023, 4051 publications were identified after we searched the Scopus database. This paper highlights key trend solutions for cloud computing security utilizing machine learning and deep learning, such as anomaly detection, security automation, and emerging technology's role. However, challenges such as data privacy, scalability, and explainability, among others, are also identified as challenges of using machine learning and deep learning for cloud security. The findings of this paper reveal that deep learning and machine learning for cloud computing security are emerging research areas. Future research directions may include addressing these challenges when utilizing machine learning and deep learning for cloud security. Additionally, exploring the development of algorithms and techniques that comply with relevant laws and regulations is essential for effective implementation in this domain.","doi":"10.1007\/s10462-024-10776-5","cleaned_title":"research trends in deep learning and machine learning for cloud computing security","cleaned_abstract":"deep learning and machine learning show effectiveness in identifying and addressing cloud security threats. despite the large number of articles published in this field, there remains a dearth of comprehensive reviews that synthesize the techniques, trends, and challenges of using deep learning and machine learning for cloud computing security. accordingly, this paper aims to provide the most updated statistics on the development and research in cloud computing security utilizing deep learning and machine learning. up to the middle of december 2023, 4051 publications were identified after we searched the scopus database. this paper highlights key trend solutions for cloud computing security utilizing machine learning and deep learning, such as anomaly detection, security automation, and emerging technology's role. however, challenges such as data privacy, scalability, and explainability, among others, are also identified as challenges of using machine learning and deep learning for cloud security. the findings of this paper reveal that deep learning and machine learning for cloud computing security are emerging research areas. future research directions may include addressing these challenges when utilizing machine learning and deep learning for cloud security. additionally, exploring the development of algorithms and techniques that comply with relevant laws and regulations is essential for effective implementation in this domain.","key_phrases":["deep learning and machine learning show effectiveness","cloud security threats","the large number","articles","this field","a dearth","comprehensive reviews","that","the techniques","trends","challenges","deep learning","machine learning","cloud computing security","this paper","the most updated statistics","the development","research","cloud computing security","deep learning","machine learning","the middle","december","4051 publications","we","the scopus database","this paper","key trend solutions","cloud computing security","machine learning","deep learning","anomaly detection","security automation","emerging technology's role","challenges","data privacy","scalability","explainability","others","challenges","machine learning","deep learning","cloud security","the findings","this paper","deep learning and machine learning","cloud computing security","research areas","future research directions","these challenges","machine learning","deep learning","cloud security","the development","algorithms","techniques","that","relevant laws","regulations","effective implementation","this domain","the middle of december 2023","4051","anomaly detection"]},{"title":"Distributed Deep Reinforcement Learning: A Survey and a Multi-player Multi-agent Learning Toolbox","authors":["Qiyue Yin","Tongtong Yu","Shengqi Shen","Jun Yang","Meijing Zhao","Wancheng Ni","Kaiqi Huang","Bin Liang","Liang Wang"],"abstract":"With the breakthrough of AlphaGo, deep reinforcement learning has become a recognized technique for solving sequential decision-making problems. Despite its reputation, data inefficiency caused by its trial and error learning mechanism makes deep reinforcement learning difficult to apply in a wide range of areas. Many methods have been developed for sample efficient deep reinforcement learning, such as environment modelling, experience transfer, and distributed modifications, among which distributed deep reinforcement learning has shown its potential in various applications, such as human-computer gaming and intelligent transportation. In this paper, we conclude the state of this exciting field, by comparing the classical distributed deep reinforcement learning methods and studying important components to achieve efficient distributed learning, covering single player single agent distributed deep reinforcement learning to the most complex multiple players multiple agents distributed deep reinforcement learning. Furthermore, we review recently released toolboxes that help to realize distributed deep reinforcement learning without many modifications of their non-distributed versions. By analysing their strengths and weaknesses, a multi-player multi-agent distributed deep reinforcement learning toolbox is developed and released, which is further validated on Wargame, a complex environment, showing the usability of the proposed toolbox for multiple players and multiple agents distributed deep reinforcement learning under complex games. Finally, we try to point out challenges and future trends, hoping that this brief review can provide a guide or a spark for researchers who are interested in distributed deep reinforcement learning.","doi":"10.1007\/s11633-023-1454-4","cleaned_title":"distributed deep reinforcement learning: a survey and a multi-player multi-agent learning toolbox","cleaned_abstract":"with the breakthrough of alphago, deep reinforcement learning has become a recognized technique for solving sequential decision-making problems. despite its reputation, data inefficiency caused by its trial and error learning mechanism makes deep reinforcement learning difficult to apply in a wide range of areas. many methods have been developed for sample efficient deep reinforcement learning, such as environment modelling, experience transfer, and distributed modifications, among which distributed deep reinforcement learning has shown its potential in various applications, such as human-computer gaming and intelligent transportation. in this paper, we conclude the state of this exciting field, by comparing the classical distributed deep reinforcement learning methods and studying important components to achieve efficient distributed learning, covering single player single agent distributed deep reinforcement learning to the most complex multiple players multiple agents distributed deep reinforcement learning. furthermore, we review recently released toolboxes that help to realize distributed deep reinforcement learning without many modifications of their non-distributed versions. by analysing their strengths and weaknesses, a multi-player multi-agent distributed deep reinforcement learning toolbox is developed and released, which is further validated on wargame, a complex environment, showing the usability of the proposed toolbox for multiple players and multiple agents distributed deep reinforcement learning under complex games. finally, we try to point out challenges and future trends, hoping that this brief review can provide a guide or a spark for researchers who are interested in distributed deep reinforcement learning.","key_phrases":["the breakthrough","alphago","deep reinforcement learning","a recognized technique","sequential decision-making problems","its reputation","data inefficiency","its trial and error learning mechanism","deep reinforcement learning","a wide range","areas","many methods","sample efficient deep reinforcement learning","environment modelling","experience transfer","modifications","which","deep reinforcement learning","its potential","various applications","human-computer gaming","intelligent transportation","this paper","we","the state","this exciting field","the classical distributed deep reinforcement learning methods","important components","efficient distributed learning","single player single agent","deep reinforcement learning","the most complex multiple players","multiple agents","deep reinforcement learning","we","toolboxes","that","distributed deep reinforcement learning","many modifications","their non-distributed versions","their strengths","weaknesses","-","agent","which","wargame","a complex environment","the usability","the proposed toolbox","multiple players","multiple agents","deep reinforcement learning","complex games","we","challenges","future trends","this brief review","a guide","a spark","researchers","who","distributed deep reinforcement learning"]},{"title":"Deep learning in rheumatological image interpretation","authors":["Berend C. Stoel","Marius Staring","Monique Reijnierse","Annette H. M. van der Helm-van Mil"],"abstract":"Artificial intelligence techniques, specifically deep learning, have already affected daily life in a wide range of areas. Likewise, initial applications have been explored in rheumatology. Deep learning might not easily surpass the accuracy of classic techniques when performing classification or regression on low-dimensional numerical data. With images as input, however, deep learning has become so successful that it has already outperformed the majority of conventional image-processing techniques developed during the past 50 years. As with any new imaging technology, rheumatologists and radiologists need to consider adapting their arsenal of diagnostic, prognostic and monitoring tools, and even their clinical role and collaborations. This adaptation requires a basic understanding of the technical background of deep learning, to efficiently utilize its benefits but also to recognize its drawbacks and pitfalls, as blindly relying on deep learning might be at odds with its capabilities. To facilitate such an understanding, it is necessary to provide an overview of deep-learning techniques for automatic image analysis in detecting, quantifying, predicting and monitoring rheumatic diseases, and of currently published deep-learning applications in radiological imaging for rheumatology, with critical assessment of possible limitations, errors and confounders, and conceivable consequences for rheumatologists and radiologists in clinical practice.","doi":"10.1038\/s41584-023-01074-5","cleaned_title":"deep learning in rheumatological image interpretation","cleaned_abstract":"artificial intelligence techniques, specifically deep learning, have already affected daily life in a wide range of areas. likewise, initial applications have been explored in rheumatology. deep learning might not easily surpass the accuracy of classic techniques when performing classification or regression on low-dimensional numerical data. with images as input, however, deep learning has become so successful that it has already outperformed the majority of conventional image-processing techniques developed during the past 50 years. as with any new imaging technology, rheumatologists and radiologists need to consider adapting their arsenal of diagnostic, prognostic and monitoring tools, and even their clinical role and collaborations. this adaptation requires a basic understanding of the technical background of deep learning, to efficiently utilize its benefits but also to recognize its drawbacks and pitfalls, as blindly relying on deep learning might be at odds with its capabilities. to facilitate such an understanding, it is necessary to provide an overview of deep-learning techniques for automatic image analysis in detecting, quantifying, predicting and monitoring rheumatic diseases, and of currently published deep-learning applications in radiological imaging for rheumatology, with critical assessment of possible limitations, errors and confounders, and conceivable consequences for rheumatologists and radiologists in clinical practice.","key_phrases":["artificial intelligence techniques","specifically deep learning","daily life","a wide range","areas","initial applications","rheumatology","deep learning","the accuracy","classic techniques","classification","regression","low-dimensional numerical data","images","input","deep learning","it","the majority","conventional image-processing techniques","the past 50 years","any new imaging technology","rheumatologists","radiologists","their arsenal","diagnostic, prognostic and monitoring tools","even their clinical role","collaborations","this adaptation","a basic understanding","the technical background","deep learning","its benefits","its drawbacks","pitfalls","deep learning","odds","its capabilities","such an understanding","it","an overview","deep-learning techniques","automatic image analysis","rheumatic diseases","currently published deep-learning applications","radiological imaging","rheumatology","critical assessment","possible limitations","errors","confounders","conceivable consequences","rheumatologists","radiologists","clinical practice","the past 50 years"]},{"title":"Loss of plasticity in deep continual learning","authors":["Shibhansh Dohare","J. Fernando Hernandez-Garcia","Qingfeng Lan","Parash Rahman","A. Rupam Mahmood","Richard S. Sutton"],"abstract":"Artificial neural networks, deep-learning methods and the backpropagation algorithm1 form the foundation of modern machine learning and artificial intelligence. These methods are almost always used in two phases, one in which the weights of the network are updated and one in which the weights are held constant while the network is used or evaluated. This contrasts with natural learning and many applications, which require continual learning. It has been unclear whether or not deep learning methods work in continual learning settings. Here we show that they do not\u2014that standard deep-learning methods gradually lose plasticity in continual-learning settings until they learn no better than a shallow network. We show such loss of plasticity using the classic ImageNet dataset and reinforcement-learning problems across a wide range of variations in the network and the learning algorithm. Plasticity is maintained indefinitely only by algorithms that continually inject diversity into the network, such as our continual\u00a0backpropagation algorithm, a variation of backpropagation in which a small fraction of less-used units are continually and randomly reinitialized. Our results indicate that methods based on gradient descent are not enough\u2014that sustained deep learning requires a random, non-gradient component to maintain variability and plasticity.","doi":"10.1038\/s41586-024-07711-7","cleaned_title":"loss of plasticity in deep continual learning","cleaned_abstract":"artificial neural networks, deep-learning methods and the backpropagation algorithm1 form the foundation of modern machine learning and artificial intelligence. these methods are almost always used in two phases, one in which the weights of the network are updated and one in which the weights are held constant while the network is used or evaluated. this contrasts with natural learning and many applications, which require continual learning. it has been unclear whether or not deep learning methods work in continual learning settings. here we show that they do not\u2014that standard deep-learning methods gradually lose plasticity in continual-learning settings until they learn no better than a shallow network. we show such loss of plasticity using the classic imagenet dataset and reinforcement-learning problems across a wide range of variations in the network and the learning algorithm. plasticity is maintained indefinitely only by algorithms that continually inject diversity into the network, such as our continual backpropagation algorithm, a variation of backpropagation in which a small fraction of less-used units are continually and randomly reinitialized. our results indicate that methods based on gradient descent are not enough\u2014that sustained deep learning requires a random, non-gradient component to maintain variability and plasticity.","key_phrases":["artificial neural networks","deep-learning methods","the backpropagation algorithm1 form","the foundation","modern machine learning","artificial intelligence","these methods","two phases","which","the weights","the network","which","the weights","the network","this","natural learning","many applications","which","continual learning","it","not deep learning methods","continual learning settings","we","they","that standard deep-learning methods","plasticity","continual-learning settings","they","a shallow network","we","such loss","plasticity","the classic imagenet dataset and reinforcement-learning problems","a wide range","variations","the network","the learning algorithm","plasticity","algorithms","that","diversity","the network","our continual backpropagation algorithm","a variation","backpropagation","which","a small fraction","less-used units","our results","methods","gradient descent","that","deep learning","a random, non-gradient component","variability","plasticity","algorithm1","two","one"]},{"title":"Comparative approach on crop detection using machine learning and deep learning techniques","authors":["V. Nithya","M. S. Josephine","V. Jeyabalaraja"],"abstract":"Agriculture is an expanding area of study. Crop prediction in agriculture is highly dependent on soil and environmental factors, such as rainfall, humidity, and temperature. Previously, farmers had the authority to select the crop to be farmed, oversee its development, and ascertain the optimal harvest time. The farming community is facing challenges in sustaining its practices due to the swift alterations in climatic conditions. Therefore, machine learning algorithms have replaced traditional methods in predicting agricultural productivity in recent years. To guarantee optimal precision through a specific machine learning approach. Authors extend their approach not limited to Machine Learning but also with Deep Learning Techniques. We use machine and deep learning algorithms to predict crop outcomes accurately. In this proposed model, we utilise machine learning algorithms such as Naive Bayes, decision tree, and KNN. It is worth noting that the decision tree algorithm demonstrates superior performance compared to the other algorithms, achieving an accuracy rate of 83%. In order to enhance the precision, we have suggested implementing a deep learning technique, specifically a convolutional neural network, to identify the crops. Achieving an accuracy of 93.54% was made possible by implementing this advanced deep-learning model.","doi":"10.1007\/s13198-024-02483-9","cleaned_title":"comparative approach on crop detection using machine learning and deep learning techniques","cleaned_abstract":"agriculture is an expanding area of study. crop prediction in agriculture is highly dependent on soil and environmental factors, such as rainfall, humidity, and temperature. previously, farmers had the authority to select the crop to be farmed, oversee its development, and ascertain the optimal harvest time. the farming community is facing challenges in sustaining its practices due to the swift alterations in climatic conditions. therefore, machine learning algorithms have replaced traditional methods in predicting agricultural productivity in recent years. to guarantee optimal precision through a specific machine learning approach. authors extend their approach not limited to machine learning but also with deep learning techniques. we use machine and deep learning algorithms to predict crop outcomes accurately. in this proposed model, we utilise machine learning algorithms such as naive bayes, decision tree, and knn. it is worth noting that the decision tree algorithm demonstrates superior performance compared to the other algorithms, achieving an accuracy rate of 83%. in order to enhance the precision, we have suggested implementing a deep learning technique, specifically a convolutional neural network, to identify the crops. achieving an accuracy of 93.54% was made possible by implementing this advanced deep-learning model.","key_phrases":["agriculture","an expanding area","study","crop prediction","agriculture","soil and environmental factors","rainfall","humidity","temperature","farmers","the authority","the crop","its development","the optimal harvest time","the farming community","challenges","its practices","the swift alterations","climatic conditions","machine learning algorithms","traditional methods","agricultural productivity","recent years","optimal precision","a specific machine learning approach","authors","their approach","machine learning","deep learning techniques","we","machine","deep learning","algorithms","crop outcomes","this proposed model","we","machine learning algorithms","naive bayes","decision tree","knn","it","the decision tree","algorithm","superior performance","the other algorithms","an accuracy rate","83%","order","the precision","we","a deep learning technique","specifically a convolutional neural network","the crops","an accuracy","93.54%","this advanced deep-learning model","recent years","83%","93.54%"]},{"title":"A systematic review on machine learning and deep learning techniques in the effective diagnosis of Alzheimer\u2019s disease","authors":["Akhilesh Deep Arya","Sourabh Singh Verma","Prasun Chakarabarti","Tulika Chakrabarti","Ahmed A. Elngar","Ali-Mohammad Kamali","Mohammad Nami"],"abstract":"Alzheimer\u2019s disease (AD) is a brain-related disease in which the condition of the patient gets worse with time. AD is not a curable disease by any medication. It is impossible to halt the death of brain cells, but with the help of medication, the effects of AD can be delayed. As not all MCI patients will suffer from AD, it is required to accurately diagnose whether a mild cognitive impaired (MCI) patient will convert to AD (namely MCI converter MCI-C) or not (namely MCI non-converter MCI-NC), during early diagnosis. There are two modalities, positron emission tomography (PET) and magnetic resonance image (MRI), used by a physician for the diagnosis of Alzheimer\u2019s disease. Machine learning and deep learning perform exceptionally well in the field of computer vision where there is a requirement to extract information from high-dimensional data. Researchers use deep learning models in the field of medicine for diagnosis, prognosis, and even to predict the future health of the patient under medication. This study is a systematic review of publications using machine learning and deep learning methods for early classification of normal cognitive (NC) and Alzheimer\u2019s disease (AD).This study is an effort to provide the details of the two most commonly used modalities PET and MRI for the identification of AD, and to evaluate the performance of both modalities while working with different classifiers.","doi":"10.1186\/s40708-023-00195-7","cleaned_title":"a systematic review on machine learning and deep learning techniques in the effective diagnosis of alzheimer\u2019s disease","cleaned_abstract":"alzheimer\u2019s disease (ad) is a brain-related disease in which the condition of the patient gets worse with time. ad is not a curable disease by any medication. it is impossible to halt the death of brain cells, but with the help of medication, the effects of ad can be delayed. as not all mci patients will suffer from ad, it is required to accurately diagnose whether a mild cognitive impaired (mci) patient will convert to ad (namely mci converter mci-c) or not (namely mci non-converter mci-nc), during early diagnosis. there are two modalities, positron emission tomography (pet) and magnetic resonance image (mri), used by a physician for the diagnosis of alzheimer\u2019s disease. machine learning and deep learning perform exceptionally well in the field of computer vision where there is a requirement to extract information from high-dimensional data. researchers use deep learning models in the field of medicine for diagnosis, prognosis, and even to predict the future health of the patient under medication. this study is a systematic review of publications using machine learning and deep learning methods for early classification of normal cognitive (nc) and alzheimer\u2019s disease (ad).this study is an effort to provide the details of the two most commonly used modalities pet and mri for the identification of ad, and to evaluate the performance of both modalities while working with different classifiers.","key_phrases":["alzheimer\u2019s disease","ad","a brain-related disease","which","the condition","the patient","time","ad","a curable disease","any medication","it","the death","brain cells","the help","medication","the effects","ad","not all mci patients","ad","it","a mild cognitive impaired (mci) patient","ad","namely mci converter mci-c","namely mci non-converter mci-nc","early diagnosis","two modalities","positron emission tomography","pet","magnetic resonance image","mri","a physician","the diagnosis","alzheimer\u2019s disease","machine learning","deep learning","the field","computer vision","a requirement","information","high-dimensional data","researchers","deep learning models","the field","medicine","diagnosis","prognosis","the future health","the patient","medication","this study","a systematic review","publications","machine learning","deep learning methods","early classification","(nc","ad).this study","an effort","the details","the two most commonly used modalities","the identification","ad","the performance","both modalities","different classifiers","mci","mci","mci","mci","mci","mci-nc","two","two"]},{"title":"Exploring the connection between deep learning and learning assessments: a cross-disciplinary engineering education perspective","authors":["Sabrina Fawzia","Azharul Karim"],"abstract":"It is widely accepted that student learning is significantly affected by assessment methods, but a concrete relationship has not been established in the context of multidisciplinary engineering education. Students make a physiological investment and internalize learning (deep learning) if they see high value in their learning. They persist despite challenges and take delight in accomplishing their work. As student deep learning is affected by the assessment system, it is important to explore the relationship between assessment systems and factors affecting deep learning. This study identifies the factors associated with deep learning and examines the relationships between different assessment systems those factors. A conceptual model is proposed, and a structured questionnaire was designed and directed to 600 Queensland University of Technology (QUT) multidisciplinary engineering students, with 243 responses received. The gathered data were analyzed using both SPSS and SEM. Exploratory factor analysis revealed that deep learning is strongly associated with learning environment and course design and content. Strong influence of both summative and formative assessment on learning was established in this study. Engineering educators can facilitate deep learning by adopting both assessment types simultaneously to make the learning process more effective. The proposed theoretical model related to the deep learning concept can support the key practices and modern learning methodologies currently adopted to enhance the learning and teaching process.","doi":"10.1057\/s41599-023-02542-9","cleaned_title":"exploring the connection between deep learning and learning assessments: a cross-disciplinary engineering education perspective","cleaned_abstract":"it is widely accepted that student learning is significantly affected by assessment methods, but a concrete relationship has not been established in the context of multidisciplinary engineering education. students make a physiological investment and internalize learning (deep learning) if they see high value in their learning. they persist despite challenges and take delight in accomplishing their work. as student deep learning is affected by the assessment system, it is important to explore the relationship between assessment systems and factors affecting deep learning. this study identifies the factors associated with deep learning and examines the relationships between different assessment systems those factors. a conceptual model is proposed, and a structured questionnaire was designed and directed to 600 queensland university of technology (qut) multidisciplinary engineering students, with 243 responses received. the gathered data were analyzed using both spss and sem. exploratory factor analysis revealed that deep learning is strongly associated with learning environment and course design and content. strong influence of both summative and formative assessment on learning was established in this study. engineering educators can facilitate deep learning by adopting both assessment types simultaneously to make the learning process more effective. the proposed theoretical model related to the deep learning concept can support the key practices and modern learning methodologies currently adopted to enhance the learning and teaching process.","key_phrases":["it","that student learning","assessment methods","a concrete relationship","the context","multidisciplinary engineering education","students","a physiological investment","they","high value","their learning","they","challenges","delight","their work","student deep learning","the assessment system","it","the relationship","assessment systems","factors","deep learning","this study","the factors","deep learning","the relationships","those factors","a conceptual model","a structured questionnaire","technology","243 responses","the gathered data","sem","exploratory factor analysis","deep learning","environment","course design","content","strong influence","both summative and formative assessment","learning","this study","engineering educators","deep learning","both assessment types","the learning process","the proposed theoretical model","the deep learning concept","the key practices","modern learning methodologies","the learning and teaching process","600 queensland university of technology","243","spss"]},{"title":"Integrating QSAR modelling and deep learning in drug discovery: the emergence of deep QSAR","authors":["Alexander Tropsha","Olexandr Isayev","Alexandre Varnek","Gisbert Schneider","Artem Cherkasov"],"abstract":"Quantitative structure\u2013activity relationship (QSAR) modelling, an approach that was introduced 60 years ago, is widely used in computer-aided drug design. In recent years, progress in artificial intelligence techniques, such as deep learning, the rapid growth of databases of molecules for virtual screening and dramatic improvements in computational power have supported the emergence of a new field of QSAR applications that we term \u2018deep QSAR\u2019. Marking a decade from the pioneering applications of deep QSAR to tasks involved in small-molecule drug discovery, we herein describe key advances in the field, including deep generative and reinforcement learning approaches in molecular design, deep learning models for synthetic planning and the application of deep QSAR models in structure-based virtual screening. We also reflect on the emergence of quantum computing, which promises to further accelerate deep QSAR applications and the need for open-source and democratized resources to support computer-aided drug design.","doi":"10.1038\/s41573-023-00832-0","cleaned_title":"integrating qsar modelling and deep learning in drug discovery: the emergence of deep qsar","cleaned_abstract":"quantitative structure\u2013activity relationship (qsar) modelling, an approach that was introduced 60 years ago, is widely used in computer-aided drug design. in recent years, progress in artificial intelligence techniques, such as deep learning, the rapid growth of databases of molecules for virtual screening and dramatic improvements in computational power have supported the emergence of a new field of qsar applications that we term \u2018deep qsar\u2019. marking a decade from the pioneering applications of deep qsar to tasks involved in small-molecule drug discovery, we herein describe key advances in the field, including deep generative and reinforcement learning approaches in molecular design, deep learning models for synthetic planning and the application of deep qsar models in structure-based virtual screening. we also reflect on the emergence of quantum computing, which promises to further accelerate deep qsar applications and the need for open-source and democratized resources to support computer-aided drug design.","key_phrases":["quantitative structure","activity relationship","qsar","modelling","an approach","that","computer-aided drug design","recent years","progress","artificial intelligence techniques","deep learning","the rapid growth","databases","molecules","virtual screening","dramatic improvements","computational power","the emergence","a new field","qsar applications","we","a decade","the pioneering applications","deep qsar","tasks","small-molecule drug discovery","we","key advances","the field","deep generative and reinforcement learning approaches","molecular design","deep learning models","synthetic planning","the application","deep qsar models","structure-based virtual screening","we","the emergence","quantum computing","which","deep qsar applications","the need","open-source","democratized resources","computer-aided drug design","60 years ago","recent years","quantum"]},{"title":"Deep learning for water quality","authors":["Wei Zhi","Alison P. Appling","Heather E. Golden","Joel Podgorski","Li Li"],"abstract":"Understanding and predicting the quality of inland waters are challenging, particularly in the context of intensifying climate extremes expected in the future. These challenges arise partly due to complex processes that regulate water quality, and arduous and expensive data collection that exacerbate the issue of data scarcity. Traditional process-based and statistical models often fall short in predicting water quality. In this Review, we posit that deep learning represents an underutilized yet promising approach that can unravel intricate structures and relationships in high-dimensional data. We demonstrate that deep learning methods can help address data scarcity by filling temporal and spatial gaps and aid in formulating and testing hypotheses via identifying influential drivers of water quality. This Review highlights the strengths and limitations of deep learning methods relative to traditional approaches, and underscores its potential as an emerging and indispensable approach in overcoming challenges and discovering new knowledge in water-quality sciences.","doi":"10.1038\/s44221-024-00202-z","cleaned_title":"deep learning for water quality","cleaned_abstract":"understanding and predicting the quality of inland waters are challenging, particularly in the context of intensifying climate extremes expected in the future. these challenges arise partly due to complex processes that regulate water quality, and arduous and expensive data collection that exacerbate the issue of data scarcity. traditional process-based and statistical models often fall short in predicting water quality. in this review, we posit that deep learning represents an underutilized yet promising approach that can unravel intricate structures and relationships in high-dimensional data. we demonstrate that deep learning methods can help address data scarcity by filling temporal and spatial gaps and aid in formulating and testing hypotheses via identifying influential drivers of water quality. this review highlights the strengths and limitations of deep learning methods relative to traditional approaches, and underscores its potential as an emerging and indispensable approach in overcoming challenges and discovering new knowledge in water-quality sciences.","key_phrases":["understanding","the quality","inland waters","the context","climate extremes","the future","these challenges","complex processes","that","water quality","arduous and expensive data collection","that","the issue","data scarcity","traditional process-based and statistical models","water quality","this review","we","deep learning","an underutilized yet promising approach","that","intricate structures","relationships","high-dimensional data","we","deep learning methods","data scarcity","temporal and spatial gaps","aid","formulating and testing hypotheses","influential drivers","water quality","this review","the strengths","limitations","deep learning methods","traditional approaches","its potential","an emerging and indispensable approach","challenges","new knowledge","water-quality sciences"]},{"title":"Comparative Analysis of Machine Learning, Ensemble Learning and Deep Learning Classifiers for Parkinson\u2019s Disease Detection","authors":["Palak Goyal","Rinkle Rani"],"abstract":"A progressive neurodegenerative ailment called Parkinson's disease (PD) is marked by the death of dopamine-producing cells in the substantia nigra area of the brain. The exact etiology of PD remains elusive, but it is believed to involve the presence of Lewy bodies, abnormal protein aggregates, in affected brain regions, leading to the mobile symptoms of PD. Hence, as the management of PD continues to evolve, there is a growing demand for the establishment of a descriptive system that enables the early detection of PD. In this study, we conducted an extensive analysis using machine learning, ensemble learning, and deep learning models with different hyperparameters to develop accurate classification models for PD prediction. To enhance classifier performance and address overfitting, we employed principal component analysis (PCA) for feature selection along with various preprocessing techniques. The dataset used consisted of voice samples, comprising 188 PD patients and 64 normal individuals. Our results demonstrated that the Random Forest (RF) model with accuracy of 82.37% outperformed the other base classifiers Among the ensemble classifiers, the LGBM model exhibited the highest accuracy of 85.90% when compared to both base and ensemble classifiers. Notably, the deep learning model has 91.33% training accuracy and 85.02% testing accuracy, suggesting that deep learning models perform comparably equivalent on small datasets compared to machine learning classifiers. Overall, our findings underscore the effectiveness of machine learning, ensemble techniques and deep learning models in accurately predicting PD.","doi":"10.1007\/s42979-023-02368-x","cleaned_title":"comparative analysis of machine learning, ensemble learning and deep learning classifiers for parkinson\u2019s disease detection","cleaned_abstract":"a progressive neurodegenerative ailment called parkinson's disease (pd) is marked by the death of dopamine-producing cells in the substantia nigra area of the brain. the exact etiology of pd remains elusive, but it is believed to involve the presence of lewy bodies, abnormal protein aggregates, in affected brain regions, leading to the mobile symptoms of pd. hence, as the management of pd continues to evolve, there is a growing demand for the establishment of a descriptive system that enables the early detection of pd. in this study, we conducted an extensive analysis using machine learning, ensemble learning, and deep learning models with different hyperparameters to develop accurate classification models for pd prediction. to enhance classifier performance and address overfitting, we employed principal component analysis (pca) for feature selection along with various preprocessing techniques. the dataset used consisted of voice samples, comprising 188 pd patients and 64 normal individuals. our results demonstrated that the random forest (rf) model with accuracy of 82.37% outperformed the other base classifiers among the ensemble classifiers, the lgbm model exhibited the highest accuracy of 85.90% when compared to both base and ensemble classifiers. notably, the deep learning model has 91.33% training accuracy and 85.02% testing accuracy, suggesting that deep learning models perform comparably equivalent on small datasets compared to machine learning classifiers. overall, our findings underscore the effectiveness of machine learning, ensemble techniques and deep learning models in accurately predicting pd.","key_phrases":["a progressive neurodegenerative ailment","parkinson's disease","pd","the death","dopamine-producing cells","the substantia nigra area","the brain","the exact etiology","pd","it","the presence","lewy bodies","abnormal protein aggregates","affected brain regions","the mobile symptoms","pd","the management","pd","a growing demand","the establishment","a descriptive system","that","the early detection","pd","this study","we","an extensive analysis","machine learning","ensemble learning","deep learning models","different hyperparameters","accurate classification models","pd prediction","classifier performance","address overfitting","we","principal component analysis","pca","feature selection","various preprocessing techniques","the dataset","voice samples","188 pd patients","64 normal individuals","our results","the random forest","(rf) model","accuracy","82.37%","the other base classifiers","the ensemble classifiers","the lgbm model","the highest accuracy","85.90%","both base and ensemble classifiers","the deep learning model","91.33% training accuracy","85.02% testing accuracy","deep learning models","small datasets","machine learning classifiers","our findings","the effectiveness","machine learning","ensemble techniques","deep learning models","pd","188","64","82.37%","85.90%","91.33%","85.02%"]},{"title":"Employing deep learning and transfer learning for accurate brain tumor detection","authors":["Sandeep Kumar Mathivanan","Sridevi Sonaimuthu","Sankar Murugesan","Hariharan Rajadurai","Basu Dev Shivahare","Mohd Asif Shah"],"abstract":"Artificial intelligence-powered deep learning methods are being used to diagnose brain tumors with high accuracy, owing to their ability to process large amounts of data. Magnetic resonance imaging stands as the gold standard for brain tumor diagnosis using machine vision, surpassing computed tomography, ultrasound, and X-ray imaging in its effectiveness. Despite this, brain tumor diagnosis remains a challenging endeavour due to the intricate structure of the brain. This study delves into the potential of deep transfer learning architectures to elevate the accuracy of brain tumor diagnosis. Transfer learning is a machine learning technique that allows us to repurpose pre-trained models on new tasks. This can be particularly useful for medical imaging tasks, where labelled data is often scarce. Four distinct transfer learning architectures were assessed in this study: ResNet152, VGG19, DenseNet169, and MobileNetv3. The models were trained and validated on a dataset from benchmark database: Kaggle. Five-fold cross validation was adopted for training and testing. To enhance the balance of the dataset and improve the performance of the models, image enhancement techniques were applied to the data for the four categories: pituitary, normal, meningioma, and glioma. MobileNetv3 achieved the highest accuracy of 99.75%, significantly outperforming other existing methods. This demonstrates the potential of deep transfer learning architectures to revolutionize the field of brain tumor diagnosis.","doi":"10.1038\/s41598-024-57970-7","cleaned_title":"employing deep learning and transfer learning for accurate brain tumor detection","cleaned_abstract":"artificial intelligence-powered deep learning methods are being used to diagnose brain tumors with high accuracy, owing to their ability to process large amounts of data. magnetic resonance imaging stands as the gold standard for brain tumor diagnosis using machine vision, surpassing computed tomography, ultrasound, and x-ray imaging in its effectiveness. despite this, brain tumor diagnosis remains a challenging endeavour due to the intricate structure of the brain. this study delves into the potential of deep transfer learning architectures to elevate the accuracy of brain tumor diagnosis. transfer learning is a machine learning technique that allows us to repurpose pre-trained models on new tasks. this can be particularly useful for medical imaging tasks, where labelled data is often scarce. four distinct transfer learning architectures were assessed in this study: resnet152, vgg19, densenet169, and mobilenetv3. the models were trained and validated on a dataset from benchmark database: kaggle. five-fold cross validation was adopted for training and testing. to enhance the balance of the dataset and improve the performance of the models, image enhancement techniques were applied to the data for the four categories: pituitary, normal, meningioma, and glioma. mobilenetv3 achieved the highest accuracy of 99.75%, significantly outperforming other existing methods. this demonstrates the potential of deep transfer learning architectures to revolutionize the field of brain tumor diagnosis.","key_phrases":["artificial intelligence-powered deep learning methods","brain tumors","high accuracy","their ability","large amounts","data","magnetic resonance imaging","the gold standard","brain tumor diagnosis","machine vision","computed tomography","ultrasound","x","-ray imaging","its effectiveness","this","brain tumor diagnosis","a challenging endeavour","the intricate structure","the brain","this study","the potential","architectures","the accuracy","brain tumor diagnosis","transfer learning","a machine learning technique","that","us","pre-trained models","new tasks","this","medical imaging tasks","labelled data","four distinct transfer learning architectures","this study","resnet152","vgg19","densenet169","mobilenetv3","the models","a dataset","benchmark database","kaggle","five-fold cross validation","training","testing","the balance","the dataset","the performance","the models","image enhancement techniques","the data","the four categories","glioma","mobilenetv3","the highest accuracy","99.75%","other existing methods","this","the potential","architectures","the field","brain tumor diagnosis","four","mobilenetv3","five-fold","four","mobilenetv3","99.75%"]},{"title":"Detecting Suicidality in Arabic Tweets Using Machine Learning and Deep Learning Techniques","authors":["Asma Abdulsalam","Areej Alhothali","Saleh Al-Ghamdi"],"abstract":"Social media platforms have revolutionized traditional communication techniques by allowing people to connect instantaneously, openly, and frequently. As people use social media to share personal stories and express their opinions, negative emotions such as thoughts of death, self-harm, and hardship are commonly expressed, particularly among younger generations. Accordingly, the use of social media to detect suicidality may help provide proper intervention that will ultimately deter the spread of self-harm and suicidal ideation on social media. To investigate the automated detection of suicidal thoughts in Arabic tweets, we developed a novel Arabic suicidal tweet dataset, examined several machine learning models trained on word frequency and embedding features, and investigated the performance of pre-trained deep learning models in identifying suicidal sentiment. The results indicate that the support vector machine trained on character n-gram features yields the best performance among conventional machine learning models, with an accuracy of 86% and F1 score of 79%. In the subsequent deep learning experiment, AraBert outperformed all other machine and deep learning models with an accuracy of 91% and F1-score of 88%, significantly improving the detection of suicidal ideation in the dataset. To the best of our knowledge, this study represents the first attempt to compile an Arabic suicidality detection dataset from Twitter and to use deep learning to detect suicidal sentiment in Arabic posts.","doi":"10.1007\/s13369-024-08767-3","cleaned_title":"detecting suicidality in arabic tweets using machine learning and deep learning techniques","cleaned_abstract":"social media platforms have revolutionized traditional communication techniques by allowing people to connect instantaneously, openly, and frequently. as people use social media to share personal stories and express their opinions, negative emotions such as thoughts of death, self-harm, and hardship are commonly expressed, particularly among younger generations. accordingly, the use of social media to detect suicidality may help provide proper intervention that will ultimately deter the spread of self-harm and suicidal ideation on social media. to investigate the automated detection of suicidal thoughts in arabic tweets, we developed a novel arabic suicidal tweet dataset, examined several machine learning models trained on word frequency and embedding features, and investigated the performance of pre-trained deep learning models in identifying suicidal sentiment. the results indicate that the support vector machine trained on character n-gram features yields the best performance among conventional machine learning models, with an accuracy of 86% and f1 score of 79%. in the subsequent deep learning experiment, arabert outperformed all other machine and deep learning models with an accuracy of 91% and f1-score of 88%, significantly improving the detection of suicidal ideation in the dataset. to the best of our knowledge, this study represents the first attempt to compile an arabic suicidality detection dataset from twitter and to use deep learning to detect suicidal sentiment in arabic posts.","key_phrases":["social media platforms","traditional communication techniques","people","people","social media","personal stories","their opinions","negative emotions","thoughts","death","self-harm","hardship","younger generations","the use","social media","suicidality","proper intervention","that","the spread","self-harm","suicidal ideation","social media","the automated detection","suicidal thoughts","arabic tweets","we","a novel arabic suicidal tweet dataset","several machine learning models","word frequency","features","the performance","pre-trained deep learning models","suicidal sentiment","the results","the support vector machine","character n-gram features","the best performance","conventional machine learning models","an accuracy","86%","f1 score","79%","the subsequent deep learning experiment","arabert","all other machine","deep learning models","an accuracy","91%","f1-score","88%","the detection","suicidal ideation","the dataset","our knowledge","this study","the first attempt","an arabic suicidality detection","twitter","deep learning","suicidal sentiment","arabic posts","arabic","arabic","86%","79%","91%","88%","first","arabic","arabic"]},{"title":"Deep learning for code generation: a survey","authors":["Huangzhao Zhang","Kechi Zhang","Zhuo Li","Jia Li","Jia Li","Yongmin Li","Yunfei Zhao","Yuqi Zhu","Fang Liu","Ge Li","Zhi Jin"],"abstract":"In the past decade, thanks to the powerfulness of deep-learning techniques, we have witnessed a whole new era of automated code generation. To sort out developments, we have conducted a comprehensive review of solutions to deep learning-based code generation. In this survey, we generally formalize the pipeline and procedure of code generation and categorize existing solutions according to taxonomy from perspectives of architecture, model-agnostic enhancing strategy, metrics, and tasks. In addition, we outline the challenges faced by current dominant large models and list several plausible directions for future research. We hope that this survey may provide handy guidance to understanding, utilizing, and developing deep learning-based code-generation techniques for researchers and practitioners.","doi":"10.1007\/s11432-023-3956-3","cleaned_title":"deep learning for code generation: a survey","cleaned_abstract":"in the past decade, thanks to the powerfulness of deep-learning techniques, we have witnessed a whole new era of automated code generation. to sort out developments, we have conducted a comprehensive review of solutions to deep learning-based code generation. in this survey, we generally formalize the pipeline and procedure of code generation and categorize existing solutions according to taxonomy from perspectives of architecture, model-agnostic enhancing strategy, metrics, and tasks. in addition, we outline the challenges faced by current dominant large models and list several plausible directions for future research. we hope that this survey may provide handy guidance to understanding, utilizing, and developing deep learning-based code-generation techniques for researchers and practitioners.","key_phrases":["the past decade","the powerfulness","deep-learning techniques","we","a whole new era","automated code generation","developments","we","a comprehensive review","solutions","deep learning-based code generation","this survey","we","the pipeline","procedure","code generation","existing solutions","taxonomy","perspectives","architecture","model-agnostic enhancing strategy","metrics","tasks","addition","we","the challenges","current dominant large models","several plausible directions","future research","we","this survey","handy guidance","understanding","deep learning-based code-generation techniques","researchers","practitioners","the past decade"]},{"title":"Privacy enhanced course recommendations through deep learning in Federated Learning environments","authors":["Chandra Sekhar Kolli","Sreenivasu Seelamanthula","Venkata Krishna Reddy V","Padamata Ramesh Babu","Mule Rama Krishna Reddy","Babu Rao Gumpina"],"abstract":"The increasing concerns around data security and privacy among users have significantly pushed the interest of the research community towards developing privacy-preserving recommendation systems. Amidst this backdrop, our study introduces a novel course recommendation methodology leveraging Federated Learning (FL) coupled with advanced Deep Learning techniques. This method executes the recommendation process across local nodes through several stages, including agglomerative matrix formulation, course clustering, bi-level matching, identification of learner-preferred courses, and ultimately, course recommendation. Notably, course clustering is achieved through Deep Fuzzy Clustering (DFC), while Deep Convolutional Neural Networks (DCNN) are employed for the recommendation phase. The efficacy of our DFC-DCNN-FL approach is rigorously evaluated based on several metrics: accuracy, False Positive Rate (FPR), loss function, Mean Square Error (MSE), Root MSE (RMSE), and Mean Average Precision (MAP). The results demonstrate remarkable performance with scores of 0.909, 0.116, 0.126, 0.291, 0.539, and 0.925, respectively.","doi":"10.1007\/s41870-024-02087-3","cleaned_title":"privacy enhanced course recommendations through deep learning in federated learning environments","cleaned_abstract":"the increasing concerns around data security and privacy among users have significantly pushed the interest of the research community towards developing privacy-preserving recommendation systems. amidst this backdrop, our study introduces a novel course recommendation methodology leveraging federated learning (fl) coupled with advanced deep learning techniques. this method executes the recommendation process across local nodes through several stages, including agglomerative matrix formulation, course clustering, bi-level matching, identification of learner-preferred courses, and ultimately, course recommendation. notably, course clustering is achieved through deep fuzzy clustering (dfc), while deep convolutional neural networks (dcnn) are employed for the recommendation phase. the efficacy of our dfc-dcnn-fl approach is rigorously evaluated based on several metrics: accuracy, false positive rate (fpr), loss function, mean square error (mse), root mse (rmse), and mean average precision (map). the results demonstrate remarkable performance with scores of 0.909, 0.116, 0.126, 0.291, 0.539, and 0.925, respectively.","key_phrases":["the increasing concerns","data security","privacy","users","the interest","the research community","privacy-preserving recommendation systems","this backdrop","our study","a novel course recommendation methodology","federated learning","advanced deep learning techniques","this method","the recommendation process","local nodes","several stages","agglomerative matrix formulation","bi-level matching","identification","learner-preferred courses","ultimately, course recommendation","course clustering","deep fuzzy clustering","dfc","deep convolutional neural networks","dcnn","the recommendation phase","the efficacy","our dfc-dcnn-fl approach","several metrics","accuracy","false positive rate","fpr","loss function","square error","mse","root mse","rmse","average precision","(map","the results","remarkable performance","scores","rmse","0.909","0.116","0.126","0.291","0.539","0.925"]},{"title":"DEEP-squared: deep learning powered De-scattering with Excitation Patterning","authors":["Navodini Wijethilake","Mithunjha Anandakumar","Cheng Zheng","Peter T. C. So","Murat Yildirim","Dushan N. Wadduwage"],"abstract":"Limited throughput is a key challenge in in vivo deep tissue imaging using nonlinear optical microscopy. Point scanning multiphoton microscopy, the current gold standard, is slow especially compared to the widefield imaging modalities used for optically cleared or thin specimens. We recently introduced \u201cDe-scattering with Excitation Patterning\u201d or \u201cDEEP\u201d as a widefield alternative to point-scanning geometries. Using patterned multiphoton excitation, DEEP encodes spatial information inside tissue before scattering. However, to de-scatter at typical depths, hundreds of such patterned excitations were needed. In this work, we present DEEP2, a deep learning-based model that can de-scatter images from just tens of patterned excitations instead of hundreds. Consequently, we improve DEEP\u2019s throughput by almost an order of magnitude. We demonstrate our method in multiple numerical and experimental imaging studies, including in vivo cortical vasculature imaging up to 4 scattering lengths deep in live mice.","doi":"10.1038\/s41377-023-01248-6","cleaned_title":"deep-squared: deep learning powered de-scattering with excitation patterning","cleaned_abstract":"limited throughput is a key challenge in in vivo deep tissue imaging using nonlinear optical microscopy. point scanning multiphoton microscopy, the current gold standard, is slow especially compared to the widefield imaging modalities used for optically cleared or thin specimens. we recently introduced \u201cde-scattering with excitation patterning\u201d or \u201cdeep\u201d as a widefield alternative to point-scanning geometries. using patterned multiphoton excitation, deep encodes spatial information inside tissue before scattering. however, to de-scatter at typical depths, hundreds of such patterned excitations were needed. in this work, we present deep2, a deep learning-based model that can de-scatter images from just tens of patterned excitations instead of hundreds. consequently, we improve deep\u2019s throughput by almost an order of magnitude. we demonstrate our method in multiple numerical and experimental imaging studies, including in vivo cortical vasculature imaging up to 4 scattering lengths deep in live mice.","key_phrases":["limited throughput","a key challenge","vivo deep tissue","nonlinear optical microscopy","point","multiphoton microscopy","the current gold standard","the widefield imaging modalities","optically cleared or thin specimens","we","de","a widefield alternative","point-scanning geometries","patterned multiphoton excitation","tissue","de","-","scatter","typical depths","hundreds","such patterned excitations","this work","we","deep2","a deep learning-based model","that","de-scatter images","just tens","patterned excitations","hundreds","we","deep\u2019s throughput","almost an order","magnitude","we","our method","multiple numerical and experimental imaging studies","vivo cortical vasculature","scattering lengths","live mice","multiphoton","hundreds","deep2","just tens","hundreds","4"]},{"title":"Harnessing deep learning for population genetic inference","authors":["Xin Huang","Aigerim Rymbekova","Olga Dolgova","Oscar Lao","Martin Kuhlwilm"],"abstract":"In population genetics, the emergence of large-scale genomic data for various species and populations has provided new opportunities to understand the evolutionary forces that drive genetic diversity using statistical inference. However, the era of population genomics presents new challenges in analysing the massive amounts of genomes and variants. Deep learning has demonstrated state-of-the-art performance for numerous applications involving large-scale data. Recently, deep learning approaches have gained popularity in population genetics; facilitated by the advent of massive genomic data sets, powerful computational hardware and complex deep learning architectures, they have been used to identify population structure, infer demographic history and investigate natural selection. Here, we introduce common deep learning architectures and provide comprehensive guidelines for implementing deep learning models for population genetic inference. We also discuss current challenges and future directions for applying deep learning in population genetics, focusing on efficiency, robustness and interpretability.","doi":"10.1038\/s41576-023-00636-3","cleaned_title":"harnessing deep learning for population genetic inference","cleaned_abstract":"in population genetics, the emergence of large-scale genomic data for various species and populations has provided new opportunities to understand the evolutionary forces that drive genetic diversity using statistical inference. however, the era of population genomics presents new challenges in analysing the massive amounts of genomes and variants. deep learning has demonstrated state-of-the-art performance for numerous applications involving large-scale data. recently, deep learning approaches have gained popularity in population genetics; facilitated by the advent of massive genomic data sets, powerful computational hardware and complex deep learning architectures, they have been used to identify population structure, infer demographic history and investigate natural selection. here, we introduce common deep learning architectures and provide comprehensive guidelines for implementing deep learning models for population genetic inference. we also discuss current challenges and future directions for applying deep learning in population genetics, focusing on efficiency, robustness and interpretability.","key_phrases":["population genetics","the emergence","large-scale genomic data","various species","populations","new opportunities","the evolutionary forces","that","genetic diversity","statistical inference","the era","population genomics","new challenges","the massive amounts","genomes","variants","deep learning","the-art","numerous applications","large-scale data","deep learning approaches","popularity","population genetics","the advent","massive genomic data sets","powerful computational hardware","complex deep learning architectures","they","population structure","demographic history","natural selection","we","common deep learning architectures","comprehensive guidelines","deep learning models","population genetic inference","we","current challenges","future directions","deep learning","population genetics","efficiency","robustness","interpretability"]},{"title":"An enhanced deep learning method for multi-class brain tumor classification using deep transfer learning","authors":["Sohaib Asif","Ming Zhao","Fengxiao Tang","Yusen Zhu"],"abstract":"Multi-class brain tumor classification is an important area of research in the field of medical imaging because of the different tumor characteristics. One such challenging problem is the multiclass classification of brain tumors using MR images. Since accuracy is critical in classification, computer vision researchers are introducing a number of techniques; However, achieving high accuracy remains challenging when classifying brain images. Early diagnosis of brain tumor types can activate timely treatment, thereby improving the patient\u2019s chances of survival. In recent years, deep learning models have achieved promising results, especially in classifying brain tumors to help neurologists. This work proposes a deep transfer learning model that accelerates brain tumor detection using MR imaging. In this paper, five popular deep learning architectures are utilized to develop a system for diagnosing brain tumors. The architectures used is this paper are Xception, DenseNet201, DenseNet121, ResNet152V2, and InceptionResNetV2. The final layer of these architectures has been modified with our deep dense block and softmax layer as the output layer to improve the classification accuracy. This article presents two main experiments to assess the effectiveness of the proposed model. First, three-class results using images from patients with glioma, meningioma, and pituitary are discussed. Second, the results of four classes are discussed using images of glioma, meningioma, pituitary and healthy patients. The results show that the proposed model based on Xception architecture is the most suitable deep learning model for detecting brain tumors. It achieves a classification accuracy of 99.67% on the 3-class dataset and 95.87% on the 4-class dataset, which is better than the state-of-the-art methods. In conclusion, the proposed model can provide radiologists with an automated medical diagnostic system to make fast and accurate decisions.","doi":"10.1007\/s11042-023-14828-w","cleaned_title":"an enhanced deep learning method for multi-class brain tumor classification using deep transfer learning","cleaned_abstract":"multi-class brain tumor classification is an important area of research in the field of medical imaging because of the different tumor characteristics. one such challenging problem is the multiclass classification of brain tumors using mr images. since accuracy is critical in classification, computer vision researchers are introducing a number of techniques; however, achieving high accuracy remains challenging when classifying brain images. early diagnosis of brain tumor types can activate timely treatment, thereby improving the patient\u2019s chances of survival. in recent years, deep learning models have achieved promising results, especially in classifying brain tumors to help neurologists. this work proposes a deep transfer learning model that accelerates brain tumor detection using mr imaging. in this paper, five popular deep learning architectures are utilized to develop a system for diagnosing brain tumors. the architectures used is this paper are xception, densenet201, densenet121, resnet152v2, and inceptionresnetv2. the final layer of these architectures has been modified with our deep dense block and softmax layer as the output layer to improve the classification accuracy. this article presents two main experiments to assess the effectiveness of the proposed model. first, three-class results using images from patients with glioma, meningioma, and pituitary are discussed. second, the results of four classes are discussed using images of glioma, meningioma, pituitary and healthy patients. the results show that the proposed model based on xception architecture is the most suitable deep learning model for detecting brain tumors. it achieves a classification accuracy of 99.67% on the 3-class dataset and 95.87% on the 4-class dataset, which is better than the state-of-the-art methods. in conclusion, the proposed model can provide radiologists with an automated medical diagnostic system to make fast and accurate decisions.","key_phrases":["multi-class brain tumor classification","an important area","research","the field","medical imaging","the different tumor characteristics","one such challenging problem","the multiclass classification","brain tumors","mr images","accuracy","classification","computer vision researchers","a number","techniques","high accuracy","brain images","early diagnosis","brain tumor types","timely treatment","the patient\u2019s chances","survival","recent years","deep learning models","promising results","brain tumors","neurologists","this work","a deep transfer learning model","that","brain tumor detection","mr imaging","this paper","five popular deep learning architectures","a system","brain tumors","the architectures","this paper","xception","densenet201","densenet121","resnet152v2","the final layer","these architectures","our deep dense block","softmax layer","the output layer","the classification accuracy","this article","two main experiments","the effectiveness","the proposed model","first, three-class results","images","patients","glioma","pituitary","the results","four classes","images","glioma","patients","the results","the proposed model","xception architecture","the most suitable deep learning model","brain tumors","it","a classification accuracy","99.67%","the 3-class dataset","95.87%","the 4-class dataset","which","the-art","conclusion","the proposed model","radiologists","an automated medical diagnostic system","fast and accurate decisions","one","recent years","five","inceptionresnetv2","two","first","three","glioma","second","four","glioma","99.67%","3","95.87%","4"]},{"title":"A Procedural Constructive Learning Mechanism with Deep Reinforcement Learning for Cognitive Agents","authors":["Leonardo de\u00a0Lellis\u00a0Rossi","Eric Rohmer","Paula Dornhofer\u00a0Paro\u00a0Costa","Esther\u00a0Luna Colombini","Alexandre da\u00a0Silva\u00a0Sim\u00f5es","Ricardo\u00a0Ribeiro Gudwin"],"abstract":"Recent advancements in AI and deep learning have created a growing demand for artificial agents capable of performing tasks within increasingly complex environments. To address the challenges associated with continuous learning constraints and knowledge capacity in this context, cognitive architectures inspired by human cognition have gained significance. This study contributes to existing research by introducing a cognitive-attentional system employing a constructive neural network-based learning approach for continuous acquisition of procedural knowledge. We replace an incremental tabular Reinforcement Learning algorithm with a constructive neural network deep reinforcement learning mechanism for continuous sensorimotor knowledge acquisition, thereby enhancing the overall learning capacity. The primary emphasis of this modification centers on optimizing memory utilization and reducing training time. Our study presents a learning strategy that amalgamates deep reinforcement learning with procedural learning, mirroring the incremental learning process observed in human sensorimotor development. This approach is embedded within the CONAIM cognitive-attentional architecture, leveraging the cognitive tools of CST. The proposed learning mechanism allows the model to dynamically create and modify elements in its procedural memory, facilitating the reuse of previously acquired functions and procedures. Additionally, it equips the model with the capability to combine learned elements to effectively adapt to complex scenarios. A constructive neural network was employed, initiating with an initial hidden layer comprising one neuron. However, it possesses the capacity to adapt its internal architecture in response to its performance in procedural and sensorimotor learning tasks, inserting new hidden layers or neurons. Experimentation conducted through simulations involving a humanoid robot demonstrates the successful resolution of tasks that were previously unsolved through incremental knowledge acquisition. Throughout the training phase, the constructive agent achieved a minimum of 40% greater rewards and executed 8% more actions when compared to other agents. In the subsequent testing phase, the constructive agent exhibited a 15% increase in the number of actions performed in contrast to its counterparts.","doi":"10.1007\/s10846-024-02064-9","cleaned_title":"a procedural constructive learning mechanism with deep reinforcement learning for cognitive agents","cleaned_abstract":"recent advancements in ai and deep learning have created a growing demand for artificial agents capable of performing tasks within increasingly complex environments. to address the challenges associated with continuous learning constraints and knowledge capacity in this context, cognitive architectures inspired by human cognition have gained significance. this study contributes to existing research by introducing a cognitive-attentional system employing a constructive neural network-based learning approach for continuous acquisition of procedural knowledge. we replace an incremental tabular reinforcement learning algorithm with a constructive neural network deep reinforcement learning mechanism for continuous sensorimotor knowledge acquisition, thereby enhancing the overall learning capacity. the primary emphasis of this modification centers on optimizing memory utilization and reducing training time. our study presents a learning strategy that amalgamates deep reinforcement learning with procedural learning, mirroring the incremental learning process observed in human sensorimotor development. this approach is embedded within the conaim cognitive-attentional architecture, leveraging the cognitive tools of cst. the proposed learning mechanism allows the model to dynamically create and modify elements in its procedural memory, facilitating the reuse of previously acquired functions and procedures. additionally, it equips the model with the capability to combine learned elements to effectively adapt to complex scenarios. a constructive neural network was employed, initiating with an initial hidden layer comprising one neuron. however, it possesses the capacity to adapt its internal architecture in response to its performance in procedural and sensorimotor learning tasks, inserting new hidden layers or neurons. experimentation conducted through simulations involving a humanoid robot demonstrates the successful resolution of tasks that were previously unsolved through incremental knowledge acquisition. throughout the training phase, the constructive agent achieved a minimum of 40% greater rewards and executed 8% more actions when compared to other agents. in the subsequent testing phase, the constructive agent exhibited a 15% increase in the number of actions performed in contrast to its counterparts.","key_phrases":["recent advancements","ai","deep learning","a growing demand","artificial agents","tasks","increasingly complex environments","the challenges","continuous learning constraints","knowledge capacity","this context","cognitive architectures","human cognition","significance","this study","existing research","a cognitive-attentional system","a constructive neural network-based learning approach","continuous acquisition","procedural knowledge","we","an incremental tabular reinforcement learning algorithm","a constructive neural network deep reinforcement learning mechanism","continuous sensorimotor knowledge acquisition","the overall learning capacity","the primary emphasis","this modification centers","memory utilization","training time","our study","a learning strategy","that","deep reinforcement learning","procedural learning","the incremental learning process","human sensorimotor development","this approach","the conaim cognitive-attentional architecture","the cognitive tools","cst","the proposed learning mechanism","the model","elements","its procedural memory","the reuse","previously acquired functions","procedures","it","the model","the capability","elements","complex scenarios","a constructive neural network","an initial hidden layer","one neuron","it","the capacity","its internal architecture","response","its performance","procedural and sensorimotor learning tasks","new hidden layers","neurons","experimentation","simulations","a humanoid robot","the successful resolution","tasks","that","incremental knowledge acquisition","the training phase","the constructive agent","a minimum","40% greater rewards","8% more actions","other agents","the subsequent testing phase","the constructive agent","a 15% increase","the number","actions","contrast","its counterparts","one","40%","8%","15%"]},{"title":"Classification of Different Plant Species Using Deep Learning and Machine Learning Algorithms","authors":["Siddharth Singh Chouhan","Uday Pratap Singh","Utkarsh Sharma","Sanjeev Jain"],"abstract":"In the present situation, a lot of research has been directed towards the potency of plants. These natural resources contain characteristics valuable in combat against a number of diseases. But due to lack of familiarity of these plants among human beings, an appropriate advantage of their significance cannot be drawn away. Plants also shares the certain similar characteristics of leaves like color, texture, shape or size, making them hard to classify them among others. So, to eradicate this problem, a deep learning model has been used for the purpose for classification of different plants species captured in real-time using internet of things practice. Six different plants namely Ashwagandha, Black Pepper, Garlic, Ginger, Basil, and Turmeric has been selected for this purpose. Our proposed convolutional neural network (CNN) model achieved higher performance with an accuracy of 99% when compared with other benchmark deep learning models. Also, to analyze the performance of deep learning versus machine learning models like logistic regression, decision tree, random forest, Gaussian na\u00efve Bayes, support vector machine results were evaluated and when compared CNN outperforms against all machine learning models. The future study will be directed towards the automated plant growth estimation.","doi":"10.1007\/s11277-024-11374-y","cleaned_title":"classification of different plant species using deep learning and machine learning algorithms","cleaned_abstract":"in the present situation, a lot of research has been directed towards the potency of plants. these natural resources contain characteristics valuable in combat against a number of diseases. but due to lack of familiarity of these plants among human beings, an appropriate advantage of their significance cannot be drawn away. plants also shares the certain similar characteristics of leaves like color, texture, shape or size, making them hard to classify them among others. so, to eradicate this problem, a deep learning model has been used for the purpose for classification of different plants species captured in real-time using internet of things practice. six different plants namely ashwagandha, black pepper, garlic, ginger, basil, and turmeric has been selected for this purpose. our proposed convolutional neural network (cnn) model achieved higher performance with an accuracy of 99% when compared with other benchmark deep learning models. also, to analyze the performance of deep learning versus machine learning models like logistic regression, decision tree, random forest, gaussian na\u00efve bayes, support vector machine results were evaluated and when compared cnn outperforms against all machine learning models. the future study will be directed towards the automated plant growth estimation.","key_phrases":["the present situation","a lot","research","the potency","plants","these natural resources","characteristics","combat","a number","diseases","lack","familiarity","these plants","human beings","an appropriate advantage","their significance","plants","the certain similar characteristics","leaves","color","texture","shape","size","them","them","others","this problem","a deep learning model","the purpose","classification","different plants species","real-time","internet","six different plants","namely ashwagandha","black pepper","garlic","ginger","basil","turmeric","this purpose","our proposed convolutional neural network (cnn) model","higher performance","an accuracy","99%","other benchmark deep learning models","the performance","deep learning","machine learning models","logistic regression","decision tree","random forest","gaussian na\u00efve bayes","support vector machine results","cnn","outperforms","all machine learning models","the future study","the automated plant growth estimation","six","basil","cnn","99%","gaussian na\u00efve bayes","cnn"]},{"title":"A deep learning model for anti-inflammatory peptides identification based on deep variational autoencoder and contrastive learning","authors":["Yujie Xu","Shengli Zhang","Feng Zhu","Yunyun Liang"],"abstract":"As a class of biologically active molecules with significant immunomodulatory and anti-inflammatory effects, anti-inflammatory peptides have important application value in the medical and biotechnology fields due to their unique biological functions. Research on the identification of anti-inflammatory peptides provides important theoretical foundations and practical value for a deeper understanding of the biological mechanisms of inflammation and immune regulation, as well as for the development of new drugs and biotechnological applications. Therefore, it is necessary to develop more advanced computational models for identifying anti-inflammatory peptides. In this study, we propose a deep learning model named DAC-AIPs based on variational autoencoder and contrastive learning for accurate identification of anti-inflammatory peptides. In the sequence encoding part, the incorporation of multi-hot encoding helps capture richer sequence information. The autoencoder, composed of convolutional layers and linear layers, can learn latent features and reconstruct features, with variational inference enhancing the representation capability of latent features. Additionally, the introduction of contrastive learning aims to improve the model's classification ability. Through cross-validation and independent dataset testing experiments, DAC-AIPs achieves superior performance compared to existing state-of-the-art models. In cross-validation, the classification accuracy of DAC-AIPs reached around 88%, which is 7% higher than previous models. Furthermore, various ablation experiments and interpretability experiments validate the effectiveness of DAC-AIPs. Finally, a user-friendly online predictor is designed to enhance the practicality of the model, and the server is freely accessible at http:\/\/dac-aips.online.","doi":"10.1038\/s41598-024-69419-y","cleaned_title":"a deep learning model for anti-inflammatory peptides identification based on deep variational autoencoder and contrastive learning","cleaned_abstract":"as a class of biologically active molecules with significant immunomodulatory and anti-inflammatory effects, anti-inflammatory peptides have important application value in the medical and biotechnology fields due to their unique biological functions. research on the identification of anti-inflammatory peptides provides important theoretical foundations and practical value for a deeper understanding of the biological mechanisms of inflammation and immune regulation, as well as for the development of new drugs and biotechnological applications. therefore, it is necessary to develop more advanced computational models for identifying anti-inflammatory peptides. in this study, we propose a deep learning model named dac-aips based on variational autoencoder and contrastive learning for accurate identification of anti-inflammatory peptides. in the sequence encoding part, the incorporation of multi-hot encoding helps capture richer sequence information. the autoencoder, composed of convolutional layers and linear layers, can learn latent features and reconstruct features, with variational inference enhancing the representation capability of latent features. additionally, the introduction of contrastive learning aims to improve the model's classification ability. through cross-validation and independent dataset testing experiments, dac-aips achieves superior performance compared to existing state-of-the-art models. in cross-validation, the classification accuracy of dac-aips reached around 88%, which is 7% higher than previous models. furthermore, various ablation experiments and interpretability experiments validate the effectiveness of dac-aips. finally, a user-friendly online predictor is designed to enhance the practicality of the model, and the server is freely accessible at http:\/\/dac-aips.online.","key_phrases":["a class","biologically active molecules","anti-inflammatory effects","anti-inflammatory peptides","important application value","the medical and biotechnology fields","their unique biological functions","research","the identification","anti-inflammatory peptides","important theoretical foundations","practical value","a deeper understanding","the biological mechanisms","inflammation","immune regulation","the development","new drugs","biotechnological applications","it","more advanced computational models","anti-inflammatory peptides","this study","we","a deep learning model","dac-aips","variational autoencoder","contrastive learning","accurate identification","anti-inflammatory peptides","the sequence","part","the incorporation","multi-hot encoding","richer sequence information","the autoencoder","convolutional layers","linear layers","latent features","features","variational inference","the representation capability","latent features","the introduction","contrastive learning","the model's classification ability","cross-validation and independent dataset testing experiments","dac-aips","superior performance","the-art","-","validation","the classification accuracy","dac-aips","around 88%","which","previous models","various ablation experiments","interpretability experiments","the effectiveness","dac-aips","a user-friendly online predictor","the practicality","the model","the server","http:\/\/dac-aips.online","linear","around 88%","7%"]},{"title":"Fraud Detection Using Machine Learning and Deep Learning","authors":["Akash Gandhar","Kapil Gupta","Aman Kumar Pandey","Dharm Raj"],"abstract":"Detecting fraudulent activities is a major worry for businesses and financial organizations because they can result in significant financial losses and reputational harm. Traditional fraud detection a method frequently depend on present rules and patterns that skilled scammer can easily circumvent. Machine learning and deep learning algorithms have surfaced as promising methods for detecting fraud in order to handle this problem. Authors present a thorough overview of the most recent ML and DL techniques for fraud identification in this article. These approaches are classified based on their fundamental tactics, which include supervised learning, unsupervised learning, and reinforcement learning. We review recent developments in each area, as well as their strengths and weaknesses. Additionally, we draw attention to some of the major problems with imbalanced datasets, adversarial assaults, and the interpretability of models as well as other important research tasks and difficulties in fraud detection. We also stress the value of feature science and data pre-processing techniques in enhancing the effectiveness of scam detection systems. Finally, we show a case study on the use of DL and ML techniques in the financial sector for fraud detection. Authors show how these algorithms can successfully identify fraudulent transactions, minimize false positives, and keep high precision and scalability. The overall aim of this article is to provide a comprehensive evaluation of the most cutting-edge ML and DL techniques for fraud identification and to shed light on potential future paths for this field of study.","doi":"10.1007\/s42979-024-02772-x","cleaned_title":"fraud detection using machine learning and deep learning","cleaned_abstract":"detecting fraudulent activities is a major worry for businesses and financial organizations because they can result in significant financial losses and reputational harm. traditional fraud detection a method frequently depend on present rules and patterns that skilled scammer can easily circumvent. machine learning and deep learning algorithms have surfaced as promising methods for detecting fraud in order to handle this problem. authors present a thorough overview of the most recent ml and dl techniques for fraud identification in this article. these approaches are classified based on their fundamental tactics, which include supervised learning, unsupervised learning, and reinforcement learning. we review recent developments in each area, as well as their strengths and weaknesses. additionally, we draw attention to some of the major problems with imbalanced datasets, adversarial assaults, and the interpretability of models as well as other important research tasks and difficulties in fraud detection. we also stress the value of feature science and data pre-processing techniques in enhancing the effectiveness of scam detection systems. finally, we show a case study on the use of dl and ml techniques in the financial sector for fraud detection. authors show how these algorithms can successfully identify fraudulent transactions, minimize false positives, and keep high precision and scalability. the overall aim of this article is to provide a comprehensive evaluation of the most cutting-edge ml and dl techniques for fraud identification and to shed light on potential future paths for this field of study.","key_phrases":["fraudulent activities","a major worry","businesses","financial organizations","they","significant financial losses","reputational harm","traditional fraud detection","a method","present rules","patterns","skilled scammer","machine learning","deep learning algorithms","promising methods","fraud","order","this problem","authors","a thorough overview","the most recent ml and dl techniques","fraud identification","this article","these approaches","their fundamental tactics","which","supervised learning","unsupervised learning","reinforcement learning","we","recent developments","each area","their strengths","weaknesses","we","attention","some","the major problems","imbalanced datasets","adversarial assaults","the interpretability","models","other important research tasks","difficulties","fraud detection","we","the value","feature science and data pre-processing techniques","the effectiveness","scam detection systems","we","a case study","the use","dl and ml techniques","the financial sector","fraud detection","authors","these algorithms","fraudulent transactions","false positives","high precision","scalability","the overall aim","this article","a comprehensive evaluation","the most cutting-edge ml and dl techniques","fraud identification","light","potential future paths","this field","study"]},{"title":"Robot autonomous grasping and assembly skill learning based on deep reinforcement learning","authors":["Chengjun Chen","Hao Zhang","Yong Pan","Dongnian Li"],"abstract":"This paper proposes a deep reinforcement learning-based framework for robot autonomous grasping and assembly skill learning. Meanwhile, a deep Q-learning-based robot grasping skill learning algorithm and a PPO-based robot assembly skill learning algorithm are presented, where a priori knowledge information is introduced to optimize the grasping action and reduce the training time and interaction data needed by the assembly strategy learning algorithm. Besides, a grasping constraint reward function and an assembly constraint reward function are designed to evaluate the robot grasping and assembly quality effectively. Finally, the effectiveness of the proposed framework and algorithms was verified in both simulated and real environments, and the average success rate of grasping in both environments was up to 90%. Under a peg-in-hole assembly tolerance of 3\u00a0mm, the assembly success rate was 86.7% and 73.3% in the simulated environment and the physical environment, respectively.","doi":"10.1007\/s00170-024-13004-0","cleaned_title":"robot autonomous grasping and assembly skill learning based on deep reinforcement learning","cleaned_abstract":"this paper proposes a deep reinforcement learning-based framework for robot autonomous grasping and assembly skill learning. meanwhile, a deep q-learning-based robot grasping skill learning algorithm and a ppo-based robot assembly skill learning algorithm are presented, where a priori knowledge information is introduced to optimize the grasping action and reduce the training time and interaction data needed by the assembly strategy learning algorithm. besides, a grasping constraint reward function and an assembly constraint reward function are designed to evaluate the robot grasping and assembly quality effectively. finally, the effectiveness of the proposed framework and algorithms was verified in both simulated and real environments, and the average success rate of grasping in both environments was up to 90%. under a peg-in-hole assembly tolerance of 3 mm, the assembly success rate was 86.7% and 73.3% in the simulated environment and the physical environment, respectively.","key_phrases":["this paper","a deep reinforcement learning-based framework","robot autonomous grasping and assembly skill learning","a deep q-learning-based robot grasping skill","algorithm","a ppo-based robot assembly skill learning algorithm","a priori knowledge information","the grasping action","the training time","interaction data","the assembly strategy","a grasping constraint reward function","an assembly constraint reward function","the robot grasping and assembly quality","the effectiveness","the proposed framework","algorithms","both simulated and real environments","the average success rate","both environments","up to 90%","hole","the assembly success rate","86.7%","73.3%","the simulated environment","the physical environment","up to 90%","3 mm","86.7%","73.3%"]},{"title":"Enabling business sustainability for stock market data using machine learning and deep learning approaches","authors":["S. Divyashree","Christy Jackson Joshua","Abdul Quadir Md","Senthilkumar Mohan","A. Sheik Abdullah","Ummul Hanan Mohamad","Nisreen Innab","Ali Ahmadian"],"abstract":"This paper introduces AlphaVision, an innovative decision support model designed for stock price prediction by seamlessly integrating real-time news updates and Return on Investment (ROI) values, utilizing various machine learning and deep learning approaches. The research investigates the application of these techniques to enhance the effectiveness of stock trading and investment decisions by accurately anticipating stock prices and providing valuable insights to investors and businesses. The study begins by analyzing the complexities and challenges of stock market analysis, considering factors like political, macroeconomic, and legal issues that contribute to market volatility. To address these challenges, we proposed the methodology called AlphaVision, which incorporates various machine learning algorithms, including Decision Trees, Random Forest, Na\u00efve Bayes, Boosting, K-Nearest Neighbors, and Support Vector Machine, alongside deep learning models such as Multi-layer Perceptron (MLP), Artificial Neural Networks, and Recurrent Neural Networks. The effectiveness of each model is evaluated based on their accuracy in predicting stock prices. Experimental results revealed that the MLP model achieved the highest accuracy of approximately 92%, outperforming other deep learning models. The Random Forest algorithm also demonstrated promising results with an accuracy of around 84.6%. These findings indicate the potential of machine learning and deep learning techniques in improving stock market analysis and prediction. The AlphaVision methodology presented in this research empowers investors and businesses with valuable tools to make informed investment decisions and navigate the complexities of the stock market. By accurately forecasting stock prices based on news updates and ROI values, the model contributes to better financial management and business sustainability. The integration of machine learning and deep learning approaches offers a promising solution for enhancing stock market analysis and prediction. Future research will focus on extracting more relevant financial features to further improve the model\u2019s accuracy. By advancing decision support models for stock price prediction, researchers and practitioners can foster better investment strategies and foster economic growth. The proposed model holds potential to revolutionize stock trading and investment practices, enabling more informed and profitable decision-making in the financial sector.","doi":"10.1007\/s10479-024-06118-x","cleaned_title":"enabling business sustainability for stock market data using machine learning and deep learning approaches","cleaned_abstract":"this paper introduces alphavision, an innovative decision support model designed for stock price prediction by seamlessly integrating real-time news updates and return on investment (roi) values, utilizing various machine learning and deep learning approaches. the research investigates the application of these techniques to enhance the effectiveness of stock trading and investment decisions by accurately anticipating stock prices and providing valuable insights to investors and businesses. the study begins by analyzing the complexities and challenges of stock market analysis, considering factors like political, macroeconomic, and legal issues that contribute to market volatility. to address these challenges, we proposed the methodology called alphavision, which incorporates various machine learning algorithms, including decision trees, random forest, na\u00efve bayes, boosting, k-nearest neighbors, and support vector machine, alongside deep learning models such as multi-layer perceptron (mlp), artificial neural networks, and recurrent neural networks. the effectiveness of each model is evaluated based on their accuracy in predicting stock prices. experimental results revealed that the mlp model achieved the highest accuracy of approximately 92%, outperforming other deep learning models. the random forest algorithm also demonstrated promising results with an accuracy of around 84.6%. these findings indicate the potential of machine learning and deep learning techniques in improving stock market analysis and prediction. the alphavision methodology presented in this research empowers investors and businesses with valuable tools to make informed investment decisions and navigate the complexities of the stock market. by accurately forecasting stock prices based on news updates and roi values, the model contributes to better financial management and business sustainability. the integration of machine learning and deep learning approaches offers a promising solution for enhancing stock market analysis and prediction. future research will focus on extracting more relevant financial features to further improve the model\u2019s accuracy. by advancing decision support models for stock price prediction, researchers and practitioners can foster better investment strategies and foster economic growth. the proposed model holds potential to revolutionize stock trading and investment practices, enabling more informed and profitable decision-making in the financial sector.","key_phrases":["this paper","alphavision","an innovative decision support model","stock price prediction","real-time news updates","return","roi","various machine learning","deep learning approaches","the research","the application","these techniques","the effectiveness","stock trading and investment decisions","stock prices","valuable insights","investors","businesses","the study","the complexities","challenges","stock market analysis","factors","political, macroeconomic, and legal issues","that","market volatility","these challenges","we","the methodology","alphavision","which","various machine learning algorithms","decision trees","random forest","na\u00efve bayes","k-nearest neighbors","vector machine","deep learning models","multi-layer perceptron","mlp","artificial neural networks","neural networks","the effectiveness","each model","their accuracy","stock prices","experimental results","the mlp model","the highest accuracy","approximately 92%","other deep learning models","the random forest algorithm","promising results","an accuracy","around 84.6%","these findings","the potential","machine learning","deep learning techniques","stock market analysis","prediction","the alphavision methodology","this research","investors","businesses","valuable tools","informed investment decisions","the complexities","the stock market","stock prices","news updates","roi values","the model","better financial management","business sustainability","the integration","machine learning","deep learning approaches","a promising solution","stock market analysis","prediction","future research","more relevant financial features","the model\u2019s accuracy","decision support models","stock price prediction","researchers","practitioners","better investment strategies","foster economic growth","the proposed model","potential","stock trading","investment practices","more informed and profitable decision-making","the financial sector","approximately 92%","around 84.6%"]},{"title":"Deep learning for lungs cancer detection: a review","authors":["Rabia Javed","Tahir Abbas","Ali Haider Khan","Ali Daud","Amal Bukhari","Riad Alharbey"],"abstract":"Although lung cancer has been recognized to be the deadliest type of cancer, a good prognosis and efficient treatment depend on early detection. Medical practitioners\u2019 burden is reduced by deep learning techniques, especially Deep Convolutional Neural Networks (DCNN), which are essential in automating the diagnosis and classification of diseases. In this study, we use a variety of medical imaging modalities, including X-rays, WSI, CT scans, and MRI, to thoroughly investigate the use of deep learning techniques in the field of lung cancer diagnosis and classification. This study conducts a comprehensive Systematic Literature Review (SLR) using deep learning techniques for lung cancer research, providing a comprehensive overview of the methodology, cutting-edge developments, quality assessments, and customized deep learning approaches. It presents data from reputable journals and concentrates on the years 2015\u20132024. Deep learning techniques solve the difficulty of manually identifying and selecting abstract features from lung cancer images. This study includes a wide range of deep learning methods for classifying lung cancer but focuses especially on the most popular method, the Convolutional Neural Network (CNN). CNN can achieve maximum accuracy because of its multi-layer structure, automatic learning of weights, and capacity to communicate local weights. Various algorithms are shown with performance measures like precision, accuracy, specificity, sensitivity, and AUC; CNN consistently shows the greatest accuracy. The findings highlight the important contributions of DCNN in improving lung cancer detection and classification, making them an invaluable resource for researchers looking to gain a greater knowledge of deep learning\u2019s function in medical applications.","doi":"10.1007\/s10462-024-10807-1","cleaned_title":"deep learning for lungs cancer detection: a review","cleaned_abstract":"although lung cancer has been recognized to be the deadliest type of cancer, a good prognosis and efficient treatment depend on early detection. medical practitioners\u2019 burden is reduced by deep learning techniques, especially deep convolutional neural networks (dcnn), which are essential in automating the diagnosis and classification of diseases. in this study, we use a variety of medical imaging modalities, including x-rays, wsi, ct scans, and mri, to thoroughly investigate the use of deep learning techniques in the field of lung cancer diagnosis and classification. this study conducts a comprehensive systematic literature review (slr) using deep learning techniques for lung cancer research, providing a comprehensive overview of the methodology, cutting-edge developments, quality assessments, and customized deep learning approaches. it presents data from reputable journals and concentrates on the years 2015\u20132024. deep learning techniques solve the difficulty of manually identifying and selecting abstract features from lung cancer images. this study includes a wide range of deep learning methods for classifying lung cancer but focuses especially on the most popular method, the convolutional neural network (cnn). cnn can achieve maximum accuracy because of its multi-layer structure, automatic learning of weights, and capacity to communicate local weights. various algorithms are shown with performance measures like precision, accuracy, specificity, sensitivity, and auc; cnn consistently shows the greatest accuracy. the findings highlight the important contributions of dcnn in improving lung cancer detection and classification, making them an invaluable resource for researchers looking to gain a greater knowledge of deep learning\u2019s function in medical applications.","key_phrases":["lung cancer","the deadliest type","cancer","a good prognosis and efficient treatment","early detection","medical practitioners\u2019 burden","deep learning techniques","especially deep convolutional neural networks","dcnn","which","the diagnosis","classification","diseases","this study","we","a variety","medical imaging modalities","x","-","rays","wsi","ct scans","mri","the use","deep learning techniques","the field","lung cancer diagnosis","classification","this study","a comprehensive systematic literature review","slr","deep learning techniques","lung cancer research","a comprehensive overview","the methodology","cutting-edge developments","quality assessments","customized deep learning approaches","it","data","reputable journals","concentrates","the years","deep learning techniques","the difficulty","abstract features","lung cancer images","this study","a wide range","deep learning methods","lung cancer","the most popular method","the convolutional neural network","cnn","cnn","maximum accuracy","its multi-layer structure","automatic learning","weights","capacity","local weights","various algorithms","performance measures","precision","accuracy","specificity","sensitivity","auc","cnn","the greatest accuracy","the findings","the important contributions","dcnn","lung cancer detection","classification","them","researchers","a greater knowledge","deep learning\u2019s function","medical applications","the years 2015\u20132024","cnn","cnn","cnn"]},{"title":"A brief review of hypernetworks in deep learning","authors":["Vinod Kumar Chauhan","Jiandong Zhou","Ping Lu","Soheila Molaei","David A. Clifton"],"abstract":"Hypernetworks, or hypernets for short, are neural networks that generate weights for another neural network, known as the target network. They have emerged as a powerful deep learning technique that allows for greater flexibility, adaptability, dynamism, faster training, information sharing, and model compression. Hypernets have shown promising results in a variety of deep learning problems, including continual learning, causal inference, transfer learning, weight pruning, uncertainty quantification, zero-shot learning, natural language processing, and reinforcement learning. Despite their success across different problem settings, there is currently no comprehensive review available to inform researchers about the latest developments and to assist in utilizing hypernets. To fill this gap, we review the progress in hypernets. We present an illustrative example of training deep neural networks using hypernets and propose categorizing hypernets based on five design criteria: inputs, outputs, variability of inputs and outputs, and the architecture of hypernets. We also review applications of hypernets across different deep learning problem settings, followed by a discussion of general scenarios where hypernets can be effectively employed. Finally, we discuss the challenges and future directions that remain underexplored in the field of hypernets. We believe that hypernetworks have the potential to revolutionize the field of deep learning. They offer a new way to design and train neural networks, and they have the potential to improve the performance of deep learning models on a variety of tasks. Through this review, we aim to inspire further advancements in deep learning through hypernetworks.","doi":"10.1007\/s10462-024-10862-8","cleaned_title":"a brief review of hypernetworks in deep learning","cleaned_abstract":"hypernetworks, or hypernets for short, are neural networks that generate weights for another neural network, known as the target network. they have emerged as a powerful deep learning technique that allows for greater flexibility, adaptability, dynamism, faster training, information sharing, and model compression. hypernets have shown promising results in a variety of deep learning problems, including continual learning, causal inference, transfer learning, weight pruning, uncertainty quantification, zero-shot learning, natural language processing, and reinforcement learning. despite their success across different problem settings, there is currently no comprehensive review available to inform researchers about the latest developments and to assist in utilizing hypernets. to fill this gap, we review the progress in hypernets. we present an illustrative example of training deep neural networks using hypernets and propose categorizing hypernets based on five design criteria: inputs, outputs, variability of inputs and outputs, and the architecture of hypernets. we also review applications of hypernets across different deep learning problem settings, followed by a discussion of general scenarios where hypernets can be effectively employed. finally, we discuss the challenges and future directions that remain underexplored in the field of hypernets. we believe that hypernetworks have the potential to revolutionize the field of deep learning. they offer a new way to design and train neural networks, and they have the potential to improve the performance of deep learning models on a variety of tasks. through this review, we aim to inspire further advancements in deep learning through hypernetworks.","key_phrases":["hypernetworks","hypernets","neural networks","that","weights","another neural network","the target network","they","a powerful deep learning technique","that","greater flexibility","adaptability","dynamism","faster training","information sharing","model compression","hypernets","promising results","a variety","deep learning problems","continual learning","causal inference","transfer learning","weight pruning","uncertainty quantification","zero-shot learning","natural language processing","reinforcement learning","their success","different problem settings","no comprehensive review","researchers","the latest developments","hypernets","this gap","we","the progress","hypernets","we","an illustrative example","deep neural networks","hypernets","categorizing hypernets","five design criteria","inputs","outputs","variability","inputs","outputs","the architecture","hypernets","we","applications","hypernets","different deep learning problem settings","a discussion","general scenarios","hypernets","we","the challenges","future directions","that","the field","hypernets","we","hypernetworks","the potential","the field","deep learning","they","a new way","neural networks","they","the potential","the performance","deep learning models","a variety","tasks","this review","we","further advancements","deep learning","hypernetworks","zero","five"]},{"title":"Learning Dynamic Batch-Graph Representation for Deep Representation Learning","authors":["Xixi Wang","Bo Jiang","Xiao Wang","Bin Luo"],"abstract":"Recently, batch-based image data representation has been demonstrated to be effective for context-enhanced image representation. The core issue for this task is capturing the dependences of image samples within each mini-batch and conducting message communication among different samples. Existing approaches mainly adopt self-attention or local self-attention models (on patch dimension) for this task which fail to fully exploit the intrinsic relationships of samples within mini-batch and also be sensitive to noises and outliers. To address this issue, in this paper, we propose a flexible Dynamic Batch-Graph Representation (DyBGR) model, to automatically explore the intrinsic relationship of samples for contextual sample representation. Specifically, DyBGR first represents the mini-batch with a graph (termed batch-graph) in which nodes represent image samples and edges encode the dependences of images. This graph is dynamically learned with the constraint of similarity, sparseness and semantic correlation. Upon this, DyBGR exchanges the sample (node) information on the batch-graph to update each node representation. Note that, both batch-graph learning and information propagation are jointly optimized to boost their respective performance. Furthermore, in practical, DyBGR model can be implemented via a simple plug-and-play block (named DyBGR block) which thus can be potentially integrated into any mini-batch based deep representation learning schemes. Extensive experiments on deep metric learning tasks demonstrate the effectiveness of DyBGR. We will release the code at https:\/\/github.com\/SissiW\/DyBGR.","doi":"10.1007\/s11263-024-02175-8","cleaned_title":"learning dynamic batch-graph representation for deep representation learning","cleaned_abstract":"recently, batch-based image data representation has been demonstrated to be effective for context-enhanced image representation. the core issue for this task is capturing the dependences of image samples within each mini-batch and conducting message communication among different samples. existing approaches mainly adopt self-attention or local self-attention models (on patch dimension) for this task which fail to fully exploit the intrinsic relationships of samples within mini-batch and also be sensitive to noises and outliers. to address this issue, in this paper, we propose a flexible dynamic batch-graph representation (dybgr) model, to automatically explore the intrinsic relationship of samples for contextual sample representation. specifically, dybgr first represents the mini-batch with a graph (termed batch-graph) in which nodes represent image samples and edges encode the dependences of images. this graph is dynamically learned with the constraint of similarity, sparseness and semantic correlation. upon this, dybgr exchanges the sample (node) information on the batch-graph to update each node representation. note that, both batch-graph learning and information propagation are jointly optimized to boost their respective performance. furthermore, in practical, dybgr model can be implemented via a simple plug-and-play block (named dybgr block) which thus can be potentially integrated into any mini-batch based deep representation learning schemes. extensive experiments on deep metric learning tasks demonstrate the effectiveness of dybgr. we will release the code at https:\/\/github.com\/sissiw\/dybgr.","key_phrases":["batch-based image data representation","context-enhanced image representation","the core issue","this task","the dependences","image samples","each mini","-","batch","message communication","different samples","existing approaches","self-attention","local self-attention models","patch dimension","this task","which","the intrinsic relationships","samples","mini","-","batch","noises","outliers","this issue","this paper","we","a flexible dynamic batch-graph representation (dybgr) model","the intrinsic relationship","samples","contextual sample representation","the mini","-","batch","a graph","termed batch-graph","which","nodes","image samples","edges","the dependences","images","this graph","the constraint","similarity","sparseness","semantic correlation","this","the batch-graph","each node representation","both batch-graph learning and information propagation","their respective performance","practical, dybgr model","a simple plug-and-play block","dybgr block","which","any mini-batch based deep representation learning schemes","extensive experiments","deep metric learning tasks","the effectiveness","dybgr","we","the code","https:\/\/github.com\/sissiw\/dybgr","first"]},{"title":"Relay learning: a physically secure framework for clinical multi-site deep learning","authors":["Zi-Hao Bo","Yuchen Guo","Jinhao Lyu","Hengrui Liang","Jianxing He","Shijie Deng","Feng Xu","Xin Lou","Qionghai Dai"],"abstract":"Big data serves as the cornerstone for constructing real-world deep learning systems across various domains. In medicine and healthcare, a single clinical site lacks sufficient data, thus necessitating the involvement of multiple sites. Unfortunately, concerns regarding data security and privacy hinder the sharing and reuse of data across sites. Existing approaches to multi-site clinical learning heavily depend on the security of the network firewall and system implementation. To address this issue, we propose Relay Learning, a secure deep-learning framework that physically isolates clinical data from external intruders while still leveraging the benefits of multi-site big data. We demonstrate the efficacy of Relay Learning in three medical tasks of different diseases and anatomical structures, including structure segmentation of retina fundus, mediastinum tumors diagnosis, and brain midline localization. We evaluate Relay Learning by comparing its performance to alternative solutions through multi-site validation and external validation. Incorporating a total of 41,038 medical images from 21 medical hosts, including 7 external hosts, with non-uniform distributions, we observe significant performance improvements with Relay Learning across all three tasks. Specifically, it achieves an average performance increase of 44.4%, 24.2%, and 36.7% for retinal fundus segmentation, mediastinum tumor diagnosis, and brain midline localization, respectively. Remarkably, Relay Learning even outperforms central learning on external test sets. In the meanwhile, Relay Learning keeps data sovereignty locally without cross-site network connections. We anticipate that Relay Learning will revolutionize clinical multi-site collaboration and reshape the landscape of healthcare in the future.","doi":"10.1038\/s41746-023-00934-4","cleaned_title":"relay learning: a physically secure framework for clinical multi-site deep learning","cleaned_abstract":"big data serves as the cornerstone for constructing real-world deep learning systems across various domains. in medicine and healthcare, a single clinical site lacks sufficient data, thus necessitating the involvement of multiple sites. unfortunately, concerns regarding data security and privacy hinder the sharing and reuse of data across sites. existing approaches to multi-site clinical learning heavily depend on the security of the network firewall and system implementation. to address this issue, we propose relay learning, a secure deep-learning framework that physically isolates clinical data from external intruders while still leveraging the benefits of multi-site big data. we demonstrate the efficacy of relay learning in three medical tasks of different diseases and anatomical structures, including structure segmentation of retina fundus, mediastinum tumors diagnosis, and brain midline localization. we evaluate relay learning by comparing its performance to alternative solutions through multi-site validation and external validation. incorporating a total of 41,038 medical images from 21 medical hosts, including 7 external hosts, with non-uniform distributions, we observe significant performance improvements with relay learning across all three tasks. specifically, it achieves an average performance increase of 44.4%, 24.2%, and 36.7% for retinal fundus segmentation, mediastinum tumor diagnosis, and brain midline localization, respectively. remarkably, relay learning even outperforms central learning on external test sets. in the meanwhile, relay learning keeps data sovereignty locally without cross-site network connections. we anticipate that relay learning will revolutionize clinical multi-site collaboration and reshape the landscape of healthcare in the future.","key_phrases":["big data","the cornerstone","real-world deep learning systems","various domains","medicine","healthcare","a single clinical site","sufficient data","the involvement","multiple sites","concerns","data security","privacy","the sharing","reuse","data","sites","existing approaches","multi-site clinical learning","the security","the network firewall and system implementation","this issue","we","relay learning","a secure deep-learning framework","that","clinical data","external intruders","the benefits","multi-site big data","we","the efficacy","three medical tasks","different diseases","anatomical structures","structure segmentation","retina fundus","mediastinum tumors diagnosis","brain midline localization","we","relay","its performance","solutions","multi-site validation","external validation","a total","41,038 medical images","21 medical hosts","7 external hosts","non-uniform distributions","we","significant performance improvements","relay","all three tasks","it","an average performance increase","44.4%","24.2%","36.7%","retinal fundus segmentation","mediastinum tumor diagnosis","brain midline localization","remarkably, relay","central learning","external test sets","the meanwhile","relay learning","data sovereignty","cross-site network connections","we","relay learning","clinical multi-site collaboration","the landscape","healthcare","the future","three","41,038","21","7","three","44.4%","24.2%","36.7%"]},{"title":"Predicting Apple Plant Diseases in Orchards Using Machine Learning and Deep Learning Algorithms","authors":["Imtiaz Ahmed","Pramod Kumar Yadav"],"abstract":"Apple cultivation in the Kashmir Valley is a cornerstone of the region\u2019s agriculture, contributing significantly to the economy through substantial annual apple exports. This study explores the application of machine learning and deep learning algorithms for predicting apple plant diseases in orchards. By leveraging advanced computational techniques, the research aims to enhance early detection and diagnosis of diseases, thereby enabling proactive disease management. The study utilizes a dataset comprising diverse environmental and plant health factors to train and validate the models. Key highlights include the comparative analysis of machine learning and deep learning approaches, the identification of optimal feature sets, and the assessment of model performance. The findings contribute to the development of efficient and accurate tools for precision agriculture, facilitating timely intervention and sustainable orchard management. The apple industry in Kashmir faces a significant challenge due to the prevalence of various diseases affecting apple trees. One prominent disease that adversely impacts apple yields in the region is the Apple Scab, caused by the fungus Venturia inadequacies. Apple Scab is characterized by dark, scaly lesions on leaves, fruit, and twigs, leading to defoliation and reduced fruit quality. The disease thrives in cool and humid conditions, which are prevalent in the Kashmir Valley. This study addresses the limitations of traditional, labor-intensive, and time-consuming laboratory methods for diagnosing apple plant diseases. The goal is to provide an accurate and efficient deep learning-based system for the prompt identification and prediction of foliar diseases in Kashmiri apple plants. Our study begins involves the creation of a dataset annotated by experts containing approximately 10,000 high-quality RGB images that illustrate key symptoms associated with foliar diseases. In the next step, an approach to deep learning that utilizes convolutional neural networks (CNNs) was developed. Comparative analysis five different deep learning algorithms, including Faster R-CNN, showed that the method was effective in detecting apple diseases in real time. The proposed framework, when tested, achieves state-of-the-art results with a remarkable 92% accuracy in identifying apple plant diseases. A new dataset is presented that includes samples of leaves from Kashmiri apple plants that have three different illnesses. The findings hold promise for revolutionizing orchard management practices, ultimately benefiting apple growers and sustaining the thriving apple industry in the Kashmir Valley.","doi":"10.1007\/s42979-024-02959-2","cleaned_title":"predicting apple plant diseases in orchards using machine learning and deep learning algorithms","cleaned_abstract":"apple cultivation in the kashmir valley is a cornerstone of the region\u2019s agriculture, contributing significantly to the economy through substantial annual apple exports. this study explores the application of machine learning and deep learning algorithms for predicting apple plant diseases in orchards. by leveraging advanced computational techniques, the research aims to enhance early detection and diagnosis of diseases, thereby enabling proactive disease management. the study utilizes a dataset comprising diverse environmental and plant health factors to train and validate the models. key highlights include the comparative analysis of machine learning and deep learning approaches, the identification of optimal feature sets, and the assessment of model performance. the findings contribute to the development of efficient and accurate tools for precision agriculture, facilitating timely intervention and sustainable orchard management. the apple industry in kashmir faces a significant challenge due to the prevalence of various diseases affecting apple trees. one prominent disease that adversely impacts apple yields in the region is the apple scab, caused by the fungus venturia inadequacies. apple scab is characterized by dark, scaly lesions on leaves, fruit, and twigs, leading to defoliation and reduced fruit quality. the disease thrives in cool and humid conditions, which are prevalent in the kashmir valley. this study addresses the limitations of traditional, labor-intensive, and time-consuming laboratory methods for diagnosing apple plant diseases. the goal is to provide an accurate and efficient deep learning-based system for the prompt identification and prediction of foliar diseases in kashmiri apple plants. our study begins involves the creation of a dataset annotated by experts containing approximately 10,000 high-quality rgb images that illustrate key symptoms associated with foliar diseases. in the next step, an approach to deep learning that utilizes convolutional neural networks (cnns) was developed. comparative analysis five different deep learning algorithms, including faster r-cnn, showed that the method was effective in detecting apple diseases in real time. the proposed framework, when tested, achieves state-of-the-art results with a remarkable 92% accuracy in identifying apple plant diseases. a new dataset is presented that includes samples of leaves from kashmiri apple plants that have three different illnesses. the findings hold promise for revolutionizing orchard management practices, ultimately benefiting apple growers and sustaining the thriving apple industry in the kashmir valley.","key_phrases":["apple cultivation","the kashmir valley","a cornerstone","the region\u2019s agriculture","the economy","substantial annual apple exports","this study","the application","machine learning","deep learning algorithms","apple plant diseases","orchards","advanced computational techniques","the research","early detection","diagnosis","diseases","proactive disease management","the study","a dataset","diverse environmental and plant health factors","the models","key highlights","the comparative analysis","machine learning","deep learning approaches","the identification","optimal feature sets","the assessment","model performance","the findings","the development","efficient and accurate tools","precision agriculture","timely intervention","sustainable orchard management","the apple industry","kashmir","a significant challenge","the prevalence","various diseases","apple trees","one prominent disease","that","apple yields","the region","the apple scab","the fungus venturia inadequacies","apple scab","dark","scaly lesions","leaves","fruit","twigs","defoliation","reduced fruit quality","the disease","cool and humid conditions","which","the kashmir valley","this study","the limitations","traditional, labor-intensive, and time-consuming laboratory methods","apple plant diseases","the goal","an accurate and efficient deep learning-based system","the prompt identification","prediction","foliar diseases","kashmiri apple plants","our study","the creation","a dataset","experts","approximately 10,000 high-quality rgb images","that","key symptoms","foliar diseases","the next step","an approach","deep learning","that","convolutional neural networks","cnns","comparative analysis","five different deep learning algorithms","faster r-cnn","the method","apple diseases","real time","the proposed framework","the-art","a remarkable 92% accuracy","apple plant diseases","a new dataset","that","samples","leaves","kashmiri apple plants","that","three different illnesses","the findings","promise","orchard management practices","apple growers","the thriving apple industry","the kashmir valley","the kashmir valley","annual","kashmir","one","apple scab","the kashmir valley","10,000","five","92%","three","the kashmir valley"]},{"title":"Deep Learning zur  Kariesdiagnostik","authors":["Norbert Kr\u00e4mer","Roland Frankenberger"],"abstract":"Deep-Learning-Modelle spielen auch in der Zahnheilkunde eine zunehmend gr\u00f6\u00dfere Rolle und werden in unterschiedlichen Feldern eingesetzt. Vor diesem Hintergrund wurde in der vorliegenden Literatur\u00fcbersicht eine systematische \u00dcbersichtsarbeit einer internationalen Autorengruppe vorgestellt, die Deep-Learning-Modelle zur Kariesdiagnostik analysierte und bewertete. Sie kam zu dem Schluss, dass in einer zunehmenden Anzahl von Studien die Kariesdiagnostik mithilfe von Deep-Learning-Modellen unterst\u00fctzt wird. Die dokumentierte Genauigkeit erscheint vielversprechend, w\u00e4hrend Studien- und Berichtsqualit\u00e4t derzeit unzureichend sind, um weiterf\u00fchrende Analysen durchzuf\u00fchren. Bei verbesserter Datenlage k\u00f6nnten jedoch k\u00fcnftig Deep-Learning-Modelle als Hilfsmittel f\u00fcr Entscheidungen \u00fcber das Vorhandensein von kari\u00f6sen L\u00e4sionen herangezogen werden.","doi":"10.1007\/s44190-023-0647-4","cleaned_title":"deep learning zur kariesdiagnostik","cleaned_abstract":"deep-learning-modelle spielen auch in der zahnheilkunde eine zunehmend gr\u00f6\u00dfere rolle und werden in unterschiedlichen feldern eingesetzt. vor diesem hintergrund wurde in der vorliegenden literatur\u00fcbersicht eine systematische \u00fcbersichtsarbeit einer internationalen autorengruppe vorgestellt, die deep-learning-modelle zur kariesdiagnostik analysierte und bewertete. sie kam zu dem schluss, dass in einer zunehmenden anzahl von studien die kariesdiagnostik mithilfe von deep-learning-modellen unterst\u00fctzt wird. die dokumentierte genauigkeit erscheint vielversprechend, w\u00e4hrend studien- und berichtsqualit\u00e4t derzeit unzureichend sind, um weiterf\u00fchrende analysen durchzuf\u00fchren. bei verbesserter datenlage k\u00f6nnten jedoch k\u00fcnftig deep-learning-modelle als hilfsmittel f\u00fcr entscheidungen \u00fcber das vorhandensein von kari\u00f6sen l\u00e4sionen herangezogen werden.","key_phrases":["deep-learning-modelle spielen","unterschiedlichen feldern eingesetzt","vor diesem hintergrund wurde","der vorliegenden literatur\u00fcbersicht eine systematische \u00fcbersichtsarbeit einer internationalen autorengruppe","deep-learning-modelle zur kariesdiagnostik analysierte","bewertete","sie kam zu dem schluss","dass","einer zunehmenden","anzahl von studien die kariesdiagnostik mithilfe von deep-learning-modellen unterst\u00fctzt wird","die dokumentierte genauigkeit erscheint vielversprechend","studien-","und berichtsqualit\u00e4t derzeit unzureichend sind","analysen durchzuf\u00fchren","bei verbesserter datenlage","jedoch k\u00fcnftig deep-learning-modelle als hilfsmittel f\u00fcr entscheidungen \u00fcber das vorhandensein von kari\u00f6sen l\u00e4sionen herangezogen werden","unterschiedlichen feldern eingesetzt","diesem","hintergrund wurde","literatur\u00fcbersicht","kam","dem","anzahl von studien","k\u00f6nnten jedoch","das vorhandensein von kari\u00f6sen"]},{"title":"Deep doubly robust outcome weighted learning","authors":["Xiaotong Jiang","Xin Zhou","Michael R. Kosorok"],"abstract":"Precision medicine is a framework that adapts treatment strategies to a patient\u2019s individual characteristics and provides helpful clinical decision support. Existing research has been extended to various situations but high-dimensional data have not yet been fully incorporated into the paradigm. We propose a new precision medicine approach called deep doubly robust outcome weighted learning (DDROWL) that can handle big and complex data. This is a machine learning tool that directly estimates the optimal decision rule and achieves the best of three worlds: deep learning, double robustness, and residual weighted learning. Two architectures have been implemented in the proposed method, a fully-connected feedforward neural network and the Deep Kernel Learning model, a Gaussian process with deep learning-filtered inputs. We compare and discuss the performance and limitation of different methods through a range of simulations. Using longitudinal and brain imaging data from patients with Alzheimer\u2019s disease, we demonstrate the application of the proposed method in real-world clinical practice. With the implementation of deep learning, the proposed method can expand the influence of precision medicine to high-dimensional abundant data with greater flexibility and computational power.","doi":"10.1007\/s10994-023-06484-w","cleaned_title":"deep doubly robust outcome weighted learning","cleaned_abstract":"precision medicine is a framework that adapts treatment strategies to a patient\u2019s individual characteristics and provides helpful clinical decision support. existing research has been extended to various situations but high-dimensional data have not yet been fully incorporated into the paradigm. we propose a new precision medicine approach called deep doubly robust outcome weighted learning (ddrowl) that can handle big and complex data. this is a machine learning tool that directly estimates the optimal decision rule and achieves the best of three worlds: deep learning, double robustness, and residual weighted learning. two architectures have been implemented in the proposed method, a fully-connected feedforward neural network and the deep kernel learning model, a gaussian process with deep learning-filtered inputs. we compare and discuss the performance and limitation of different methods through a range of simulations. using longitudinal and brain imaging data from patients with alzheimer\u2019s disease, we demonstrate the application of the proposed method in real-world clinical practice. with the implementation of deep learning, the proposed method can expand the influence of precision medicine to high-dimensional abundant data with greater flexibility and computational power.","key_phrases":["precision medicine","a framework","that","treatment strategies","a patient\u2019s individual characteristics","helpful clinical decision support","existing research","various situations","high-dimensional data","the paradigm","we","a new precision medicine approach","deep doubly robust outcome","ddrowl","that","big and complex data","this","a machine learning tool","that","the optimal decision rule","three worlds","deep learning","double robustness","residual weighted learning","two architectures","the proposed method","a fully-connected feedforward neural network","the deep kernel learning model","a gaussian process","deep learning-filtered inputs","we","the performance","limitation","different methods","a range","simulations","longitudinal and brain imaging data","patients","disease","we","the application","the proposed method","real-world clinical practice","the implementation","deep learning","the proposed method","the influence","precision medicine","high-dimensional abundant data","greater flexibility","computational power","three","two"]},{"title":"Topological deep learning: a review of an emerging paradigm","authors":["Ali Zia","Abdelwahed Khamis","James Nichols","Usman Bashir Tayab","Zeeshan Hayder","Vivien Rolland","Eric Stone","Lars Petersson"],"abstract":"Topological deep learning (TDL) is an emerging area that combines the principles of Topological data analysis (TDA) with deep learning techniques. TDA provides insight into data shape; it obtains global descriptions of multi-dimensional data whilst exhibiting robustness to deformation and noise. Such properties are desirable in deep learning pipelines, but they are typically obtained using non-TDA strategies. This is partly caused by the difficulty of combining TDA constructs (e.g. barcode and persistence diagrams) with current deep learning algorithms. Fortunately, we are now witnessing a growth of deep learning applications embracing topologically-guided components. In this survey, we review the nascent field of topological deep learning by first revisiting the core concepts of TDA. We then explore how the use of TDA techniques has evolved over time to support deep learning frameworks, and how they can be integrated into different aspects of deep learning. Furthermore, we touch on TDA usage for analyzing existing deep models; deep topological analytics. Finally, we discuss the challenges and future prospects of topological deep learning.","doi":"10.1007\/s10462-024-10710-9","cleaned_title":"topological deep learning: a review of an emerging paradigm","cleaned_abstract":"topological deep learning (tdl) is an emerging area that combines the principles of topological data analysis (tda) with deep learning techniques. tda provides insight into data shape; it obtains global descriptions of multi-dimensional data whilst exhibiting robustness to deformation and noise. such properties are desirable in deep learning pipelines, but they are typically obtained using non-tda strategies. this is partly caused by the difficulty of combining tda constructs (e.g. barcode and persistence diagrams) with current deep learning algorithms. fortunately, we are now witnessing a growth of deep learning applications embracing topologically-guided components. in this survey, we review the nascent field of topological deep learning by first revisiting the core concepts of tda. we then explore how the use of tda techniques has evolved over time to support deep learning frameworks, and how they can be integrated into different aspects of deep learning. furthermore, we touch on tda usage for analyzing existing deep models; deep topological analytics. finally, we discuss the challenges and future prospects of topological deep learning.","key_phrases":["topological deep learning","tdl","an emerging area","that","the principles","topological data analysis","tda","deep learning techniques","tda","insight","data shape","it","global descriptions","multi-dimensional data","robustness","deformation","noise","such properties","deep learning pipelines","they","non-tda strategies","this","the difficulty","tda constructs","e.g. barcode and persistence diagrams","current deep learning algorithms","we","a growth","deep learning applications","topologically-guided components","this survey","we","the nascent field","topological deep learning","the core concepts","tda","we","the use","tda techniques","time","deep learning frameworks","they","different aspects","deep learning","we","tda usage","existing deep models","deep topological analytics","we","the challenges","future prospects","topological deep learning","first"]},{"title":"Deep learning in computational mechanics: a review","authors":["Leon Herrmann","Stefan Kollmannsberger"],"abstract":"The rapid growth of deep learning research, including within the field of computational mechanics, has resulted in an extensive and diverse body of literature. To help researchers identify key concepts and promising methodologies within this field, we provide an overview of deep learning in deterministic computational mechanics. Five main categories are identified and explored: simulation substitution, simulation enhancement, discretizations as neural networks, generative approaches, and deep reinforcement learning. This review focuses on deep learning methods rather than applications for computational mechanics, thereby enabling researchers to explore this field more effectively. As such, the review is not necessarily aimed at researchers with extensive knowledge of deep learning\u2014instead, the primary audience is researchers on the verge of entering this field or those attempting to gain an overview of deep learning in computational mechanics. The discussed concepts are, therefore, explained as simple as possible.","doi":"10.1007\/s00466-023-02434-4","cleaned_title":"deep learning in computational mechanics: a review","cleaned_abstract":"the rapid growth of deep learning research, including within the field of computational mechanics, has resulted in an extensive and diverse body of literature. to help researchers identify key concepts and promising methodologies within this field, we provide an overview of deep learning in deterministic computational mechanics. five main categories are identified and explored: simulation substitution, simulation enhancement, discretizations as neural networks, generative approaches, and deep reinforcement learning. this review focuses on deep learning methods rather than applications for computational mechanics, thereby enabling researchers to explore this field more effectively. as such, the review is not necessarily aimed at researchers with extensive knowledge of deep learning\u2014instead, the primary audience is researchers on the verge of entering this field or those attempting to gain an overview of deep learning in computational mechanics. the discussed concepts are, therefore, explained as simple as possible.","key_phrases":["the rapid growth","deep learning research","the field","computational mechanics","an extensive and diverse body","literature","researchers","key concepts","methodologies","this field","we","an overview","deep learning","deterministic computational mechanics","five main categories","simulation substitution","simulation enhancement","discretizations","neural networks","generative approaches","deep reinforcement learning","this review","deep learning methods","applications","computational mechanics","researchers","this field","the review","researchers","extensive knowledge","deep learning","the primary audience","researchers","the verge","this field","those","an overview","deep learning","computational mechanics","the discussed concepts","five"]},{"title":"OpBench: an operator-level GPU benchmark for deep learning","authors":["Qingwen Gu","Bo Fan","Zhengning Liu","Kaicheng Cao","Songhai Zhang","Shimin Hu"],"abstract":"Operators (such as Conv and ReLU) play an important role in deep neural networks. Every neural network is composed of a series of differentiable operators. However, existing AI benchmarks mainly focus on accessing model training and inference performance of deep learning systems on specific models. To help GPU hardware find computing bottlenecks and intuitively evaluate GPU performance on specific deep learning tasks, this paper focuses on evaluating GPU performance at the operator level. We statistically analyze the information of operators on 12 representative deep learning models from six prominent AI tasks and provide an operator dataset to show the different importance of various types of operators in different networks. An operator-level benchmark, OpBench, is proposed on the basis of this dataset, allowing users to choose from a given range of models and set the input sizes according to their demands. This benchmark offers a detailed operator-level performance report for AI and hardware developers. We also evaluate four GPU models on OpBench and find that their performances differ on various types of operators and are not fully consistent with the performance metric FLOPS (floating point operations per second).","doi":"10.1007\/s11432-023-3989-3","cleaned_title":"opbench: an operator-level gpu benchmark for deep learning","cleaned_abstract":"operators (such as conv and relu) play an important role in deep neural networks. every neural network is composed of a series of differentiable operators. however, existing ai benchmarks mainly focus on accessing model training and inference performance of deep learning systems on specific models. to help gpu hardware find computing bottlenecks and intuitively evaluate gpu performance on specific deep learning tasks, this paper focuses on evaluating gpu performance at the operator level. we statistically analyze the information of operators on 12 representative deep learning models from six prominent ai tasks and provide an operator dataset to show the different importance of various types of operators in different networks. an operator-level benchmark, opbench, is proposed on the basis of this dataset, allowing users to choose from a given range of models and set the input sizes according to their demands. this benchmark offers a detailed operator-level performance report for ai and hardware developers. we also evaluate four gpu models on opbench and find that their performances differ on various types of operators and are not fully consistent with the performance metric flops (floating point operations per second).","key_phrases":["operators","relu","an important role","deep neural networks","every neural network","a series","differentiable operators","existing ai benchmarks","model training and inference performance","deep learning systems","specific models","gpu hardware","computing bottlenecks","gpu performance","specific deep learning tasks","this paper","gpu performance","the operator level","we","the information","operators","12 representative deep learning models","six prominent ai tasks","an operator dataset","the different importance","various types","operators","different networks","an operator-level benchmark, opbench","the basis","this dataset","users","a given range","models","the input sizes","their demands","this benchmark","a detailed operator-level performance report","ai and hardware developers","we","four gpu models","opbench","their performances","various types","operators","the performance metric flops","floating point operations","12","six","four","second"]},{"title":"Diabetes detection based on machine learning and deep learning approaches","authors":["Boon Feng Wee","Saaveethya Sivakumar","King Hann Lim","W. K. Wong","Filbert H. Juwono"],"abstract":"The increasing number of diabetes individuals in the globe has alarmed the medical sector to seek alternatives to improve their medical technologies. Machine learning and deep learning approaches are active research in developing intelligent and efficient diabetes detection systems. This study profoundly investigates and discusses the impacts of the latest machine learning and deep learning approaches in diabetes identification\/classifications. It is observed that diabetes data are limited in availability. Available databases comprise lab-based and invasive test measurements. Investigating anthropometric measurements and non-invasive tests must be performed to create a cost-effective yet high-performance solution. Several findings showed the possibility of reconstructing the detection models based on anthropometric measurements and non-invasive medical indicators. This study investigated the consequences of oversampling techniques and data dimensionality reduction through feature selection approaches. The future direction is highlighted in the research of feature selection approaches to improve the accuracy and reliability of diabetes identifications.","doi":"10.1007\/s11042-023-16407-5","cleaned_title":"diabetes detection based on machine learning and deep learning approaches","cleaned_abstract":"the increasing number of diabetes individuals in the globe has alarmed the medical sector to seek alternatives to improve their medical technologies. machine learning and deep learning approaches are active research in developing intelligent and efficient diabetes detection systems. this study profoundly investigates and discusses the impacts of the latest machine learning and deep learning approaches in diabetes identification\/classifications. it is observed that diabetes data are limited in availability. available databases comprise lab-based and invasive test measurements. investigating anthropometric measurements and non-invasive tests must be performed to create a cost-effective yet high-performance solution. several findings showed the possibility of reconstructing the detection models based on anthropometric measurements and non-invasive medical indicators. this study investigated the consequences of oversampling techniques and data dimensionality reduction through feature selection approaches. the future direction is highlighted in the research of feature selection approaches to improve the accuracy and reliability of diabetes identifications.","key_phrases":["the increasing number","diabetes individuals","the globe","the medical sector","alternatives","their medical technologies","machine learning","deep learning approaches","active research","intelligent and efficient diabetes detection systems","this study profoundly investigates","the impacts","the latest machine learning","deep learning approaches","diabetes identification\/classifications","it","diabetes data","availability","available databases","lab-based and invasive test measurements","anthropometric measurements","non-invasive tests","a cost-effective yet high-performance solution","several findings","the possibility","the detection models","anthropometric measurements","non-invasive medical indicators","this study","the consequences","techniques and data dimensionality reduction","feature selection approaches","the future direction","the research","feature selection approaches","the accuracy","reliability","diabetes identifications"]},{"title":"Deep-kidney: an effective deep learning framework for chronic kidney disease prediction","authors":["Dina Saif","Amany M. Sarhan","Nada M. Elshennawy"],"abstract":"Chronic kidney disease (CKD) is one of today\u2019s most serious illnesses. Because this disease usually does not manifest itself until the kidney is severely damaged, early detection saves many people\u2019s lives. Therefore, the contribution of the current paper is proposing three predictive models to predict CKD possible occurrence within 6 or 12 months before disease existence namely; convolutional neural network (CNN), long short-term memory (LSTM) model, and deep ensemble model. The deep ensemble model fuses three base deep learning classifiers (CNN, LSTM, and LSTM-BLSTM) using majority voting technique. To evaluate the performance of the proposed models, several experiments were conducted on two different public datasets. Among the predictive models and the reached results, the deep ensemble model is superior to all the other models, with an accuracy of 0.993 and 0.992 for the 6-month data and 12-month data predictions, respectively.","doi":"10.1007\/s13755-023-00261-8","cleaned_title":"deep-kidney: an effective deep learning framework for chronic kidney disease prediction","cleaned_abstract":"chronic kidney disease (ckd) is one of today\u2019s most serious illnesses. because this disease usually does not manifest itself until the kidney is severely damaged, early detection saves many people\u2019s lives. therefore, the contribution of the current paper is proposing three predictive models to predict ckd possible occurrence within 6 or 12 months before disease existence namely; convolutional neural network (cnn), long short-term memory (lstm) model, and deep ensemble model. the deep ensemble model fuses three base deep learning classifiers (cnn, lstm, and lstm-blstm) using majority voting technique. to evaluate the performance of the proposed models, several experiments were conducted on two different public datasets. among the predictive models and the reached results, the deep ensemble model is superior to all the other models, with an accuracy of 0.993 and 0.992 for the 6-month data and 12-month data predictions, respectively.","key_phrases":["chronic kidney disease","ckd","today\u2019s most serious illnesses","this disease","itself","the kidney","early detection","many people\u2019s lives","the contribution","the current paper","three predictive models","possible occurrence","disease existence","convolutional neural network","cnn","long short-term memory","lstm) model","deep ensemble model","the deep ensemble model","three base deep learning classifiers","cnn","lstm","lstm-blstm","majority voting technique","the performance","the proposed models","several experiments","two different public datasets","the predictive models","the reached results","the deep ensemble model","all the other models","an accuracy","the 6-month data","12-month data predictions","today","three","6 or 12 months","cnn","three","cnn","two","0.993","0.992","6-month","12-month"]},{"title":"Predicting Potato Crop Yield with Machine Learning and Deep Learning for Sustainable Agriculture","authors":["El-Sayed M. El-Kenawy","Amel Ali Alhussan","Nima Khodadadi","Seyedali Mirjalili","Marwa M. Eid"],"abstract":"Potatoes are an important crop in the world; they are the main source of food for a large number of people globally and also provide an income for many people. The true forecasting of potato yields is a determining factor for the rational use and maximization of agricultural practices, responsible management of the resources, and wider regions\u2019 food security. The latest discoveries in machine learning and deep learning provide new directions to yield prediction models more accurately and sparingly. From the study, we evaluated different types of predictive models, including K-nearest neighbors (KNN), gradient boosting, XGBoost, and multilayer perceptron that use machine learning, as well as graph neural networks (GNNs), gated recurrent units (GRUs), and long short-term memory networks (LSTM), which are popular in deep learning models. These models are evaluated on the basis of some performance measures like mean squared error (MSE), root mean squared error (RMSE), and mean absolute error (MAE) to know how much they accurately predict the potato yields. The terminal results show that although gradient boosting and XGBoost algorithms are good at potato yield prediction, GNNs and LSTMs not only have the advantage of high accuracy but also capture the complex spatial and temporal patterns in the data. Gradient boosting resulted in an MSE of 0.03438 and an R2 of 0.49168, while XGBoost had an MSE of 0.03583 and an R2 of 0.35106. Out of all deep learning models, GNNs displayed an MSE of 0.02363 and an R2 of 0.51719, excelling in the overall performance. LSTMs and GRUs were reported to be very promising as well, with LSTMs comprehending an MSE of 0.03177 and GRUs grabbing an MSE of 0.03150. These findings underscore the potential of advanced predictive models to support sustainable agricultural practices and informed decision-making in the context of potato farming.","doi":"10.1007\/s11540-024-09753-w","cleaned_title":"predicting potato crop yield with machine learning and deep learning for sustainable agriculture","cleaned_abstract":"potatoes are an important crop in the world; they are the main source of food for a large number of people globally and also provide an income for many people. the true forecasting of potato yields is a determining factor for the rational use and maximization of agricultural practices, responsible management of the resources, and wider regions\u2019 food security. the latest discoveries in machine learning and deep learning provide new directions to yield prediction models more accurately and sparingly. from the study, we evaluated different types of predictive models, including k-nearest neighbors (knn), gradient boosting, xgboost, and multilayer perceptron that use machine learning, as well as graph neural networks (gnns), gated recurrent units (grus), and long short-term memory networks (lstm), which are popular in deep learning models. these models are evaluated on the basis of some performance measures like mean squared error (mse), root mean squared error (rmse), and mean absolute error (mae) to know how much they accurately predict the potato yields. the terminal results show that although gradient boosting and xgboost algorithms are good at potato yield prediction, gnns and lstms not only have the advantage of high accuracy but also capture the complex spatial and temporal patterns in the data. gradient boosting resulted in an mse of 0.03438 and an r2 of 0.49168, while xgboost had an mse of 0.03583 and an r2 of 0.35106. out of all deep learning models, gnns displayed an mse of 0.02363 and an r2 of 0.51719, excelling in the overall performance. lstms and grus were reported to be very promising as well, with lstms comprehending an mse of 0.03177 and grus grabbing an mse of 0.03150. these findings underscore the potential of advanced predictive models to support sustainable agricultural practices and informed decision-making in the context of potato farming.","key_phrases":["potatoes","an important crop","the world","they","the main source","food","a large number","people","an income","many people","the true forecasting","potato yields","a determining factor","the rational use","maximization","agricultural practices","responsible management","the resources","wider regions\u2019 food security","the latest discoveries","machine learning","deep learning","new directions","prediction models","the study","we","different types","predictive models","k-nearest neighbors","gradient boosting","xgboost","multilayer perceptron","that","machine learning","graph neural networks","gnns","grus","lstm","which","deep learning models","these models","the basis","some performance measures","mean squared error","mse","root mean squared error","rmse","absolute error","mae","they","the potato yields","the terminal results","gradient boosting","xgboost algorithms","potato yield prediction","gnns","the advantage","high accuracy","the complex spatial and temporal patterns","the data","an mse","an r2","xgboost","an mse","an r2","all deep learning models","gnns","an mse","an r2","the overall performance","grus","an mse","an mse","these findings","the potential","advanced predictive models","sustainable agricultural practices","informed decision-making","the context","potato farming","0.03438","0.49168","0.03583","0.02363","0.51719","0.03177","0.03150"]},{"title":"Prediction of glycopeptide fragment mass spectra by deep learning","authors":["Yi Yang","Qun Fang"],"abstract":"Deep learning has achieved a notable success in mass spectrometry-based proteomics and is now emerging in glycoproteomics. While various deep learning models can predict fragment mass spectra of peptides with good accuracy, they cannot cope with the non-linear glycan structure in an intact glycopeptide. Herein, we present DeepGlyco, a deep learning-based approach for the prediction of fragment spectra of intact glycopeptides. Our model adopts tree-structured long-short term memory networks to process the glycan moiety and a graph neural network architecture to incorporate potential fragmentation pathways of a specific glycan structure. This feature is beneficial to model explainability and differentiation ability of glycan structural isomers. We further demonstrate that predicted spectral libraries can be used for data-independent acquisition glycoproteomics as a supplement for library completeness. We expect that this work will provide a valuable deep learning resource for glycoproteomics.","doi":"10.1038\/s41467-024-46771-1","cleaned_title":"prediction of glycopeptide fragment mass spectra by deep learning","cleaned_abstract":"deep learning has achieved a notable success in mass spectrometry-based proteomics and is now emerging in glycoproteomics. while various deep learning models can predict fragment mass spectra of peptides with good accuracy, they cannot cope with the non-linear glycan structure in an intact glycopeptide. herein, we present deepglyco, a deep learning-based approach for the prediction of fragment spectra of intact glycopeptides. our model adopts tree-structured long-short term memory networks to process the glycan moiety and a graph neural network architecture to incorporate potential fragmentation pathways of a specific glycan structure. this feature is beneficial to model explainability and differentiation ability of glycan structural isomers. we further demonstrate that predicted spectral libraries can be used for data-independent acquisition glycoproteomics as a supplement for library completeness. we expect that this work will provide a valuable deep learning resource for glycoproteomics.","key_phrases":["deep learning","a notable success","mass spectrometry-based proteomics","glycoproteomics","various deep learning models","fragment mass spectra","peptides","good accuracy","they","the non-linear glycan structure","an intact glycopeptide","we","deepglyco","a deep learning-based approach","the prediction","fragment spectra","intact glycopeptides","our model","tree-structured long-short term memory networks","the glycan moiety","a graph neural network architecture","potential fragmentation pathways","a specific glycan structure","this feature","model explainability","differentiation ability","glycan structural isomers","we","spectral libraries","data-independent acquisition glycoproteomics","a supplement","library completeness","we","this work","a valuable deep learning resource","glycoproteomics"]},{"title":"Deep learning application in diagnosing breast cancer recurrence","authors":["Zeinab Jam","Amir Albadvi","Alireza Atashi"],"abstract":"Patients' lives can always be saved when diseases, especially special diseases, are detected early. The chances of a patient surviving can be increased by early detection. Breast cancer is one of the deadliest and common cancers. After recovering from breast cancer, patients are always worried about recurrence and return. The use of modern technology, however, can help predict disease recurrence at an early stage, allowing patients to receive treatment sooner.Significant strides have been achieved in deep learning, demonstrating strong performance in handling unstructured data challenges. However, when it comes to predicting tabular data, deep learning hasn't quite matched its success with unstructured data. Presently, ensemble models relying on gradient-boosted decision trees (GBDT) are frequently favored for tabular data prediction tasks. Typically, these GBDT-based models outshine deep learning approaches.Many novel deep learning techniques are emerging for handling tabular data. TabNet, for instance, mirrors decision tree feature selection within a neural network framework. AutoInt addresses high dimensionality by condensing data through embedding layers. Tab Transformer adapts the transformer model, generating text representations for categorical attributes. Despite their innovation, these methods remain less recognized compared to those for image and text data processing.In this study, 158 different characteristics of 5142 breast cancer patients from 1997 to 2019 were examined. We aim to evaluate deep learning techniques effectiveness in detecting breast cancer recurrence. Through examination of evaluation metrics, it becomes evident that deep learning approaches applied to tabular data surpass traditional machine learning algorithms, even when dealing with imbalanced datasets. Ultimately, the results derived from each algorithm analyzed and concluded with a review and comparison of the findings.","doi":"10.1007\/s11042-024-19423-1","cleaned_title":"deep learning application in diagnosing breast cancer recurrence","cleaned_abstract":"patients' lives can always be saved when diseases, especially special diseases, are detected early. the chances of a patient surviving can be increased by early detection. breast cancer is one of the deadliest and common cancers. after recovering from breast cancer, patients are always worried about recurrence and return. the use of modern technology, however, can help predict disease recurrence at an early stage, allowing patients to receive treatment sooner.significant strides have been achieved in deep learning, demonstrating strong performance in handling unstructured data challenges. however, when it comes to predicting tabular data, deep learning hasn't quite matched its success with unstructured data. presently, ensemble models relying on gradient-boosted decision trees (gbdt) are frequently favored for tabular data prediction tasks. typically, these gbdt-based models outshine deep learning approaches.many novel deep learning techniques are emerging for handling tabular data. tabnet, for instance, mirrors decision tree feature selection within a neural network framework. autoint addresses high dimensionality by condensing data through embedding layers. tab transformer adapts the transformer model, generating text representations for categorical attributes. despite their innovation, these methods remain less recognized compared to those for image and text data processing.in this study, 158 different characteristics of 5142 breast cancer patients from 1997 to 2019 were examined. we aim to evaluate deep learning techniques effectiveness in detecting breast cancer recurrence. through examination of evaluation metrics, it becomes evident that deep learning approaches applied to tabular data surpass traditional machine learning algorithms, even when dealing with imbalanced datasets. ultimately, the results derived from each algorithm analyzed and concluded with a review and comparison of the findings.","key_phrases":["patients' lives","diseases","especially special diseases","the chances","a patient surviving","early detection","breast cancer","the deadliest and common cancers","breast cancer","patients","recurrence","return","the use","modern technology","disease recurrence","an early stage","patients","treatment sooner.significant strides","deep learning","strong performance","unstructured data challenges","it","tabular data","deep learning","its success","unstructured data","ensemble models","gradient-boosted decision trees","gbdt","tabular data prediction tasks","these gbdt-based models","approaches.many novel deep learning techniques","tabular data","tabnet","instance","mirrors decision tree","a neural network framework","autoint addresses high dimensionality","data","embedding layers","tab transformer","the transformer model","text representations","categorical attributes","their innovation","these methods","those","image","text data","158 different characteristics","5142 breast cancer patients","we","deep learning techniques effectiveness","breast cancer recurrence","examination","evaluation metrics","it","deep learning approaches","data surpass traditional machine learning algorithms","imbalanced datasets","the results","each algorithm","a review","comparison","the findings","one","sooner.significant","autoint","158","5142","from 1997 to 2019"]},{"title":"CNN-Transformer: A deep learning method for automatically identifying learning engagement","authors":["Yan Xiong","Guo Xinya","Junjie Xu"],"abstract":"Learning engagement is an essential indication to define students' learning pacification in the class, and its automated identification technique is the foundation for exploring how to effectively explain the motive of learning impact modifications and making intelligent teaching choices. Current research have demonstrated that there is a direct link between learning engagement and emotional investment and behavioural investment, and it is appropriate and required to apply artificial intelligence to perform autonomous assessment. Unfortunately, the number of relevant research is limited, and the features of learning engagement in certain contexts have not been thoroughly examined. In this research, we highlight the features of a particular application scenario of learning engagement: the application scenario of learning engagement has to incorporate both the coarse-grained information of human body position and the fine-grained information of facial expressions. On the basis of this analysis, a fine-grained learning participation recognition model that suppresses background clutter information is presented. This model can effectively extract coarse and fine-grained information to improve the recognition of learning participation in real-world teaching situations. Particularly, the CNN-Transformer model suggested in this study employs CNN to extract fine-grained information of facial expressions and Transformer to recover coarse-grained information of human body position. Simultaneously, we gathered and categorised real teaching data based on the features of learning engagement situations and enhanced the data quality via crowdsourcing and expert verification. The experimental findings indicate that the CNN-Transformer model can accurately predict the learning engagement of unknown participants with a 92.9% rate of accuracy. Comparative trials reveal that the model's recognition impact is much greater than that of other sophisticated deep learning approaches. Our research offers a framework for future work on deep learning approaches in learning engagement settings.","doi":"10.1007\/s10639-023-12058-z","cleaned_title":"cnn-transformer: a deep learning method for automatically identifying learning engagement","cleaned_abstract":"learning engagement is an essential indication to define students' learning pacification in the class, and its automated identification technique is the foundation for exploring how to effectively explain the motive of learning impact modifications and making intelligent teaching choices. current research have demonstrated that there is a direct link between learning engagement and emotional investment and behavioural investment, and it is appropriate and required to apply artificial intelligence to perform autonomous assessment. unfortunately, the number of relevant research is limited, and the features of learning engagement in certain contexts have not been thoroughly examined. in this research, we highlight the features of a particular application scenario of learning engagement: the application scenario of learning engagement has to incorporate both the coarse-grained information of human body position and the fine-grained information of facial expressions. on the basis of this analysis, a fine-grained learning participation recognition model that suppresses background clutter information is presented. this model can effectively extract coarse and fine-grained information to improve the recognition of learning participation in real-world teaching situations. particularly, the cnn-transformer model suggested in this study employs cnn to extract fine-grained information of facial expressions and transformer to recover coarse-grained information of human body position. simultaneously, we gathered and categorised real teaching data based on the features of learning engagement situations and enhanced the data quality via crowdsourcing and expert verification. the experimental findings indicate that the cnn-transformer model can accurately predict the learning engagement of unknown participants with a 92.9% rate of accuracy. comparative trials reveal that the model's recognition impact is much greater than that of other sophisticated deep learning approaches. our research offers a framework for future work on deep learning approaches in learning engagement settings.","key_phrases":["learning engagement","an essential indication","students' learning pacification","the class","its automated identification technique","the foundation","the motive","impact modifications","intelligent teaching choices","current research","a direct link","engagement","emotional investment","behavioural investment","it","artificial intelligence","autonomous assessment","the number","relevant research","the features","engagement","certain contexts","this research","we","the features","a particular application scenario","learning engagement","the application scenario","learning engagement","both the coarse-grained information","human body position","the fine-grained information","facial expressions","the basis","this analysis","a fine-grained learning participation recognition model","that","background clutter information","this model","coarse and fine-grained information","the recognition","participation","real-world teaching situations","the cnn-transformer model","this study","cnn","fine-grained information","facial expressions","transformer","coarse-grained information","human body position","we","real teaching data","the features","engagement situations","the data quality","crowdsourcing and expert verification","the experimental findings","the cnn-transformer model","the learning engagement","unknown participants","a 92.9% rate","accuracy","comparative trials","the model's recognition impact","that","other sophisticated deep learning approaches","our research","a framework","future work","deep learning approaches","engagement settings","cnn","cnn","cnn","92.9%"]},{"title":"Deep learning based features extraction for facial gender classification using ensemble of machine learning technique","authors":["Fazal Waris","Feipeng Da","Shanghuan Liu"],"abstract":"Accurate and efficient gender recognition is an essential for many applications such as surveillance, security, and biometrics. Recently, deep learning techniques have made remarkable advancements in feature extraction and have become extensively implemented in various applications, including gender classification. However, despite the numerous studies conducted on the problem, correctly recognizing robust and essential features from face images and efficiently distinguishing them with high accuracy in the wild is still a challenging task for real-world applications. This article proposes an approach that combines deep learning and soft voting-based ensemble model to perform automatic gender classification with high accuracy in an unconstrained environment. In the proposed technique, a novel deep convolutional neural network (DCNN) was designed to extract 128 high-quality and accurate features from face images. The StandardScaler method was then used to pre-process these extracted features, and finally, these preprocessed features were classified with soft voting ensemble learning model combining the outputs from several machine learning classifiers such as random forest (RF), support vector machine (SVM), linear discriminant analysis (LDA), logistic regression (LR), gradient boosting classifier (GBC) and XGBoost to improve the prediction accuracy. The experimental study was performed on the UTK, label faces in the wild (LFW), Adience and FEI datasets. The results attained evidently show that the proposed approach outperforms all current approaches in terms of accuracy across all datasets.","doi":"10.1007\/s00530-024-01399-5","cleaned_title":"deep learning based features extraction for facial gender classification using ensemble of machine learning technique","cleaned_abstract":"accurate and efficient gender recognition is an essential for many applications such as surveillance, security, and biometrics. recently, deep learning techniques have made remarkable advancements in feature extraction and have become extensively implemented in various applications, including gender classification. however, despite the numerous studies conducted on the problem, correctly recognizing robust and essential features from face images and efficiently distinguishing them with high accuracy in the wild is still a challenging task for real-world applications. this article proposes an approach that combines deep learning and soft voting-based ensemble model to perform automatic gender classification with high accuracy in an unconstrained environment. in the proposed technique, a novel deep convolutional neural network (dcnn) was designed to extract 128 high-quality and accurate features from face images. the standardscaler method was then used to pre-process these extracted features, and finally, these preprocessed features were classified with soft voting ensemble learning model combining the outputs from several machine learning classifiers such as random forest (rf), support vector machine (svm), linear discriminant analysis (lda), logistic regression (lr), gradient boosting classifier (gbc) and xgboost to improve the prediction accuracy. the experimental study was performed on the utk, label faces in the wild (lfw), adience and fei datasets. the results attained evidently show that the proposed approach outperforms all current approaches in terms of accuracy across all datasets.","key_phrases":["accurate and efficient gender recognition","many applications","surveillance","security","biometrics","deep learning techniques","remarkable advancements","feature extraction","various applications","gender classification","the numerous studies","the problem","robust and essential features","face images","them","high accuracy","the wild","a challenging task","real-world applications","this article","an approach","that","deep learning","soft voting-based ensemble model","automatic gender classification","high accuracy","an unconstrained environment","the proposed technique","a novel deep convolutional neural network","dcnn","128 high-quality and accurate features","face images","the standardscaler method","these extracted features","these preprocessed features","soft voting ensemble learning model","the outputs","several machine learning classifiers","random forest","rf","vector machine","svm","linear","discriminant analysis","logistic regression","lr","gradient","classifier","gbc","xgboost","the prediction accuracy","the experimental study","the utk","the results","the proposed approach","all current approaches","terms","accuracy","all datasets","128","linear","gbc"]},{"title":"Ensemble of Deep Learning Architectures with Machine Learning for Pneumonia Classification Using Chest X-rays","authors":["Rupali Vyas","Deepak Rao Khadatkar"],"abstract":"Pneumonia is a severe health concern, particularly for vulnerable groups, needing early and correct classification for optimal treatment. This study addresses the use of deep learning combined with machine learning classifiers (DLxMLCs) for pneumonia classification from chest X-ray (CXR) images. We deployed modified VGG19, ResNet50V2, and DenseNet121 models for feature extraction, followed by five machine learning classifiers (logistic regression, support vector machine, decision tree, random forest, artificial neural network). The approach we suggested displayed remarkable accuracy, with VGG19 and DenseNet121 models obtaining 99.98% accuracy when combined with random forest or decision tree classifiers. ResNet50V2 achieved 99.25% accuracy with random forest. These results illustrate the advantages of merging deep learning models with machine learning classifiers in boosting the speedy and accurate identification of pneumonia. The study underlines the potential of DLxMLC systems in enhancing diagnostic accuracy and efficiency. By integrating these models into clinical practice, healthcare practitioners could greatly boost patient care and results. Future research should focus on refining these models and exploring their application to other medical imaging tasks, as well as including explainability methodologies to better understand their decision-making processes and build trust in their clinical use. This technique promises promising breakthroughs in medical imaging and patient management.","doi":"10.1007\/s10278-024-01201-y","cleaned_title":"ensemble of deep learning architectures with machine learning for pneumonia classification using chest x-rays","cleaned_abstract":"pneumonia is a severe health concern, particularly for vulnerable groups, needing early and correct classification for optimal treatment. this study addresses the use of deep learning combined with machine learning classifiers (dlxmlcs) for pneumonia classification from chest x-ray (cxr) images. we deployed modified vgg19, resnet50v2, and densenet121 models for feature extraction, followed by five machine learning classifiers (logistic regression, support vector machine, decision tree, random forest, artificial neural network). the approach we suggested displayed remarkable accuracy, with vgg19 and densenet121 models obtaining 99.98% accuracy when combined with random forest or decision tree classifiers. resnet50v2 achieved 99.25% accuracy with random forest. these results illustrate the advantages of merging deep learning models with machine learning classifiers in boosting the speedy and accurate identification of pneumonia. the study underlines the potential of dlxmlc systems in enhancing diagnostic accuracy and efficiency. by integrating these models into clinical practice, healthcare practitioners could greatly boost patient care and results. future research should focus on refining these models and exploring their application to other medical imaging tasks, as well as including explainability methodologies to better understand their decision-making processes and build trust in their clinical use. this technique promises promising breakthroughs in medical imaging and patient management.","key_phrases":["pneumonia","a severe health concern","vulnerable groups","early and correct classification","optimal treatment","this study","the use","deep learning","machine learning classifiers","dlxmlcs","pneumonia classification","chest x","cxr","we","modified vgg19","densenet121 models","feature extraction","five machine learning classifiers","logistic regression","vector machine","decision tree","random forest","artificial neural network","the approach","we","remarkable accuracy","vgg19 and densenet121 models","99.98% accuracy","random forest","decision tree classifiers","resnet50v2","99.25% accuracy","random forest","these results","the advantages","deep learning models","machine learning classifiers","the speedy and accurate identification","pneumonia","the study","the potential","dlxmlc systems","diagnostic accuracy","efficiency","these models","clinical practice","healthcare practitioners","patient care","results","future research","these models","their application","other medical imaging tasks","explainability methodologies","their decision-making processes","trust","their clinical use","this technique","breakthroughs","medical imaging and patient management","five","99.98%","99.25%"]},{"title":"Instance segmentation on distributed deep learning big data cluster","authors":["Mohammed Elhmadany","Islam Elmadah","Hossam E. Abdelmunim"],"abstract":"Distributed deep learning is a promising approach for training and deploying large and complex deep learning models. This paper presents a comprehensive workflow for deploying and optimizing the YOLACT instance segmentation model as on big data clusters. OpenVINO, a toolkit known for its high-speed data processing and ability to optimize deep learning models for deployment on a variety of devices, was used to optimize the YOLACT model. The model is then run on a big data cluster using BigDL, a distributed deep learning library for Apache Spark. BigDL provides a high-level programming interface for defining and training deep neural networks, making it suitable for large-scale deep learning applications. In distributed deep learning, input data is divided and distributed across multiple machines for parallel processing. This approach offers several advantages, including the ability to handle very large data that can be stored in a distributed manner, scalability to decrease processing time by increasing the number of workers, and fault tolerance. The proposed workflow was evaluated on virtual machines and Azure Databricks, a cloud-based platform for big data analytics. The results indicated that the workflow can scale to large datasets and deliver high performance on Azure Databricks. This study explores the benefits and challenges of using distributed deep learning on big data clusters for instance segmentation. Popular distributed deep learning frameworks are discussed, and BigDL is chosen. Overall, this study highlights the practicality of distributed deep learning for deploying and scaling sophisticated deep learning models on big data clusters.","doi":"10.1186\/s40537-023-00871-9","cleaned_title":"instance segmentation on distributed deep learning big data cluster","cleaned_abstract":"distributed deep learning is a promising approach for training and deploying large and complex deep learning models. this paper presents a comprehensive workflow for deploying and optimizing the yolact instance segmentation model as on big data clusters. openvino, a toolkit known for its high-speed data processing and ability to optimize deep learning models for deployment on a variety of devices, was used to optimize the yolact model. the model is then run on a big data cluster using bigdl, a distributed deep learning library for apache spark. bigdl provides a high-level programming interface for defining and training deep neural networks, making it suitable for large-scale deep learning applications. in distributed deep learning, input data is divided and distributed across multiple machines for parallel processing. this approach offers several advantages, including the ability to handle very large data that can be stored in a distributed manner, scalability to decrease processing time by increasing the number of workers, and fault tolerance. the proposed workflow was evaluated on virtual machines and azure databricks, a cloud-based platform for big data analytics. the results indicated that the workflow can scale to large datasets and deliver high performance on azure databricks. this study explores the benefits and challenges of using distributed deep learning on big data clusters for instance segmentation. popular distributed deep learning frameworks are discussed, and bigdl is chosen. overall, this study highlights the practicality of distributed deep learning for deploying and scaling sophisticated deep learning models on big data clusters.","key_phrases":["distributed deep learning","a promising approach","training","large and complex deep learning models","this paper","a comprehensive workflow","the yolact instance segmentation model","big data clusters","a toolkit","its high-speed data processing","ability","deep learning models","deployment","a variety","devices","the yolact model","the model","a big data cluster","bigdl","a distributed deep learning library","apache spark","a high-level programming interface","deep neural networks","it","large-scale deep learning applications","distributed deep learning","input data","multiple machines","parallel processing","this approach","several advantages","the ability","very large data","that","a distributed manner","scalability","processing time","the number","workers","tolerance","the proposed workflow","virtual machines","azure databricks","a cloud-based platform","big data analytics","the results","the workflow","large datasets","high performance","azure databricks","this study","the benefits","challenges","distributed deep learning","big data clusters","instance segmentation","popular distributed deep learning frameworks","bigdl","this study","the practicality","distributed deep learning","sophisticated deep learning models","big data clusters"]},{"title":"Deep learning-based personalized learning recommendation system design for \"T++\" Guzheng Pedagogy","authors":["Xingyue Wang"],"abstract":"This study investigates the development and impact of a deep learning-based personalized learning recommendation system designed specifically for 'T++' Guzheng pedagogy. In the realm of music education, particularly in the context of the traditional Chinese Guzheng instrument, technology-driven personalization has the potential to revolutionize learning experiences. The research involves data collection, algorithm development, and integration to create a system that tailors Guzheng learning materials to individual students' skill levels and preferences. Results indicate high recommendation accuracy, increased user satisfaction, and positive engagement rates. This study contributes to the intersection of technology, music education, and cultural preservation, showcasing the promise of personalized learning in the realm of Guzheng pedagogy.","doi":"10.1007\/s41870-024-01871-5","cleaned_title":"deep learning-based personalized learning recommendation system design for \"t++\" guzheng pedagogy","cleaned_abstract":"this study investigates the development and impact of a deep learning-based personalized learning recommendation system designed specifically for 't++' guzheng pedagogy. in the realm of music education, particularly in the context of the traditional chinese guzheng instrument, technology-driven personalization has the potential to revolutionize learning experiences. the research involves data collection, algorithm development, and integration to create a system that tailors guzheng learning materials to individual students' skill levels and preferences. results indicate high recommendation accuracy, increased user satisfaction, and positive engagement rates. this study contributes to the intersection of technology, music education, and cultural preservation, showcasing the promise of personalized learning in the realm of guzheng pedagogy.","key_phrases":["this study","the development","impact","a deep learning-based personalized learning recommendation system","t++' guzheng pedagogy","the realm","music education","the context","the traditional chinese guzheng instrument, technology-driven personalization","the potential","learning experiences","the research","data collection","algorithm development","integration","a system","that","guzheng learning materials","individual students' skill levels","preferences","results","high recommendation accuracy","increased user satisfaction","positive engagement rates","this study","the intersection","technology","music education","cultural preservation","the promise","personalized learning","the realm","guzheng pedagogy","chinese","guzheng","guzheng"]},{"title":"Deep learning in two-dimensional materials: Characterization, prediction, and design","authors":["Xinqin Meng","Chengbing Qin","Xilong Liang","Guofeng Zhang","Ruiyun Chen","Jianyong Hu","Zhichun Yang","Jianzhong Huo","Liantuan Xiao","Suotang Jia"],"abstract":"Since the isolation of graphene, two-dimensional (2D) materials have attracted increasing interest because of their excellent chemical and physical properties, as well as promising applications. Nonetheless, particular challenges persist in their further development, particularly in the effective identification of diverse 2D materials, the domains of large-scale and high-precision characterization, also intelligent function prediction and design. These issues are mainly solved by computational techniques, such as density function theory and molecular dynamic simulation, which require powerful computational resources and high time consumption. The booming deep learning methods in recent years offer innovative insights and tools to address these challenges. This review comprehensively outlines the current progress of deep learning within the realm of 2D materials. Firstly, we will briefly introduce the basic concepts of deep learning and commonly used architectures, including convolutional neural and generative adversarial networks, as well as U-net models. Then, the characterization of 2D materials by deep learning methods will be discussed, including defects and materials identification, as well as automatic thickness characterization. Thirdly, the research progress for predicting the unique properties of 2D materials, involving electronic, mechanical, and thermodynamic features, will be evaluated succinctly. Lately, the current works on the inverse design of functional 2D materials will be presented. At last, we will look forward to the application prospects and opportunities of deep learning in other aspects of 2D materials. This review may offer some guidance to boost the understanding and employing novel 2D materials.","doi":"10.1007\/s11467-024-1394-7","cleaned_title":"deep learning in two-dimensional materials: characterization, prediction, and design","cleaned_abstract":"since the isolation of graphene, two-dimensional (2d) materials have attracted increasing interest because of their excellent chemical and physical properties, as well as promising applications. nonetheless, particular challenges persist in their further development, particularly in the effective identification of diverse 2d materials, the domains of large-scale and high-precision characterization, also intelligent function prediction and design. these issues are mainly solved by computational techniques, such as density function theory and molecular dynamic simulation, which require powerful computational resources and high time consumption. the booming deep learning methods in recent years offer innovative insights and tools to address these challenges. this review comprehensively outlines the current progress of deep learning within the realm of 2d materials. firstly, we will briefly introduce the basic concepts of deep learning and commonly used architectures, including convolutional neural and generative adversarial networks, as well as u-net models. then, the characterization of 2d materials by deep learning methods will be discussed, including defects and materials identification, as well as automatic thickness characterization. thirdly, the research progress for predicting the unique properties of 2d materials, involving electronic, mechanical, and thermodynamic features, will be evaluated succinctly. lately, the current works on the inverse design of functional 2d materials will be presented. at last, we will look forward to the application prospects and opportunities of deep learning in other aspects of 2d materials. this review may offer some guidance to boost the understanding and employing novel 2d materials.","key_phrases":["the isolation","graphene","two-dimensional (2d) materials","increasing interest","their excellent chemical and physical properties","promising applications","particular challenges","their further development","the effective identification","diverse 2d materials","the domains","large-scale and high-precision characterization","also intelligent function prediction","design","these issues","computational techniques","density function theory","molecular dynamic simulation","which","powerful computational resources","high time consumption","the booming deep learning methods","recent years","innovative insights","tools","these challenges","this review","the current progress","deep learning","the realm","2d materials","we","the basic concepts","deep learning","commonly used architectures","convolutional neural and generative adversarial networks","u-net models","the characterization","2d materials","deep learning methods","defects and materials identification","automatic thickness characterization","the research progress","the unique properties","2d materials","electronic, mechanical, and thermodynamic features","the inverse design","functional 2d materials","we","the application prospects","opportunities","deep learning","other aspects","2d materials","this review","some guidance","the understanding and employing novel 2d materials","two","2d","2d","recent years","2d","firstly","2d","thirdly","2d","2d","2d","2d"]},{"title":"Revitalizing Arabic Character Classification: Unleashing the Power of Deep Learning with Transfer Learning and Data Augmentation Techniques","authors":["Marwa Amara","Nadia  Smairi","Sami  Mnasri","Abdelmalek  Zidouri"],"abstract":"Deep learning techniques have demonstrated remarkable success in various domains, including character classification tasks. However, the performance of deep learning models heavily relies on the availability of large-annotated datasets. This research work is motivated by the need to overcome the difficulties associated with handwritten Arabic character recognition, and the constraints provided by limited training data. It is also motivated by the need to enhance the model\u2019s generalizability to unknown characteristics and to improve the accuracy of deep learning character classification models. To overcome this limitation, we apply transfer learning and data augmentation strategies to improve the character classification using deep learning. The proposed model transfers knowledge from previously trained models via transfer learning, addresses data scarcity, and reflects generalizable properties. Indeed, we utilize a VGG16-ImageNet transfer learning model, which is systematically enhanced through data augmentation across three distinct models: pre-trained ImageNet weights with a frozen backbone, pre-trained ImageNet weights with a fine-tuned backbone, and initiating with a randomly initialized. In each case, data augmentation plays a critical role. Our experimental results show that better precision and recall values were recorded for most classes in our dataset, which indicates the model\u2019s ability to accurately identify instances of each character. Moreover, when applying our method to the IFHCDB and HACDB datasets, we observed an impressive recognition accuracy of 96.01% and 97.15%, respectively. This clearly indicates that involving transfer learning and data augmentation significantly improves the performance of deep learning models, especially for small size training datasets.","doi":"10.1007\/s13369-024-08818-9","cleaned_title":"revitalizing arabic character classification: unleashing the power of deep learning with transfer learning and data augmentation techniques","cleaned_abstract":"deep learning techniques have demonstrated remarkable success in various domains, including character classification tasks. however, the performance of deep learning models heavily relies on the availability of large-annotated datasets. this research work is motivated by the need to overcome the difficulties associated with handwritten arabic character recognition, and the constraints provided by limited training data. it is also motivated by the need to enhance the model\u2019s generalizability to unknown characteristics and to improve the accuracy of deep learning character classification models. to overcome this limitation, we apply transfer learning and data augmentation strategies to improve the character classification using deep learning. the proposed model transfers knowledge from previously trained models via transfer learning, addresses data scarcity, and reflects generalizable properties. indeed, we utilize a vgg16-imagenet transfer learning model, which is systematically enhanced through data augmentation across three distinct models: pre-trained imagenet weights with a frozen backbone, pre-trained imagenet weights with a fine-tuned backbone, and initiating with a randomly initialized. in each case, data augmentation plays a critical role. our experimental results show that better precision and recall values were recorded for most classes in our dataset, which indicates the model\u2019s ability to accurately identify instances of each character. moreover, when applying our method to the ifhcdb and hacdb datasets, we observed an impressive recognition accuracy of 96.01% and 97.15%, respectively. this clearly indicates that involving transfer learning and data augmentation significantly improves the performance of deep learning models, especially for small size training datasets.","key_phrases":["deep learning techniques","remarkable success","various domains","character classification tasks","the performance","deep learning models","the availability","large-annotated datasets","this research work","the need","the difficulties","handwritten arabic character recognition","the constraints","limited training data","it","the need","the model\u2019s generalizability","unknown characteristics","the accuracy","deep learning character classification models","this limitation","we","transfer learning and data augmentation strategies","the character classification","deep learning","the proposed model transfers","previously trained models","transfer learning","addresses data scarcity","generalizable properties","we","a vgg16-imagenet transfer learning model","which","data augmentation","three distinct models","pre-trained imagenet weights","a frozen backbone","pre-trained imagenet weights","a fine-tuned backbone","each case","data augmentation","a critical role","our experimental results","better precision","recall values","most classes","our dataset","which","the model\u2019s ability","instances","each character","our method","the ifhcdb","hacdb datasets","we","an impressive recognition accuracy","96.01%","97.15%","this","transfer learning and data augmentation","the performance","deep learning models","small size training datasets","recognition","three","96.01%","97.15%"]},{"title":"A systematic survey of fuzzy deep learning for uncertain medical data","authors":["Yuanhang Zheng","Zeshui Xu","Tong Wu","Zhang Yi"],"abstract":"Intelligent medical industry is in a rapid stage of development around the world, followed by are the expanding market size and basic theories of intelligent medical diagnosis and decision-making. Deep learning models have achieved good practical results in medical domain. However, traditional deep learning is almost calculated and developed by crisp values, while imprecise, uncertain, and vague medical data is common in the process of diagnosis and treatment. It is important and significant to review the contributions of fuzzy deep learning for uncertain medical data, because fuzzy deep learning that originated from fuzzy sets, can effectively deal with uncertain and inaccurate information, providing new viewpoints for alleviating the presence of noise, artifact or high dimensional unstructured information in uncertain medical data. Therefore, taking focus on the intersection of both different fuzzy deep learning models and several types of uncertain medical data, the paper first constructs four types of frameworks of fuzzy deep learning models used for uncertain medical data, and investigates the status from three aspects: fuzzy deep learning models, uncertain medical data and application scenarios. Then the performance evaluation metrics of fuzzy deep learning models are analyzed in details. This work has some original points: (1) four types of frameworks of applying fuzzy deep learning models for uncertain medical data are first proposed. (2) Seven fuzzy deep learning models, five types of uncertain medical data, and five application scenarios are reviewed in details, respectively. (3) The advantages, challenges, and future research directions of fuzzy deep learning for uncertain medical data are critically analyzed, providing valuable suggestions for further deep research.","doi":"10.1007\/s10462-024-10871-7","cleaned_title":"a systematic survey of fuzzy deep learning for uncertain medical data","cleaned_abstract":"intelligent medical industry is in a rapid stage of development around the world, followed by are the expanding market size and basic theories of intelligent medical diagnosis and decision-making. deep learning models have achieved good practical results in medical domain. however, traditional deep learning is almost calculated and developed by crisp values, while imprecise, uncertain, and vague medical data is common in the process of diagnosis and treatment. it is important and significant to review the contributions of fuzzy deep learning for uncertain medical data, because fuzzy deep learning that originated from fuzzy sets, can effectively deal with uncertain and inaccurate information, providing new viewpoints for alleviating the presence of noise, artifact or high dimensional unstructured information in uncertain medical data. therefore, taking focus on the intersection of both different fuzzy deep learning models and several types of uncertain medical data, the paper first constructs four types of frameworks of fuzzy deep learning models used for uncertain medical data, and investigates the status from three aspects: fuzzy deep learning models, uncertain medical data and application scenarios. then the performance evaluation metrics of fuzzy deep learning models are analyzed in details. this work has some original points: (1) four types of frameworks of applying fuzzy deep learning models for uncertain medical data are first proposed. (2) seven fuzzy deep learning models, five types of uncertain medical data, and five application scenarios are reviewed in details, respectively. (3) the advantages, challenges, and future research directions of fuzzy deep learning for uncertain medical data are critically analyzed, providing valuable suggestions for further deep research.","key_phrases":["intelligent medical industry","a rapid stage","development","the world","the expanding market size","basic theories","intelligent medical diagnosis","decision-making","deep learning models","good practical results","medical domain","traditional deep learning","crisp values","imprecise","vague medical data","the process","diagnosis","treatment","it","the contributions","fuzzy deep learning","uncertain medical data","fuzzy deep learning","that","fuzzy sets","uncertain and inaccurate information","new viewpoints","the presence","noise","artifact","high dimensional unstructured information","uncertain medical data","focus","the intersection","both different fuzzy deep learning models","several types","uncertain medical data","the paper","four types","frameworks","fuzzy deep learning models","uncertain medical data","the status","three aspects","fuzzy deep learning models","uncertain medical data","application scenarios","the performance evaluation metrics","fuzzy deep learning models","details","this work","some original points","(1) four types","frameworks","fuzzy deep learning models","uncertain medical data","(2) seven fuzzy deep learning models","five types","uncertain medical data","five application scenarios","details","(3) the advantages","challenges","future research directions","fuzzy deep learning","uncertain medical data","valuable suggestions","further deep research","first","four","three","1","first","2","seven","five","five","3"]},{"title":"Predicting discrete-time bifurcations with deep learning","authors":["Thomas M. Bury","Daniel Dylewsky","Chris T. Bauch","Madhur Anand","Leon Glass","Alvin Shrier","Gil Bub"],"abstract":"Many natural and man-made systems are prone to critical transitions\u2014abrupt and potentially devastating changes in dynamics. Deep learning classifiers can provide an early warning signal for critical transitions by learning generic features of bifurcations from large simulated training data sets. So far, classifiers have only been trained to predict continuous-time bifurcations, ignoring rich dynamics unique to discrete-time bifurcations. Here, we train a deep learning classifier to provide an early warning signal for the five local discrete-time bifurcations of codimension-one. We test the classifier on simulation data from discrete-time models used in physiology, economics and ecology, as well as experimental data of spontaneously beating chick-heart aggregates that undergo a period-doubling bifurcation. The classifier shows higher sensitivity and specificity than commonly used early warning signals under a wide range of noise intensities and rates of approach to the bifurcation. It also predicts the correct bifurcation in most cases, with particularly high accuracy for the period-doubling, Neimark-Sacker and fold bifurcations. Deep learning as a tool for bifurcation prediction is still in its nascence and has the potential to transform the way we monitor systems for critical transitions.","doi":"10.1038\/s41467-023-42020-z","cleaned_title":"predicting discrete-time bifurcations with deep learning","cleaned_abstract":"many natural and man-made systems are prone to critical transitions\u2014abrupt and potentially devastating changes in dynamics. deep learning classifiers can provide an early warning signal for critical transitions by learning generic features of bifurcations from large simulated training data sets. so far, classifiers have only been trained to predict continuous-time bifurcations, ignoring rich dynamics unique to discrete-time bifurcations. here, we train a deep learning classifier to provide an early warning signal for the five local discrete-time bifurcations of codimension-one. we test the classifier on simulation data from discrete-time models used in physiology, economics and ecology, as well as experimental data of spontaneously beating chick-heart aggregates that undergo a period-doubling bifurcation. the classifier shows higher sensitivity and specificity than commonly used early warning signals under a wide range of noise intensities and rates of approach to the bifurcation. it also predicts the correct bifurcation in most cases, with particularly high accuracy for the period-doubling, neimark-sacker and fold bifurcations. deep learning as a tool for bifurcation prediction is still in its nascence and has the potential to transform the way we monitor systems for critical transitions.","key_phrases":["many natural and man-made systems","critical transitions","abrupt and potentially devastating changes","dynamics","deep learning classifiers","an early warning signal","critical transitions","generic features","bifurcations","large simulated training data sets","classifiers","continuous-time bifurcations","rich dynamics","discrete-time bifurcations","we","a deep learning classifier","an early warning signal","the five local discrete-time bifurcations","we","the classifier","simulation data","discrete-time models","physiology","economics","ecology","experimental data","chick-heart aggregates","that","a period-doubling bifurcation","the classifier","higher sensitivity","specificity","early warning signals","a wide range","noise intensities","rates","approach","the bifurcation","it","the correct bifurcation","most cases","particularly high accuracy","the period-doubling, neimark-sacker and fold bifurcations","deep learning","a tool","bifurcation prediction","its nascence","the potential","the way","we","systems","critical transitions","five"]},{"title":"Machine learning and deep learning techniques for poultry tasks management: a review","authors":["Thavamani. Subramani","Vijayakumar. Jeganathan","Sruthi. Kunkuma Balasubramanian"],"abstract":"In recent years the poultry production industry has adopted automation with the help of different kinds of technological advancements like verities of monitoring and sensing tools, IoT devices, sensors, monitoring devices, and more. These advanced techniques will offer numerous advantages in poultry product production. Only with human resources based large-scale poultry production is not an easy task because the public health threads from ingesting foods with high antibiotic remains will residues a problem. Sometimes, zoonotic diseases and foodborne diseases present a crucial task to poultry producers. These repeated tasks concern massive and hazardous work accomplished within a suffering work domain, leading to high human health safety, labor cost, and sustainable risk of cross-infection through manufacturing facilities. In recent years, Artificial Intelligence technology has played a vital role in all sectors including poultry production has utilized advanced Machine Learning and Deep Learning technologies which is a subfield of AI, this technology has an acceptable potential to manage numerous challenges in the establishment of information-based farming and various task-handling systems in the poultry production sector. This article comprehensively reviews some recently established machine learning and deep learning algorithms for poultry management tasks like chicken activity monitoring, farm weather monitoring and control, weight prediction, earlier identification of diseased chickens, and more. It is divine that this proposed review work will compose a piece of helpful information for all who are interested in enhancing recognition of the future scope of utilizing the ML and DL techniques in the poultry production sectors.","doi":"10.1007\/s11042-024-18951-0","cleaned_title":"machine learning and deep learning techniques for poultry tasks management: a review","cleaned_abstract":"in recent years the poultry production industry has adopted automation with the help of different kinds of technological advancements like verities of monitoring and sensing tools, iot devices, sensors, monitoring devices, and more. these advanced techniques will offer numerous advantages in poultry product production. only with human resources based large-scale poultry production is not an easy task because the public health threads from ingesting foods with high antibiotic remains will residues a problem. sometimes, zoonotic diseases and foodborne diseases present a crucial task to poultry producers. these repeated tasks concern massive and hazardous work accomplished within a suffering work domain, leading to high human health safety, labor cost, and sustainable risk of cross-infection through manufacturing facilities. in recent years, artificial intelligence technology has played a vital role in all sectors including poultry production has utilized advanced machine learning and deep learning technologies which is a subfield of ai, this technology has an acceptable potential to manage numerous challenges in the establishment of information-based farming and various task-handling systems in the poultry production sector. this article comprehensively reviews some recently established machine learning and deep learning algorithms for poultry management tasks like chicken activity monitoring, farm weather monitoring and control, weight prediction, earlier identification of diseased chickens, and more. it is divine that this proposed review work will compose a piece of helpful information for all who are interested in enhancing recognition of the future scope of utilizing the ml and dl techniques in the poultry production sectors.","key_phrases":["recent years","the poultry production industry","automation","the help","different kinds","technological advancements","verities","monitoring","sensing tools","devices","these advanced techniques","numerous advantages","poultry product production","human resources","an easy task","the public health threads","foods","high antibiotic remains","a problem","zoonotic diseases","foodborne diseases","a crucial task","poultry producers","these repeated tasks concern massive and hazardous work","a suffering work domain","high human health safety","labor cost","sustainable risk","cross","-","infection","manufacturing facilities","recent years","artificial intelligence technology","a vital role","all sectors","poultry production","advanced machine learning","deep learning technologies","which","a subfield","ai","this technology","an acceptable potential","numerous challenges","the establishment","information-based farming","various task-handling systems","the poultry production sector","this article","some recently established machine learning","deep learning algorithms","poultry management tasks","chicken activity monitoring","farm weather monitoring","control","weight prediction","earlier identification","diseased chickens","it","this proposed review work","a piece","helpful information","all","who","recognition","the future scope","the ml and dl techniques","the poultry production sectors","recent years","recent years"]},{"title":"Explaining deep learning-based leaf disease identification","authors":["Ankit Rajpal","Rashmi Mishra","Sheetal Rajpal","Kavita","Varnika Bhatia","Naveen Kumar"],"abstract":"Crop diseases adversely affect agricultural productivity and quality. The primary cause of these diseases is the presence of biotic stresses such as fungi, viruses, and bacteria. Detecting these causes at early stages requires constant monitoring by domain experts. Technological advancements in machine learning and deep learning methods have enabled the automated identification of leaf disease-specific symptoms through image analysis. This paper proposes image-based detection of leaf diseases using various deep learning-based models. The experiment was conducted on the PlantVillage dataset, which consists of 54,305 colour leaf images (healthy and diseased) belonging to 11 crop species categorized into 38 classes. The Inception-ResNet-V2-based model achieved a 10-fold cross-validation accuracy of \\(0.9991 \\pm 0.002\\) outperforming the other deep neural architectures and surpassing the performance of existing models in recent state-of-the-art works. Each underlined model is validated on an independent cohort. The Inception-ResNet-V2-based model achieved the best 10-fold cross-validation accuracy of \\(0.9535 \\pm 0.041\\) and was found statistically significant among other deep learning-based models. However, these deep learning models are considered a black box as their leaf disease predictions are opaque to end users. To address this issue, a local interpretable framework is proposed to mark the superpixels that contribute to identifying leaf disease. These superpixels closely confirmed the annotations of the human expert.","doi":"10.1007\/s00500-024-09939-x","cleaned_title":"explaining deep learning-based leaf disease identification","cleaned_abstract":"crop diseases adversely affect agricultural productivity and quality. the primary cause of these diseases is the presence of biotic stresses such as fungi, viruses, and bacteria. detecting these causes at early stages requires constant monitoring by domain experts. technological advancements in machine learning and deep learning methods have enabled the automated identification of leaf disease-specific symptoms through image analysis. this paper proposes image-based detection of leaf diseases using various deep learning-based models. the experiment was conducted on the plantvillage dataset, which consists of 54,305 colour leaf images (healthy and diseased) belonging to 11 crop species categorized into 38 classes. the inception-resnet-v2-based model achieved a 10-fold cross-validation accuracy of \\(0.9991 \\pm 0.002\\) outperforming the other deep neural architectures and surpassing the performance of existing models in recent state-of-the-art works. each underlined model is validated on an independent cohort. the inception-resnet-v2-based model achieved the best 10-fold cross-validation accuracy of \\(0.9535 \\pm 0.041\\) and was found statistically significant among other deep learning-based models. however, these deep learning models are considered a black box as their leaf disease predictions are opaque to end users. to address this issue, a local interpretable framework is proposed to mark the superpixels that contribute to identifying leaf disease. these superpixels closely confirmed the annotations of the human expert.","key_phrases":["crop diseases","agricultural productivity","quality","the primary cause","these diseases","the presence","biotic stresses","fungi","viruses","bacteria","these causes","early stages","constant monitoring","domain experts","technological advancements","machine learning","deep learning methods","the automated identification","leaf disease-specific symptoms","image analysis","this paper","image-based detection","leaf diseases","various deep learning-based models","the experiment","the plantvillage dataset","which","54,305 colour leaf images","11 crop species","38 classes","the inception-resnet-v2-based model","a 10-fold cross-validation accuracy","\\(0.9991 \\pm 0.002\\","the other deep neural architectures","the performance","existing models","the-art","each underlined model","an independent cohort","the inception-resnet-v2-based model","the best 10-fold cross-validation accuracy","\\(0.9535 \\pm 0.041\\","other deep learning-based models","these deep learning models","a black box","their leaf disease predictions","users","this issue","a local interpretable framework","the superpixels","that","leaf disease","these superpixels","the annotations","the human expert","54,305","11","38","10-fold","\\pm 0.002\\","10-fold","\\(0.9535 \\pm 0.041\\"]},{"title":"Model-based deep reinforcement learning for accelerated learning from flow simulations","authors":["Andre Weiner","Janis Geise"],"abstract":"In recent years, deep reinforcement learning has emerged as a technique to solve closed-loop flow control problems. Employing simulation-based environments in reinforcement learning enables a priori end-to-end optimization of the control system, provides a virtual testbed for safety-critical control applications, and allows to gain a deep understanding of the control mechanisms. While reinforcement learning has been applied successfully in a number of rather simple flow control benchmarks, a major bottleneck toward real-world applications is the high computational cost and turnaround time of flow simulations. In this contribution, we demonstrate the benefits of model-based reinforcement learning for flow control applications. Specifically, we optimize the policy by alternating between trajectories sampled from flow simulations and trajectories sampled from an ensemble of environment models. The model-based learning reduces the overall training time by up to \\(85\\%\\) for the fluidic pinball test case. Even larger savings are expected for more demanding flow simulations.","doi":"10.1007\/s11012-024-01808-z","cleaned_title":"model-based deep reinforcement learning for accelerated learning from flow simulations","cleaned_abstract":"in recent years, deep reinforcement learning has emerged as a technique to solve closed-loop flow control problems. employing simulation-based environments in reinforcement learning enables a priori end-to-end optimization of the control system, provides a virtual testbed for safety-critical control applications, and allows to gain a deep understanding of the control mechanisms. while reinforcement learning has been applied successfully in a number of rather simple flow control benchmarks, a major bottleneck toward real-world applications is the high computational cost and turnaround time of flow simulations. in this contribution, we demonstrate the benefits of model-based reinforcement learning for flow control applications. specifically, we optimize the policy by alternating between trajectories sampled from flow simulations and trajectories sampled from an ensemble of environment models. the model-based learning reduces the overall training time by up to \\(85\\%\\) for the fluidic pinball test case. even larger savings are expected for more demanding flow simulations.","key_phrases":["recent years","deep reinforcement learning","a technique","closed-loop flow control problems","simulation-based environments","reinforcement learning","end","the control system","safety-critical control applications","a deep understanding","the control mechanisms","reinforcement learning","a number","rather simple flow control benchmarks","a major bottleneck","real-world applications","the high computational cost and turnaround time","flow simulations","this contribution","we","the benefits","model-based reinforcement learning","flow control applications","we","the policy","trajectories","flow simulations","trajectories","an ensemble","environment models","the model-based learning","the overall training time","\\(85\\%\\","the fluidic pinball test case","even larger savings","more demanding flow simulations","recent years"]},{"title":"Deep learning algorithms applied to computational chemistry","authors":["Abimael Guzman-Pando","Graciela Ramirez-Alonso","Carlos Arzate-Quintana","Javier Camarillo-Cisneros"],"abstract":"Recently, there has been a significant increase in the use of deep learning techniques in the molecular sciences, which have shown high performance on datasets and the ability to generalize across data. However, no model has achieved perfect performance in solving all problems, and the pros and cons of each approach remain unclear to those new to the field. Therefore, this paper aims to review deep learning algorithms that have been applied to solve molecular challenges in computational chemistry. We proposed a comprehensive categorization that encompasses two primary approaches; conventional deep learning and geometric deep learning models. This classification takes into account the distinct techniques employed by the algorithms within each approach. We present an up-to-date analysis of these algorithms, emphasizing their key features and open issues. This includes details of input descriptors, datasets used, open-source code availability, task solutions, and actual research applications, focusing on general applications rather than specific ones such as drug discovery. Furthermore, our report discusses trends and future directions in molecular algorithm design, including the input descriptors used for each deep learning model, GPU usage, training and forward processing time, model parameters, the most commonly used datasets, libraries, and optimization schemes. This information aids in identifying the most suitable algorithms for a given task. It also serves as a reference for the datasets and input data frequently used for each algorithm technique. In addition, it provides insights into the benefits and open issues of each technique, and supports the development of novel computational chemistry systems.","doi":"10.1007\/s11030-023-10771-y","cleaned_title":"deep learning algorithms applied to computational chemistry","cleaned_abstract":"recently, there has been a significant increase in the use of deep learning techniques in the molecular sciences, which have shown high performance on datasets and the ability to generalize across data. however, no model has achieved perfect performance in solving all problems, and the pros and cons of each approach remain unclear to those new to the field. therefore, this paper aims to review deep learning algorithms that have been applied to solve molecular challenges in computational chemistry. we proposed a comprehensive categorization that encompasses two primary approaches; conventional deep learning and geometric deep learning models. this classification takes into account the distinct techniques employed by the algorithms within each approach. we present an up-to-date analysis of these algorithms, emphasizing their key features and open issues. this includes details of input descriptors, datasets used, open-source code availability, task solutions, and actual research applications, focusing on general applications rather than specific ones such as drug discovery. furthermore, our report discusses trends and future directions in molecular algorithm design, including the input descriptors used for each deep learning model, gpu usage, training and forward processing time, model parameters, the most commonly used datasets, libraries, and optimization schemes. this information aids in identifying the most suitable algorithms for a given task. it also serves as a reference for the datasets and input data frequently used for each algorithm technique. in addition, it provides insights into the benefits and open issues of each technique, and supports the development of novel computational chemistry systems.","key_phrases":["a significant increase","the use","deep learning techniques","the molecular sciences","which","high performance","datasets","the ability","data","no model","perfect performance","all problems","the pros","cons","each approach","those","the field","this paper","deep learning algorithms","that","molecular challenges","computational chemistry","we","a comprehensive categorization","that","two primary approaches","conventional deep learning","geometric deep learning models","this classification","account","the distinct techniques","the algorithms","each approach","we","date","these algorithms","their key features","open issues","this","details","input descriptors","datasets","open-source code availability","task solutions","actual research applications","general applications","specific ones","drug discovery","our report","trends","future directions","molecular algorithm design","the input descriptors","each deep learning model","gpu usage","training","forward processing time","model parameters","the most commonly used datasets","libraries","optimization schemes","the most suitable algorithms","a given task","it","a reference","the datasets","input data","each algorithm technique","addition","it","insights","the benefits","open issues","each technique","the development","novel computational chemistry systems","two"]},{"title":"Deep learning for named entity recognition: a survey","authors":["Zhentao Hu","Wei Hou","Xianxing Liu"],"abstract":"Named entity recognition (NER) aims to identify the required entities and their types from unstructured text, which can be utilized for the construction of knowledge graphs. Traditional methods heavily rely on manual feature engineering and face challenges in adapting to large datasets within complex linguistic contexts. In recent years, with the development of deep learning, a plethora of NER methods based on deep learning have emerged. This paper begins by providing a succinct introduction to the definition of the problem and the limitations of traditional methods. It enumerates commonly used NER datasets suitable for deep learning methods and categorizes them into three classes based on the complexity of named entities. Then, some typical deep learning-based NER methods are summarized in detail according to the development history of deep learning models. Subsequently, an in-depth analysis and comparison of methods achieving outstanding performance on representative and widely used datasets is conducted. Furthermore, the paper reproduces and analyzes the recognition results of some typical models on three different types of typical datasets. Finally, the paper concludes by offering insights into the future trends of NER development.","doi":"10.1007\/s00521-024-09646-6","cleaned_title":"deep learning for named entity recognition: a survey","cleaned_abstract":"named entity recognition (ner) aims to identify the required entities and their types from unstructured text, which can be utilized for the construction of knowledge graphs. traditional methods heavily rely on manual feature engineering and face challenges in adapting to large datasets within complex linguistic contexts. in recent years, with the development of deep learning, a plethora of ner methods based on deep learning have emerged. this paper begins by providing a succinct introduction to the definition of the problem and the limitations of traditional methods. it enumerates commonly used ner datasets suitable for deep learning methods and categorizes them into three classes based on the complexity of named entities. then, some typical deep learning-based ner methods are summarized in detail according to the development history of deep learning models. subsequently, an in-depth analysis and comparison of methods achieving outstanding performance on representative and widely used datasets is conducted. furthermore, the paper reproduces and analyzes the recognition results of some typical models on three different types of typical datasets. finally, the paper concludes by offering insights into the future trends of ner development.","key_phrases":["entity recognition","ner","the required entities","their types","unstructured text","which","the construction","knowledge graphs","traditional methods","manual feature engineering","challenges","large datasets","complex linguistic contexts","recent years","the development","deep learning","a plethora","ner methods","deep learning","this paper","a succinct introduction","the definition","the problem","the limitations","traditional methods","it","ner datasets","deep learning methods","them","three classes","the complexity","named entities","some typical deep learning-based ner methods","detail","the development history","deep learning models","an in-depth analysis","comparison","methods","outstanding performance","representative and widely used datasets","furthermore, the paper reproduces","the recognition results","some typical models","three different types","typical datasets","the paper","insights","the future trends","ner development","ner","recent years","three","three"]},{"title":"Deep Kernel learning for reaction outcome prediction and optimization","authors":["Sukriti Singh","Jos\u00e9 Miguel Hern\u00e1ndez-Lobato"],"abstract":"Recent years have seen a rapid growth in the application of various machine learning methods for reaction outcome prediction. Deep learning models have gained popularity due to their ability to learn representations directly from the molecular structure. Gaussian processes (GPs), on the other hand, provide reliable uncertainty estimates but are unable to learn representations from the data. We combine the feature learning ability of neural networks (NNs) with uncertainty quantification of GPs in a deep kernel learning (DKL) framework to predict the reaction outcome. The DKL model is observed to obtain very good predictive performance across different input representations. It significantly outperforms standard GPs and provides comparable performance to graph neural networks, but with uncertainty estimation. Additionally, the uncertainty estimates on predictions provided by the DKL model facilitated its incorporation as a surrogate model for Bayesian optimization (BO). The proposed method, therefore, has a great potential towards accelerating reaction discovery by integrating accurate predictive models that provide reliable uncertainty estimates with BO.","doi":"10.1038\/s42004-024-01219-x","cleaned_title":"deep kernel learning for reaction outcome prediction and optimization","cleaned_abstract":"recent years have seen a rapid growth in the application of various machine learning methods for reaction outcome prediction. deep learning models have gained popularity due to their ability to learn representations directly from the molecular structure. gaussian processes (gps), on the other hand, provide reliable uncertainty estimates but are unable to learn representations from the data. we combine the feature learning ability of neural networks (nns) with uncertainty quantification of gps in a deep kernel learning (dkl) framework to predict the reaction outcome. the dkl model is observed to obtain very good predictive performance across different input representations. it significantly outperforms standard gps and provides comparable performance to graph neural networks, but with uncertainty estimation. additionally, the uncertainty estimates on predictions provided by the dkl model facilitated its incorporation as a surrogate model for bayesian optimization (bo). the proposed method, therefore, has a great potential towards accelerating reaction discovery by integrating accurate predictive models that provide reliable uncertainty estimates with bo.","key_phrases":["recent years","a rapid growth","the application","various machine learning methods","reaction outcome prediction","deep learning models","popularity","their ability","representations","the molecular structure","gaussian processes","gps","the other hand","reliable uncertainty estimates","representations","the data","we","the feature learning ability","neural networks","nns","uncertainty quantification","gps","a deep kernel learning (dkl) framework","the reaction outcome","the dkl model","very good predictive performance","different input representations","it","standard gps","comparable performance","neural networks","uncertainty estimation","the uncertainty","predictions","the dkl model","its incorporation","a surrogate model","bayesian optimization","bo","the proposed method","a great potential","reaction discovery","accurate predictive models","that","reliable uncertainty estimates","bo","recent years","gaussian"]},{"title":"Colour fusion effect on deep learning classification of uveal melanoma","authors":["Albert K. Dadzie","Sabrina P. Iddir","Mansour Abtahi","Behrouz Ebrahimi","David Le","Sanjay Ganesh","Taeyoon Son","Michael J. Heiferman","Xincheng Yao"],"abstract":"BackgroundReliable differentiation of uveal melanoma and choroidal nevi is crucial to guide appropriate treatment, preventing unnecessary procedures for benign lesions and ensuring timely treatment for potentially malignant cases. The purpose of this study is to validate deep learning classification of uveal melanoma and choroidal nevi, and to evaluate the effect of colour fusion options on the classification performance.MethodsA total of 798 ultra-widefield retinal images of 438 patients were included in this retrospective study, comprising 157 patients diagnosed with UM and 281 patients diagnosed with choroidal naevus. Colour fusion options, including early fusion, intermediate fusion and late fusion, were tested for deep learning image classification with a convolutional neural network (CNN). F1-score, accuracy and the area under the curve (AUC) of a receiver operating characteristic (ROC) were used to evaluate the classification performance.ResultsColour fusion options were observed to affect the deep learning performance significantly. For single-colour learning, the red colour image was observed to have superior performance compared to green and blue channels. For multi-colour learning, the intermediate fusion is better than early and late fusion options.ConclusionDeep learning is a promising approach for automated classification of uveal melanoma and choroidal nevi. Colour fusion options can significantly affect the classification performance.","doi":"10.1038\/s41433-024-03148-4","cleaned_title":"colour fusion effect on deep learning classification of uveal melanoma","cleaned_abstract":"backgroundreliable differentiation of uveal melanoma and choroidal nevi is crucial to guide appropriate treatment, preventing unnecessary procedures for benign lesions and ensuring timely treatment for potentially malignant cases. the purpose of this study is to validate deep learning classification of uveal melanoma and choroidal nevi, and to evaluate the effect of colour fusion options on the classification performance.methodsa total of 798 ultra-widefield retinal images of 438 patients were included in this retrospective study, comprising 157 patients diagnosed with um and 281 patients diagnosed with choroidal naevus. colour fusion options, including early fusion, intermediate fusion and late fusion, were tested for deep learning image classification with a convolutional neural network (cnn). f1-score, accuracy and the area under the curve (auc) of a receiver operating characteristic (roc) were used to evaluate the classification performance.resultscolour fusion options were observed to affect the deep learning performance significantly. for single-colour learning, the red colour image was observed to have superior performance compared to green and blue channels. for multi-colour learning, the intermediate fusion is better than early and late fusion options.conclusiondeep learning is a promising approach for automated classification of uveal melanoma and choroidal nevi. colour fusion options can significantly affect the classification performance.","key_phrases":["backgroundreliable differentiation","uveal melanoma","choroidal nevi","appropriate treatment","unnecessary procedures","benign lesions","timely treatment","potentially malignant cases","the purpose","this study","deep learning classification","uveal melanoma","choroidal","nevi","the effect","colour fusion options","the classification performance.methodsa total","798 ultra-widefield retinal images","438 patients","this retrospective study","157 patients","281 patients","choroidal naevus","colour fusion options","early fusion","intermediate fusion","late fusion","deep learning image classification","a convolutional neural network","cnn","f1-score","accuracy","the area","the curve","auc","(roc","the classification performance.resultscolour fusion options","the deep learning performance","single-colour learning","the red colour image","superior performance","green and blue channels","multi-colour learning","the intermediate fusion","early and late fusion","options.conclusiondeep learning","a promising approach","automated classification","uveal melanoma","choroidal","nevi","colour fusion options","the classification performance","798","438","157","281","cnn","roc","performance.resultscolour"]},{"title":"Predicting DNA structure using a deep learning method","authors":["Jinsen Li","Tsu-Pei Chiu","Remo Rohs"],"abstract":"Understanding the mechanisms of protein-DNA binding is critical in comprehending gene regulation. Three-dimensional DNA structure, also described as DNA shape, plays a key role in these mechanisms. In this study, we present a deep learning-based method, Deep DNAshape, that fundamentally changes the current k-mer based high-throughput prediction of DNA shape features by accurately accounting for the influence of extended flanking regions, without the need for extensive molecular simulations or structural biology experiments. By using the Deep DNAshape method, DNA structural features can be predicted for any length and number of DNA sequences in a high-throughput manner, providing an understanding of the effects of flanking regions on DNA structure in a target region of a sequence. The Deep DNAshape method provides access to the influence of distant flanking regions on a region of interest. Our findings reveal that DNA shape readout mechanisms of a core target are quantitatively affected by flanking regions, including extended flanking regions, providing valuable insights into the detailed structural readout mechanisms of protein-DNA binding. Furthermore, when incorporated in machine learning models, the features generated by Deep DNAshape improve the model prediction accuracy. Collectively, Deep DNAshape can serve as versatile and powerful tool for diverse DNA structure-related studies.","doi":"10.1038\/s41467-024-45191-5","cleaned_title":"predicting dna structure using a deep learning method","cleaned_abstract":"understanding the mechanisms of protein-dna binding is critical in comprehending gene regulation. three-dimensional dna structure, also described as dna shape, plays a key role in these mechanisms. in this study, we present a deep learning-based method, deep dnashape, that fundamentally changes the current k-mer based high-throughput prediction of dna shape features by accurately accounting for the influence of extended flanking regions, without the need for extensive molecular simulations or structural biology experiments. by using the deep dnashape method, dna structural features can be predicted for any length and number of dna sequences in a high-throughput manner, providing an understanding of the effects of flanking regions on dna structure in a target region of a sequence. the deep dnashape method provides access to the influence of distant flanking regions on a region of interest. our findings reveal that dna shape readout mechanisms of a core target are quantitatively affected by flanking regions, including extended flanking regions, providing valuable insights into the detailed structural readout mechanisms of protein-dna binding. furthermore, when incorporated in machine learning models, the features generated by deep dnashape improve the model prediction accuracy. collectively, deep dnashape can serve as versatile and powerful tool for diverse dna structure-related studies.","key_phrases":["the mechanisms","protein-dna binding","comprehending gene regulation","three-dimensional dna structure","dna shape","a key role","these mechanisms","this study","we","a deep learning-based method","deep dnashape","the current k-mer based high-throughput prediction","dna shape features","the influence","extended flanking regions","the need","extensive molecular simulations","structural biology experiments","the deep dnashape method","dna structural features","any length","number","dna sequences","a high-throughput manner","an understanding","the effects","flanking regions","dna structure","a target region","a sequence","the deep dnashape method","access","the influence","distant flanking regions","a region","interest","our findings","dna","readout mechanisms","a core target","flanking regions","extended flanking regions","valuable insights","the detailed structural readout mechanisms","protein-dna binding","machine learning models","the features","deep dnashape","the model prediction accuracy","deep dnashape","versatile and powerful tool","diverse dna","structure-related studies","three"]},{"title":"Machine learning and deep learning for classifying the justification of brain CT referrals","authors":["Jaka Poto\u010dnik","Edel Thomas","Aonghus Lawlor","Dearbhla Kearney","Eric J. Heffernan","Ronan P. Killeen","Shane J. Foley"],"abstract":"ObjectivesTo train the machine and deep\u00a0learning models to automate the justification analysis of radiology referrals in accordance with iGuide categorisation, and to determine if prediction models can generalise across multiple clinical sites and outperform human experts.MethodsAdult brain computed tomography (CT) referrals from scans performed in three CT centres in Ireland in 2020 and 2021 were retrospectively collected. Two radiographers analysed the justification of 3000 randomly selected referrals using iGuide, with two consultant radiologists analysing the referrals with disagreement. Insufficient or duplicate referrals were discarded. The inter-rater agreement among radiographers and consultants was computed. A random split (4:1) was performed to apply machine learning (ML) and deep\u00a0learning (DL) techniques to unstructured clinical indications to automate retrospective justification auditing with multi-class classification. The accuracy and macro-averaged F1 score of the best-performing classifier of each type on the training set were computed on the test set.Results42 referrals were ignored. 1909 (64.5%) referrals were justified, 811 (27.4%) were potentially justified, and 238 (8.1%) were unjustified. The agreement between radiographers (\u03ba\u2009=\u20090.268) was lower than radiologists (\u03ba\u2009=\u20090.460). The best-performing ML model was the bag-of-words-based gradient-boosting classifier achieving a 94.4% accuracy and a macro F1 of 0.94. DL models were inferior, with bi-directional long short-term memory achieving 92.3% accuracy, a macro F1 of 0.92, and outperforming multilayer perceptrons.ConclusionInterpreting unstructured clinical indications is challenging necessitating clinical decision support. ML and DL can generalise across multiple clinical sites, outperform human experts, and be used as an artificial intelligence-based iGuide interpreter when retrospectively vetting radiology referrals.Clinical relevance statementHealthcare vendors and clinical sites should consider developing and utilising artificial intelligence-enabled systems for justifying medical exposures. This would enable better implementation of imaging referral guidelines in clinical practices and reduce population dose burden, CT waiting lists, and wasteful use of resources.Key Points\n\nSignificant variations exist among human experts in interpreting unstructured clinical indications\/patient presentations.\n\n\nMachine and deep learning can automate the justification analysis of radiology referrals according to iGuide categorisation.\n\n\nMachine and deep learning can improve retrospective and prospective justification auditing\u00a0for better implementation of imaging referral guidelines.","doi":"10.1007\/s00330-024-10851-z","cleaned_title":"machine learning and deep learning for classifying the justification of brain ct referrals","cleaned_abstract":"objectivesto train the machine and deep learning models to automate the justification analysis of radiology referrals in accordance with iguide categorisation, and to determine if prediction models can generalise across multiple clinical sites and outperform human experts.methodsadult brain computed tomography (ct) referrals from scans performed in three ct centres in ireland in 2020 and 2021 were retrospectively collected. two radiographers analysed the justification of 3000 randomly selected referrals using iguide, with two consultant radiologists analysing the referrals with disagreement. insufficient or duplicate referrals were discarded. the inter-rater agreement among radiographers and consultants was computed. a random split (4:1) was performed to apply machine learning (ml) and deep learning (dl) techniques to unstructured clinical indications to automate retrospective justification auditing with multi-class classification. the accuracy and macro-averaged f1 score of the best-performing classifier of each type on the training set were computed on the test set.results42 referrals were ignored. 1909 (64.5%) referrals were justified, 811 (27.4%) were potentially justified, and 238 (8.1%) were unjustified. the agreement between radiographers (\u03ba = 0.268) was lower than radiologists (\u03ba = 0.460). the best-performing ml model was the bag-of-words-based gradient-boosting classifier achieving a 94.4% accuracy and a macro f1 of 0.94. dl models were inferior, with bi-directional long short-term memory achieving 92.3% accuracy, a macro f1 of 0.92, and outperforming multilayer perceptrons.conclusioninterpreting unstructured clinical indications is challenging necessitating clinical decision support. ml and dl can generalise across multiple clinical sites, outperform human experts, and be used as an artificial intelligence-based iguide interpreter when retrospectively vetting radiology referrals.clinical relevance statementhealthcare vendors and clinical sites should consider developing and utilising artificial intelligence-enabled systems for justifying medical exposures. this would enable better implementation of imaging referral guidelines in clinical practices and reduce population dose burden, ct waiting lists, and wasteful use of resources.key points significant variations exist among human experts in interpreting unstructured clinical indications\/patient presentations. machine and deep learning can automate the justification analysis of radiology referrals according to iguide categorisation. machine and deep learning can improve retrospective and prospective justification auditing for better implementation of imaging referral guidelines.","key_phrases":["objectivesto","the machine","deep learning models","the justification analysis","radiology referrals","accordance","iguide categorisation","prediction models","multiple clinical sites","human experts.methodsadult brain computed tomography (ct) referrals","scans","three ct centres","ireland","two radiographers","the justification","3000 randomly selected referrals","iguide","two consultant radiologists","the referrals","disagreement","insufficient or duplicate referrals","the inter-rater agreement","radiographers","consultants","a random split","machine learning","ml","deep learning","(dl) techniques","unstructured clinical indications","multi-class classification","the accuracy and macro-averaged f1 score","the best-performing classifier","each type","the training set","the test","set.results42 referrals","(64.5%","238 (8.1%","the agreement","radiographers","radiologists","\u03ba","the best-performing ml model","the bag-of-words-based gradient-boosting classifier","a 94.4% accuracy","a macro f1","dl models","bi-directional long short-term memory","92.3% accuracy","a macro f1","outperforming multilayer","necessitating clinical decision support","ml","dl","multiple clinical sites","human experts","an artificial intelligence-based iguide interpreter","radiology referrals.clinical relevance statementhealthcare vendors","clinical sites","artificial intelligence-enabled systems","medical exposures","this","better implementation","referral guidelines","clinical practices","population dose burden","ct waiting lists","wasteful use","resources.key points","significant variations","human experts","unstructured clinical indications","patient presentations","machine","deep learning","the justification analysis","radiology referrals","iguide categorisation","machine","deep learning","retrospective and prospective justification","better implementation","imaging referral guidelines","objectivesto train","three","ireland","2020","2021","two","3000","two","4:1","set.results42","1909","64.5%","811","27.4%","238","8.1%","0.268","0.460","94.4%","0.94","92.3%","0.92"]},{"title":"Applications of deep learning method of artificial intelligence in education","authors":["Fan Zhang","Xiangyu Wang","Xinhong Zhang"],"abstract":"Intersection of education and deep learning method of artificial intelligence (AI) is gradually becoming a hot research field. Education will be profoundly transformed by AI. The purpose of this review is to help education practitioners understand the research frontiers and directions of AI applications in education. This paper reviews the applications of deep learning in education and provides a visualized bibliometric analysis. The data of this paper come from Education Resources Information Center (ERIC) and Web-of-Science (WOS). These two data bases are searched to identify research of deep learning-based education applications from 2015 to 2023. CiteSpace is used to analyze the number of publications, authors, institutions, and keywords of articles that are related to deep learning and education. This paper reviews and systematically analyzes the educational applications of deep learning in the following six aspects: learning effect prediction, educational game, learning recommendation, automatic scoring, assisted teaching and medical education. Based on the visualized bibliometric analysis, the path and inflection point of research evolution can be inferred, the potential motivation of research can be analyzed, and the frontier of research can be explored. At present, AI can enable teachers to focus more on teaching and personalized interactions with students, enhancing rather than replacing human instructions.","doi":"10.1007\/s10639-024-12883-w","cleaned_title":"applications of deep learning method of artificial intelligence in education","cleaned_abstract":"intersection of education and deep learning method of artificial intelligence (ai) is gradually becoming a hot research field. education will be profoundly transformed by ai. the purpose of this review is to help education practitioners understand the research frontiers and directions of ai applications in education. this paper reviews the applications of deep learning in education and provides a visualized bibliometric analysis. the data of this paper come from education resources information center (eric) and web-of-science (wos). these two data bases are searched to identify research of deep learning-based education applications from 2015 to 2023. citespace is used to analyze the number of publications, authors, institutions, and keywords of articles that are related to deep learning and education. this paper reviews and systematically analyzes the educational applications of deep learning in the following six aspects: learning effect prediction, educational game, learning recommendation, automatic scoring, assisted teaching and medical education. based on the visualized bibliometric analysis, the path and inflection point of research evolution can be inferred, the potential motivation of research can be analyzed, and the frontier of research can be explored. at present, ai can enable teachers to focus more on teaching and personalized interactions with students, enhancing rather than replacing human instructions.","key_phrases":["intersection","education","deep learning method","artificial intelligence","a hot research field","education","ai","the purpose","this review","education practitioners","the research frontiers","directions","ai applications","education","this paper","the applications","deep learning","education","a visualized bibliometric analysis","the data","this paper","education resources information center","eric","web","science","wos","these two data bases","research","deep learning-based education applications","citespace","the number","publications","authors","institutions","keywords","articles","that","deep learning","education","this paper reviews","the educational applications","deep learning","the following six aspects","effect prediction","educational game","recommendation","automatic scoring","assisted teaching","medical education","the visualized bibliometric analysis","the path","inflection point","research evolution","the potential motivation","research","the frontier","research","present","teachers","teaching and personalized interactions","students","human instructions","eric","two","2015","citespace","six"]},{"title":"A systematic literature review of visual feature learning: deep learning techniques, applications, challenges and future directions","authors":["Mohammed Abdullahi","Olaide Nathaniel Oyelade","Armand Florentin Donfack Kana","Mustapha Aminu Bagiwa","Fatimah Binta Abdullahi","Sahalu Balarabe Junaidu","Ibrahim Iliyasu","Ajayi Ore-ofe","Haruna Chiroma"],"abstract":"Visual Feature Learning (VFL) is a critical area of research in computer vision that involves the automatic extraction of features and patterns from images and videos. The applications of VFL are vast, including object detection and recognition, facial recognition, scene understanding, medical image analysis, and autonomous vehicles. In this paper, we propose to conduct extensive systematic literature review (SLR) on VFL based on deep learning algorithms. The paper conducted an SLR covering deep learning algorithms such as Convolutional Neural Networks (CNNs), Autoencoders, and Generative Adversarial Networks (GANs) including their variants. The review highlights the importance of VFL in computer vision and the limitations of traditional feature extraction techniques. Furthermore, it provides an in-depth analysis of the strengths and weaknesses of various deep learning algorithms for solving problems in VFL. The discussion of the applications of VFL provides an insight into the impact of VFL on various industries and domains. The review also analyzed the challenges faced by VFL, such as data scarcity and quality, overfitting, generalization, interpretability, and explainability. The discussion of future directions for VFL includes hybrid techniques, unsupervised feature learning, continual learning, attention-based models, and explainable AI. These techniques aim to address the challenges faced by VFL and improve the performance of the models. The systematic literature review concludes that VFL is a rapidly evolving field with the potential to transform many industries and domains. The review highlights the need for further research in VFL and emphasizes the importance of responsible use of VFL models in various applications. The review provides valuable insights for researchers and practitioners in the field of computer vision, who can use these insights to enhance their work and ensure the responsible use of VFL models.","doi":"10.1007\/s11042-024-19823-3","cleaned_title":"a systematic literature review of visual feature learning: deep learning techniques, applications, challenges and future directions","cleaned_abstract":"visual feature learning (vfl) is a critical area of research in computer vision that involves the automatic extraction of features and patterns from images and videos. the applications of vfl are vast, including object detection and recognition, facial recognition, scene understanding, medical image analysis, and autonomous vehicles. in this paper, we propose to conduct extensive systematic literature review (slr) on vfl based on deep learning algorithms. the paper conducted an slr covering deep learning algorithms such as convolutional neural networks (cnns), autoencoders, and generative adversarial networks (gans) including their variants. the review highlights the importance of vfl in computer vision and the limitations of traditional feature extraction techniques. furthermore, it provides an in-depth analysis of the strengths and weaknesses of various deep learning algorithms for solving problems in vfl. the discussion of the applications of vfl provides an insight into the impact of vfl on various industries and domains. the review also analyzed the challenges faced by vfl, such as data scarcity and quality, overfitting, generalization, interpretability, and explainability. the discussion of future directions for vfl includes hybrid techniques, unsupervised feature learning, continual learning, attention-based models, and explainable ai. these techniques aim to address the challenges faced by vfl and improve the performance of the models. the systematic literature review concludes that vfl is a rapidly evolving field with the potential to transform many industries and domains. the review highlights the need for further research in vfl and emphasizes the importance of responsible use of vfl models in various applications. the review provides valuable insights for researchers and practitioners in the field of computer vision, who can use these insights to enhance their work and ensure the responsible use of vfl models.","key_phrases":["vfl","a critical area","research","computer vision","that","the automatic extraction","features","patterns","images","videos","the applications","vfl","object detection","recognition","facial recognition","scene understanding","medical image analysis","autonomous vehicles","this paper","we","extensive systematic literature review","slr","vfl","deep learning algorithms","the paper","an slr","deep learning algorithms","convolutional neural networks","cnns","autoencoders","generative adversarial networks","gans","their variants","the review","the importance","vfl","computer vision","the limitations","traditional feature extraction techniques","it","an in-depth analysis","the strengths","weaknesses","various deep learning algorithms","problems","vfl","the discussion","the applications","vfl","an insight","the impact","vfl","various industries","domains","the review","the challenges","vfl","data scarcity","quality","overfitting","generalization","interpretability","explainability","the discussion","future directions","vfl","hybrid techniques","feature learning","continual learning","attention-based models","explainable ai","these techniques","the challenges","vfl","the performance","the models","the systematic literature review","vfl","a rapidly evolving field","the potential","many industries","domains","the review","the need","further research","vfl","the importance","responsible use","vfl models","various applications","the review","valuable insights","researchers","practitioners","the field","computer vision","who","these insights","their work","the responsible use","vfl models","vfl","vfl","vfl","vfl","vfl","vfl","vfl","vfl","vfl","vfl","vfl","vfl","vfl","vfl"]},{"title":"Deep\u00a0Learning\u00a0Challenges\u00a0and\u00a0Prospects\u00a0in\u00a0Wireless\u00a0Sensor\u00a0Network\u00a0Deployment","authors":["Yaner Qiu","Liyun Ma","Rahul Priyadarshi"],"abstract":"This paper explores the transformative integration of deep learning applications in the deployment of Wireless Sensor Networks (WSNs). As WSNs continue to play a pivotal role in diverse domains, the infusion of deep learning techniques offers unprecedented opportunities for enhanced data processing, analysis, and decision-making. The research problem addressed in this paper revolves around navigating the challenges associated with incorporating deep learning into WSN deployment. The methodology involves an extensive literature review, highlighting the increasing role of deep learning in addressing WSN challenges. Key findings underscore the potential improvements in energy efficiency, data processing speed, and accuracy achieved through deep learning-empowered WSNs. The implications of this research extend to diverse applications, including environmental monitoring, healthcare, industrial systems, and smart agriculture. As we delve into the future research agenda, the paper identifies the need for further exploration in areas such as adaptability to dynamic environments, privacy-preserving optimizations, and scalable deep learning models tailored to the unique constraints of WSNs.","doi":"10.1007\/s11831-024-10079-6","cleaned_title":"deep learning challenges and prospects in wireless sensor network deployment","cleaned_abstract":"this paper explores the transformative integration of deep learning applications in the deployment of wireless sensor networks (wsns). as wsns continue to play a pivotal role in diverse domains, the infusion of deep learning techniques offers unprecedented opportunities for enhanced data processing, analysis, and decision-making. the research problem addressed in this paper revolves around navigating the challenges associated with incorporating deep learning into wsn deployment. the methodology involves an extensive literature review, highlighting the increasing role of deep learning in addressing wsn challenges. key findings underscore the potential improvements in energy efficiency, data processing speed, and accuracy achieved through deep learning-empowered wsns. the implications of this research extend to diverse applications, including environmental monitoring, healthcare, industrial systems, and smart agriculture. as we delve into the future research agenda, the paper identifies the need for further exploration in areas such as adaptability to dynamic environments, privacy-preserving optimizations, and scalable deep learning models tailored to the unique constraints of wsns.","key_phrases":["this paper","the transformative integration","deep learning applications","the deployment","wireless sensor networks","wsns","wsns","a pivotal role","diverse domains","the infusion","deep learning techniques","unprecedented opportunities","enhanced data processing","analysis","decision-making","the research problem","this paper","the challenges","deep learning","wsn deployment","the methodology","an extensive literature review","the increasing role","deep learning","wsn challenges","key findings","the potential improvements","energy efficiency","data processing speed","accuracy","deep learning-empowered wsns","the implications","this research extend","diverse applications","environmental monitoring","healthcare","industrial systems","smart agriculture","we","the future research agenda","the paper","the need","further exploration","areas","adaptability","dynamic environments","privacy-preserving optimizations","scalable deep learning models","the unique constraints","wsns","wsns"]},{"title":"Wheat crop classification using deep learning","authors":["Harmandeep Singh Gill","Bikramjit Singh Bath","Rajanbir Singh","Amarinder Singh Riar"],"abstract":"Crop yield forecasting is becoming more essential in the present environment, when food security must be maintained despite climate, population, and climate change concerns. Machine learning is a useful decision-making tool for predicting agricultural yields, as well as for deciding what crops to plant and what to do throughout the crop\u2019s growth season. To aid agricultural production prediction studies, a number of machine learning methods have been used. Wheat is a significant food source in India, particularly in the north. The wheat crop is categorised using deep learning techniques in the proposed research. The suggested system uses deep learning CNN, RNN, and LSTM applications to classify wheat crops. The results showed that the test accuracy ranged from \\(85\\%\\) to 95.68\\(\\%\\) for varietal level classification. Hence, the proposed approach results are accurate and reliable, encouraging the deployment of such an approach in practice.","doi":"10.1007\/s11042-024-18617-x","cleaned_title":"wheat crop classification using deep learning","cleaned_abstract":"crop yield forecasting is becoming more essential in the present environment, when food security must be maintained despite climate, population, and climate change concerns. machine learning is a useful decision-making tool for predicting agricultural yields, as well as for deciding what crops to plant and what to do throughout the crop\u2019s growth season. to aid agricultural production prediction studies, a number of machine learning methods have been used. wheat is a significant food source in india, particularly in the north. the wheat crop is categorised using deep learning techniques in the proposed research. the suggested system uses deep learning cnn, rnn, and lstm applications to classify wheat crops. the results showed that the test accuracy ranged from \\(85\\%\\) to 95.68\\(\\%\\) for varietal level classification. hence, the proposed approach results are accurate and reliable, encouraging the deployment of such an approach in practice.","key_phrases":["crop yield forecasting","the present environment","food security","climate","population","climate change concerns","machine learning","a useful decision-making tool","agricultural yields","what","crops","what","the crop\u2019s growth season","agricultural production prediction studies","a number","machine learning methods","wheat","a significant food source","india","the north","the wheat crop","deep learning techniques","the proposed research","the suggested system","deep learning cnn","lstm applications","wheat crops","the results","the test accuracy","\\(85\\%\\","varietal level classification","the proposed approach results","the deployment","such an approach","practice","india","cnn"]},{"title":"HARNet in deep learning approach\u2014a systematic survey","authors":["Neelam Sanjeev Kumar","G. Deepika","V. Goutham","B. Buvaneswari","R. Vijaya Kumar Reddy","Sanjeevkumar Angadi","C. Dhanamjayulu","Ravikumar Chinthaginjala","Faruq Mohammad","Baseem Khan"],"abstract":"A comprehensive examination of human action recognition (HAR) methodologies situated at the convergence of deep learning and computer vision is the subject of this article. We examine the progression from handcrafted feature-based approaches to end-to-end learning, with a particular focus on the significance of large-scale datasets. By classifying research paradigms, such as temporal modelling and spatial features, our proposed taxonomy illuminates the merits and drawbacks of each. We specifically present HARNet, an architecture for Multi-Model Deep Learning that integrates recurrent and convolutional neural networks while utilizing attention mechanisms to improve accuracy and robustness. The VideoMAE v2 method (https:\/\/github.com\/OpenGVLab\/VideoMAEv2) has been utilized as a case study to illustrate practical implementations and obstacles. For researchers and practitioners interested in gaining a comprehensive understanding of the most recent advancements in HAR as they relate to computer vision and deep learning, this survey is an invaluable resource.","doi":"10.1038\/s41598-024-58074-y","cleaned_title":"harnet in deep learning approach\u2014a systematic survey","cleaned_abstract":"a comprehensive examination of human action recognition (har) methodologies situated at the convergence of deep learning and computer vision is the subject of this article. we examine the progression from handcrafted feature-based approaches to end-to-end learning, with a particular focus on the significance of large-scale datasets. by classifying research paradigms, such as temporal modelling and spatial features, our proposed taxonomy illuminates the merits and drawbacks of each. we specifically present harnet, an architecture for multi-model deep learning that integrates recurrent and convolutional neural networks while utilizing attention mechanisms to improve accuracy and robustness. the videomae v2 method (https:\/\/github.com\/opengvlab\/videomaev2) has been utilized as a case study to illustrate practical implementations and obstacles. for researchers and practitioners interested in gaining a comprehensive understanding of the most recent advancements in har as they relate to computer vision and deep learning, this survey is an invaluable resource.","key_phrases":["a comprehensive examination","human action recognition (har) methodologies","the convergence","deep learning and computer vision","the subject","this article","we","the progression","handcrafted feature-based approaches","end","learning","a particular focus","the significance","large-scale datasets","research paradigms","temporal modelling","spatial features","our proposed taxonomy","the merits","drawbacks","each","we","an architecture","multi-model deep learning","that","recurrent and convolutional neural networks","attention mechanisms","accuracy","robustness","the videomae v2 method","https:\/\/github.com\/opengvlab\/videomaev2","a case study","practical implementations","obstacles","researchers","practitioners","a comprehensive understanding","the most recent advancements","har","they","computer vision","deep learning","this survey","an invaluable resource"]},{"title":"Deep learning bulk spacetime from boundary optical conductivity","authors":["Byoungjoon Ahn","Hyun-Sik Jeong","Keun-Young Kim","Kwan Yun"],"abstract":"We employ a deep learning method to deduce the bulk spacetime from boundary optical conductivity. We apply the neural ordinary differential equation technique, tailored for continuous functions such as the metric, to the typical class of holographic condensed matter models featuring broken translations: linear-axion models. We successfully extract the bulk metric from the boundary holographic optical conductivity. Furthermore, as an example for real material, we use experimental optical conductivity of UPd2Al3, a representative of heavy fermion metals in strongly correlated electron systems, and construct the corresponding bulk metric. To our knowledge, our work is the first illustration of deep learning bulk spacetime from boundary holographic or experimental conductivity data.","doi":"10.1007\/JHEP03(2024)141","cleaned_title":"deep learning bulk spacetime from boundary optical conductivity","cleaned_abstract":"we employ a deep learning method to deduce the bulk spacetime from boundary optical conductivity. we apply the neural ordinary differential equation technique, tailored for continuous functions such as the metric, to the typical class of holographic condensed matter models featuring broken translations: linear-axion models. we successfully extract the bulk metric from the boundary holographic optical conductivity. furthermore, as an example for real material, we use experimental optical conductivity of upd2al3, a representative of heavy fermion metals in strongly correlated electron systems, and construct the corresponding bulk metric. to our knowledge, our work is the first illustration of deep learning bulk spacetime from boundary holographic or experimental conductivity data.","key_phrases":["we","a deep learning method","the bulk spacetime","boundary optical conductivity","we","the neural ordinary differential equation technique","continuous functions","the metric","the typical class","holographic condensed matter models","broken translations","linear-axion models","we","the bulk metric","the boundary holographic optical conductivity","an example","real material","we","experimental optical conductivity","upd2al3","a representative","heavy fermion metals","strongly correlated electron systems","the corresponding bulk metric","our knowledge","our work","the first illustration","deep learning bulk spacetime","boundary holographic or experimental conductivity data","linear","first"]},{"title":"Deep learning: systematic review, models, challenges, and research directions","authors":["Tala Talaei Khoei","Hadjar Ould Slimane","Naima Kaabouch"],"abstract":"The current development in deep learning is witnessing an exponential transition into automation applications. This automation transition can provide a promising framework for higher performance and lower complexity. This ongoing transition undergoes several rapid changes, resulting in the processing of the data by several studies, while it may lead to time-consuming and costly models. Thus, to address these challenges, several studies have been conducted to investigate deep learning techniques; however, they mostly focused on specific learning approaches, such as supervised deep learning. In addition, these studies did not comprehensively investigate other deep learning techniques, such as deep unsupervised and deep reinforcement learning techniques. Moreover, the majority of these studies neglect to discuss some main methodologies in deep learning, such as transfer learning, federated learning, and online learning. Therefore, motivated by the limitations of the existing studies, this study summarizes the deep learning techniques into supervised, unsupervised, reinforcement, and hybrid learning-based models. In addition to address each category, a brief description of these categories and their models is provided. Some of the critical topics in deep learning, namely, transfer, federated, and online learning models, are explored and discussed in detail. Finally, challenges and future directions are outlined to provide wider outlooks for future researchers.","doi":"10.1007\/s00521-023-08957-4","cleaned_title":"deep learning: systematic review, models, challenges, and research directions","cleaned_abstract":"the current development in deep learning is witnessing an exponential transition into automation applications. this automation transition can provide a promising framework for higher performance and lower complexity. this ongoing transition undergoes several rapid changes, resulting in the processing of the data by several studies, while it may lead to time-consuming and costly models. thus, to address these challenges, several studies have been conducted to investigate deep learning techniques; however, they mostly focused on specific learning approaches, such as supervised deep learning. in addition, these studies did not comprehensively investigate other deep learning techniques, such as deep unsupervised and deep reinforcement learning techniques. moreover, the majority of these studies neglect to discuss some main methodologies in deep learning, such as transfer learning, federated learning, and online learning. therefore, motivated by the limitations of the existing studies, this study summarizes the deep learning techniques into supervised, unsupervised, reinforcement, and hybrid learning-based models. in addition to address each category, a brief description of these categories and their models is provided. some of the critical topics in deep learning, namely, transfer, federated, and online learning models, are explored and discussed in detail. finally, challenges and future directions are outlined to provide wider outlooks for future researchers.","key_phrases":["the current development","deep learning","an exponential transition","automation applications","this automation transition","a promising framework","higher performance","lower complexity","this ongoing transition","several rapid changes","the processing","the data","several studies","it","time-consuming and costly models","these challenges","several studies","deep learning techniques","they","specific learning approaches","supervised deep learning","addition","these studies","other deep learning techniques","deep unsupervised and deep reinforcement learning techniques","the majority","these studies","some main methodologies","deep learning","transfer learning","federated learning","online learning","the limitations","the existing studies","this study","the deep learning techniques","supervised, unsupervised, reinforcement, and hybrid learning-based models","addition","each category","a brief description","these categories","their models","some","the critical topics","deep learning","namely, transfer","online learning models","detail","challenges","future directions","wider outlooks","future researchers"]},{"title":"Colour fusion effect on deep learning classification of uveal melanoma","authors":["Albert K. Dadzie","Sabrina P. Iddir","Mansour Abtahi","Behrouz Ebrahimi","David Le","Sanjay Ganesh","Taeyoon Son","Michael J. Heiferman","Xincheng Yao"],"abstract":"BackgroundReliable differentiation of uveal melanoma and choroidal nevi is crucial to guide appropriate treatment, preventing unnecessary procedures for benign lesions and ensuring timely treatment for potentially malignant cases. The purpose of this study is to validate deep learning classification of uveal melanoma and choroidal nevi, and to evaluate the effect of colour fusion options on the classification performance.MethodsA total of 798 ultra-widefield retinal images of 438 patients were included in this retrospective study, comprising 157 patients diagnosed with UM and 281 patients diagnosed with choroidal naevus. Colour fusion options, including early fusion, intermediate fusion and late fusion, were tested for deep learning image classification with a convolutional neural network (CNN). F1-score, accuracy and the area under the curve (AUC) of a receiver operating characteristic (ROC) were used to evaluate the classification performance.ResultsColour fusion options were observed to affect the deep learning performance significantly. For single-colour learning, the red colour image was observed to have superior performance compared to green and blue channels. For multi-colour learning, the intermediate fusion is better than early and late fusion options.ConclusionDeep learning is a promising approach for automated classification of uveal melanoma and choroidal nevi. Colour fusion options can significantly affect the classification performance.","doi":"10.1038\/s41433-024-03148-4","cleaned_title":"colour fusion effect on deep learning classification of uveal melanoma","cleaned_abstract":"backgroundreliable differentiation of uveal melanoma and choroidal nevi is crucial to guide appropriate treatment, preventing unnecessary procedures for benign lesions and ensuring timely treatment for potentially malignant cases. the purpose of this study is to validate deep learning classification of uveal melanoma and choroidal nevi, and to evaluate the effect of colour fusion options on the classification performance.methodsa total of 798 ultra-widefield retinal images of 438 patients were included in this retrospective study, comprising 157 patients diagnosed with um and 281 patients diagnosed with choroidal naevus. colour fusion options, including early fusion, intermediate fusion and late fusion, were tested for deep learning image classification with a convolutional neural network (cnn). f1-score, accuracy and the area under the curve (auc) of a receiver operating characteristic (roc) were used to evaluate the classification performance.resultscolour fusion options were observed to affect the deep learning performance significantly. for single-colour learning, the red colour image was observed to have superior performance compared to green and blue channels. for multi-colour learning, the intermediate fusion is better than early and late fusion options.conclusiondeep learning is a promising approach for automated classification of uveal melanoma and choroidal nevi. colour fusion options can significantly affect the classification performance.","key_phrases":["backgroundreliable differentiation","uveal melanoma","choroidal nevi","appropriate treatment","unnecessary procedures","benign lesions","timely treatment","potentially malignant cases","the purpose","this study","deep learning classification","uveal melanoma","choroidal","nevi","the effect","colour fusion options","the classification performance.methodsa total","798 ultra-widefield retinal images","438 patients","this retrospective study","157 patients","281 patients","choroidal naevus","colour fusion options","early fusion","intermediate fusion","late fusion","deep learning image classification","a convolutional neural network","cnn","f1-score","accuracy","the area","the curve","auc","(roc","the classification performance.resultscolour fusion options","the deep learning performance","single-colour learning","the red colour image","superior performance","green and blue channels","multi-colour learning","the intermediate fusion","early and late fusion","options.conclusiondeep learning","a promising approach","automated classification","uveal melanoma","choroidal","nevi","colour fusion options","the classification performance","798","438","157","281","cnn","roc","performance.resultscolour"]},{"title":"Machine learning and deep learning for classifying the justification of brain CT referrals","authors":["Jaka Poto\u010dnik","Edel Thomas","Aonghus Lawlor","Dearbhla Kearney","Eric J. Heffernan","Ronan P. Killeen","Shane J. Foley"],"abstract":"ObjectivesTo train the machine and deep\u00a0learning models to automate the justification analysis of radiology referrals in accordance with iGuide categorisation, and to determine if prediction models can generalise across multiple clinical sites and outperform human experts.MethodsAdult brain computed tomography (CT) referrals from scans performed in three CT centres in Ireland in 2020 and 2021 were retrospectively collected. Two radiographers analysed the justification of 3000 randomly selected referrals using iGuide, with two consultant radiologists analysing the referrals with disagreement. Insufficient or duplicate referrals were discarded. The inter-rater agreement among radiographers and consultants was computed. A random split (4:1) was performed to apply machine learning (ML) and deep\u00a0learning (DL) techniques to unstructured clinical indications to automate retrospective justification auditing with multi-class classification. The accuracy and macro-averaged F1 score of the best-performing classifier of each type on the training set were computed on the test set.Results42 referrals were ignored. 1909 (64.5%) referrals were justified, 811 (27.4%) were potentially justified, and 238 (8.1%) were unjustified. The agreement between radiographers (\u03ba\u2009=\u20090.268) was lower than radiologists (\u03ba\u2009=\u20090.460). The best-performing ML model was the bag-of-words-based gradient-boosting classifier achieving a 94.4% accuracy and a macro F1 of 0.94. DL models were inferior, with bi-directional long short-term memory achieving 92.3% accuracy, a macro F1 of 0.92, and outperforming multilayer perceptrons.ConclusionInterpreting unstructured clinical indications is challenging necessitating clinical decision support. ML and DL can generalise across multiple clinical sites, outperform human experts, and be used as an artificial intelligence-based iGuide interpreter when retrospectively vetting radiology referrals.Clinical relevance statementHealthcare vendors and clinical sites should consider developing and utilising artificial intelligence-enabled systems for justifying medical exposures. This would enable better implementation of imaging referral guidelines in clinical practices and reduce population dose burden, CT waiting lists, and wasteful use of resources.Key Points\n\nSignificant variations exist among human experts in interpreting unstructured clinical indications\/patient presentations.\n\n\nMachine and deep learning can automate the justification analysis of radiology referrals according to iGuide categorisation.\n\n\nMachine and deep learning can improve retrospective and prospective justification auditing\u00a0for better implementation of imaging referral guidelines.","doi":"10.1007\/s00330-024-10851-z","cleaned_title":"machine learning and deep learning for classifying the justification of brain ct referrals","cleaned_abstract":"objectivesto train the machine and deep learning models to automate the justification analysis of radiology referrals in accordance with iguide categorisation, and to determine if prediction models can generalise across multiple clinical sites and outperform human experts.methodsadult brain computed tomography (ct) referrals from scans performed in three ct centres in ireland in 2020 and 2021 were retrospectively collected. two radiographers analysed the justification of 3000 randomly selected referrals using iguide, with two consultant radiologists analysing the referrals with disagreement. insufficient or duplicate referrals were discarded. the inter-rater agreement among radiographers and consultants was computed. a random split (4:1) was performed to apply machine learning (ml) and deep learning (dl) techniques to unstructured clinical indications to automate retrospective justification auditing with multi-class classification. the accuracy and macro-averaged f1 score of the best-performing classifier of each type on the training set were computed on the test set.results42 referrals were ignored. 1909 (64.5%) referrals were justified, 811 (27.4%) were potentially justified, and 238 (8.1%) were unjustified. the agreement between radiographers (\u03ba = 0.268) was lower than radiologists (\u03ba = 0.460). the best-performing ml model was the bag-of-words-based gradient-boosting classifier achieving a 94.4% accuracy and a macro f1 of 0.94. dl models were inferior, with bi-directional long short-term memory achieving 92.3% accuracy, a macro f1 of 0.92, and outperforming multilayer perceptrons.conclusioninterpreting unstructured clinical indications is challenging necessitating clinical decision support. ml and dl can generalise across multiple clinical sites, outperform human experts, and be used as an artificial intelligence-based iguide interpreter when retrospectively vetting radiology referrals.clinical relevance statementhealthcare vendors and clinical sites should consider developing and utilising artificial intelligence-enabled systems for justifying medical exposures. this would enable better implementation of imaging referral guidelines in clinical practices and reduce population dose burden, ct waiting lists, and wasteful use of resources.key points significant variations exist among human experts in interpreting unstructured clinical indications\/patient presentations. machine and deep learning can automate the justification analysis of radiology referrals according to iguide categorisation. machine and deep learning can improve retrospective and prospective justification auditing for better implementation of imaging referral guidelines.","key_phrases":["objectivesto","the machine","deep learning models","the justification analysis","radiology referrals","accordance","iguide categorisation","prediction models","multiple clinical sites","human experts.methodsadult brain computed tomography (ct) referrals","scans","three ct centres","ireland","two radiographers","the justification","3000 randomly selected referrals","iguide","two consultant radiologists","the referrals","disagreement","insufficient or duplicate referrals","the inter-rater agreement","radiographers","consultants","a random split","machine learning","ml","deep learning","(dl) techniques","unstructured clinical indications","multi-class classification","the accuracy and macro-averaged f1 score","the best-performing classifier","each type","the training set","the test","set.results42 referrals","(64.5%","238 (8.1%","the agreement","radiographers","radiologists","\u03ba","the best-performing ml model","the bag-of-words-based gradient-boosting classifier","a 94.4% accuracy","a macro f1","dl models","bi-directional long short-term memory","92.3% accuracy","a macro f1","outperforming multilayer","necessitating clinical decision support","ml","dl","multiple clinical sites","human experts","an artificial intelligence-based iguide interpreter","radiology referrals.clinical relevance statementhealthcare vendors","clinical sites","artificial intelligence-enabled systems","medical exposures","this","better implementation","referral guidelines","clinical practices","population dose burden","ct waiting lists","wasteful use","resources.key points","significant variations","human experts","unstructured clinical indications","patient presentations","machine","deep learning","the justification analysis","radiology referrals","iguide categorisation","machine","deep learning","retrospective and prospective justification","better implementation","imaging referral guidelines","objectivesto train","three","ireland","2020","2021","two","3000","two","4:1","set.results42","1909","64.5%","811","27.4%","238","8.1%","0.268","0.460","94.4%","0.94","92.3%","0.92"]},{"title":"General deep learning framework for emissivity engineering","authors":["Shilv Yu","Peng Zhou","Wang Xi","Zihe Chen","Yuheng Deng","Xiaobing Luo","Wangnan Li","Junichiro Shiomi","Run Hu"],"abstract":"Wavelength-selective thermal emitters (WS-TEs) have been frequently designed to achieve desired target emissivity spectra, as a typical emissivity engineering, for broad applications such as thermal camouflage, radiative cooling, and gas sensing, etc. However, previous designs require prior knowledge of materials or structures for different applications and the designed WS-TEs usually vary from applications to applications in terms of materials and structures, thus lacking of a general design framework for emissivity engineering across different applications. Moreover, previous designs fail to tackle the simultaneous design of both materials and structures, as they either fix materials to design structures or fix structures to select suitable materials. Herein, we employ the deep Q-learning network algorithm, a reinforcement learning method based on deep learning framework, to design multilayer WS-TEs. To demonstrate the general validity, three WS-TEs are designed for various applications, including thermal camouflage, radiative cooling and gas sensing, which are then fabricated and measured. The merits of the deep Q-learning algorithm include that it can (1) offer a general design framework for WS-TEs beyond one-dimensional multilayer structures; (2) autonomously select suitable materials from a self-built material library and (3) autonomously optimize structural parameters for the target emissivity spectra. The present framework is demonstrated to be feasible and efficient in designing WS-TEs across different applications, and the design parameters are highly scalable in materials, structures, dimensions, and the target functions, offering a general framework for emissivity engineering and paving the way for efficient design of nonlinear optimization problems beyond thermal metamaterials.","doi":"10.1038\/s41377-023-01341-w","cleaned_title":"general deep learning framework for emissivity engineering","cleaned_abstract":"wavelength-selective thermal emitters (ws-tes) have been frequently designed to achieve desired target emissivity spectra, as a typical emissivity engineering, for broad applications such as thermal camouflage, radiative cooling, and gas sensing, etc. however, previous designs require prior knowledge of materials or structures for different applications and the designed ws-tes usually vary from applications to applications in terms of materials and structures, thus lacking of a general design framework for emissivity engineering across different applications. moreover, previous designs fail to tackle the simultaneous design of both materials and structures, as they either fix materials to design structures or fix structures to select suitable materials. herein, we employ the deep q-learning network algorithm, a reinforcement learning method based on deep learning framework, to design multilayer ws-tes. to demonstrate the general validity, three ws-tes are designed for various applications, including thermal camouflage, radiative cooling and gas sensing, which are then fabricated and measured. the merits of the deep q-learning algorithm include that it can (1) offer a general design framework for ws-tes beyond one-dimensional multilayer structures; (2) autonomously select suitable materials from a self-built material library and (3) autonomously optimize structural parameters for the target emissivity spectra. the present framework is demonstrated to be feasible and efficient in designing ws-tes across different applications, and the design parameters are highly scalable in materials, structures, dimensions, and the target functions, offering a general framework for emissivity engineering and paving the way for efficient design of nonlinear optimization problems beyond thermal metamaterials.","key_phrases":["wavelength-selective thermal emitters","ws-tes","desired target emissivity spectra","a typical emissivity engineering","broad applications","thermal camouflage","cooling","gas sensing","previous designs","prior knowledge","materials","structures","different applications","the designed ws-tes","applications","applications","terms","materials","structures","a general design framework","emissivity engineering","different applications","previous designs","the simultaneous design","both materials","structures","they","materials","structures","structures","suitable materials","we","the deep q-learning network algorithm","a reinforcement learning method","deep learning framework","multilayer ws-tes","the general validity","three ws-tes","various applications","thermal camouflage","gas sensing","which","the merits","the deep q-learning algorithm","it","a general design framework","ws-tes","one-dimensional multilayer structures","suitable materials","a self-built material library","structural parameters","the target emissivity spectra","the present framework","ws-tes","different applications","the design parameters","materials","structures","dimensions","the target functions","a general framework","emissivity engineering","the way","efficient design","nonlinear optimization problems","thermal metamaterials","three","1","2","3"]},{"title":"Forecasting VIX using Bayesian deep learning","authors":["H\u00e9ctor J. Hort\u00faa","Andr\u00e9s Mora-Valencia"],"abstract":"Recently, deep learning techniques are gradually replacing traditional statistical and machine learning models as the first choice for price forecasting tasks. In this paper, we leverage probabilistic deep learning for inferring the volatility index VIX. We employ the probabilistic counterpart of WaveNet, Temporal Convolutional Network (TCN), and Transformers. We show that TCN outperforms all models with an RMSE around 0.189. In addition, it has been well known that modern neural networks provide inaccurate uncertainty estimates. For solving this problem, we use the standard deviation scaling to calibrate the networks. Furthermore, we found out that MNF with Gaussian prior outperforms Reparameterization Trick and Flipout models in terms of precision and uncertainty predictions. Finally, we claim that MNF with Cauchy and LogUniform prior distributions yield well-calibrated TCN, and Transformer and WaveNet networks being the former that best infer the VIX values for one and five-step-ahead forecasting, and the probabilistic Transformer model yields an adequate forecasting for the COVID-19 pandemic period.","doi":"10.1007\/s41060-024-00562-5","cleaned_title":"forecasting vix using bayesian deep learning","cleaned_abstract":"recently, deep learning techniques are gradually replacing traditional statistical and machine learning models as the first choice for price forecasting tasks. in this paper, we leverage probabilistic deep learning for inferring the volatility index vix. we employ the probabilistic counterpart of wavenet, temporal convolutional network (tcn), and transformers. we show that tcn outperforms all models with an rmse around 0.189. in addition, it has been well known that modern neural networks provide inaccurate uncertainty estimates. for solving this problem, we use the standard deviation scaling to calibrate the networks. furthermore, we found out that mnf with gaussian prior outperforms reparameterization trick and flipout models in terms of precision and uncertainty predictions. finally, we claim that mnf with cauchy and loguniform prior distributions yield well-calibrated tcn, and transformer and wavenet networks being the former that best infer the vix values for one and five-step-ahead forecasting, and the probabilistic transformer model yields an adequate forecasting for the covid-19 pandemic period.","key_phrases":["deep learning techniques","traditional statistical and machine learning models","the first choice","price forecasting tasks","this paper","we","probabilistic deep learning","the volatility index vix","we","the probabilistic counterpart","wavenet","temporal convolutional network","tcn","transformers","we","tcn","all models","an rmse","addition","it","modern neural networks","inaccurate uncertainty estimates","this problem","we","the standard deviation","the networks","we","that mnf","prior outperforms reparameterization trick","flipout","terms","precision and uncertainty predictions","we","cauchy","prior distributions","well-calibrated tcn","the vix values","one and five-step-ahead forecasting","the probabilistic transformer model","an adequate forecasting","the covid-19 pandemic period","first","rmse","0.189","one","five","covid-19"]},{"title":"Review of Deep Learning Techniques for Neurological Disorders Detection","authors":["Akhilesh Kumar Tripathi","Rafeeq Ahmed","Arvind Kumar Tiwari"],"abstract":"Neurological disease is one of the most common types of dementia that predominantly concerns the elderly. In clinical approaches, identifying its premature stages is complicated, and no biomarker is comprehended to be thorough in witnessing neurological disorders in their earlier stages. Deep learning approaches have attracted much attention in the scientific community using scanned images. They differ from simple machine learning (ML) algorithms in that they study the most favorable depiction of untreated images. Deep learning is helpful in the neuroimaging analysis of neurological diseases with subtle and dispersed changes because it can discover abstract and complicated patterns. The current study discusses a vital part of deep learning and looks at past work that has been used to switch between different ML algorithms that can predict neurological diseases. Convolution Neural Networks, Generative Adversarial Network, Recurrent Neural Network, Deep Belief Network, Auto Encoder, and other algorithms for Alzheimer\u2019s illness prediction have been considered. Many publications on preprocessing methods, such as scaling, correction, stripping, and normalizing, have been evaluated.","doi":"10.1007\/s11277-024-11464-x","cleaned_title":"review of deep learning techniques for neurological disorders detection","cleaned_abstract":"neurological disease is one of the most common types of dementia that predominantly concerns the elderly. in clinical approaches, identifying its premature stages is complicated, and no biomarker is comprehended to be thorough in witnessing neurological disorders in their earlier stages. deep learning approaches have attracted much attention in the scientific community using scanned images. they differ from simple machine learning (ml) algorithms in that they study the most favorable depiction of untreated images. deep learning is helpful in the neuroimaging analysis of neurological diseases with subtle and dispersed changes because it can discover abstract and complicated patterns. the current study discusses a vital part of deep learning and looks at past work that has been used to switch between different ml algorithms that can predict neurological diseases. convolution neural networks, generative adversarial network, recurrent neural network, deep belief network, auto encoder, and other algorithms for alzheimer\u2019s illness prediction have been considered. many publications on preprocessing methods, such as scaling, correction, stripping, and normalizing, have been evaluated.","key_phrases":["neurological disease","the most common types","dementia","that","clinical approaches","its premature stages","no biomarker","neurological disorders","their earlier stages","deep learning approaches","much attention","the scientific community","scanned images","they","simple machine learning","ml","they","the most favorable depiction","untreated images","deep learning","the neuroimaging analysis","neurological diseases","subtle and dispersed changes","it","abstract and complicated patterns","the current study","a vital part","deep learning","past work","that","different ml algorithms","that","neurological diseases","convolution neural networks","generative adversarial network","recurrent neural network","deep belief network","auto encoder","other algorithms","alzheimer\u2019s illness prediction","many publications","preprocessing methods","scaling","correction"]},{"title":"Enhancing Cardiovascular Health Monitoring Through IoT and Deep Learning Technologies","authors":["Huu-Hoa Nguyen","Tri-Thuc Vo"],"abstract":"Monitoring cardiovascular conditions is crucial in healthcare due to their significant impact on overall wellness and their role in mitigating heart-related diseases. To address this pressing issue, the research community has introduced various methodologies, among which deep learning approaches have shown notable effectiveness. Despite this potential, creating effective deep learning models tailored to time-series health data remains challenging. These challenges include processing vast amounts of data from IoT devices, building and selecting optimal deep learning models with appropriate parameters, and designing and implementing reliable systems for cardiovascular health monitoring. In response, our research introduces an advanced cardiovascular health monitoring system that takes advantage of wearable IoT and deep learning technologies to enhance healthcare. It features a multi-layered architecture, where each layer serves a specific function and integrates closely with the others. This integration enhances the system\u2019s overall functionality and reliability. The system efficiently integrates processes from health data collection through deep learning analysis to the delivery of timely health alerts. A critical feature of this system is the targeted deep learning model, selected from six potential algorithms based on experiments with data from IoT-enabled smartwatches. The selection process involves an in-depth evaluation of the models\u2019 performance, leading to the choice of the most effective model for system implementation. Our results highlight the system\u2019s effectiveness in monitoring cardiovascular health, underscoring its potential to enhance personalized healthcare, particularly for individuals with cardiovascular conditions, through advanced monitoring technologies.","doi":"10.1007\/s42979-024-02962-7","cleaned_title":"enhancing cardiovascular health monitoring through iot and deep learning technologies","cleaned_abstract":"monitoring cardiovascular conditions is crucial in healthcare due to their significant impact on overall wellness and their role in mitigating heart-related diseases. to address this pressing issue, the research community has introduced various methodologies, among which deep learning approaches have shown notable effectiveness. despite this potential, creating effective deep learning models tailored to time-series health data remains challenging. these challenges include processing vast amounts of data from iot devices, building and selecting optimal deep learning models with appropriate parameters, and designing and implementing reliable systems for cardiovascular health monitoring. in response, our research introduces an advanced cardiovascular health monitoring system that takes advantage of wearable iot and deep learning technologies to enhance healthcare. it features a multi-layered architecture, where each layer serves a specific function and integrates closely with the others. this integration enhances the system\u2019s overall functionality and reliability. the system efficiently integrates processes from health data collection through deep learning analysis to the delivery of timely health alerts. a critical feature of this system is the targeted deep learning model, selected from six potential algorithms based on experiments with data from iot-enabled smartwatches. the selection process involves an in-depth evaluation of the models\u2019 performance, leading to the choice of the most effective model for system implementation. our results highlight the system\u2019s effectiveness in monitoring cardiovascular health, underscoring its potential to enhance personalized healthcare, particularly for individuals with cardiovascular conditions, through advanced monitoring technologies.","key_phrases":["cardiovascular conditions","healthcare","their significant impact","overall wellness","their role","heart-related diseases","this pressing issue","the research community","various methodologies","which","deep learning approaches","notable effectiveness","this potential","effective deep learning models","time-series health data","these challenges","vast amounts","data","iot devices","optimal deep learning models","appropriate parameters","reliable systems","cardiovascular health monitoring","response","our research","an advanced cardiovascular health monitoring system","that","advantage","wearable iot and deep learning technologies","healthcare","it","a multi-layered architecture","each layer","a specific function","integrates","the others","this integration","the system\u2019s overall functionality","reliability","the system","processes","health data collection","deep learning analysis","the delivery","timely health alerts","a critical feature","this system","the targeted deep learning model","six potential algorithms","experiments","data","iot-enabled smartwatches","the selection process","-depth","the models\u2019 performance","the choice","the most effective model","system implementation","our results","the system\u2019s effectiveness","cardiovascular health","its potential","personalized healthcare","individuals","cardiovascular conditions","advanced monitoring technologies","six"]},{"title":"Image recognition algorithm based on hybrid deep learning","authors":["Tang Xiangdong"],"abstract":"As we embrace the information age, our lives have experienced revolutionary transformations. With the continuous advancement of computer technology, data sharing and information exchange have become increasingly robust. Consequently, the widespread adoption of hybrid deep learning algorithms has been prioritized. The emergence of deep learning is intricately linked to the progress of artificial intelligence, with deep learning serving as a tangible manifestation of AI. Deep learning algorithms are an important part of the field of robotics research and development. In image recognition, deep learning algorithms are playing an irreplaceable role. Based on the technological breakthrough of convolutional neural network, deep learning is unique in image recognition. In addition, aspects such as speech recognition, body function monitoring, and sports data analysis all have deep learning, and they are advancing all the way with unstoppable development momentum. Of course, with the current development of computer technology, image data also shines on its basis. Not only has it achieved a substantial surpass in quantity, but its types and styles are coming out in a variety of colors. On the basis of this series of developments, we are faced with a problem: traditional recognition algorithms cannot meet our next development requirements. Therefore, a new type of algorithm to deal with this problem has emerged: the research of image recognition algorithms based on hybrid deep learning. In the research of this article, we will focus on comparing its various algorithms to find out their advantages and disadvantages. So as to promote the further development of image recognition.","doi":"10.1007\/s13198-023-02134-5","cleaned_title":"image recognition algorithm based on hybrid deep learning","cleaned_abstract":"as we embrace the information age, our lives have experienced revolutionary transformations. with the continuous advancement of computer technology, data sharing and information exchange have become increasingly robust. consequently, the widespread adoption of hybrid deep learning algorithms has been prioritized. the emergence of deep learning is intricately linked to the progress of artificial intelligence, with deep learning serving as a tangible manifestation of ai. deep learning algorithms are an important part of the field of robotics research and development. in image recognition, deep learning algorithms are playing an irreplaceable role. based on the technological breakthrough of convolutional neural network, deep learning is unique in image recognition. in addition, aspects such as speech recognition, body function monitoring, and sports data analysis all have deep learning, and they are advancing all the way with unstoppable development momentum. of course, with the current development of computer technology, image data also shines on its basis. not only has it achieved a substantial surpass in quantity, but its types and styles are coming out in a variety of colors. on the basis of this series of developments, we are faced with a problem: traditional recognition algorithms cannot meet our next development requirements. therefore, a new type of algorithm to deal with this problem has emerged: the research of image recognition algorithms based on hybrid deep learning. in the research of this article, we will focus on comparing its various algorithms to find out their advantages and disadvantages. so as to promote the further development of image recognition.","key_phrases":["we","the information age","our lives","revolutionary transformations","the continuous advancement","computer technology","data sharing","information exchange","the widespread adoption","hybrid deep learning algorithms","the emergence","deep learning","the progress","artificial intelligence","deep learning","a tangible manifestation","ai","deep learning algorithms","an important part","the field","robotics research","development","image recognition","deep learning algorithms","an irreplaceable role","the technological breakthrough","convolutional neural network","deep learning","image recognition","addition","aspects","speech recognition","body function monitoring","sports data analysis","all","deep learning","they","unstoppable development momentum","course","the current development","computer technology","image data","its basis","it","a substantial surpass","quantity","its types","styles","a variety","colors","the basis","this series","developments","we","a problem","traditional recognition algorithms","our next development requirements","a new type","algorithm","this problem","the research","image recognition algorithms","hybrid deep learning","the research","this article","we","its various algorithms","their advantages","disadvantages","the further development","image recognition"]},{"title":"Deep Ensemble learning and quantum machine learning approach for Alzheimer\u2019s disease detection","authors":["Abebech Jenber Belay","Yelkal Mulualem Walle","Melaku Bitew Haile"],"abstract":"Alzheimer disease (AD) is among the most chronic neurodegenerative diseases that threaten global public health. The prevalence of Alzheimer disease and consequently the increased risk of spread all over the world pose a vital threat to human safekeeping. Early diagnosis of AD is a suitable action for timely intervention and medication, which may increase the prognosis and quality of life for affected individuals. Quantum computing provides a more efficient model for different disease classification tasks than classical machine learning approaches. The full potential of quantum computing is not applied to Alzheimer\u2019s disease classification tasks as expected. In this study, we proposed an ensemble deep learning model based on quantum machine learning classifiers to classify Alzheimer\u2019s disease. The Alzheimer\u2019s disease Neuroimaging Initiative I and Alzheimer\u2019s disease Neuroimaging Initiative II datasets are merged for the AD disease classification. We combined important features extracted based on the customized version of VGG16 and ResNet50 models from the merged images then feed these features to the Quantum Machine Learning classifier to classify them as non-demented, mild demented, moderate demented, and very mild demented. We evaluate the performance of our model by using six metrics; accuracy, the area under the curve, F1-score, precision, and recall. The result validates that the proposed model outperforms several state-of-the-art methods for detecting Alzheimer\u2019s disease by registering an accuracy of 99.89 and 98.37 F1-score.","doi":"10.1038\/s41598-024-61452-1","cleaned_title":"deep ensemble learning and quantum machine learning approach for alzheimer\u2019s disease detection","cleaned_abstract":"alzheimer disease (ad) is among the most chronic neurodegenerative diseases that threaten global public health. the prevalence of alzheimer disease and consequently the increased risk of spread all over the world pose a vital threat to human safekeeping. early diagnosis of ad is a suitable action for timely intervention and medication, which may increase the prognosis and quality of life for affected individuals. quantum computing provides a more efficient model for different disease classification tasks than classical machine learning approaches. the full potential of quantum computing is not applied to alzheimer\u2019s disease classification tasks as expected. in this study, we proposed an ensemble deep learning model based on quantum machine learning classifiers to classify alzheimer\u2019s disease. the alzheimer\u2019s disease neuroimaging initiative i and alzheimer\u2019s disease neuroimaging initiative ii datasets are merged for the ad disease classification. we combined important features extracted based on the customized version of vgg16 and resnet50 models from the merged images then feed these features to the quantum machine learning classifier to classify them as non-demented, mild demented, moderate demented, and very mild demented. we evaluate the performance of our model by using six metrics; accuracy, the area under the curve, f1-score, precision, and recall. the result validates that the proposed model outperforms several state-of-the-art methods for detecting alzheimer\u2019s disease by registering an accuracy of 99.89 and 98.37 f1-score.","key_phrases":["alzheimer disease","ad","the most chronic neurodegenerative diseases","that","global public health","the prevalence","alzheimer disease","consequently the increased risk","spread","the world","a vital threat","human safekeeping","early diagnosis","ad","a suitable action","timely intervention","medication","which","the prognosis","quality","life","affected individuals","quantum computing","a more efficient model","different disease classification tasks","classical machine learning approaches","the full potential","quantum computing","alzheimer\u2019s disease classification tasks","this study","we","an ensemble deep learning model","quantum machine learning classifiers","alzheimer\u2019s disease","i","alzheimer","disease","neuroimaging initiative ii datasets","the ad disease classification","we","important features","the customized version","vgg16 and resnet50 models","the merged images","these features","the quantum machine","classifier","them","we","the performance","our model","six metrics","accuracy","the area","the curve","f1-score","precision","the result","the proposed model","the-art","alzheimer\u2019s disease","an accuracy","98.37 f1-score","quantum","quantum","quantum","resnet50","quantum","six","99.89","98.37"]},{"title":"A Review on Machine Learning and Deep Learning Based Systems for the Diagnosis of Brain Cancer","authors":["Prottoy Saha","Shanta Kumar Das","Rudra Das"],"abstract":"Brain cancer is a disease of the brain caused by a brain tumor. A brain tumor is the development of cells in the brain that grow in an unregulated and unnatural manner. Patients may suffer irreversible brain damage or even death if these tumors are not detected and treated properly. As with all types of treatment, Positional information and tumor size are critical for conventional systems. Thus, establishing a meticulous and automated approach to providing information to medical practitioners is required. With machine learning, deep learning, and several imaging modalities, physicians may now more reliably detect tumor types in a shorter period. The paper aims to provide an overview of newly developed systems that use machine learning and deep learning approaches to analyze various medical imaging modalities in the case of diagnosing brain tumors. Datasets used by the authors, dataset partitioning strategies, and different performance evaluation matrices are also described in this paper. To better understand the policy of categorization, we propose a taxonomy here where we have categorized deep learning and machine learning based systems with respect to single classifier, multiple classifiers, single dataset and multiple dataset. Finally, we focus on the challenges of deep learning algorithms for brain tumor classification and possible future trends in this field.","doi":"10.1007\/s42979-023-02360-5","cleaned_title":"a review on machine learning and deep learning based systems for the diagnosis of brain cancer","cleaned_abstract":"brain cancer is a disease of the brain caused by a brain tumor. a brain tumor is the development of cells in the brain that grow in an unregulated and unnatural manner. patients may suffer irreversible brain damage or even death if these tumors are not detected and treated properly. as with all types of treatment, positional information and tumor size are critical for conventional systems. thus, establishing a meticulous and automated approach to providing information to medical practitioners is required. with machine learning, deep learning, and several imaging modalities, physicians may now more reliably detect tumor types in a shorter period. the paper aims to provide an overview of newly developed systems that use machine learning and deep learning approaches to analyze various medical imaging modalities in the case of diagnosing brain tumors. datasets used by the authors, dataset partitioning strategies, and different performance evaluation matrices are also described in this paper. to better understand the policy of categorization, we propose a taxonomy here where we have categorized deep learning and machine learning based systems with respect to single classifier, multiple classifiers, single dataset and multiple dataset. finally, we focus on the challenges of deep learning algorithms for brain tumor classification and possible future trends in this field.","key_phrases":["brain cancer","a disease","the brain","a brain tumor","a brain tumor","the development","cells","the brain","that","an unregulated and unnatural manner","patients","irreversible brain damage","even death","these tumors","all types","treatment","positional information","tumor size","conventional systems","a meticulous and automated approach","information","medical practitioners","machine learning","deep learning","several imaging modalities","physicians","tumor types","a shorter period","the paper","an overview","newly developed systems","that","machine learning","approaches","various medical imaging modalities","the case","brain tumors","datasets","the authors","partitioning strategies","different performance evaluation matrices","this paper","the policy","categorization","we","a taxonomy","we","deep learning and machine learning based systems","respect","single classifier","multiple classifiers","single dataset","multiple dataset","we","the challenges","deep learning algorithms","brain tumor classification","possible future trends","this field"]},{"title":"Ensemble deep learning for Alzheimer\u2019s disease characterization and estimation","authors":["M. Tanveer","T. Goel","R. Sharma","A. K. Malik","I. Beheshti","J. Del Ser","P. N. Suganthan","C. T. Lin"],"abstract":"Alzheimer\u2019s disease, which is characterized by a continual deterioration of cognitive abilities in older people, is the most common form of dementia. Neuroimaging data, for example, from magnetic resonance imaging and positron emission tomography, enable identification of the structural and functional changes caused by Alzheimer\u2019s disease in the brain. Diagnosing Alzheimer\u2019s disease is critical in medical settings, as it supports early intervention and treatment planning and contributes to expanding our knowledge of the dynamics of Alzheimer\u2019s disease in the brain. Lately, ensemble deep learning has become popular for enhancing the performance and reliability of Alzheimer\u2019s disease diagnosis. These models combine several deep neural networks to increase a prediction\u2019s robustness. Here we revisit key developments of ensemble deep learning, connecting its design\u2014the type of ensemble, its heterogeneity and data modalities\u2014with its application to AD diagnosis using neuroimaging and genetic data. Trends and challenges are discussed thoroughly to assess where our knowledge in this area stands.","doi":"10.1038\/s44220-024-00237-x","cleaned_title":"ensemble deep learning for alzheimer\u2019s disease characterization and estimation","cleaned_abstract":"alzheimer\u2019s disease, which is characterized by a continual deterioration of cognitive abilities in older people, is the most common form of dementia. neuroimaging data, for example, from magnetic resonance imaging and positron emission tomography, enable identification of the structural and functional changes caused by alzheimer\u2019s disease in the brain. diagnosing alzheimer\u2019s disease is critical in medical settings, as it supports early intervention and treatment planning and contributes to expanding our knowledge of the dynamics of alzheimer\u2019s disease in the brain. lately, ensemble deep learning has become popular for enhancing the performance and reliability of alzheimer\u2019s disease diagnosis. these models combine several deep neural networks to increase a prediction\u2019s robustness. here we revisit key developments of ensemble deep learning, connecting its design\u2014the type of ensemble, its heterogeneity and data modalities\u2014with its application to ad diagnosis using neuroimaging and genetic data. trends and challenges are discussed thoroughly to assess where our knowledge in this area stands.","key_phrases":["alzheimer\u2019s disease","which","a continual deterioration","cognitive abilities","older people","the most common form","dementia","neuroimaging data","example","magnetic resonance imaging","positron emission tomography","identification","the structural and functional changes","alzheimer\u2019s disease","the brain","alzheimer\u2019s disease","medical settings","it","early intervention","treatment planning","our knowledge","the dynamics","alzheimer\u2019s disease","the brain","ensemble deep learning","the performance","reliability","alzheimer\u2019s disease diagnosis","these models","several deep neural networks","a prediction\u2019s robustness","we","key developments","ensemble deep learning","its design","the type","ensemble","its heterogeneity and data modalities","its application","ad diagnosis","neuroimaging and genetic data","trends","challenges","our knowledge","this area"]},{"title":"Prediction of crop yield in India using machine learning and hybrid deep learning models","authors":["Krithikha Sanju Saravanan","Velammal Bhagavathiappan"],"abstract":"Crop yield prediction is one of the burgeoning research areas in the agriculture domain. The crop yield forecasting models are developed to enhance productivity with improved decision-making strategies. The highly efficient crop yield forecasting model assists farmers in determining when, what and how much to plant on their cultivable land. The main objective of the proposed research work is to build a high efficacious crop yield prediction model based on the data available for the period of 21\u00a0years from 1997 to 2017 using machine learning and hybrid deep learning approaches. Two prediction models have been proposed in this research work to predict the crop yield accurately. The first model is a machine learning-based model which uses the CatBoost regression model and its hyperparameters are tuned which improves the performance of the yield prediction using the Optuna framework. The second model is the hybrid deep learning model which uses spatio-temporal attention-based convolutional neural network (STACNN) for extracting the features and the bidirectional long short-term memory (BiLSTM) model for predicting the crop yield effectively. The proposed models are evaluated using the error metrics and compared with the latest contemporary models. From the evaluation results, it is shown that the proposed models significantly outperform all other existing models and CatBoost regression model slightly performs better than the STACNN-BiLSTM model, with the R-squared value of 0.99.","doi":"10.1007\/s11600-024-01312-8","cleaned_title":"prediction of crop yield in india using machine learning and hybrid deep learning models","cleaned_abstract":"crop yield prediction is one of the burgeoning research areas in the agriculture domain. the crop yield forecasting models are developed to enhance productivity with improved decision-making strategies. the highly efficient crop yield forecasting model assists farmers in determining when, what and how much to plant on their cultivable land. the main objective of the proposed research work is to build a high efficacious crop yield prediction model based on the data available for the period of 21 years from 1997 to 2017 using machine learning and hybrid deep learning approaches. two prediction models have been proposed in this research work to predict the crop yield accurately. the first model is a machine learning-based model which uses the catboost regression model and its hyperparameters are tuned which improves the performance of the yield prediction using the optuna framework. the second model is the hybrid deep learning model which uses spatio-temporal attention-based convolutional neural network (stacnn) for extracting the features and the bidirectional long short-term memory (bilstm) model for predicting the crop yield effectively. the proposed models are evaluated using the error metrics and compared with the latest contemporary models. from the evaluation results, it is shown that the proposed models significantly outperform all other existing models and catboost regression model slightly performs better than the stacnn-bilstm model, with the r-squared value of 0.99.","key_phrases":["crop yield prediction","the burgeoning research areas","the agriculture domain","the crop yield forecasting models","productivity","improved decision-making strategies","the highly efficient crop yield forecasting model","farmers","their cultivable land","the main objective","the proposed research work","a high efficacious crop yield prediction model","the data","the period","21 years","machine learning","hybrid deep learning approaches","two prediction models","this research work","the first model","a machine learning-based model","which","the catboost regression model","which","the performance","the yield prediction","the optuna framework","the second model","the hybrid deep learning model","which","spatio-temporal attention-based convolutional neural network","stacnn","the features","the bidirectional long short-term memory (bilstm) model","the crop yield","the proposed models","the error metrics","the latest contemporary models","the evaluation results","it","the proposed models","all other existing models","catboost regression model","the stacnn-bilstm model","the r-squared value","the period of 21 years from 1997 to 2017","two","first","second","stacnn","stacnn","0.99"]},{"title":"Deep learning CT reconstruction improves liver metastases detection","authors":["Achraf Kanan","Bruno Pereira","Constance Hordonneau","Lucie Cassagnes","El\u00e9onore Pouget","L\u00e9on Appolinaire Tianhoun","Beno\u00eet Chauveau","Beno\u00eet Magnin"],"abstract":"ObjectivesDetection of liver metastases is crucial for guiding oncological management. Computed tomography through iterative reconstructions is widely used in this indication but has certain limitations. Deep learning image reconstructions (DLIR) use deep neural networks to achieve a significant noise reduction compared to iterative reconstructions. While reports have demonstrated improvements in image quality, their impact on liver metastases detection remains unclear. Our main objective was to determine whether DLIR affects the number of detected liver metastasis. Our secondary objective was to compare metastases conspicuity between the two reconstruction methods.MethodsCT images of 121 patients with liver metastases were reconstructed using a 50% adaptive statistical iterative reconstruction (50%-ASiR-V), and three levels of DLIR (DLIR-low, DLIR-medium, and DLIR-high). For each reconstruction, two double-blinded radiologists counted up to a maximum of ten metastases. Visibility and contour definitions were also assessed. Comparisons between methods for continuous parameters were performed using mixed models.ResultsA higher number of metastases was detected by one reader with DLIR-high: 7 (2\u201310) (median (Q\u2081\u2013Q\u2083); total 733) versus 5 (2\u201310), respectively for DLIR-medium, DLIR-low, and ASiR-V (p\u2009<\u20090.001). Ten patents were detected with more metastases with DLIR-high simultaneously by both readers and a third reader for confirmation. Metastases visibility and contour definition were better with DLIR than ASiR-V.ConclusionDLIR-high enhanced the detection and visibility of liver metastases compared to ASiR-V, and also increased the number of liver metastases detected.Critical relevance statementDeep learning-based reconstruction at high strength allowed an increase in liver metastases detection compared to hybrid iterative reconstruction and can be used in clinical oncology imaging to help overcome the limitations of CT.Key Points\n\nDetection of liver metastases is crucial but limited with standard CT reconstructions.\n\n\nMore liver metastases were detected with deep-learning CT reconstruction compared to iterative reconstruction.\n\n\nDeep learning reconstructions are suitable for hepatic metastases staging and follow-up.\n\nGraphical Abstract","doi":"10.1186\/s13244-024-01753-1","cleaned_title":"deep learning ct reconstruction improves liver metastases detection","cleaned_abstract":"objectivesdetection of liver metastases is crucial for guiding oncological management. computed tomography through iterative reconstructions is widely used in this indication but has certain limitations. deep learning image reconstructions (dlir) use deep neural networks to achieve a significant noise reduction compared to iterative reconstructions. while reports have demonstrated improvements in image quality, their impact on liver metastases detection remains unclear. our main objective was to determine whether dlir affects the number of detected liver metastasis. our secondary objective was to compare metastases conspicuity between the two reconstruction methods.methodsct images of 121 patients with liver metastases were reconstructed using a 50% adaptive statistical iterative reconstruction (50%-asir-v), and three levels of dlir (dlir-low, dlir-medium, and dlir-high). for each reconstruction, two double-blinded radiologists counted up to a maximum of ten metastases. visibility and contour definitions were also assessed. comparisons between methods for continuous parameters were performed using mixed models.resultsa higher number of metastases was detected by one reader with dlir-high: 7 (2\u201310) (median (q\u2081\u2013q\u2083); total 733) versus 5 (2\u201310), respectively for dlir-medium, dlir-low, and asir-v (p < 0.001). ten patents were detected with more metastases with dlir-high simultaneously by both readers and a third reader for confirmation. metastases visibility and contour definition were better with dlir than asir-v.conclusiondlir-high enhanced the detection and visibility of liver metastases compared to asir-v, and also increased the number of liver metastases detected.critical relevance statementdeep learning-based reconstruction at high strength allowed an increase in liver metastases detection compared to hybrid iterative reconstruction and can be used in clinical oncology imaging to help overcome the limitations of ct.key points detection of liver metastases is crucial but limited with standard ct reconstructions. more liver metastases were detected with deep-learning ct reconstruction compared to iterative reconstruction. deep learning reconstructions are suitable for hepatic metastases staging and follow-up. graphical abstract","key_phrases":["objectivesdetection","liver metastases","oncological management","computed tomography","iterative reconstructions","this indication","certain limitations","image reconstructions","dlir","deep neural networks","a significant noise reduction","iterative reconstructions","reports","improvements","image quality","their impact","liver metastases detection","our main objective","dlir","the number","detected liver metastasis","our secondary objective","metastases","conspicuity","the two reconstruction methods.methodsct images","121 patients","liver metastases","a 50% adaptive statistical iterative reconstruction","three levels","dlir","each reconstruction","two double-blinded radiologists","a maximum","ten metastases","visibility","contour definitions","comparisons","methods","continuous parameters","mixed models.resultsa higher number","metastases","one reader","median","(q\u2081\u2013q\u2083","p","ten patents","more metastases","both readers","a third reader","confirmation","metastases visibility","contour definition","dlir","the detection","visibility","liver metastases","asir-v","the number","liver metastases","detected.critical relevance","high strength","an increase","liver metastases detection","hybrid iterative reconstruction","clinical oncology imaging","the limitations","ct.key","points detection","liver metastases","standard ct reconstructions","more liver metastases","deep-learning ct reconstruction","iterative reconstruction","deep learning reconstructions","hepatic metastases","follow-up","graphical abstract","two","121","50%","50%-asir","three","two","ten","one","7","2\u201310","733","5","2\u201310","ten","third","ct.key"]},{"title":"Predicting Renal Toxicity of Compounds with Deep Learning and Machine Learning Methods","authors":["Bitopan Mazumdar","Pankaj Kumar Deva Sarma","Hridoy Jyoti Mahanta"],"abstract":"Renal toxicity prediction plays a vital role in drug discovery and clinical practice, as it helps to identify potentially harmful compounds and mitigate adverse effects on the renal system. Compound with inherent renal-toxic potential is one of the major concerns for drug development as it leads to failure in drug discovery. Predicting nephrotoxic probabilities of a compound at an early stage can be effective for reducing the drug failure rate. Hence, it is crucial to develop a mechanism to analyze the renal toxicity of a drug-candidate optimally and quickly. To mitigate the risks associated with renal toxicity, predictive models leveraging machine learning and deep learning techniques have gained significant attention. In this study, 287 human renal-toxic drugs and 278 non-renal-toxic drugs were collected to develop a deep learning model and 27 machine learning models using 8 kinds of fingerprints and Rdkit descriptors. The deep neural network model shows better generalization scores on five-fold cross-validation and Extra-tree model shows better performance score on test data. Structural alerts, specific chemical substructures associated with renal toxicity, offer a valuable tool for early toxicity assessment. Therefore, the substructures of renal toxic compounds were studied by applying association rule mining technique based on frequent itemset patterns. A method has been proposed for generating structural alerts and 10 structural alerts have been generated using the method.","doi":"10.1007\/s42979-023-02258-2","cleaned_title":"predicting renal toxicity of compounds with deep learning and machine learning methods","cleaned_abstract":"renal toxicity prediction plays a vital role in drug discovery and clinical practice, as it helps to identify potentially harmful compounds and mitigate adverse effects on the renal system. compound with inherent renal-toxic potential is one of the major concerns for drug development as it leads to failure in drug discovery. predicting nephrotoxic probabilities of a compound at an early stage can be effective for reducing the drug failure rate. hence, it is crucial to develop a mechanism to analyze the renal toxicity of a drug-candidate optimally and quickly. to mitigate the risks associated with renal toxicity, predictive models leveraging machine learning and deep learning techniques have gained significant attention. in this study, 287 human renal-toxic drugs and 278 non-renal-toxic drugs were collected to develop a deep learning model and 27 machine learning models using 8 kinds of fingerprints and rdkit descriptors. the deep neural network model shows better generalization scores on five-fold cross-validation and extra-tree model shows better performance score on test data. structural alerts, specific chemical substructures associated with renal toxicity, offer a valuable tool for early toxicity assessment. therefore, the substructures of renal toxic compounds were studied by applying association rule mining technique based on frequent itemset patterns. a method has been proposed for generating structural alerts and 10 structural alerts have been generated using the method.","key_phrases":["renal toxicity prediction","a vital role","drug discovery","clinical practice","it","potentially harmful compounds","adverse effects","the renal system","inherent renal-toxic potential","the major concerns","drug development","it","failure","drug discovery","nephrotoxic probabilities","a compound","an early stage","the drug failure rate","it","a mechanism","the renal toxicity","a drug-candidate","the risks","renal toxicity","predictive models","machine learning","deep learning techniques","significant attention","this study","287 human renal-toxic drugs","278 non-renal-toxic drugs","a deep learning model","27 machine learning models","8 kinds","fingerprints","rdkit descriptors","the deep neural network model","better generalization scores","five-fold cross-validation and extra-tree model","better performance score","test data","structural alerts","specific chemical substructures","renal toxicity","a valuable tool","early toxicity assessment","the substructures","renal toxic compounds","association rule mining technique","frequent itemset patterns","a method","structural alerts","10 structural alerts","the method","287","278","27","8","five-fold","10"]},{"title":"A deep ensemble learning method for cherry classification","authors":["Kiyas Kayaalp"],"abstract":"In many agricultural products, information technologies are utilized in classification processes at the desired quality. It is undesirable to mix different types of cherries, especially in export-type cherries. In this study on cherries, one of the important export products of Turkey, the classification of cherry species was carried out with ensemble learning methods. In this study, a new dataset consisting of 3570 images of seven different cherry species grown in Isparta region was created. The generated new dataset was trained with six different deep learning models with pre-learning on the original and incremental dataset. As a result of the training with incremental data, the best result was obtained from the DenseNet169 model with an accuracy of 99.57%. The two deep learning models with the best results were transferred to ensemble learning and a 100% accuracy rate was obtained with the Maximum Voting model.","doi":"10.1007\/s00217-024-04490-3","cleaned_title":"a deep ensemble learning method for cherry classification","cleaned_abstract":"in many agricultural products, information technologies are utilized in classification processes at the desired quality. it is undesirable to mix different types of cherries, especially in export-type cherries. in this study on cherries, one of the important export products of turkey, the classification of cherry species was carried out with ensemble learning methods. in this study, a new dataset consisting of 3570 images of seven different cherry species grown in isparta region was created. the generated new dataset was trained with six different deep learning models with pre-learning on the original and incremental dataset. as a result of the training with incremental data, the best result was obtained from the densenet169 model with an accuracy of 99.57%. the two deep learning models with the best results were transferred to ensemble learning and a 100% accuracy rate was obtained with the maximum voting model.","key_phrases":["many agricultural products","information technologies","classification processes","the desired quality","it","different types","cherries","export-type cherries","this study","cherries","the important export products","turkey","the classification","cherry species","ensemble learning methods","this study","a new dataset","3570 images","seven different cherry species","isparta region","the generated new dataset","six different deep learning models","pre","the original and incremental dataset","a result","the training","incremental data","the best result","the densenet169 model","an accuracy","99.57%","the two deep learning models","the best results","ensemble learning","a 100% accuracy rate","the maximum voting model","one","3570","seven","isparta","six","99.57%","two","100%"]},{"title":"The application of deep learning technology in integrated circuit design","authors":["Lihua Dai","Ben Wang","Xuemin Cheng","Qin Wang","Xinsen Ni"],"abstract":"This study addresses the intricate challenge of circuit layout optimization central to integrated circuit (IC) design, where the primary goals involve attaining an optimal balance among power consumption, performance metrics, and chip area (collectively known as PPA optimization). The complexity of this task, evolving into a multidimensional problem under multiple constraints, necessitates the exploration of advanced methodologies. In response to these challenges, our research introduces deep learning technology as an innovative strategy to revolutionize circuit layout optimization. Specifically, we employ Convolutional Neural Networks (CNNs) in developing an optimized layout strategy, a performance prediction model, and a system for fault detection and real-time monitoring. These methodologies leverage the capacity of deep learning models to learn from high-dimensional data representations and handle multiple constraints effectively. Extensive case studies and rigorous experimental validations demonstrate the efficacy of our proposed deep learning-driven approaches. The results highlight significant enhancements in optimization efficiency, with an average power consumption reduction of 120% and latency decrease by 1.5%. Furthermore, the predictive capabilities are markedly improved, evidenced by a reduction in the average absolute error for power predictions to 3%. Comparative analyses conclusively illustrate the superiority of deep learning methodologies over conventional techniques across several dimensions. Our findings underscore the potential of deep learning in achieving higher accuracy in predictions, demonstrating stronger generalization abilities, facilitating superior design quality, and ultimately enhancing user satisfaction. These advancements not only validate the applicability of deep learning in IC design optimization but also pave the way for future advancements in addressing the multidimensional challenges inherent to circuit layout optimization.","doi":"10.1186\/s42162-024-00380-w","cleaned_title":"the application of deep learning technology in integrated circuit design","cleaned_abstract":"this study addresses the intricate challenge of circuit layout optimization central to integrated circuit (ic) design, where the primary goals involve attaining an optimal balance among power consumption, performance metrics, and chip area (collectively known as ppa optimization). the complexity of this task, evolving into a multidimensional problem under multiple constraints, necessitates the exploration of advanced methodologies. in response to these challenges, our research introduces deep learning technology as an innovative strategy to revolutionize circuit layout optimization. specifically, we employ convolutional neural networks (cnns) in developing an optimized layout strategy, a performance prediction model, and a system for fault detection and real-time monitoring. these methodologies leverage the capacity of deep learning models to learn from high-dimensional data representations and handle multiple constraints effectively. extensive case studies and rigorous experimental validations demonstrate the efficacy of our proposed deep learning-driven approaches. the results highlight significant enhancements in optimization efficiency, with an average power consumption reduction of 120% and latency decrease by 1.5%. furthermore, the predictive capabilities are markedly improved, evidenced by a reduction in the average absolute error for power predictions to 3%. comparative analyses conclusively illustrate the superiority of deep learning methodologies over conventional techniques across several dimensions. our findings underscore the potential of deep learning in achieving higher accuracy in predictions, demonstrating stronger generalization abilities, facilitating superior design quality, and ultimately enhancing user satisfaction. these advancements not only validate the applicability of deep learning in ic design optimization but also pave the way for future advancements in addressing the multidimensional challenges inherent to circuit layout optimization.","key_phrases":["this study","the intricate challenge","circuit layout optimization","integrated circuit (ic) design","the primary goals","an optimal balance","power consumption","performance metrics","chip area","ppa optimization","the complexity","this task","a multidimensional problem","multiple constraints","the exploration","advanced methodologies","response","these challenges","our research","deep learning technology","an innovative strategy","circuit layout optimization","we","convolutional neural networks","cnns","an optimized layout strategy","a performance prediction model","a system","fault detection","real-time monitoring","these methodologies","the capacity","deep learning models","high-dimensional data representations","multiple constraints","extensive case studies","rigorous experimental validations","the efficacy","our proposed deep learning-driven approaches","the results","significant enhancements","optimization efficiency","an average power consumption reduction","120% and latency decrease","1.5%","the predictive capabilities","a reduction","the average absolute error","power predictions","3%","comparative analyses","the superiority","deep learning methodologies","conventional techniques","several dimensions","our findings","the potential","deep learning","higher accuracy","predictions","stronger generalization abilities","superior design quality","user satisfaction","these advancements","the applicability","deep learning","design optimization","the way","future advancements","the multidimensional challenges","layout optimization","120%","1.5%","3%"]},{"title":"Alzheimer Disease Detection Using MRI: Deep Learning Review","authors":["Pallavi Saikia","Sanjib Kumar Kalita"],"abstract":"Deep learning for Alzheimer disease detection using MRI is an emerging area of research in medical image processing. With the advent of new technologies based on methods of Deep Learning, medical diagnosis of certain diseases has become possible. Alzheimer\u2019s is a disease which till date has no cure but the progression of the disease can be slowed down or a person who might develop Alzheimer in future can be treated if early prediction of it is possible. Early prediction of the disease benefits medical professionals a lot for the correct diagnosis. Medical professionals label Alzheimer patients based on the progression of the disease as AD (Alzheimer\u2019s), CN (cognitive impairment) and MCI (mild cognitive impairment). In literature many Deep Learning models are used for the early detection of Alzheimer\u2019s disease. Though there are many image modalities, MRI images being non-invasive are considered best for these types of medical experiments. In the present study, we have studied the evolution of Alzheimer\u2019s disease over time, research gaps, challenges towards building advanced models, possible recommendations to overcome those challenges and determining the best performance model. We have focussed on an exhaustive and comprehensive survey of very deep learning-based research papers on Alzheimer\u2019s disease detection. The present work will benefit researchers by providing a clear direction for future scope in Alzheimer disease detection and analysis.","doi":"10.1007\/s42979-024-02868-4","cleaned_title":"alzheimer disease detection using mri: deep learning review","cleaned_abstract":"deep learning for alzheimer disease detection using mri is an emerging area of research in medical image processing. with the advent of new technologies based on methods of deep learning, medical diagnosis of certain diseases has become possible. alzheimer\u2019s is a disease which till date has no cure but the progression of the disease can be slowed down or a person who might develop alzheimer in future can be treated if early prediction of it is possible. early prediction of the disease benefits medical professionals a lot for the correct diagnosis. medical professionals label alzheimer patients based on the progression of the disease as ad (alzheimer\u2019s), cn (cognitive impairment) and mci (mild cognitive impairment). in literature many deep learning models are used for the early detection of alzheimer\u2019s disease. though there are many image modalities, mri images being non-invasive are considered best for these types of medical experiments. in the present study, we have studied the evolution of alzheimer\u2019s disease over time, research gaps, challenges towards building advanced models, possible recommendations to overcome those challenges and determining the best performance model. we have focussed on an exhaustive and comprehensive survey of very deep learning-based research papers on alzheimer\u2019s disease detection. the present work will benefit researchers by providing a clear direction for future scope in alzheimer disease detection and analysis.","key_phrases":["deep learning","alzheimer disease detection","mri","an emerging area","research","medical image processing","the advent","new technologies","methods","deep learning","medical diagnosis","certain diseases","alzheimer\u2019s","a disease","which","date","no cure","the progression","the disease","a person","who","alzheimer","future","early prediction","it","early prediction","medical professionals","the correct diagnosis","medical professionals","alzheimer patients","the progression","the disease","ad","alzheimer","(mild cognitive impairment","literature","many deep learning models","the early detection","alzheimer\u2019s disease","many image modalities","mri images","these types","medical experiments","the present study","we","the evolution","alzheimer\u2019s disease","time","advanced models","possible recommendations","those challenges","the best performance model","we","an exhaustive and comprehensive survey","very deep learning-based research papers","alzheimer\u2019s disease detection","the present work","researchers","a clear direction","future scope","alzheimer disease detection","analysis","mci"]},{"title":"Enhancing Cardiovascular Health Monitoring Through IoT and Deep Learning Technologies","authors":["Huu-Hoa Nguyen","Tri-Thuc Vo"],"abstract":"Monitoring cardiovascular conditions is crucial in healthcare due to their significant impact on overall wellness and their role in mitigating heart-related diseases. To address this pressing issue, the research community has introduced various methodologies, among which deep learning approaches have shown notable effectiveness. Despite this potential, creating effective deep learning models tailored to time-series health data remains challenging. These challenges include processing vast amounts of data from IoT devices, building and selecting optimal deep learning models with appropriate parameters, and designing and implementing reliable systems for cardiovascular health monitoring. In response, our research introduces an advanced cardiovascular health monitoring system that takes advantage of wearable IoT and deep learning technologies to enhance healthcare. It features a multi-layered architecture, where each layer serves a specific function and integrates closely with the others. This integration enhances the system\u2019s overall functionality and reliability. The system efficiently integrates processes from health data collection through deep learning analysis to the delivery of timely health alerts. A critical feature of this system is the targeted deep learning model, selected from six potential algorithms based on experiments with data from IoT-enabled smartwatches. The selection process involves an in-depth evaluation of the models\u2019 performance, leading to the choice of the most effective model for system implementation. Our results highlight the system\u2019s effectiveness in monitoring cardiovascular health, underscoring its potential to enhance personalized healthcare, particularly for individuals with cardiovascular conditions, through advanced monitoring technologies.","doi":"10.1007\/s42979-024-02962-7","cleaned_title":"enhancing cardiovascular health monitoring through iot and deep learning technologies","cleaned_abstract":"monitoring cardiovascular conditions is crucial in healthcare due to their significant impact on overall wellness and their role in mitigating heart-related diseases. to address this pressing issue, the research community has introduced various methodologies, among which deep learning approaches have shown notable effectiveness. despite this potential, creating effective deep learning models tailored to time-series health data remains challenging. these challenges include processing vast amounts of data from iot devices, building and selecting optimal deep learning models with appropriate parameters, and designing and implementing reliable systems for cardiovascular health monitoring. in response, our research introduces an advanced cardiovascular health monitoring system that takes advantage of wearable iot and deep learning technologies to enhance healthcare. it features a multi-layered architecture, where each layer serves a specific function and integrates closely with the others. this integration enhances the system\u2019s overall functionality and reliability. the system efficiently integrates processes from health data collection through deep learning analysis to the delivery of timely health alerts. a critical feature of this system is the targeted deep learning model, selected from six potential algorithms based on experiments with data from iot-enabled smartwatches. the selection process involves an in-depth evaluation of the models\u2019 performance, leading to the choice of the most effective model for system implementation. our results highlight the system\u2019s effectiveness in monitoring cardiovascular health, underscoring its potential to enhance personalized healthcare, particularly for individuals with cardiovascular conditions, through advanced monitoring technologies.","key_phrases":["cardiovascular conditions","healthcare","their significant impact","overall wellness","their role","heart-related diseases","this pressing issue","the research community","various methodologies","which","deep learning approaches","notable effectiveness","this potential","effective deep learning models","time-series health data","these challenges","vast amounts","data","iot devices","optimal deep learning models","appropriate parameters","reliable systems","cardiovascular health monitoring","response","our research","an advanced cardiovascular health monitoring system","that","advantage","wearable iot and deep learning technologies","healthcare","it","a multi-layered architecture","each layer","a specific function","integrates","the others","this integration","the system\u2019s overall functionality","reliability","the system","processes","health data collection","deep learning analysis","the delivery","timely health alerts","a critical feature","this system","the targeted deep learning model","six potential algorithms","experiments","data","iot-enabled smartwatches","the selection process","-depth","the models\u2019 performance","the choice","the most effective model","system implementation","our results","the system\u2019s effectiveness","cardiovascular health","its potential","personalized healthcare","individuals","cardiovascular conditions","advanced monitoring technologies","six"]},{"title":"Medical Image Analysis Through Deep Learning Techniques: A Comprehensive Survey","authors":["K. Balasamy","V. Seethalakshmi","S. Suganyadevi"],"abstract":"Deep learning has been the subject of a significant amount of research interest in the development of novel algorithms for deep learning algorithms and medical image processing have proven very effective in a number of medical imaging tasks to help illness identification and diagnosis. The shortage of large-sized datasets that are also adequately annotated is a key barrier that is preventing the continued advancement of deep learning models used in medical image analysis, despite the effectiveness of these models. Over the course of the previous 5\u00a0years, a great number of research have concentrated on finding solutions to this problem. In this work, we present a complete overview of the use of deep learning techniques in a variety of medical image analysis tasks by reviewing and summarizing the current research that have been conducted in this area. In particular, we place an emphasis on the most recent developments and contributions of state-of-the-art semi-supervised and unsupervised deep learning in medical image analysis. These advancements and contributions are shortened based on various application scenarios, which include image registration, segmentation, classification and detection. In addition to this, we explore the significant technological obstacles that lie ahead and provide some potential answers for the ongoing study.","doi":"10.1007\/s11277-024-11428-1","cleaned_title":"medical image analysis through deep learning techniques: a comprehensive survey","cleaned_abstract":"deep learning has been the subject of a significant amount of research interest in the development of novel algorithms for deep learning algorithms and medical image processing have proven very effective in a number of medical imaging tasks to help illness identification and diagnosis. the shortage of large-sized datasets that are also adequately annotated is a key barrier that is preventing the continued advancement of deep learning models used in medical image analysis, despite the effectiveness of these models. over the course of the previous 5 years, a great number of research have concentrated on finding solutions to this problem. in this work, we present a complete overview of the use of deep learning techniques in a variety of medical image analysis tasks by reviewing and summarizing the current research that have been conducted in this area. in particular, we place an emphasis on the most recent developments and contributions of state-of-the-art semi-supervised and unsupervised deep learning in medical image analysis. these advancements and contributions are shortened based on various application scenarios, which include image registration, segmentation, classification and detection. in addition to this, we explore the significant technological obstacles that lie ahead and provide some potential answers for the ongoing study.","key_phrases":["deep learning","the subject","a significant amount","research interest","the development","novel algorithms","deep learning algorithms","medical image processing","a number","medical imaging tasks","identification","diagnosis","the shortage","large-sized datasets","that","a key barrier","that","the continued advancement","deep learning models","medical image analysis","the effectiveness","these models","the course","the previous 5 years","a great number","research","solutions","this problem","this work","we","a complete overview","the use","deep learning techniques","a variety","medical image analysis tasks","the current research","that","this area","we","an emphasis","the most recent developments","contributions","state","the-art","deep learning","medical image analysis","these advancements","contributions","various application scenarios","which","image registration","segmentation","classification","detection","addition","this","we","the significant technological obstacles","that","some potential answers","the ongoing study","the previous 5 years"]},{"title":"Deep-SEA: a deep learning based patient specific multi-modality post-cancer survival estimation architecture","authors":["Ibtihaj Ahmad","Saleem Riaz"],"abstract":"Cancer survival estimation is essential for post-cancer patient care, cancer management policy building, and the development of tailored treatment plans. Existing survival estimation methods use censored data; therefore, standard machine learning methods can not be used directly. Some censoring-based semi-machine learning methods have recently been proposed; however, these methods pose challenges. They are less patient-specific and non-linear. Furthermore, they rely on single-modality features. These drawbacks result in lower survival estimation performance. To address these issues, this work proposes a framework named Deep-SEA. Compared to the state-of-the-art, Deep-SEA uses multi-modality features, i.e., clinical, radiology, and histology features. These features are analyzed with statistical methods, and only significant features are selected. Then, the baseline hazard of the Cox model is estimated using Breslow\u2019s estimator, which is optimized using stochastic gradient descent. Finally, the risk function, i.e., the parameters of our model, are estimated via an ANN with time as additional input. ANN makes it non-linear while training on the patient-specific features makes it more patient-specific than the state-of-the-art. We train and evaluate Deep-SEA on five datasets, including head, neck, and colorectal-liver cancer. We have achieved a Concordance-index (C-index) score of 0.7181, the highest compared to the state-of-the-art. Results and ablation studies on Deep-SEA suggest that the proposed method improves cancer survival estimation and can be applied to other estimations, such as cancer recurrence estimation.","doi":"10.1007\/s10489-024-05794-3","cleaned_title":"deep-sea: a deep learning based patient specific multi-modality post-cancer survival estimation architecture","cleaned_abstract":"cancer survival estimation is essential for post-cancer patient care, cancer management policy building, and the development of tailored treatment plans. existing survival estimation methods use censored data; therefore, standard machine learning methods can not be used directly. some censoring-based semi-machine learning methods have recently been proposed; however, these methods pose challenges. they are less patient-specific and non-linear. furthermore, they rely on single-modality features. these drawbacks result in lower survival estimation performance. to address these issues, this work proposes a framework named deep-sea. compared to the state-of-the-art, deep-sea uses multi-modality features, i.e., clinical, radiology, and histology features. these features are analyzed with statistical methods, and only significant features are selected. then, the baseline hazard of the cox model is estimated using breslow\u2019s estimator, which is optimized using stochastic gradient descent. finally, the risk function, i.e., the parameters of our model, are estimated via an ann with time as additional input. ann makes it non-linear while training on the patient-specific features makes it more patient-specific than the state-of-the-art. we train and evaluate deep-sea on five datasets, including head, neck, and colorectal-liver cancer. we have achieved a concordance-index (c-index) score of 0.7181, the highest compared to the state-of-the-art. results and ablation studies on deep-sea suggest that the proposed method improves cancer survival estimation and can be applied to other estimations, such as cancer recurrence estimation.","key_phrases":["cancer survival estimation","post-cancer patient care","cancer management policy building","the development","tailored treatment plans","existing survival estimation methods","censored data","standard machine learning methods","some censoring-based semi-machine learning methods","these methods","challenges","they","they","single-modality features","these drawbacks","lower survival estimation performance","these issues","this work","a framework","deep-sea","the-art","deep-sea","these features","statistical methods","only significant features","the baseline hazard","the cox model","breslow\u2019s estimator","which","stochastic gradient descent","the risk function","the parameters","our model","an ann","time","additional input","ann","it","while training","the patient-specific features","it","the state","the-art","we","deep-sea","five datasets","head","neck","colorectal-liver cancer","we","c-index","the state","the-art","results","ablation studies","deep-sea","the proposed method","cancer survival estimation","other estimations","cancer recurrence estimation","ann","five","0.7181"]},{"title":"A survey on deep learning and machine learning techniques over histopathology image based Osteosarcoma Detection","authors":["K. V. Deepak","R. Bharanidharan"],"abstract":"Osteosarcoma is a common type of cancer that occurs in the cells and spreads to the bones. Osteosarcoma can develop due to genetic mutations, but most cases are not inherited. It often starts at the ends of long bones in the arms and legs during periods of rapid growth. Osteosarcoma can be diagnosed by Histopathological examination using microscopic images. In recent trends, pathological images are increasingly being examined by computer-intelligent approaches, also known as computational pathology. This field utilizes machine learning and deep learning algorithms to analyze digital pathology slides. Histopathology image based osteosarcoma detection is not the subject of any review papers. So, this study examines historical and current literature to present a concise review of Histopathology image based Osteosarcoma detection. This review discussed the types, clinical diagnosis, and modern and future treatment methods of Osteosarcoma. Also, a review of machine learning (ML) and deep learning (DL) approaches for histopathology image osteosarcoma detection is analyzed. Sophisticated methods and the future scope of osteosarcoma detection have also been discussed. Comparative analysis and results of reviewed papers were also analyzed in this review study. This article aims to generate fresh concepts for creating more potent treatment alternatives.","doi":"10.1007\/s11042-024-19554-5","cleaned_title":"a survey on deep learning and machine learning techniques over histopathology image based osteosarcoma detection","cleaned_abstract":"osteosarcoma is a common type of cancer that occurs in the cells and spreads to the bones. osteosarcoma can develop due to genetic mutations, but most cases are not inherited. it often starts at the ends of long bones in the arms and legs during periods of rapid growth. osteosarcoma can be diagnosed by histopathological examination using microscopic images. in recent trends, pathological images are increasingly being examined by computer-intelligent approaches, also known as computational pathology. this field utilizes machine learning and deep learning algorithms to analyze digital pathology slides. histopathology image based osteosarcoma detection is not the subject of any review papers. so, this study examines historical and current literature to present a concise review of histopathology image based osteosarcoma detection. this review discussed the types, clinical diagnosis, and modern and future treatment methods of osteosarcoma. also, a review of machine learning (ml) and deep learning (dl) approaches for histopathology image osteosarcoma detection is analyzed. sophisticated methods and the future scope of osteosarcoma detection have also been discussed. comparative analysis and results of reviewed papers were also analyzed in this review study. this article aims to generate fresh concepts for creating more potent treatment alternatives.","key_phrases":["osteosarcoma","a common type","cancer","that","the cells","the bones","osteosarcoma","genetic mutations","most cases","it","the ends","long bones","the arms","legs","periods","rapid growth","osteosarcoma","histopathological examination","microscopic images","recent trends","pathological images","computer-intelligent approaches","computational pathology","this field","machine learning","deep learning algorithms","digital pathology slides","osteosarcoma detection","the subject","any review papers","this study","historical and current literature","a concise review","osteosarcoma detection","this review","the types","clinical diagnosis","modern and future treatment methods","osteosarcoma","a review","machine learning","ml","dl","histopathology image osteosarcoma detection","sophisticated methods","the future scope","osteosarcoma detection","comparative analysis","results","reviewed papers","this review study","this article","fresh concepts","more potent treatment alternatives","osteosarcoma","osteosarcoma","osteosarcoma","osteosarcoma","osteosarcoma","osteosarcoma","osteosarcoma","osteosarcoma"]},{"title":"Prediction of crop yield in India using machine learning and hybrid deep learning models","authors":["Krithikha Sanju Saravanan","Velammal Bhagavathiappan"],"abstract":"Crop yield prediction is one of the burgeoning research areas in the agriculture domain. The crop yield forecasting models are developed to enhance productivity with improved decision-making strategies. The highly efficient crop yield forecasting model assists farmers in determining when, what and how much to plant on their cultivable land. The main objective of the proposed research work is to build a high efficacious crop yield prediction model based on the data available for the period of 21\u00a0years from 1997 to 2017 using machine learning and hybrid deep learning approaches. Two prediction models have been proposed in this research work to predict the crop yield accurately. The first model is a machine learning-based model which uses the CatBoost regression model and its hyperparameters are tuned which improves the performance of the yield prediction using the Optuna framework. The second model is the hybrid deep learning model which uses spatio-temporal attention-based convolutional neural network (STACNN) for extracting the features and the bidirectional long short-term memory (BiLSTM) model for predicting the crop yield effectively. The proposed models are evaluated using the error metrics and compared with the latest contemporary models. From the evaluation results, it is shown that the proposed models significantly outperform all other existing models and CatBoost regression model slightly performs better than the STACNN-BiLSTM model, with the R-squared value of 0.99.","doi":"10.1007\/s11600-024-01312-8","cleaned_title":"prediction of crop yield in india using machine learning and hybrid deep learning models","cleaned_abstract":"crop yield prediction is one of the burgeoning research areas in the agriculture domain. the crop yield forecasting models are developed to enhance productivity with improved decision-making strategies. the highly efficient crop yield forecasting model assists farmers in determining when, what and how much to plant on their cultivable land. the main objective of the proposed research work is to build a high efficacious crop yield prediction model based on the data available for the period of 21 years from 1997 to 2017 using machine learning and hybrid deep learning approaches. two prediction models have been proposed in this research work to predict the crop yield accurately. the first model is a machine learning-based model which uses the catboost regression model and its hyperparameters are tuned which improves the performance of the yield prediction using the optuna framework. the second model is the hybrid deep learning model which uses spatio-temporal attention-based convolutional neural network (stacnn) for extracting the features and the bidirectional long short-term memory (bilstm) model for predicting the crop yield effectively. the proposed models are evaluated using the error metrics and compared with the latest contemporary models. from the evaluation results, it is shown that the proposed models significantly outperform all other existing models and catboost regression model slightly performs better than the stacnn-bilstm model, with the r-squared value of 0.99.","key_phrases":["crop yield prediction","the burgeoning research areas","the agriculture domain","the crop yield forecasting models","productivity","improved decision-making strategies","the highly efficient crop yield forecasting model","farmers","their cultivable land","the main objective","the proposed research work","a high efficacious crop yield prediction model","the data","the period","21 years","machine learning","hybrid deep learning approaches","two prediction models","this research work","the first model","a machine learning-based model","which","the catboost regression model","which","the performance","the yield prediction","the optuna framework","the second model","the hybrid deep learning model","which","spatio-temporal attention-based convolutional neural network","stacnn","the features","the bidirectional long short-term memory (bilstm) model","the crop yield","the proposed models","the error metrics","the latest contemporary models","the evaluation results","it","the proposed models","all other existing models","catboost regression model","the stacnn-bilstm model","the r-squared value","the period of 21 years from 1997 to 2017","two","first","second","stacnn","stacnn","0.99"]},{"title":"Learning sparse and smooth functions by deep Sigmoid nets","authors":["Xia Liu"],"abstract":"To pursue the outperformance of deep nets in learning, we construct a deep net with three hidden layers and prove that, implementing the empirical risk minimization (ERM) on this deep net, the estimator can theoretically realize the optimal learning rates without the classical saturation problem. In other words, deepening the networks with only three hidden layers can overcome the saturation and not degrade the optimal learning rates. The obtained results underlie the success of deep nets and provide a theoretical guidance for deep learning.","doi":"10.1007\/s11766-023-4309-4","cleaned_title":"learning sparse and smooth functions by deep sigmoid nets","cleaned_abstract":"to pursue the outperformance of deep nets in learning, we construct a deep net with three hidden layers and prove that, implementing the empirical risk minimization (erm) on this deep net, the estimator can theoretically realize the optimal learning rates without the classical saturation problem. in other words, deepening the networks with only three hidden layers can overcome the saturation and not degrade the optimal learning rates. the obtained results underlie the success of deep nets and provide a theoretical guidance for deep learning.","key_phrases":["the outperformance","deep nets","we","a deep net","three hidden layers","the empirical risk minimization","erm","this deep net","the estimator","the optimal learning rates","the classical saturation problem","other words","the networks","only three hidden layers","the saturation","the optimal learning rates","the obtained results","the success","deep nets","a theoretical guidance","deep learning","three","only three"]},{"title":"Parametric RSigELU: a new trainable activation function for deep learning","authors":["Serhat Kili\u00e7arslan","Mete Celik"],"abstract":"Activation functions are used to extract meaningful relationships from real-world problems with the help of deep learning models. Thus, the development of activation functions which affect deep learning models\u2019 performances is of great interest to researchers. In the literature, mostly, nonlinear activation functions are preferred since linear activation functions limit the learning performances of the deep learning models. Non-linear activation functions can be classified as fixed-parameter and trainable activation functions based on whether the activation function parameter is fixed (i.e., user-given) or modified during the training process of deep learning models. The parameters of the fixed-parameter activation functions should be specified before the deep learning model training process. However, it takes too much time to determine appropriate function parameter values and can cause the slow convergence of the deep learning model. In contrast, trainable activation functions whose parameters are updated in each iteration of deep learning models training process achieve faster and better convergence by obtaining the most suitable parameter values for the datasets and deep learning architectures. This study proposes parametric RSigELU (P+RSigELU) trainable activation functions, such as P+RSigELU Single (P+RSigELUS) and P+RSigELU Double (P+RSigELUD), to improve the performance of fixed-parameter activation function of RSigELU. The performances of the proposed trainable activation functions were evaluated on the benchmark datasets of MNIST, CIFAR-10, and CIFAR-100 datasets. Results show that the proposed activation functions outperforms PReLU, PELU, ALISA, P+FELU, PSigmoid, and GELU activation functions found in the literature. To access the codes of the activation function; https:\/\/github.com\/serhatklc\/P-RsigELU-Activation-Function.","doi":"10.1007\/s00521-024-09538-9","cleaned_title":"parametric rsigelu: a new trainable activation function for deep learning","cleaned_abstract":"activation functions are used to extract meaningful relationships from real-world problems with the help of deep learning models. thus, the development of activation functions which affect deep learning models\u2019 performances is of great interest to researchers. in the literature, mostly, nonlinear activation functions are preferred since linear activation functions limit the learning performances of the deep learning models. non-linear activation functions can be classified as fixed-parameter and trainable activation functions based on whether the activation function parameter is fixed (i.e., user-given) or modified during the training process of deep learning models. the parameters of the fixed-parameter activation functions should be specified before the deep learning model training process. however, it takes too much time to determine appropriate function parameter values and can cause the slow convergence of the deep learning model. in contrast, trainable activation functions whose parameters are updated in each iteration of deep learning models training process achieve faster and better convergence by obtaining the most suitable parameter values for the datasets and deep learning architectures. this study proposes parametric rsigelu (p+rsigelu) trainable activation functions, such as p+rsigelu single (p+rsigelus) and p+rsigelu double (p+rsigelud), to improve the performance of fixed-parameter activation function of rsigelu. the performances of the proposed trainable activation functions were evaluated on the benchmark datasets of mnist, cifar-10, and cifar-100 datasets. results show that the proposed activation functions outperforms prelu, pelu, alisa, p+felu, psigmoid, and gelu activation functions found in the literature. to access the codes of the activation function; https:\/\/github.com\/serhatklc\/p-rsigelu-activation-function.","key_phrases":["activation functions","meaningful relationships","real-world problems","the help","deep learning models","the development","activation functions","which","deep learning models\u2019 performances","great interest","researchers","the literature","nonlinear activation functions","linear activation functions","the learning performances","the deep learning models","non-linear activation functions","fixed-parameter and trainable activation functions","the activation function parameter","the training process","deep learning models","the parameters","the fixed-parameter activation functions","the deep learning model training process","it","too much time","appropriate function parameter values","the slow convergence","the deep learning model","contrast","trainable activation functions","whose parameters","each iteration","deep learning models training process","faster and better convergence","the most suitable parameter values","the datasets","deep learning architectures","this study","parametric rsigelu (p+rsigelu) trainable activation functions","p+rsigelu single (p+rsigelus","(p+rsigelud","the performance","fixed-parameter activation function","rsigelu","the performances","the proposed trainable activation functions","the benchmark datasets","mnist","cifar-10","cifar-100 datasets","results","the proposed activation functions","prelu","pelu","alisa","p+felu","psigmoid","gelu activation functions","the literature","the codes","the activation function","linear","p+rsigelu","cifar-10","prelu","pelu","alisa, p+felu"]},{"title":"Deep-learning-enabled antibiotic discovery through molecular de-extinction","authors":["Fangping Wan","Marcelo D. T. Torres","Jacqueline Peng","Cesar de la Fuente-Nunez"],"abstract":"Molecular de-extinction aims at resurrecting molecules to solve antibiotic resistance and other present-day biological and biomedical problems. Here we show that deep learning can be used to mine the proteomes of all available extinct organisms for the discovery of antibiotic peptides. We trained ensembles of deep-learning models consisting of a peptide-sequence encoder coupled with neural networks for the prediction of antimicrobial activity and used it to mine 10,311,899 peptides. The models predicted 37,176 sequences with broad-spectrum antimicrobial activity, 11,035 of which were not found in extant organisms. We synthesized 69 peptides and experimentally confirmed their activity against bacterial pathogens. Most peptides killed bacteria by depolarizing their cytoplasmic membrane, contrary to known antimicrobial peptides, which tend to target the outer membrane. Notably, lead compounds (including mammuthusin-2 from the woolly mammoth, elephasin-2 from the straight-tusked elephant, hydrodamin-1 from the ancient sea cow, mylodonin-2 from the giant sloth and megalocerin-1 from the extinct giant elk) showed anti-infective activity in mice with skin abscess or thigh infections. Molecular de-extinction aided by deep learning may accelerate the discovery of therapeutic molecules.","doi":"10.1038\/s41551-024-01201-x","cleaned_title":"deep-learning-enabled antibiotic discovery through molecular de-extinction","cleaned_abstract":"molecular de-extinction aims at resurrecting molecules to solve antibiotic resistance and other present-day biological and biomedical problems. here we show that deep learning can be used to mine the proteomes of all available extinct organisms for the discovery of antibiotic peptides. we trained ensembles of deep-learning models consisting of a peptide-sequence encoder coupled with neural networks for the prediction of antimicrobial activity and used it to mine 10,311,899 peptides. the models predicted 37,176 sequences with broad-spectrum antimicrobial activity, 11,035 of which were not found in extant organisms. we synthesized 69 peptides and experimentally confirmed their activity against bacterial pathogens. most peptides killed bacteria by depolarizing their cytoplasmic membrane, contrary to known antimicrobial peptides, which tend to target the outer membrane. notably, lead compounds (including mammuthusin-2 from the woolly mammoth, elephasin-2 from the straight-tusked elephant, hydrodamin-1 from the ancient sea cow, mylodonin-2 from the giant sloth and megalocerin-1 from the extinct giant elk) showed anti-infective activity in mice with skin abscess or thigh infections. molecular de-extinction aided by deep learning may accelerate the discovery of therapeutic molecules.","key_phrases":["-","extinction","molecules","antibiotic resistance","other present-day biological and biomedical problems","we","deep learning","the proteomes","all available extinct organisms","the discovery","antibiotic peptides","we","ensembles","deep-learning models","a peptide-sequence encoder","neural networks","the prediction","antimicrobial activity","it","mine","10,311,899 peptides","the models","37,176 sequences","broad-spectrum antimicrobial activity","which","extant organisms","we","69 peptides","their activity","bacterial pathogens","most peptides","bacteria","their cytoplasmic membrane","known antimicrobial peptides","which","the outer membrane","lead compounds","the woolly mammoth","the straight-tusked elephant","hydrodamin-1","the ancient sea cow","the giant sloth","megalocerin-1","the extinct giant","elk","anti-infective activity","mice","skin abscess","thigh infections","molecular de-extinction","deep learning","the discovery","therapeutic molecules","10,311,899","37,176","11,035","69"]},{"title":"Investigations on machine learning, deep learning, and longitudinal regression methods for global greenhouse gases predictions","authors":["S. D. Yazd","N. Gharib","J. F. Derakhshandeh"],"abstract":"Combating climate change is one of the key topics and concerns that our community is currently facing these days. Since a few decades ago, greenhouse gases emissions gradually started to increase. Thus, the researchers attempted to find a permanent solution for this challenge. In this paper, different methods of machine learning and deep learning models are applied to evaluate their effectiveness and accuracy in predicting greenhouse gases emissions. To increase the accuracy of the assessment, the data of 101 countries over a period of 31\u00a0years (1991\u20132021) from the official World Bank sources are considered. In this study, therefore, a range of matrices are analyzed including Mean Squared Error, Root Mean Squared Error, Mean Absolute Error, p value, and correlation coefficient for each model. The results demonstrate that machine learning models typically overtake the deep learning models with the support vector regression polynomial model. Besides, the statistical findings of longitudinal regression analysis reveal that by increasing cereal yield, and permanent cropland areas the greenhouse gas emissions are significantly increase (p value\u2009=\u20090.000) and (p value\u2009=\u20090.06) respectively; however, increasing in renewable energy consumption and forest areas will lead to decreasing in greenhouse gas emissions (p value\u2009=\u20090.000) and (p value\u2009=\u20090.07) respectively.","doi":"10.1007\/s13762-024-06014-8","cleaned_title":"investigations on machine learning, deep learning, and longitudinal regression methods for global greenhouse gases predictions","cleaned_abstract":"combating climate change is one of the key topics and concerns that our community is currently facing these days. since a few decades ago, greenhouse gases emissions gradually started to increase. thus, the researchers attempted to find a permanent solution for this challenge. in this paper, different methods of machine learning and deep learning models are applied to evaluate their effectiveness and accuracy in predicting greenhouse gases emissions. to increase the accuracy of the assessment, the data of 101 countries over a period of 31 years (1991\u20132021) from the official world bank sources are considered. in this study, therefore, a range of matrices are analyzed including mean squared error, root mean squared error, mean absolute error, p value, and correlation coefficient for each model. the results demonstrate that machine learning models typically overtake the deep learning models with the support vector regression polynomial model. besides, the statistical findings of longitudinal regression analysis reveal that by increasing cereal yield, and permanent cropland areas the greenhouse gas emissions are significantly increase (p value = 0.000) and (p value = 0.06) respectively; however, increasing in renewable energy consumption and forest areas will lead to decreasing in greenhouse gas emissions (p value = 0.000) and (p value = 0.07) respectively.","key_phrases":["climate change","the key topics","concerns","our community","these days","greenhouse gases emissions","the researchers","a permanent solution","this challenge","this paper","different methods","machine learning","deep learning models","their effectiveness","accuracy","greenhouse gases emissions","the accuracy","the assessment","the data","101 countries","a period","31 years","the official world bank sources","this study","a range","matrices","mean squared error","root mean squared error","absolute error","p value","correlation","each model","the results","machine learning models","the deep learning models","the support vector regression polynomial model","the statistical findings","longitudinal regression analysis","cereal yield","the greenhouse gas emissions","(p value","renewable energy consumption","forest areas","greenhouse gas emissions","p value","(p value","one","these days","a few decades ago","101","a period of 31 years","0.000","0.06","0.000","0.07"]},{"title":"Deep ensemble transfer learning framework for COVID-19 Arabic text identification via deep active learning and text data augmentation","authors":["Abdullah Y. Muaad","Hanumanthappa Jayappa Davanagere","Jamil Hussain","Mugahed A. Al-antari"],"abstract":"Since the declaration of COVID-19 as an epidemic by the World Health Organization in September 2019, the task of monitoring and managing the spread of misinformation related to COVID-19 on social media has become increasingly challenging. Particularly, when it comes to Arabic text recognition, tracking and identifying misleading information regarding COVID-19 on social media platforms presents significant difficulties. The detection of such text is crucial in order to safeguard our communities from the dissemination of false rumors and to establish a reliable framework for text detection. This research paper introduces a novel deep ensemble learning framework that aims to recognize ten distinct categories of Arabic text related to COVID-19, including rumors, restrictions, celebrity news, informational news, plans, requests, advice, personal anecdotes, and others. To build our framework, we leverage a dataset called ArCOVID-19Vac (Dataset1), which consists of 10,000 text samples. In addition, the DAL technique is employed to automatically annotate new text samples acquired for Dataset2. To further expand our datasets, we employ back translation and random insertion augmentation strategies, resulting in Datasets3 and Datasets4, each containing 24,000 text samples. By merging the original and augmented datasets, we create Dataset5, which comprises a total of 39,000 text samples. The final text prediction is carried out using three transformer-based BERT models through ensemble transfer learning. Our proposed ensemble framework is evaluated using each dataset independently, and it demonstrates promising results, particularly when utilizing the largest dataset (Dataset5), achieving an accuracy of 93%, precision of 92%, recall of 93%, and an F1-score of 91%. Furthermore, our proposed model exhibits performance improvements of 27%, 18%, 2%, and 1% when utilizing Datasets2, 3, 4, and 5, respectively. The comprehensive experimental results demonstrate that our ensemble framework outperforms other state-of-the-art AI-based models. The encouraging performance of our framework in accurately identifying Arabic text has the potential to enhance decision-making processes regarding the identification of misleading information and to facilitate the development of strategies to combat such issues in the future.","doi":"10.1007\/s11042-024-18487-3","cleaned_title":"deep ensemble transfer learning framework for covid-19 arabic text identification via deep active learning and text data augmentation","cleaned_abstract":"since the declaration of covid-19 as an epidemic by the world health organization in september 2019, the task of monitoring and managing the spread of misinformation related to covid-19 on social media has become increasingly challenging. particularly, when it comes to arabic text recognition, tracking and identifying misleading information regarding covid-19 on social media platforms presents significant difficulties. the detection of such text is crucial in order to safeguard our communities from the dissemination of false rumors and to establish a reliable framework for text detection. this research paper introduces a novel deep ensemble learning framework that aims to recognize ten distinct categories of arabic text related to covid-19, including rumors, restrictions, celebrity news, informational news, plans, requests, advice, personal anecdotes, and others. to build our framework, we leverage a dataset called arcovid-19vac (dataset1), which consists of 10,000 text samples. in addition, the dal technique is employed to automatically annotate new text samples acquired for dataset2. to further expand our datasets, we employ back translation and random insertion augmentation strategies, resulting in datasets3 and datasets4, each containing 24,000 text samples. by merging the original and augmented datasets, we create dataset5, which comprises a total of 39,000 text samples. the final text prediction is carried out using three transformer-based bert models through ensemble transfer learning. our proposed ensemble framework is evaluated using each dataset independently, and it demonstrates promising results, particularly when utilizing the largest dataset (dataset5), achieving an accuracy of 93%, precision of 92%, recall of 93%, and an f1-score of 91%. furthermore, our proposed model exhibits performance improvements of 27%, 18%, 2%, and 1% when utilizing datasets2, 3, 4, and 5, respectively. the comprehensive experimental results demonstrate that our ensemble framework outperforms other state-of-the-art ai-based models. the encouraging performance of our framework in accurately identifying arabic text has the potential to enhance decision-making processes regarding the identification of misleading information and to facilitate the development of strategies to combat such issues in the future.","key_phrases":["the declaration","covid-19","an epidemic","the world health organization","september","the task","monitoring","the spread","misinformation","covid-19","social media","it","arabic text recognition","tracking","misleading information","covid-19","social media platforms","significant difficulties","the detection","such text","order","our communities","the dissemination","false rumors","a reliable framework","text detection","this research paper","a novel deep ensemble learning framework","that","ten distinct categories","arabic text","covid-19","rumors","restrictions","celebrity news","informational news","plans","requests","advice","personal anecdotes","others","our framework","we","a dataset","dataset1","which","10,000 text samples","addition","the dal technique","new text samples","dataset2","our datasets","we","back translation and random insertion augmentation strategies","datasets3","datasets4","each","24,000 text samples","the original and augmented datasets","we","dataset5","which","a total","39,000 text samples","the final text prediction","three transformer-based bert models","ensemble transfer learning","our proposed ensemble framework","each dataset","it","promising results","the largest dataset","dataset5","an accuracy","93%","precision","92%","recall","93%","an f1-score","91%","our proposed model","performance improvements","27%","18%","2%","1%","datasets2","the comprehensive experimental results","our ensemble framework","the-art","the encouraging performance","our framework","arabic text","the potential","decision-making processes","the identification","information","the development","strategies","such issues","the future","covid-19","the world health organization","september 2019","covid-19","arabic","recognition","covid-19","ten","arabic","covid-19","dataset1","10,000","dataset2","datasets3","24,000","39,000","three","93%","92%","93%","91%","27%","18%","2%","1%","datasets2","3","4","5","arabic"]},{"title":"A comprehensive review of image denoising in deep learning","authors":["Rusul Sabah Jebur","Mohd Hazli Bin Mohamed Zabil","Dalal Adulmohsin Hammood","Lim Kok Cheng"],"abstract":"Deep learning has gained significant interest in image denoising, but there are notable distinctions in the types of deep learning methods used. Discriminative learning is suitable for handling Gaussian noise, while optimization models are effective in estimating real noise. However, there is limited research that summarizes the different deep learning techniques for image denoising. This paper conducts a comprehensive review of techniques and methods used for image denoising and identifying challenges associated with existing approaches. In this paper, a comparative study of deep techniques is offered in image denoising. The study conducted a comprehensive review of 68 papers on image denoising published between 2018 and 2023, providing a detailed analysis of the field\u2019s progress and methodologies over a period of 5 years. Through its literature review, the paper provides a comprehensive summary of image denoising in deep learning, including machine learning methods for image denoising, CNNs for image denoising, additive white noisy-image denoising, real noisy image denoising, blind denoising, hybrid noisy images, state- of-the-art methods for image denoising with deep learning, salt and pepper noise, non-linear filters for digital color images. The main objective of this paper is to provide a comprehensive overview of various approaches used for image denoising, each of which has been explored and developed based on individual research studies. The paper aims to discuss these approaches in a systematic and organized manner, comparing their strengths and weaknesses to provide insights for future research in the field.","doi":"10.1007\/s11042-023-17468-2","cleaned_title":"a comprehensive review of image denoising in deep learning","cleaned_abstract":"deep learning has gained significant interest in image denoising, but there are notable distinctions in the types of deep learning methods used. discriminative learning is suitable for handling gaussian noise, while optimization models are effective in estimating real noise. however, there is limited research that summarizes the different deep learning techniques for image denoising. this paper conducts a comprehensive review of techniques and methods used for image denoising and identifying challenges associated with existing approaches. in this paper, a comparative study of deep techniques is offered in image denoising. the study conducted a comprehensive review of 68 papers on image denoising published between 2018 and 2023, providing a detailed analysis of the field\u2019s progress and methodologies over a period of 5 years. through its literature review, the paper provides a comprehensive summary of image denoising in deep learning, including machine learning methods for image denoising, cnns for image denoising, additive white noisy-image denoising, real noisy image denoising, blind denoising, hybrid noisy images, state- of-the-art methods for image denoising with deep learning, salt and pepper noise, non-linear filters for digital color images. the main objective of this paper is to provide a comprehensive overview of various approaches used for image denoising, each of which has been explored and developed based on individual research studies. the paper aims to discuss these approaches in a systematic and organized manner, comparing their strengths and weaknesses to provide insights for future research in the field.","key_phrases":["deep learning","significant interest","image denoising","notable distinctions","the types","deep learning methods","discriminative learning","gaussian noise","optimization models","real noise","limited research","that","the different deep learning techniques","image denoising","this paper","a comprehensive review","techniques","methods","image denoising","challenges","existing approaches","this paper","a comparative study","deep techniques","image denoising","the study","a comprehensive review","68 papers","image denoising","a detailed analysis","the field\u2019s progress","methodologies","a period","5 years","its literature review","the paper","a comprehensive summary","image","deep learning","machine learning methods","image denoising","cnns","image denoising","additive white noisy-image denoising","real noisy image denoising","blind denoising","hybrid noisy images","state-","the-art","methods","image","deep learning","salt and pepper noise","non-linear filters","digital color images","the main objective","this paper","a comprehensive overview","various approaches","image denoising","each","which","individual research studies","the paper","these approaches","a systematic and organized manner","their strengths","weaknesses","insights","future research","the field","68","between 2018 and 2023","5 years"]},{"title":"Fisheye freshness detection using common deep learning algorithms and machine learning methods with a developed mobile application","authors":["Muslume Beyza Yildiz","Elham Tahsin Yasin","Murat Koklu"],"abstract":"AbstractFish is commonly ingested as a source of protein and essential nutrients for humans. To fully benefit from the proteins and substances in fish it is crucial to ensure its freshness. If fish is stored for an extended period, its freshness deteriorates. Determining the freshness of fish can be done by examining its eyes, smell, skin, and gills. In this study, artificial intelligence techniques are employed to assess fish freshness. The author\u2019s objective is to evaluate the freshness of fish by analyzing its eye characteristics. To achieve this, we have developed a combination of deep and machine learning models that accurately classify the freshness of fish. Furthermore, an application that utilizes both deep learning and machine learning, to instantly detect the freshness of any given fish sample was created. Two deep learning algorithms (SqueezeNet, and VGG19) were implemented to extract features from image data. Additionally, five machine learning models to classify the freshness levels of fish samples were applied. Machine learning models include (k-NN, RF, SVM, LR, and ANN). Based on the results, it can be inferred that employing the VGG19 model for feature selection in conjunction with an Artificial Neural Network (ANN) for classification yields the most favorable success rate of 77.3% for the FFE dataset.Graphical Abstract","doi":"10.1007\/s00217-024-04493-0","cleaned_title":"fisheye freshness detection using common deep learning algorithms and machine learning methods with a developed mobile application","cleaned_abstract":"abstractfish is commonly ingested as a source of protein and essential nutrients for humans. to fully benefit from the proteins and substances in fish it is crucial to ensure its freshness. if fish is stored for an extended period, its freshness deteriorates. determining the freshness of fish can be done by examining its eyes, smell, skin, and gills. in this study, artificial intelligence techniques are employed to assess fish freshness. the author\u2019s objective is to evaluate the freshness of fish by analyzing its eye characteristics. to achieve this, we have developed a combination of deep and machine learning models that accurately classify the freshness of fish. furthermore, an application that utilizes both deep learning and machine learning, to instantly detect the freshness of any given fish sample was created. two deep learning algorithms (squeezenet, and vgg19) were implemented to extract features from image data. additionally, five machine learning models to classify the freshness levels of fish samples were applied. machine learning models include (k-nn, rf, svm, lr, and ann). based on the results, it can be inferred that employing the vgg19 model for feature selection in conjunction with an artificial neural network (ann) for classification yields the most favorable success rate of 77.3% for the ffe dataset.graphical abstract","key_phrases":["abstractfish","a source","protein","essential nutrients","humans","the proteins","substances","fish","it","its freshness","fish","an extended period","its freshness","the freshness","fish","its eyes","smell","skin","gills","this study","artificial intelligence techniques","fish freshness","the author\u2019s objective","the freshness","fish","its eye characteristics","this","we","a combination","deep and machine learning models","that","the freshness","fish","an application","that","both deep learning","machine learning","the freshness","any given fish sample","two deep learning algorithms","squeezenet","vgg19","features","image data","five machine learning models","the freshness levels","fish samples","machine learning models","k",", rf, svm","lr","ann","the results","it","the vgg19 model","feature selection","conjunction","an artificial neural network","ann","classification yields","the most favorable success rate","77.3%","the ffe dataset.graphical abstract","two","five","77.3%"]},{"title":"Deep learning for transesophageal echocardiography view classification","authors":["Kirsten R. Steffner","Matthew Christensen","George Gill","Michael Bowdish","Justin Rhee","Abirami Kumaresan","Bryan He","James Zou","David Ouyang"],"abstract":"Transesophageal echocardiography (TEE) imaging is a vital tool used in the evaluation of complex cardiac pathology and the management of cardiac surgery patients. A key limitation to the application of deep learning strategies to intraoperative and intraprocedural TEE data is the complexity and unstructured nature of these images. In the present study, we developed a deep learning-based, multi-category TEE view classification model that can be used to add structure to intraoperative and intraprocedural TEE imaging data. More specifically, we trained a convolutional neural network (CNN) to predict standardized TEE views using labeled intraoperative and intraprocedural TEE videos from Cedars-Sinai Medical Center (CSMC). We externally validated our model on intraoperative TEE videos from Stanford University Medical Center (SUMC). Accuracy of our model was high across all labeled views. The highest performance was achieved for the Trans-Gastric Left Ventricular Short Axis View (area under the receiver operating curve [AUC]\u2009=\u20090.971 at CSMC, 0.957 at SUMC), the Mid-Esophageal Long Axis View (AUC\u2009=\u20090.954 at CSMC, 0.905 at SUMC), the Mid-Esophageal Aortic Valve Short Axis View (AUC\u2009=\u20090.946 at CSMC, 0.898 at SUMC), and the Mid-Esophageal 4-Chamber View (AUC\u2009=\u20090.939 at CSMC, 0.902 at SUMC). Ultimately, we demonstrate that our deep learning model can accurately classify standardized TEE views, which will facilitate further downstream deep learning analyses for intraoperative and intraprocedural TEE imaging.","doi":"10.1038\/s41598-023-50735-8","cleaned_title":"deep learning for transesophageal echocardiography view classification","cleaned_abstract":"transesophageal echocardiography (tee) imaging is a vital tool used in the evaluation of complex cardiac pathology and the management of cardiac surgery patients. a key limitation to the application of deep learning strategies to intraoperative and intraprocedural tee data is the complexity and unstructured nature of these images. in the present study, we developed a deep learning-based, multi-category tee view classification model that can be used to add structure to intraoperative and intraprocedural tee imaging data. more specifically, we trained a convolutional neural network (cnn) to predict standardized tee views using labeled intraoperative and intraprocedural tee videos from cedars-sinai medical center (csmc). we externally validated our model on intraoperative tee videos from stanford university medical center (sumc). accuracy of our model was high across all labeled views. the highest performance was achieved for the trans-gastric left ventricular short axis view (area under the receiver operating curve [auc] = 0.971 at csmc, 0.957 at sumc), the mid-esophageal long axis view (auc = 0.954 at csmc, 0.905 at sumc), the mid-esophageal aortic valve short axis view (auc = 0.946 at csmc, 0.898 at sumc), and the mid-esophageal 4-chamber view (auc = 0.939 at csmc, 0.902 at sumc). ultimately, we demonstrate that our deep learning model can accurately classify standardized tee views, which will facilitate further downstream deep learning analyses for intraoperative and intraprocedural tee imaging.","key_phrases":["transesophageal echocardiography","(tee) imaging","a vital tool","the evaluation","complex cardiac pathology","the management","cardiac surgery patients","a key limitation","the application","deep learning strategies","to intraoperative and intraprocedural tee data","the complexity","unstructured nature","these images","the present study","we","a deep learning-based, multi-category tee view classification model","that","structure","intraoperative and intraprocedural tee imaging data","we","a convolutional neural network","cnn","standardized tee views","labeled intraoperative and intraprocedural tee videos","cedars-sinai medical center","csmc","we","our model","intraoperative tee videos","stanford university medical center","sumc","accuracy","our model","all labeled views","the highest performance","ventricular short axis view","area","the receiver operating curve","csmc","sumc",", the mid-esophageal long axis view","auc =","csmc","sumc","the mid-esophageal aortic valve short axis view","csmc","sumc","the mid-esophageal 4-chamber view","csmc","sumc","we","our deep learning model","standardized tee views","which","further downstream deep learning analyses","intraoperative and intraprocedural tee imaging","cnn","stanford university medical center","0.971","0.957","0.954","0.905","0.946","0.898","4","0.939","0.902"]},{"title":"Automated optical inspection based on synthetic mechanisms combining deep learning and machine learning","authors":["Chung-Ming Lo","Ting-Yi Lin"],"abstract":"The quality inspection of products before delivery plays a critical role in ensuring manufacturing quality. Quick and accurate inspection of samples is realized by highly automated inspection based on pattern recognition in smart manufacturing. Conventional ensemble methods have been demonstrated to be effective for defect detection. This study further proposed synthetic mechanisms based on using various features and learning classifiers. A database of 6000 sample images of printed circuit board (PCB) connectors collected from factories was compiled. A novel confidence synthesis mechanism was proposed to prescreen images using deep learning features. Spatially connected texture features were then used to reclassify images with low reliabilities. The synthetic mechanism was found to outperform a single classifier. In particular, the highest improvement in accuracy (from 96.00 to 97.83%) was obtained using the confidence-based synthesis. The synthetic mechanism can be used to achieve high accuracy in defect detection and make automation in smart manufacturing more practicable.","doi":"10.1007\/s10845-024-02474-4","cleaned_title":"automated optical inspection based on synthetic mechanisms combining deep learning and machine learning","cleaned_abstract":"the quality inspection of products before delivery plays a critical role in ensuring manufacturing quality. quick and accurate inspection of samples is realized by highly automated inspection based on pattern recognition in smart manufacturing. conventional ensemble methods have been demonstrated to be effective for defect detection. this study further proposed synthetic mechanisms based on using various features and learning classifiers. a database of 6000 sample images of printed circuit board (pcb) connectors collected from factories was compiled. a novel confidence synthesis mechanism was proposed to prescreen images using deep learning features. spatially connected texture features were then used to reclassify images with low reliabilities. the synthetic mechanism was found to outperform a single classifier. in particular, the highest improvement in accuracy (from 96.00 to 97.83%) was obtained using the confidence-based synthesis. the synthetic mechanism can be used to achieve high accuracy in defect detection and make automation in smart manufacturing more practicable.","key_phrases":["the quality inspection","products","delivery","a critical role","manufacturing quality","quick and accurate inspection","samples","highly automated inspection","pattern recognition","smart manufacturing","conventional ensemble methods","defect detection","this study","further proposed synthetic mechanisms","various features","learning classifiers","a database","6000 sample images","printed circuit board (pcb) connectors","factories","a novel confidence synthesis mechanism","images","deep learning features","spatially connected texture features","images","low reliabilities","the synthetic mechanism","a single classifier","the highest improvement","accuracy","97.83%","the confidence-based synthesis","the synthetic mechanism","high accuracy","defect detection","automation","smart manufacturing","6000","96.00","97.83%"]},{"title":"Deep ensemble transfer learning framework for COVID-19 Arabic text identification via deep active learning and text data augmentation","authors":["Abdullah Y. Muaad","Hanumanthappa Jayappa Davanagere","Jamil Hussain","Mugahed A. Al-antari"],"abstract":"Since the declaration of COVID-19 as an epidemic by the World Health Organization in September 2019, the task of monitoring and managing the spread of misinformation related to COVID-19 on social media has become increasingly challenging. Particularly, when it comes to Arabic text recognition, tracking and identifying misleading information regarding COVID-19 on social media platforms presents significant difficulties. The detection of such text is crucial in order to safeguard our communities from the dissemination of false rumors and to establish a reliable framework for text detection. This research paper introduces a novel deep ensemble learning framework that aims to recognize ten distinct categories of Arabic text related to COVID-19, including rumors, restrictions, celebrity news, informational news, plans, requests, advice, personal anecdotes, and others. To build our framework, we leverage a dataset called ArCOVID-19Vac (Dataset1), which consists of 10,000 text samples. In addition, the DAL technique is employed to automatically annotate new text samples acquired for Dataset2. To further expand our datasets, we employ back translation and random insertion augmentation strategies, resulting in Datasets3 and Datasets4, each containing 24,000 text samples. By merging the original and augmented datasets, we create Dataset5, which comprises a total of 39,000 text samples. The final text prediction is carried out using three transformer-based BERT models through ensemble transfer learning. Our proposed ensemble framework is evaluated using each dataset independently, and it demonstrates promising results, particularly when utilizing the largest dataset (Dataset5), achieving an accuracy of 93%, precision of 92%, recall of 93%, and an F1-score of 91%. Furthermore, our proposed model exhibits performance improvements of 27%, 18%, 2%, and 1% when utilizing Datasets2, 3, 4, and 5, respectively. The comprehensive experimental results demonstrate that our ensemble framework outperforms other state-of-the-art AI-based models. The encouraging performance of our framework in accurately identifying Arabic text has the potential to enhance decision-making processes regarding the identification of misleading information and to facilitate the development of strategies to combat such issues in the future.","doi":"10.1007\/s11042-024-18487-3","cleaned_title":"deep ensemble transfer learning framework for covid-19 arabic text identification via deep active learning and text data augmentation","cleaned_abstract":"since the declaration of covid-19 as an epidemic by the world health organization in september 2019, the task of monitoring and managing the spread of misinformation related to covid-19 on social media has become increasingly challenging. particularly, when it comes to arabic text recognition, tracking and identifying misleading information regarding covid-19 on social media platforms presents significant difficulties. the detection of such text is crucial in order to safeguard our communities from the dissemination of false rumors and to establish a reliable framework for text detection. this research paper introduces a novel deep ensemble learning framework that aims to recognize ten distinct categories of arabic text related to covid-19, including rumors, restrictions, celebrity news, informational news, plans, requests, advice, personal anecdotes, and others. to build our framework, we leverage a dataset called arcovid-19vac (dataset1), which consists of 10,000 text samples. in addition, the dal technique is employed to automatically annotate new text samples acquired for dataset2. to further expand our datasets, we employ back translation and random insertion augmentation strategies, resulting in datasets3 and datasets4, each containing 24,000 text samples. by merging the original and augmented datasets, we create dataset5, which comprises a total of 39,000 text samples. the final text prediction is carried out using three transformer-based bert models through ensemble transfer learning. our proposed ensemble framework is evaluated using each dataset independently, and it demonstrates promising results, particularly when utilizing the largest dataset (dataset5), achieving an accuracy of 93%, precision of 92%, recall of 93%, and an f1-score of 91%. furthermore, our proposed model exhibits performance improvements of 27%, 18%, 2%, and 1% when utilizing datasets2, 3, 4, and 5, respectively. the comprehensive experimental results demonstrate that our ensemble framework outperforms other state-of-the-art ai-based models. the encouraging performance of our framework in accurately identifying arabic text has the potential to enhance decision-making processes regarding the identification of misleading information and to facilitate the development of strategies to combat such issues in the future.","key_phrases":["the declaration","covid-19","an epidemic","the world health organization","september","the task","monitoring","the spread","misinformation","covid-19","social media","it","arabic text recognition","tracking","misleading information","covid-19","social media platforms","significant difficulties","the detection","such text","order","our communities","the dissemination","false rumors","a reliable framework","text detection","this research paper","a novel deep ensemble learning framework","that","ten distinct categories","arabic text","covid-19","rumors","restrictions","celebrity news","informational news","plans","requests","advice","personal anecdotes","others","our framework","we","a dataset","dataset1","which","10,000 text samples","addition","the dal technique","new text samples","dataset2","our datasets","we","back translation and random insertion augmentation strategies","datasets3","datasets4","each","24,000 text samples","the original and augmented datasets","we","dataset5","which","a total","39,000 text samples","the final text prediction","three transformer-based bert models","ensemble transfer learning","our proposed ensemble framework","each dataset","it","promising results","the largest dataset","dataset5","an accuracy","93%","precision","92%","recall","93%","an f1-score","91%","our proposed model","performance improvements","27%","18%","2%","1%","datasets2","the comprehensive experimental results","our ensemble framework","the-art","the encouraging performance","our framework","arabic text","the potential","decision-making processes","the identification","information","the development","strategies","such issues","the future","covid-19","the world health organization","september 2019","covid-19","arabic","recognition","covid-19","ten","arabic","covid-19","dataset1","10,000","dataset2","datasets3","24,000","39,000","three","93%","92%","93%","91%","27%","18%","2%","1%","datasets2","3","4","5","arabic"]},{"title":"A comprehensive review of image denoising in deep learning","authors":["Rusul Sabah Jebur","Mohd Hazli Bin Mohamed Zabil","Dalal Adulmohsin Hammood","Lim Kok Cheng"],"abstract":"Deep learning has gained significant interest in image denoising, but there are notable distinctions in the types of deep learning methods used. Discriminative learning is suitable for handling Gaussian noise, while optimization models are effective in estimating real noise. However, there is limited research that summarizes the different deep learning techniques for image denoising. This paper conducts a comprehensive review of techniques and methods used for image denoising and identifying challenges associated with existing approaches. In this paper, a comparative study of deep techniques is offered in image denoising. The study conducted a comprehensive review of 68 papers on image denoising published between 2018 and 2023, providing a detailed analysis of the field\u2019s progress and methodologies over a period of 5 years. Through its literature review, the paper provides a comprehensive summary of image denoising in deep learning, including machine learning methods for image denoising, CNNs for image denoising, additive white noisy-image denoising, real noisy image denoising, blind denoising, hybrid noisy images, state- of-the-art methods for image denoising with deep learning, salt and pepper noise, non-linear filters for digital color images. The main objective of this paper is to provide a comprehensive overview of various approaches used for image denoising, each of which has been explored and developed based on individual research studies. The paper aims to discuss these approaches in a systematic and organized manner, comparing their strengths and weaknesses to provide insights for future research in the field.","doi":"10.1007\/s11042-023-17468-2","cleaned_title":"a comprehensive review of image denoising in deep learning","cleaned_abstract":"deep learning has gained significant interest in image denoising, but there are notable distinctions in the types of deep learning methods used. discriminative learning is suitable for handling gaussian noise, while optimization models are effective in estimating real noise. however, there is limited research that summarizes the different deep learning techniques for image denoising. this paper conducts a comprehensive review of techniques and methods used for image denoising and identifying challenges associated with existing approaches. in this paper, a comparative study of deep techniques is offered in image denoising. the study conducted a comprehensive review of 68 papers on image denoising published between 2018 and 2023, providing a detailed analysis of the field\u2019s progress and methodologies over a period of 5 years. through its literature review, the paper provides a comprehensive summary of image denoising in deep learning, including machine learning methods for image denoising, cnns for image denoising, additive white noisy-image denoising, real noisy image denoising, blind denoising, hybrid noisy images, state- of-the-art methods for image denoising with deep learning, salt and pepper noise, non-linear filters for digital color images. the main objective of this paper is to provide a comprehensive overview of various approaches used for image denoising, each of which has been explored and developed based on individual research studies. the paper aims to discuss these approaches in a systematic and organized manner, comparing their strengths and weaknesses to provide insights for future research in the field.","key_phrases":["deep learning","significant interest","image denoising","notable distinctions","the types","deep learning methods","discriminative learning","gaussian noise","optimization models","real noise","limited research","that","the different deep learning techniques","image denoising","this paper","a comprehensive review","techniques","methods","image denoising","challenges","existing approaches","this paper","a comparative study","deep techniques","image denoising","the study","a comprehensive review","68 papers","image denoising","a detailed analysis","the field\u2019s progress","methodologies","a period","5 years","its literature review","the paper","a comprehensive summary","image","deep learning","machine learning methods","image denoising","cnns","image denoising","additive white noisy-image denoising","real noisy image denoising","blind denoising","hybrid noisy images","state-","the-art","methods","image","deep learning","salt and pepper noise","non-linear filters","digital color images","the main objective","this paper","a comprehensive overview","various approaches","image denoising","each","which","individual research studies","the paper","these approaches","a systematic and organized manner","their strengths","weaknesses","insights","future research","the field","68","between 2018 and 2023","5 years"]},{"title":"Fisheye freshness detection using common deep learning algorithms and machine learning methods with a developed mobile application","authors":["Muslume Beyza Yildiz","Elham Tahsin Yasin","Murat Koklu"],"abstract":"AbstractFish is commonly ingested as a source of protein and essential nutrients for humans. To fully benefit from the proteins and substances in fish it is crucial to ensure its freshness. If fish is stored for an extended period, its freshness deteriorates. Determining the freshness of fish can be done by examining its eyes, smell, skin, and gills. In this study, artificial intelligence techniques are employed to assess fish freshness. The author\u2019s objective is to evaluate the freshness of fish by analyzing its eye characteristics. To achieve this, we have developed a combination of deep and machine learning models that accurately classify the freshness of fish. Furthermore, an application that utilizes both deep learning and machine learning, to instantly detect the freshness of any given fish sample was created. Two deep learning algorithms (SqueezeNet, and VGG19) were implemented to extract features from image data. Additionally, five machine learning models to classify the freshness levels of fish samples were applied. Machine learning models include (k-NN, RF, SVM, LR, and ANN). Based on the results, it can be inferred that employing the VGG19 model for feature selection in conjunction with an Artificial Neural Network (ANN) for classification yields the most favorable success rate of 77.3% for the FFE dataset.Graphical Abstract","doi":"10.1007\/s00217-024-04493-0","cleaned_title":"fisheye freshness detection using common deep learning algorithms and machine learning methods with a developed mobile application","cleaned_abstract":"abstractfish is commonly ingested as a source of protein and essential nutrients for humans. to fully benefit from the proteins and substances in fish it is crucial to ensure its freshness. if fish is stored for an extended period, its freshness deteriorates. determining the freshness of fish can be done by examining its eyes, smell, skin, and gills. in this study, artificial intelligence techniques are employed to assess fish freshness. the author\u2019s objective is to evaluate the freshness of fish by analyzing its eye characteristics. to achieve this, we have developed a combination of deep and machine learning models that accurately classify the freshness of fish. furthermore, an application that utilizes both deep learning and machine learning, to instantly detect the freshness of any given fish sample was created. two deep learning algorithms (squeezenet, and vgg19) were implemented to extract features from image data. additionally, five machine learning models to classify the freshness levels of fish samples were applied. machine learning models include (k-nn, rf, svm, lr, and ann). based on the results, it can be inferred that employing the vgg19 model for feature selection in conjunction with an artificial neural network (ann) for classification yields the most favorable success rate of 77.3% for the ffe dataset.graphical abstract","key_phrases":["abstractfish","a source","protein","essential nutrients","humans","the proteins","substances","fish","it","its freshness","fish","an extended period","its freshness","the freshness","fish","its eyes","smell","skin","gills","this study","artificial intelligence techniques","fish freshness","the author\u2019s objective","the freshness","fish","its eye characteristics","this","we","a combination","deep and machine learning models","that","the freshness","fish","an application","that","both deep learning","machine learning","the freshness","any given fish sample","two deep learning algorithms","squeezenet","vgg19","features","image data","five machine learning models","the freshness levels","fish samples","machine learning models","k",", rf, svm","lr","ann","the results","it","the vgg19 model","feature selection","conjunction","an artificial neural network","ann","classification yields","the most favorable success rate","77.3%","the ffe dataset.graphical abstract","two","five","77.3%"]},{"title":"Deep learning for transesophageal echocardiography view classification","authors":["Kirsten R. Steffner","Matthew Christensen","George Gill","Michael Bowdish","Justin Rhee","Abirami Kumaresan","Bryan He","James Zou","David Ouyang"],"abstract":"Transesophageal echocardiography (TEE) imaging is a vital tool used in the evaluation of complex cardiac pathology and the management of cardiac surgery patients. A key limitation to the application of deep learning strategies to intraoperative and intraprocedural TEE data is the complexity and unstructured nature of these images. In the present study, we developed a deep learning-based, multi-category TEE view classification model that can be used to add structure to intraoperative and intraprocedural TEE imaging data. More specifically, we trained a convolutional neural network (CNN) to predict standardized TEE views using labeled intraoperative and intraprocedural TEE videos from Cedars-Sinai Medical Center (CSMC). We externally validated our model on intraoperative TEE videos from Stanford University Medical Center (SUMC). Accuracy of our model was high across all labeled views. The highest performance was achieved for the Trans-Gastric Left Ventricular Short Axis View (area under the receiver operating curve [AUC]\u2009=\u20090.971 at CSMC, 0.957 at SUMC), the Mid-Esophageal Long Axis View (AUC\u2009=\u20090.954 at CSMC, 0.905 at SUMC), the Mid-Esophageal Aortic Valve Short Axis View (AUC\u2009=\u20090.946 at CSMC, 0.898 at SUMC), and the Mid-Esophageal 4-Chamber View (AUC\u2009=\u20090.939 at CSMC, 0.902 at SUMC). Ultimately, we demonstrate that our deep learning model can accurately classify standardized TEE views, which will facilitate further downstream deep learning analyses for intraoperative and intraprocedural TEE imaging.","doi":"10.1038\/s41598-023-50735-8","cleaned_title":"deep learning for transesophageal echocardiography view classification","cleaned_abstract":"transesophageal echocardiography (tee) imaging is a vital tool used in the evaluation of complex cardiac pathology and the management of cardiac surgery patients. a key limitation to the application of deep learning strategies to intraoperative and intraprocedural tee data is the complexity and unstructured nature of these images. in the present study, we developed a deep learning-based, multi-category tee view classification model that can be used to add structure to intraoperative and intraprocedural tee imaging data. more specifically, we trained a convolutional neural network (cnn) to predict standardized tee views using labeled intraoperative and intraprocedural tee videos from cedars-sinai medical center (csmc). we externally validated our model on intraoperative tee videos from stanford university medical center (sumc). accuracy of our model was high across all labeled views. the highest performance was achieved for the trans-gastric left ventricular short axis view (area under the receiver operating curve [auc] = 0.971 at csmc, 0.957 at sumc), the mid-esophageal long axis view (auc = 0.954 at csmc, 0.905 at sumc), the mid-esophageal aortic valve short axis view (auc = 0.946 at csmc, 0.898 at sumc), and the mid-esophageal 4-chamber view (auc = 0.939 at csmc, 0.902 at sumc). ultimately, we demonstrate that our deep learning model can accurately classify standardized tee views, which will facilitate further downstream deep learning analyses for intraoperative and intraprocedural tee imaging.","key_phrases":["transesophageal echocardiography","(tee) imaging","a vital tool","the evaluation","complex cardiac pathology","the management","cardiac surgery patients","a key limitation","the application","deep learning strategies","to intraoperative and intraprocedural tee data","the complexity","unstructured nature","these images","the present study","we","a deep learning-based, multi-category tee view classification model","that","structure","intraoperative and intraprocedural tee imaging data","we","a convolutional neural network","cnn","standardized tee views","labeled intraoperative and intraprocedural tee videos","cedars-sinai medical center","csmc","we","our model","intraoperative tee videos","stanford university medical center","sumc","accuracy","our model","all labeled views","the highest performance","ventricular short axis view","area","the receiver operating curve","csmc","sumc",", the mid-esophageal long axis view","auc =","csmc","sumc","the mid-esophageal aortic valve short axis view","csmc","sumc","the mid-esophageal 4-chamber view","csmc","sumc","we","our deep learning model","standardized tee views","which","further downstream deep learning analyses","intraoperative and intraprocedural tee imaging","cnn","stanford university medical center","0.971","0.957","0.954","0.905","0.946","0.898","4","0.939","0.902"]},{"title":"Context-aware geometric deep learning for protein sequence design","authors":["Lucien F. Krapp","Fernando A. Meireles","Luciano A. Abriata","Jean Devillard","Sarah Vacle","Maria J. Marcaida","Matteo Dal Peraro"],"abstract":"Protein design and engineering are evolving at an unprecedented pace leveraging the advances in deep learning. Current models nonetheless cannot natively consider non-protein entities within the design process. Here, we introduce a deep learning approach based solely on a geometric transformer of atomic coordinates and element names that predicts protein sequences from backbone scaffolds aware of the restraints imposed by diverse molecular environments. To validate the method, we show that it can produce highly thermostable, catalytically active enzymes with high success rates. This concept is anticipated to improve the versatility of protein design pipelines for crafting desired functions.","doi":"10.1038\/s41467-024-50571-y","cleaned_title":"context-aware geometric deep learning for protein sequence design","cleaned_abstract":"protein design and engineering are evolving at an unprecedented pace leveraging the advances in deep learning. current models nonetheless cannot natively consider non-protein entities within the design process. here, we introduce a deep learning approach based solely on a geometric transformer of atomic coordinates and element names that predicts protein sequences from backbone scaffolds aware of the restraints imposed by diverse molecular environments. to validate the method, we show that it can produce highly thermostable, catalytically active enzymes with high success rates. this concept is anticipated to improve the versatility of protein design pipelines for crafting desired functions.","key_phrases":["protein design","engineering","an unprecedented pace","the advances","deep learning","current models","non-protein entities","the design process","we","a deep learning approach","a geometric transformer","atomic coordinates","element names","that","protein sequences","backbone scaffolds","the restraints","diverse molecular environments","the method","we","it","highly thermostable, catalytically active enzymes","high success rates","this concept","the versatility","protein design pipelines","desired functions"]},{"title":"Automated optical inspection based on synthetic mechanisms combining deep learning and machine learning","authors":["Chung-Ming Lo","Ting-Yi Lin"],"abstract":"The quality inspection of products before delivery plays a critical role in ensuring manufacturing quality. Quick and accurate inspection of samples is realized by highly automated inspection based on pattern recognition in smart manufacturing. Conventional ensemble methods have been demonstrated to be effective for defect detection. This study further proposed synthetic mechanisms based on using various features and learning classifiers. A database of 6000 sample images of printed circuit board (PCB) connectors collected from factories was compiled. A novel confidence synthesis mechanism was proposed to prescreen images using deep learning features. Spatially connected texture features were then used to reclassify images with low reliabilities. The synthetic mechanism was found to outperform a single classifier. In particular, the highest improvement in accuracy (from 96.00 to 97.83%) was obtained using the confidence-based synthesis. The synthetic mechanism can be used to achieve high accuracy in defect detection and make automation in smart manufacturing more practicable.","doi":"10.1007\/s10845-024-02474-4","cleaned_title":"automated optical inspection based on synthetic mechanisms combining deep learning and machine learning","cleaned_abstract":"the quality inspection of products before delivery plays a critical role in ensuring manufacturing quality. quick and accurate inspection of samples is realized by highly automated inspection based on pattern recognition in smart manufacturing. conventional ensemble methods have been demonstrated to be effective for defect detection. this study further proposed synthetic mechanisms based on using various features and learning classifiers. a database of 6000 sample images of printed circuit board (pcb) connectors collected from factories was compiled. a novel confidence synthesis mechanism was proposed to prescreen images using deep learning features. spatially connected texture features were then used to reclassify images with low reliabilities. the synthetic mechanism was found to outperform a single classifier. in particular, the highest improvement in accuracy (from 96.00 to 97.83%) was obtained using the confidence-based synthesis. the synthetic mechanism can be used to achieve high accuracy in defect detection and make automation in smart manufacturing more practicable.","key_phrases":["the quality inspection","products","delivery","a critical role","manufacturing quality","quick and accurate inspection","samples","highly automated inspection","pattern recognition","smart manufacturing","conventional ensemble methods","defect detection","this study","further proposed synthetic mechanisms","various features","learning classifiers","a database","6000 sample images","printed circuit board (pcb) connectors","factories","a novel confidence synthesis mechanism","images","deep learning features","spatially connected texture features","images","low reliabilities","the synthetic mechanism","a single classifier","the highest improvement","accuracy","97.83%","the confidence-based synthesis","the synthetic mechanism","high accuracy","defect detection","automation","smart manufacturing","6000","96.00","97.83%"]},{"title":"Healthcare predictive analytics using machine learning and deep learning techniques: a survey","authors":["Mohammed Badawy","Nagy Ramadan","Hesham Ahmed Hefny"],"abstract":"Healthcare prediction has been a significant factor in saving lives in recent years. In the domain of health care, there is a rapid development of intelligent systems for analyzing complicated data relationships and transforming them into real information for use in the prediction process. Consequently, artificial intelligence is rapidly transforming the healthcare industry, and thus comes the role of systems depending on machine learning and deep learning in the creation of steps that diagnose and predict diseases, whether from clinical data or based on images, that provide tremendous clinical support by simulating human perception and can even diagnose diseases that are difficult to detect by human intelligence. Predictive analytics for healthcare a\u00a0critical imperative in the healthcare industry. It can significantly affect the accuracy of disease prediction, which may lead to saving patients' lives in the case of accurate and timely prediction; on the contrary, in the case of an incorrect prediction, it may endanger patients' lives. Therefore, diseases must be accurately predicted and estimated. Hence, reliable and efficient methods for healthcare predictive analysis are essential. Therefore, this paper aims to present a comprehensive survey of existing machine learning and deep learning approaches utilized in healthcare prediction and identify the inherent obstacles to applying these approaches in the healthcare domain.","doi":"10.1186\/s43067-023-00108-y","cleaned_title":"healthcare predictive analytics using machine learning and deep learning techniques: a survey","cleaned_abstract":"healthcare prediction has been a significant factor in saving lives in recent years. in the domain of health care, there is a rapid development of intelligent systems for analyzing complicated data relationships and transforming them into real information for use in the prediction process. consequently, artificial intelligence is rapidly transforming the healthcare industry, and thus comes the role of systems depending on machine learning and deep learning in the creation of steps that diagnose and predict diseases, whether from clinical data or based on images, that provide tremendous clinical support by simulating human perception and can even diagnose diseases that are difficult to detect by human intelligence. predictive analytics for healthcare a critical imperative in the healthcare industry. it can significantly affect the accuracy of disease prediction, which may lead to saving patients' lives in the case of accurate and timely prediction; on the contrary, in the case of an incorrect prediction, it may endanger patients' lives. therefore, diseases must be accurately predicted and estimated. hence, reliable and efficient methods for healthcare predictive analysis are essential. therefore, this paper aims to present a comprehensive survey of existing machine learning and deep learning approaches utilized in healthcare prediction and identify the inherent obstacles to applying these approaches in the healthcare domain.","key_phrases":["healthcare prediction","a significant factor","lives","recent years","the domain","health care","a rapid development","intelligent systems","complicated data relationships","them","real information","use","the prediction process","artificial intelligence","the healthcare industry","the role","systems","machine learning","deep learning","the creation","steps","that","diseases","clinical data","images","that","tremendous clinical support","human perception","diseases","that","human intelligence","predictive analytics","healthcare","the healthcare industry","it","the accuracy","disease prediction","which","patients' lives","the case","accurate and timely prediction","the contrary","the case","an incorrect prediction","it","patients' lives","diseases","reliable and efficient methods","healthcare predictive analysis","this paper","a comprehensive survey","existing machine learning","deep learning approaches","healthcare prediction","the inherent obstacles","these approaches","the healthcare domain","healthcare","recent years"]},{"title":"LSNet: a deep learning based method for skin lesion classification using limited samples and transfer learning","authors":["Xiaodan Deng"],"abstract":"When analyzing skin lesion image data using deep learning, the lack of a sufficient amount of effective training data poses a challenge. Although transfer learning can alleviate the problem of a small amount of data, the difference between the source data and the target data makes the tranfer learning missing some key knowledge. It is important to find the key knowledge neglected by transfer learning. This paper argues that this key knowledge is contained in challenging samples. We propose a novel method named as Limited Samples Network (LSNet) to search challenging samples and strengthen the learning of them. Specifically, LSNet utilizes patch-based structured input and employs pseudoinverse learning autoencoder to quickly obtain position-sensitive loss. Challenging samples can be obtained by searching for position-sensitive loss. Subsequently, challenging samples-augmented transfer learning is employed to enhance the classification performance of deep learning models on skin lesion datasets with limited samples. We carry comparison experiment with the existing state-of-the-art method. Experiments are carried out on the ISIC 2017, ISIC 2018 and ISIC 2019 skin lesion datasets. The results demonstrate that our training strategy significantly improves the vanilla transfer learning procedure for different types of pre-trained DCNNs. In particular, our method achieves state-of-the-art performance on different skin lesion datasets without using any extra training data.","doi":"10.1007\/s11042-023-17975-2","cleaned_title":"lsnet: a deep learning based method for skin lesion classification using limited samples and transfer learning","cleaned_abstract":"when analyzing skin lesion image data using deep learning, the lack of a sufficient amount of effective training data poses a challenge. although transfer learning can alleviate the problem of a small amount of data, the difference between the source data and the target data makes the tranfer learning missing some key knowledge. it is important to find the key knowledge neglected by transfer learning. this paper argues that this key knowledge is contained in challenging samples. we propose a novel method named as limited samples network (lsnet) to search challenging samples and strengthen the learning of them. specifically, lsnet utilizes patch-based structured input and employs pseudoinverse learning autoencoder to quickly obtain position-sensitive loss. challenging samples can be obtained by searching for position-sensitive loss. subsequently, challenging samples-augmented transfer learning is employed to enhance the classification performance of deep learning models on skin lesion datasets with limited samples. we carry comparison experiment with the existing state-of-the-art method. experiments are carried out on the isic 2017, isic 2018 and isic 2019 skin lesion datasets. the results demonstrate that our training strategy significantly improves the vanilla transfer learning procedure for different types of pre-trained dcnns. in particular, our method achieves state-of-the-art performance on different skin lesion datasets without using any extra training data.","key_phrases":["skin lesion image data","deep learning","the lack","a sufficient amount","effective training data","a challenge","transfer learning","the problem","a small amount","data","the difference","the source data","the target data","the tranfer","some key knowledge","it","the key knowledge","transfer learning","this paper","this key knowledge","challenging samples","we","a novel method","limited samples network","lsnet","challenging samples","the learning","them","lsnet","patch-based structured input","pseudoinverse","autoencoder","position-sensitive loss","challenging samples","position-sensitive loss","challenging samples-augmented transfer learning","the classification performance","deep learning models","skin lesion datasets","limited samples","we","comparison experiment","the-art","experiments","the isic","2019 skin lesion datasets","the results","our training strategy","the vanilla transfer learning procedure","different types","pre-trained dcnns","our method","the-art","different skin lesion datasets","any extra training data","2017","2018","2019"]},{"title":"Classical learning or deep learning: a study on food photo aesthetic assessment","authors":["Zhaotong Li","Zeru Zhang","Song Gao"],"abstract":"Food photo aesthetic assessment has gained increasing attention in both commercial activity and social life. However, there has been little research dedicated to the quality classification of food photos. This paper presents a study on food photo aesthetic evaluation, covering dataset collection and evaluation methods. First, a dataset of food photos was collected by web crawler from food-sharing websites, and the appropriate images were selected and labeled using a WeChat applet for binary classification. Then, food photo aesthetic assessment was evaluated using classical machine learning and deep learning methods. Different hand-crafted features, including layout, texture, color, local, and deep features, were manually extracted. Two classifiers, support vector machine and random forest, were used to establish the classical learning models. Meanwhile, three convolutional neural networks (AlexNet, VGGNet, ResNet) were applied to compare with former methods by fine-tuning the model parameters. Four quantitative metrics (accuracy, recall, precision, and f1-score) were used to evaluate the performance of food photo aesthetic assessment, with the accuracy of classical and deep learning methods being 91.09% vs 94.70%, respectively. This demonstrates that classical learning with good enough hand-crafted features is capable of producing performance close to that of CNNs. The dataset for food photo aesthetic assessment can be used as a preliminary exploration of food image aesthetics assessment from both classical learning and deep learning.","doi":"10.1007\/s11042-023-15791-2","cleaned_title":"classical learning or deep learning: a study on food photo aesthetic assessment","cleaned_abstract":"food photo aesthetic assessment has gained increasing attention in both commercial activity and social life. however, there has been little research dedicated to the quality classification of food photos. this paper presents a study on food photo aesthetic evaluation, covering dataset collection and evaluation methods. first, a dataset of food photos was collected by web crawler from food-sharing websites, and the appropriate images were selected and labeled using a wechat applet for binary classification. then, food photo aesthetic assessment was evaluated using classical machine learning and deep learning methods. different hand-crafted features, including layout, texture, color, local, and deep features, were manually extracted. two classifiers, support vector machine and random forest, were used to establish the classical learning models. meanwhile, three convolutional neural networks (alexnet, vggnet, resnet) were applied to compare with former methods by fine-tuning the model parameters. four quantitative metrics (accuracy, recall, precision, and f1-score) were used to evaluate the performance of food photo aesthetic assessment, with the accuracy of classical and deep learning methods being 91.09% vs 94.70%, respectively. this demonstrates that classical learning with good enough hand-crafted features is capable of producing performance close to that of cnns. the dataset for food photo aesthetic assessment can be used as a preliminary exploration of food image aesthetics assessment from both classical learning and deep learning.","key_phrases":["food photo aesthetic assessment","increasing attention","both commercial activity","social life","little research","the quality classification","food photos","this paper","a study","food photo aesthetic evaluation","dataset collection and evaluation methods","a dataset","food photos","web crawler","food-sharing websites","the appropriate images","a wechat applet","binary classification","food photo aesthetic assessment","classical machine learning","deep learning methods","different hand-crafted features","layout","texture","color","deep features","two classifiers","support vector machine","random forest","the classical learning models","three convolutional neural networks","alexnet","vggnet","resnet","former methods","the model parameters","four quantitative metrics","accuracy","recall","precision","f1-score","the performance","food photo aesthetic assessment","the accuracy","classical and deep learning methods","91.09%","94.70%","this","classical learning","good enough hand-crafted features","performance","that","cnns","the dataset","food photo aesthetic assessment","a preliminary exploration","food image aesthetics assessment","both classical learning","deep learning","first","two","three","four","91.09%","94.70%"]},{"title":"Deep learning in terrestrial conservation biology","authors":["Zolt\u00e1n Barta"],"abstract":"Biodiversity is being lost at an unprecedented rate on Earth. As a first step to more effectively combat this process we need efficient methods to monitor biodiversity changes. Recent technological advance can provide powerful tools (e.g.\u00a0camera traps, digital acoustic recorders, satellite imagery, social media records) that can speed up the collection of biological data. Nevertheless, the processing steps of the raw data served by these tools are still painstakingly slow. A new computer technology, deep learning based artificial intelligence, might, however, help. In this short and subjective review I oversee recent technological advances used in conservation biology, highlight problems of processing their data, shortly describe deep learning technology and show case studies of its use in conservation biology. Some of the limitations of the technology are also highlighted.","doi":"10.1007\/s42977-023-00200-4","cleaned_title":"deep learning in terrestrial conservation biology","cleaned_abstract":"biodiversity is being lost at an unprecedented rate on earth. as a first step to more effectively combat this process we need efficient methods to monitor biodiversity changes. recent technological advance can provide powerful tools (e.g. camera traps, digital acoustic recorders, satellite imagery, social media records) that can speed up the collection of biological data. nevertheless, the processing steps of the raw data served by these tools are still painstakingly slow. a new computer technology, deep learning based artificial intelligence, might, however, help. in this short and subjective review i oversee recent technological advances used in conservation biology, highlight problems of processing their data, shortly describe deep learning technology and show case studies of its use in conservation biology. some of the limitations of the technology are also highlighted.","key_phrases":["biodiversity","an unprecedented rate","earth","a first step","this process","we","efficient methods","biodiversity changes","recent technological advance","powerful tools","e.g. camera traps","digital acoustic recorders","satellite imagery","social media records","that","the collection","biological data","the processing steps","the raw data","these tools","a new computer technology","based artificial intelligence","this short and subjective review","i","recent technological advances","conservation biology","problems","their data","deep learning technology","show case studies","its use","conservation biology","some","the limitations","the technology","earth","first"]},{"title":"Medical images classification using deep learning: a survey","authors":["Rakesh Kumar","Pooja Kumbharkar","Sandeep Vanam","Sanjeev Sharma"],"abstract":"Deep learning has made significant advancements in recent years. The technology is rapidly evolving and has been used in numerous automated applications with minimal loss. With these deep learning methods, medical image analysis for disease detection can be performed with minimal errors and losses. A survey of deep learning-based medical image classification is presented in this paper. As a result of their automatic feature representations, these methods have high accuracy and precision. This paper reviews various models like CNN, Transfer learning, Long short term memory, Generative adversarial networks, and Autoencoders and their combinations for various purposes in medical image classification. The total number of papers reviewed is 158. In the study, we discussed the advantages and limitations of the methods. A discussion is provided on the various applications of medical imaging, the available datasets for medical imaging, and the evaluation metrics. We also discuss the future trends in medical imaging using artificial intelligence.","doi":"10.1007\/s11042-023-15576-7","cleaned_title":"medical images classification using deep learning: a survey","cleaned_abstract":"deep learning has made significant advancements in recent years. the technology is rapidly evolving and has been used in numerous automated applications with minimal loss. with these deep learning methods, medical image analysis for disease detection can be performed with minimal errors and losses. a survey of deep learning-based medical image classification is presented in this paper. as a result of their automatic feature representations, these methods have high accuracy and precision. this paper reviews various models like cnn, transfer learning, long short term memory, generative adversarial networks, and autoencoders and their combinations for various purposes in medical image classification. the total number of papers reviewed is 158. in the study, we discussed the advantages and limitations of the methods. a discussion is provided on the various applications of medical imaging, the available datasets for medical imaging, and the evaluation metrics. we also discuss the future trends in medical imaging using artificial intelligence.","key_phrases":["deep learning","significant advancements","recent years","the technology","numerous automated applications","minimal loss","these deep learning methods","medical image analysis","disease detection","minimal errors","losses","a survey","deep learning-based medical image classification","this paper","a result","their automatic feature representations","these methods","high accuracy","precision","this paper","various models","cnn","generative adversarial networks","autoencoders","their combinations","various purposes","medical image classification","the total number","papers","the study","we","the advantages","limitations","the methods","a discussion","the various applications","medical imaging","the available datasets","medical imaging","the evaluation metrics","we","the future trends","medical imaging","artificial intelligence","recent years","cnn","158"]},{"title":"Super-resolution Deep Learning Reconstruction Cervical Spine 1.5T MRI: Improved Interobserver Agreement in Evaluations of Neuroforaminal Stenosis Compared to Conventional Deep Learning Reconstruction","authors":["Koichiro Yasaka","Shunichi Uehara","Shimpei Kato","Yusuke Watanabe","Taku Tajima","Hiroyuki Akai","Naoki Yoshioka","Masaaki Akahane","Kuni Ohtomo","Osamu Abe","Shigeru Kiryu"],"abstract":"The aim of this study was to investigate whether super-resolution deep learning reconstruction (SR-DLR) is superior to conventional deep learning reconstruction (DLR) with respect to interobserver agreement in the evaluation of neuroforaminal stenosis using 1.5T cervical spine MRI. This retrospective study included 39 patients who underwent 1.5T cervical spine MRI. T2-weighted sagittal images were reconstructed with SR-DLR and DLR. Three blinded radiologists independently evaluated the images in terms of the degree of neuroforaminal stenosis, depictions of the vertebrae, spinal cord and neural foramina, sharpness, noise, artefacts and diagnostic acceptability. In quantitative image analyses, a fourth radiologist evaluated the signal-to-noise ratio (SNR) by placing a circular or ovoid region of interest on the spinal cord, and the edge slope based on a linear region of interest placed across the surface of the spinal cord. Interobserver agreement in the evaluations of neuroforaminal stenosis using SR-DLR and DLR was 0.422\u20130.571 and 0.410\u20130.542, respectively. The kappa values between reader 1 vs. reader 2 and reader 2 vs. reader 3 significantly differed. Two of the three readers rated depictions of the spinal cord, sharpness, and diagnostic acceptability as significantly better with SR-DLR than with DLR. Both SNR and edge slope (\/mm) were also significantly better with SR-DLR (12.9 and 6031, respectively) than with DLR (11.5 and 3741, respectively) (p\u2009<\u20090.001 for both). In conclusion, compared to DLR, SR-DLR improved interobserver agreement in the evaluations of neuroforaminal stenosis using 1.5T cervical spine MRI.","doi":"10.1007\/s10278-024-01112-y","cleaned_title":"super-resolution deep learning reconstruction cervical spine 1.5t mri: improved interobserver agreement in evaluations of neuroforaminal stenosis compared to conventional deep learning reconstruction","cleaned_abstract":"the aim of this study was to investigate whether super-resolution deep learning reconstruction (sr-dlr) is superior to conventional deep learning reconstruction (dlr) with respect to interobserver agreement in the evaluation of neuroforaminal stenosis using 1.5t cervical spine mri. this retrospective study included 39 patients who underwent 1.5t cervical spine mri. t2-weighted sagittal images were reconstructed with sr-dlr and dlr. three blinded radiologists independently evaluated the images in terms of the degree of neuroforaminal stenosis, depictions of the vertebrae, spinal cord and neural foramina, sharpness, noise, artefacts and diagnostic acceptability. in quantitative image analyses, a fourth radiologist evaluated the signal-to-noise ratio (snr) by placing a circular or ovoid region of interest on the spinal cord, and the edge slope based on a linear region of interest placed across the surface of the spinal cord. interobserver agreement in the evaluations of neuroforaminal stenosis using sr-dlr and dlr was 0.422\u20130.571 and 0.410\u20130.542, respectively. the kappa values between reader 1 vs. reader 2 and reader 2 vs. reader 3 significantly differed. two of the three readers rated depictions of the spinal cord, sharpness, and diagnostic acceptability as significantly better with sr-dlr than with dlr. both snr and edge slope (\/mm) were also significantly better with sr-dlr (12.9 and 6031, respectively) than with dlr (11.5 and 3741, respectively) (p < 0.001 for both). in conclusion, compared to dlr, sr-dlr improved interobserver agreement in the evaluations of neuroforaminal stenosis using 1.5t cervical spine mri.","key_phrases":["the aim","this study","super-resolution deep learning reconstruction","sr-dlr","conventional deep learning reconstruction","dlr","respect","interobserver agreement","the evaluation","neuroforaminal stenosis","1.5t cervical spine mri","this retrospective study","39 patients","who","1.5t cervical spine mri","t2-weighted sagittal images","sr-dlr","dlr","three blinded radiologists","the images","terms","the degree","neuroforaminal stenosis","the vertebrae","spinal cord","neural foramina","sharpness","noise","artefacts","diagnostic acceptability","quantitative image analyses","a fourth radiologist","noise","snr","a circular or ovoid region","interest","the spinal cord","the edge slope","a linear region","interest","the surface","the spinal cord","interobserver agreement","the evaluations","neuroforaminal stenosis","sr-dlr","dlr","0.422\u20130.571","the kappa","reader","reader","reader","the three readers","depictions","the spinal cord","sharpness","diagnostic acceptability","sr-dlr","dlr","edge slope","\/mm","dlr","both","conclusion","dlr","sr-dlr","the evaluations","neuroforaminal stenosis","1.5t cervical spine mri","1.5","39","1.5","three","fourth","0.422\u20130.571","0.410\u20130.542","1","2","2","3","two","three","12.9","6031","11.5","3741","1.5"]},{"title":"Accelerating three-dimensional phase-field simulations via deep learning approaches","authors":["Xuewei Zhou","Sheng Sun","Songlin Cai","Gongyu Chen","Honghui Wu","Jie Xiong","Jiaming Zhu"],"abstract":"Phase-field modeling (PFM) is a powerful but computationally expensive technique for simulating three-dimensional (3D) microstructure evolutions. Very recently, integrating machine learning into phase-field simulations provides a promising way to reduce calculation time remarkably. In this study, we propose a deep learning model that combines a convolutional autoencoder with a deep operator network to predict 3D microstructure evolution by using 2D slices of the 3D system. It is found that the deep learning model can shorten the calculation time from 37\u00a0min to 3\u00a0s after the initial training, while skipping 5-time steps, and reduce the phase-field simulation time by 31% in entire calculation of the\u00a0evolution process. Interestingly, this model achieves good accuracy in predicting 3D microstructures by utilizing only 2D information. This work demonstrates the efficiency of machine learning in accelerating phase-field simulations while maintaining high accuracy and promotes the application of PFM in fundamental studies.","doi":"10.1007\/s10853-024-10118-4","cleaned_title":"accelerating three-dimensional phase-field simulations via deep learning approaches","cleaned_abstract":"phase-field modeling (pfm) is a powerful but computationally expensive technique for simulating three-dimensional (3d) microstructure evolutions. very recently, integrating machine learning into phase-field simulations provides a promising way to reduce calculation time remarkably. in this study, we propose a deep learning model that combines a convolutional autoencoder with a deep operator network to predict 3d microstructure evolution by using 2d slices of the 3d system. it is found that the deep learning model can shorten the calculation time from 37 min to 3 s after the initial training, while skipping 5-time steps, and reduce the phase-field simulation time by 31% in entire calculation of the evolution process. interestingly, this model achieves good accuracy in predicting 3d microstructures by utilizing only 2d information. this work demonstrates the efficiency of machine learning in accelerating phase-field simulations while maintaining high accuracy and promotes the application of pfm in fundamental studies.","key_phrases":["phase-field modeling","pfm","a powerful but computationally expensive technique","three-dimensional (3d) microstructure evolutions","phase-field simulations","a promising way","calculation time","this study","we","a deep learning model","that","a convolutional autoencoder","a deep operator network","3d microstructure evolution","2d slices","the 3d system","it","the deep learning model","the calculation time","37 min","3 s","the initial training","5-time steps","the phase-field simulation time","31%","entire calculation","the evolution process","this model","good accuracy","3d microstructures","only 2d information","this work","the efficiency","accelerating phase-field simulations","high accuracy","the application","pfm","fundamental studies","three","3d","3d","2d","3d","37","3 s","5","31%","3d","2d"]},{"title":"Fostering success in online English education: Exploring the effects of ICT literacy, online learning self-efficacy, and motivation on deep learning","authors":["Wei Sun","Hong Shi"],"abstract":"This research explores how motivation and online learning self-efficacy (OLSE) act as mediators in the association between information and communication technology (ICT) literacy and deep learning within the context of online English as a foreign language (EFL) education. A sample of 372 participants were recruited on a voluntary and anonymous basis from a public university in northern China for this study. Confirmatory factor analysis\u00a0(CFA) was employed to evaluate the reliability and validity of the questionnaires. Subsequently, structural equation modeling (SEM)\u00a0was conducted to examine the hypothesized model, with both CFA and SEM conducted with AMOS 29.0. The results reveal that ICT literacy, motivation, and OLSE positively and directly predict deep learning in online education. Additionally, ICT literacy also positively predicts deep learning indirectly with motivation and OLSE being significant mediators. The findings underscore the importance of ICT literacy, motivation, and OLSE for EFL learners to achieve deep learning in online EFL education. Drawn from these findings, pedagogical implications for alleviating EFL learners\u2019 online deep learning were provided.","doi":"10.1007\/s10639-024-12827-4","cleaned_title":"fostering success in online english education: exploring the effects of ict literacy, online learning self-efficacy, and motivation on deep learning","cleaned_abstract":"this research explores how motivation and online learning self-efficacy (olse) act as mediators in the association between information and communication technology (ict) literacy and deep learning within the context of online english as a foreign language (efl) education. a sample of 372 participants were recruited on a voluntary and anonymous basis from a public university in northern china for this study. confirmatory factor analysis (cfa) was employed to evaluate the reliability and validity of the questionnaires. subsequently, structural equation modeling (sem) was conducted to examine the hypothesized model, with both cfa and sem conducted with amos 29.0. the results reveal that ict literacy, motivation, and olse positively and directly predict deep learning in online education. additionally, ict literacy also positively predicts deep learning indirectly with motivation and olse being significant mediators. the findings underscore the importance of ict literacy, motivation, and olse for efl learners to achieve deep learning in online efl education. drawn from these findings, pedagogical implications for alleviating efl learners\u2019 online deep learning were provided.","key_phrases":["this research","motivation","self-efficacy","olse","mediators","the association","information","communication technology","ict) literacy","deep learning","the context","online english","a foreign language (efl) education","a sample","372 participants","a voluntary and anonymous basis","a public university","northern china","this study","confirmatory factor analysis","cfa","the reliability","validity","the questionnaires","sem","the hypothesized model","both cfa","sem","amos","the results","ict literacy","motivation","olse","deep learning","online education","ict literacy","motivation","olse","significant mediators","the findings","the importance","ict literacy","motivation","olse","efl learners","deep learning","online efl education","these findings","pedagogical implications","efl learners","online deep learning","english","372","china","cfa","cfa","29.0"]},{"title":"Efficient socket-based data transmission method and implementation in deep learning","authors":["Xin-Jian Wei","Shu-Ping Li","Wu-Yang Yang","Xiang-Yang Zhang","Hai-Shan Li","Xin Xu","Nan Wang","Zhanbao Fu"],"abstract":"The deep learning algorithm, which has been increasingly applied in the field of petroleum geophysical prospecting, has achieved good results in improving efficiency and accuracy based on test applications. To play a greater role in actual production, these algorithm modules must be integrated into software systems and used more often in actual production projects. Deep learning frameworks, such as TensorFlow and PyTorch, basically take Python as the core architecture, while the application program mainly uses Java, C#, and other programming languages. During integration, the seismic data read by the Java and C# data interfaces must be transferred to the Python main program module. The data exchange methods between Java, C#, and Python include shared memory, shared directory, and so on. However, these methods have the disadvantages of low transmission efficiency and unsuitability for asynchronous networks. Considering the large volume of seismic data and the need for network support for deep learning, this paper proposes a method of transmitting seismic data based on Socket. By maximizing Socket\u2019s cross-network and efficient longdistance transmission, this approach solves the problem of inefficient transmission of underlying data while integrating the deep learning algorithm module into a software system. Furthermore, the actual production application shows that this method effectively solves the shortage of data transmission in shared memory, shared directory, and other modes while simultaneously improving the transmission efficiency of massive seismic data across modules at the bottom of the software.","doi":"10.1007\/s11770-024-1090-y","cleaned_title":"efficient socket-based data transmission method and implementation in deep learning","cleaned_abstract":"the deep learning algorithm, which has been increasingly applied in the field of petroleum geophysical prospecting, has achieved good results in improving efficiency and accuracy based on test applications. to play a greater role in actual production, these algorithm modules must be integrated into software systems and used more often in actual production projects. deep learning frameworks, such as tensorflow and pytorch, basically take python as the core architecture, while the application program mainly uses java, c#, and other programming languages. during integration, the seismic data read by the java and c# data interfaces must be transferred to the python main program module. the data exchange methods between java, c#, and python include shared memory, shared directory, and so on. however, these methods have the disadvantages of low transmission efficiency and unsuitability for asynchronous networks. considering the large volume of seismic data and the need for network support for deep learning, this paper proposes a method of transmitting seismic data based on socket. by maximizing socket\u2019s cross-network and efficient longdistance transmission, this approach solves the problem of inefficient transmission of underlying data while integrating the deep learning algorithm module into a software system. furthermore, the actual production application shows that this method effectively solves the shortage of data transmission in shared memory, shared directory, and other modes while simultaneously improving the transmission efficiency of massive seismic data across modules at the bottom of the software.","key_phrases":["the deep learning algorithm","which","the field","petroleum geophysical prospecting","good results","efficiency","accuracy","test applications","a greater role","actual production","these algorithm modules","software systems","actual production projects","deep learning frameworks","tensorflow","pytorch","python","the core architecture","the application program","integration","the seismic data","the java and c# data interfaces","the python main program module","the data exchange methods","java","python","shared memory","shared directory","these methods","the disadvantages","low transmission efficiency","unsuitability","asynchronous networks","the large volume","seismic data","the need","network support","deep learning","this paper","a method","seismic data","socket","socket","-","efficient longdistance transmission","this approach","the problem","inefficient transmission","underlying data","the deep learning algorithm module","a software system","the actual production application","this method","the shortage","data transmission","shared memory","shared directory","other modes","the transmission efficiency","massive seismic data","modules","the bottom","the software","#","java","#","java","#"]},{"title":"Regularization by deep learning in signal processing","authors":["Carlos Ramirez Villamarin","Erwin Suazo","Tamer Oraby"],"abstract":"In this paper, we explore a new idea of using deep learning representations as a principle for regularization in inverse problems for digital signal processing. Specifically, we consider the standard variational formulation, where a composite function encodes a fidelity term that quantifies the proximity of the candidate solution to the observations (under a physical process), and a second regularization term that constrains the space of solutions according to some prior knowledge. In this work, we investigate deep learning representations as a means of fulfilling the role of this second (regularization) term. Several numerical examples are presented for signal restoration under different degradation processes, showing successful recovery under the proposed methodology. Moreover, one of these examples uses real data on energy usage by households in London from 2012 to 2014.","doi":"10.1007\/s11760-024-03083-7","cleaned_title":"regularization by deep learning in signal processing","cleaned_abstract":"in this paper, we explore a new idea of using deep learning representations as a principle for regularization in inverse problems for digital signal processing. specifically, we consider the standard variational formulation, where a composite function encodes a fidelity term that quantifies the proximity of the candidate solution to the observations (under a physical process), and a second regularization term that constrains the space of solutions according to some prior knowledge. in this work, we investigate deep learning representations as a means of fulfilling the role of this second (regularization) term. several numerical examples are presented for signal restoration under different degradation processes, showing successful recovery under the proposed methodology. moreover, one of these examples uses real data on energy usage by households in london from 2012 to 2014.","key_phrases":["this paper","we","a new idea","deep learning representations","a principle","regularization","inverse problems","digital signal processing","we","the standard variational formulation","a composite function","a fidelity term","that","the proximity","the candidate solution","the observations","a physical process","a second regularization term","that","the space","solutions","some prior knowledge","this work","we","deep learning representations","a means","the role","this second (regularization) term","several numerical examples","signal restoration","different degradation processes","successful recovery","the proposed methodology","these examples","real data","energy usage","households","london","fidelity","second","one","london","2012","2014"]},{"title":"Network intrusion detection using feature fusion with deep learning","authors":["Abiodun Ayantayo","Amrit Kaur","Anit Kour","Xavier Schmoor","Fayyaz Shah","Ian Vickers","Paul Kearney","Mohammed M. Abdelsamea"],"abstract":"Network intrusion detection systems (NIDSs) are one of the main tools used to defend against cyber-attacks. Deep learning has shown remarkable success in network intrusion detection. However, the effect of feature fusion has yet to be explored in how to boost the performance of the deep learning model and improve its generalisation capability in NIDS. In this paper, we propose novel deep learning architectures with different feature fusion mechanisms aimed at improving the performance of the multi-classification components of NIDS. We propose three different deep learning models, which we call early-fusion, late-fusion, and late-ensemble learning models using feature fusion with fully connected deep networks. Our feature fusion mechanisms were designed to encourage deep learning models to learn relationships between different input features more efficiently and mitigate any potential bias that may occur with a particular feature type. To assess the efficacy of our deep learning solutions and make comparisons with state-of-the-art models, we employ the widely accessible UNSW-NB15 and NSL-KDD datasets specifically designed to enhance the development and evaluation of improved NIDSs. Through quantitative analysis, we demonstrate the resilience of our proposed models in effectively addressing the challenges posed by multi-classification tasks, especially in the presence of class imbalance issues. Moreover, our late-fusion and late-ensemble models showed the best generalisation behaviour (against overfitting) with similar performance on the training and validation sets.","doi":"10.1186\/s40537-023-00834-0","cleaned_title":"network intrusion detection using feature fusion with deep learning","cleaned_abstract":"network intrusion detection systems (nidss) are one of the main tools used to defend against cyber-attacks. deep learning has shown remarkable success in network intrusion detection. however, the effect of feature fusion has yet to be explored in how to boost the performance of the deep learning model and improve its generalisation capability in nids. in this paper, we propose novel deep learning architectures with different feature fusion mechanisms aimed at improving the performance of the multi-classification components of nids. we propose three different deep learning models, which we call early-fusion, late-fusion, and late-ensemble learning models using feature fusion with fully connected deep networks. our feature fusion mechanisms were designed to encourage deep learning models to learn relationships between different input features more efficiently and mitigate any potential bias that may occur with a particular feature type. to assess the efficacy of our deep learning solutions and make comparisons with state-of-the-art models, we employ the widely accessible unsw-nb15 and nsl-kdd datasets specifically designed to enhance the development and evaluation of improved nidss. through quantitative analysis, we demonstrate the resilience of our proposed models in effectively addressing the challenges posed by multi-classification tasks, especially in the presence of class imbalance issues. moreover, our late-fusion and late-ensemble models showed the best generalisation behaviour (against overfitting) with similar performance on the training and validation sets.","key_phrases":["network intrusion detection systems","nidss","the main tools","cyber-attacks","deep learning","remarkable success","network intrusion detection","the effect","feature fusion","the performance","the deep learning model","its generalisation capability","nids","this paper","we","novel deep learning architectures","different feature fusion mechanisms","the performance","the multi-classification components","nids","we","three different deep learning models","which","we","early-fusion, late-fusion, and late-ensemble learning models","feature fusion","fully connected deep networks","our feature fusion mechanisms","deep learning models","relationships","different input features","any potential bias","that","a particular feature type","the efficacy","our deep learning solutions","comparisons","the-art","we","the widely accessible unsw-nb15 and nsl-kdd datasets","the development","evaluation","improved nidss","quantitative analysis","we","the resilience","our proposed models","the challenges","multi-classification tasks","the presence","class imbalance issues","our late-fusion and late-ensemble models","the best generalisation behaviour","overfitting","similar performance","the training and validation sets","one","three"]},{"title":"Amplifying document categorization with advanced features and deep learning","authors":["M. Kavitha","K. Akila"],"abstract":"The field of natural language processing (NLP) plays a pivotal role in discerning unstructured data from diverse origins. This study employs advanced techniques rooted in machine learning and deep learning to effectively categorize news articles. Notably, deep learning models have demonstrated superior performance over traditional machine learning algorithms, rendering them a popular choice for a range of NLP tasks. The research employs feature extraction techniques to identify multiword tokens, negation words, and out-of-vocabulary words and replace them. Additionally, convolutional neural network models leverage embedding, convolutional layers, and max pooling layers to capture intricate features. For tasks requiring an understanding of dependencies among long phrases, long short-term memory models come into play. The evaluation of the proposed model hinges on training it with datasets like AG News, BBC, and 20 Newsgroup, gauging its efficacy. The study delves into the myriad challenges inherent to text classification. These challenges are thoughtfully discussed, shedding light on the intricacies of the process. Furthermore, the research furnishes comprehensive test outcomes for both conventional machine learning and deep learning models. The significance of this proposed model is that it uses a multiword expression lexicon, wordnet synset, and word embedding techniques for feature extraction. The performance of the models is increased when using these feature extraction techniques.","doi":"10.1007\/s11042-024-18483-7","cleaned_title":"amplifying document categorization with advanced features and deep learning","cleaned_abstract":"the field of natural language processing (nlp) plays a pivotal role in discerning unstructured data from diverse origins. this study employs advanced techniques rooted in machine learning and deep learning to effectively categorize news articles. notably, deep learning models have demonstrated superior performance over traditional machine learning algorithms, rendering them a popular choice for a range of nlp tasks. the research employs feature extraction techniques to identify multiword tokens, negation words, and out-of-vocabulary words and replace them. additionally, convolutional neural network models leverage embedding, convolutional layers, and max pooling layers to capture intricate features. for tasks requiring an understanding of dependencies among long phrases, long short-term memory models come into play. the evaluation of the proposed model hinges on training it with datasets like ag news, bbc, and 20 newsgroup, gauging its efficacy. the study delves into the myriad challenges inherent to text classification. these challenges are thoughtfully discussed, shedding light on the intricacies of the process. furthermore, the research furnishes comprehensive test outcomes for both conventional machine learning and deep learning models. the significance of this proposed model is that it uses a multiword expression lexicon, wordnet synset, and word embedding techniques for feature extraction. the performance of the models is increased when using these feature extraction techniques.","key_phrases":["the field","natural language processing","nlp","a pivotal role","unstructured data","diverse origins","this study","advanced techniques","machine learning","deep learning","news articles","deep learning models","superior performance","traditional machine learning algorithms","them","a popular choice","a range","nlp tasks","the research","feature extraction techniques","multiword tokens","negation words","out-of-vocabulary words","them","convolutional neural network models","max pooling layers","intricate features","tasks","an understanding","dependencies","long phrases","long short-term memory models","play","the evaluation","the proposed model hinges","it","datasets","ag news","20 newsgroup","its efficacy","the study","the myriad challenges","classification","these challenges","light","the intricacies","the process","the research","comprehensive test outcomes","both conventional machine learning","deep learning models","the significance","this proposed model","it","a multiword expression lexicon","wordnet synset","techniques","feature extraction","the performance","the models","these feature extraction techniques","multiword tokens","max","ag news","bbc","20","multiword"]},{"title":"Severity Grading of Ulcerative Colitis Using Endoscopy Images: An Ensembled Deep Learning and Transfer Learning Approach","authors":["Subhashree Mohapatra","Pukhraj Singh Jeji","Girish Kumar Pati","Janmenjoy Nayak","Manohar Mishra","Tripti Swarnkar"],"abstract":"Ulcerative colitis (UC) is a persistent condition necessitating prompt treatment to avert potential complications. Detecting UC severity aids treatment decisions. The Mayo-endoscopic subscore is a standard for UC severity grading (UCSG). Deep learning (DL) and transfer learning (TL) have enhanced severity grading, but ensemble learning\u2019s impact remains unexplored. This study designed DL-ensemble and TL-ensemble models for UCSG. Using the HyperKvasir dataset, we classified UCSG into two stages: initial and advanced. Three deep convolutional neural networks were trained from scratch for DL, and three pre-trained networks were trained for TL. UCSG was conducted using a majority voting ensemble scheme. A detailed comparative analysis evaluated individual networks. It is observed that TL models perform better than the DL models, and implementation of ensemble learning enhances the performance of both DL and TL models. Following a comprehensive assessment, it is observed that the TL-ensemble model has delivered the optimal outcome, boasting an accuracy of 90.58% and a MCC of 0.7624. This study highlights the efficacy of our methodology. TL-ensemble models, especially, excelled, providing valuable insights into automatic UCSG systems\u2019 potential enhancement. Ensemble learning offers promise for enhancing accuracy and reliability in UCSG, with implications for future research in this field.","doi":"10.1007\/s40031-024-01099-8","cleaned_title":"severity grading of ulcerative colitis using endoscopy images: an ensembled deep learning and transfer learning approach","cleaned_abstract":"ulcerative colitis (uc) is a persistent condition necessitating prompt treatment to avert potential complications. detecting uc severity aids treatment decisions. the mayo-endoscopic subscore is a standard for uc severity grading (ucsg). deep learning (dl) and transfer learning (tl) have enhanced severity grading, but ensemble learning\u2019s impact remains unexplored. this study designed dl-ensemble and tl-ensemble models for ucsg. using the hyperkvasir dataset, we classified ucsg into two stages: initial and advanced. three deep convolutional neural networks were trained from scratch for dl, and three pre-trained networks were trained for tl. ucsg was conducted using a majority voting ensemble scheme. a detailed comparative analysis evaluated individual networks. it is observed that tl models perform better than the dl models, and implementation of ensemble learning enhances the performance of both dl and tl models. following a comprehensive assessment, it is observed that the tl-ensemble model has delivered the optimal outcome, boasting an accuracy of 90.58% and a mcc of 0.7624. this study highlights the efficacy of our methodology. tl-ensemble models, especially, excelled, providing valuable insights into automatic ucsg systems\u2019 potential enhancement. ensemble learning offers promise for enhancing accuracy and reliability in ucsg, with implications for future research in this field.","key_phrases":["ulcerative colitis","uc","a persistent condition","prompt treatment","potential complications","uc severity aids treatment decisions","the mayo-endoscopic","a standard","ucsg","deep learning","dl","learning","tl","severity grading","ensemble learning\u2019s impact","this study","dl","-ensemble and tl-ensemble models","ucsg","the hyperkvasir","we","ucsg","two stages","three deep convolutional neural networks","scratch","dl","three pre-trained networks","tl. ucsg","a majority voting ensemble scheme","a detailed comparative analysis","it","tl models","the dl models","implementation","the performance","both dl and tl models","a comprehensive assessment","it","the tl-ensemble model","the optimal outcome","an accuracy","90.58%","a mcc","this study","the efficacy","our methodology","tl-ensemble models","valuable insights","automatic ucsg systems\u2019 potential enhancement","ensemble learning","promise","accuracy","reliability","ucsg","implications","future research","this field","two","three","three","ucsg","90.58%","0.7624"]},{"title":"A systematic review on smart waste biomass production using machine learning and deep learning","authors":["Wei Peng","Omid Karimi Sadaghiani"],"abstract":"The utilization of waste materials, as an energy resources, requires four main steps of production, pre-treatment, bio-refinery, and upgrading. This work reviews Machine Learning applications in the waste biomass production step. By investigating numerous related works, it is concluded that there is a considerable reviewing gap in the surveying and collecting the applications of Machine Learning in the waste biomass. To fill this gap with the current work, the kinds and resources of waste biomass as well as the role of Machine Learning and Deep Learning in their development are reviewed. Moreover, the storage and transportation of the wastes are surveyed followed by the application of Machine Learning and Deep Learning in these areas. Summarily, after analysis of numerous papers, it is concluded that Machine Learning and Deep Learning are widely utilized in waste biomass production areas to enhance the waste collecting quality and quality, improve the predictions, diminish the losses, as well as increase storage and transformation conditions.","doi":"10.1007\/s10163-023-01794-6","cleaned_title":"a systematic review on smart waste biomass production using machine learning and deep learning","cleaned_abstract":"the utilization of waste materials, as an energy resources, requires four main steps of production, pre-treatment, bio-refinery, and upgrading. this work reviews machine learning applications in the waste biomass production step. by investigating numerous related works, it is concluded that there is a considerable reviewing gap in the surveying and collecting the applications of machine learning in the waste biomass. to fill this gap with the current work, the kinds and resources of waste biomass as well as the role of machine learning and deep learning in their development are reviewed. moreover, the storage and transportation of the wastes are surveyed followed by the application of machine learning and deep learning in these areas. summarily, after analysis of numerous papers, it is concluded that machine learning and deep learning are widely utilized in waste biomass production areas to enhance the waste collecting quality and quality, improve the predictions, diminish the losses, as well as increase storage and transformation conditions.","key_phrases":["the utilization","waste materials","an energy resources","four main steps","production",", bio","refinery","upgrading","this work","machine learning applications","the waste biomass production step","numerous related works","it","a considerable reviewing gap","the surveying","the applications","machine learning","the waste biomass","this gap","the current work","the kinds","resources","waste biomass","the role","machine learning","deep learning","their development","the storage","transportation","the wastes","the application","machine learning","deep learning","these areas","analysis","numerous papers","it","machine learning","deep learning","waste biomass production areas","the waste","quality","quality","the predictions","the losses","storage and transformation conditions","four"]},{"title":"Accelerating three-dimensional phase-field simulations via deep learning approaches","authors":["Xuewei Zhou","Sheng Sun","Songlin Cai","Gongyu Chen","Honghui Wu","Jie Xiong","Jiaming Zhu"],"abstract":"Phase-field modeling (PFM) is a powerful but computationally expensive technique for simulating three-dimensional (3D) microstructure evolutions. Very recently, integrating machine learning into phase-field simulations provides a promising way to reduce calculation time remarkably. In this study, we propose a deep learning model that combines a convolutional autoencoder with a deep operator network to predict 3D microstructure evolution by using 2D slices of the 3D system. It is found that the deep learning model can shorten the calculation time from 37\u00a0min to 3\u00a0s after the initial training, while skipping 5-time steps, and reduce the phase-field simulation time by 31% in entire calculation of the\u00a0evolution process. Interestingly, this model achieves good accuracy in predicting 3D microstructures by utilizing only 2D information. This work demonstrates the efficiency of machine learning in accelerating phase-field simulations while maintaining high accuracy and promotes the application of PFM in fundamental studies.","doi":"10.1007\/s10853-024-10118-4","cleaned_title":"accelerating three-dimensional phase-field simulations via deep learning approaches","cleaned_abstract":"phase-field modeling (pfm) is a powerful but computationally expensive technique for simulating three-dimensional (3d) microstructure evolutions. very recently, integrating machine learning into phase-field simulations provides a promising way to reduce calculation time remarkably. in this study, we propose a deep learning model that combines a convolutional autoencoder with a deep operator network to predict 3d microstructure evolution by using 2d slices of the 3d system. it is found that the deep learning model can shorten the calculation time from 37 min to 3 s after the initial training, while skipping 5-time steps, and reduce the phase-field simulation time by 31% in entire calculation of the evolution process. interestingly, this model achieves good accuracy in predicting 3d microstructures by utilizing only 2d information. this work demonstrates the efficiency of machine learning in accelerating phase-field simulations while maintaining high accuracy and promotes the application of pfm in fundamental studies.","key_phrases":["phase-field modeling","pfm","a powerful but computationally expensive technique","three-dimensional (3d) microstructure evolutions","phase-field simulations","a promising way","calculation time","this study","we","a deep learning model","that","a convolutional autoencoder","a deep operator network","3d microstructure evolution","2d slices","the 3d system","it","the deep learning model","the calculation time","37 min","3 s","the initial training","5-time steps","the phase-field simulation time","31%","entire calculation","the evolution process","this model","good accuracy","3d microstructures","only 2d information","this work","the efficiency","accelerating phase-field simulations","high accuracy","the application","pfm","fundamental studies","three","3d","3d","2d","3d","37","3 s","5","31%","3d","2d"]},{"title":"Efficient socket-based data transmission method and implementation in deep learning","authors":["Xin-Jian Wei","Shu-Ping Li","Wu-Yang Yang","Xiang-Yang Zhang","Hai-Shan Li","Xin Xu","Nan Wang","Zhanbao Fu"],"abstract":"The deep learning algorithm, which has been increasingly applied in the field of petroleum geophysical prospecting, has achieved good results in improving efficiency and accuracy based on test applications. To play a greater role in actual production, these algorithm modules must be integrated into software systems and used more often in actual production projects. Deep learning frameworks, such as TensorFlow and PyTorch, basically take Python as the core architecture, while the application program mainly uses Java, C#, and other programming languages. During integration, the seismic data read by the Java and C# data interfaces must be transferred to the Python main program module. The data exchange methods between Java, C#, and Python include shared memory, shared directory, and so on. However, these methods have the disadvantages of low transmission efficiency and unsuitability for asynchronous networks. Considering the large volume of seismic data and the need for network support for deep learning, this paper proposes a method of transmitting seismic data based on Socket. By maximizing Socket\u2019s cross-network and efficient longdistance transmission, this approach solves the problem of inefficient transmission of underlying data while integrating the deep learning algorithm module into a software system. Furthermore, the actual production application shows that this method effectively solves the shortage of data transmission in shared memory, shared directory, and other modes while simultaneously improving the transmission efficiency of massive seismic data across modules at the bottom of the software.","doi":"10.1007\/s11770-024-1090-y","cleaned_title":"efficient socket-based data transmission method and implementation in deep learning","cleaned_abstract":"the deep learning algorithm, which has been increasingly applied in the field of petroleum geophysical prospecting, has achieved good results in improving efficiency and accuracy based on test applications. to play a greater role in actual production, these algorithm modules must be integrated into software systems and used more often in actual production projects. deep learning frameworks, such as tensorflow and pytorch, basically take python as the core architecture, while the application program mainly uses java, c#, and other programming languages. during integration, the seismic data read by the java and c# data interfaces must be transferred to the python main program module. the data exchange methods between java, c#, and python include shared memory, shared directory, and so on. however, these methods have the disadvantages of low transmission efficiency and unsuitability for asynchronous networks. considering the large volume of seismic data and the need for network support for deep learning, this paper proposes a method of transmitting seismic data based on socket. by maximizing socket\u2019s cross-network and efficient longdistance transmission, this approach solves the problem of inefficient transmission of underlying data while integrating the deep learning algorithm module into a software system. furthermore, the actual production application shows that this method effectively solves the shortage of data transmission in shared memory, shared directory, and other modes while simultaneously improving the transmission efficiency of massive seismic data across modules at the bottom of the software.","key_phrases":["the deep learning algorithm","which","the field","petroleum geophysical prospecting","good results","efficiency","accuracy","test applications","a greater role","actual production","these algorithm modules","software systems","actual production projects","deep learning frameworks","tensorflow","pytorch","python","the core architecture","the application program","integration","the seismic data","the java and c# data interfaces","the python main program module","the data exchange methods","java","python","shared memory","shared directory","these methods","the disadvantages","low transmission efficiency","unsuitability","asynchronous networks","the large volume","seismic data","the need","network support","deep learning","this paper","a method","seismic data","socket","socket","-","efficient longdistance transmission","this approach","the problem","inefficient transmission","underlying data","the deep learning algorithm module","a software system","the actual production application","this method","the shortage","data transmission","shared memory","shared directory","other modes","the transmission efficiency","massive seismic data","modules","the bottom","the software","#","java","#","java","#"]},{"title":"Amplifying document categorization with advanced features and deep learning","authors":["M. Kavitha","K. Akila"],"abstract":"The field of natural language processing (NLP) plays a pivotal role in discerning unstructured data from diverse origins. This study employs advanced techniques rooted in machine learning and deep learning to effectively categorize news articles. Notably, deep learning models have demonstrated superior performance over traditional machine learning algorithms, rendering them a popular choice for a range of NLP tasks. The research employs feature extraction techniques to identify multiword tokens, negation words, and out-of-vocabulary words and replace them. Additionally, convolutional neural network models leverage embedding, convolutional layers, and max pooling layers to capture intricate features. For tasks requiring an understanding of dependencies among long phrases, long short-term memory models come into play. The evaluation of the proposed model hinges on training it with datasets like AG News, BBC, and 20 Newsgroup, gauging its efficacy. The study delves into the myriad challenges inherent to text classification. These challenges are thoughtfully discussed, shedding light on the intricacies of the process. Furthermore, the research furnishes comprehensive test outcomes for both conventional machine learning and deep learning models. The significance of this proposed model is that it uses a multiword expression lexicon, wordnet synset, and word embedding techniques for feature extraction. The performance of the models is increased when using these feature extraction techniques.","doi":"10.1007\/s11042-024-18483-7","cleaned_title":"amplifying document categorization with advanced features and deep learning","cleaned_abstract":"the field of natural language processing (nlp) plays a pivotal role in discerning unstructured data from diverse origins. this study employs advanced techniques rooted in machine learning and deep learning to effectively categorize news articles. notably, deep learning models have demonstrated superior performance over traditional machine learning algorithms, rendering them a popular choice for a range of nlp tasks. the research employs feature extraction techniques to identify multiword tokens, negation words, and out-of-vocabulary words and replace them. additionally, convolutional neural network models leverage embedding, convolutional layers, and max pooling layers to capture intricate features. for tasks requiring an understanding of dependencies among long phrases, long short-term memory models come into play. the evaluation of the proposed model hinges on training it with datasets like ag news, bbc, and 20 newsgroup, gauging its efficacy. the study delves into the myriad challenges inherent to text classification. these challenges are thoughtfully discussed, shedding light on the intricacies of the process. furthermore, the research furnishes comprehensive test outcomes for both conventional machine learning and deep learning models. the significance of this proposed model is that it uses a multiword expression lexicon, wordnet synset, and word embedding techniques for feature extraction. the performance of the models is increased when using these feature extraction techniques.","key_phrases":["the field","natural language processing","nlp","a pivotal role","unstructured data","diverse origins","this study","advanced techniques","machine learning","deep learning","news articles","deep learning models","superior performance","traditional machine learning algorithms","them","a popular choice","a range","nlp tasks","the research","feature extraction techniques","multiword tokens","negation words","out-of-vocabulary words","them","convolutional neural network models","max pooling layers","intricate features","tasks","an understanding","dependencies","long phrases","long short-term memory models","play","the evaluation","the proposed model hinges","it","datasets","ag news","20 newsgroup","its efficacy","the study","the myriad challenges","classification","these challenges","light","the intricacies","the process","the research","comprehensive test outcomes","both conventional machine learning","deep learning models","the significance","this proposed model","it","a multiword expression lexicon","wordnet synset","techniques","feature extraction","the performance","the models","these feature extraction techniques","multiword tokens","max","ag news","bbc","20","multiword"]},{"title":"Severity Grading of Ulcerative Colitis Using Endoscopy Images: An Ensembled Deep Learning and Transfer Learning Approach","authors":["Subhashree Mohapatra","Pukhraj Singh Jeji","Girish Kumar Pati","Janmenjoy Nayak","Manohar Mishra","Tripti Swarnkar"],"abstract":"Ulcerative colitis (UC) is a persistent condition necessitating prompt treatment to avert potential complications. Detecting UC severity aids treatment decisions. The Mayo-endoscopic subscore is a standard for UC severity grading (UCSG). Deep learning (DL) and transfer learning (TL) have enhanced severity grading, but ensemble learning\u2019s impact remains unexplored. This study designed DL-ensemble and TL-ensemble models for UCSG. Using the HyperKvasir dataset, we classified UCSG into two stages: initial and advanced. Three deep convolutional neural networks were trained from scratch for DL, and three pre-trained networks were trained for TL. UCSG was conducted using a majority voting ensemble scheme. A detailed comparative analysis evaluated individual networks. It is observed that TL models perform better than the DL models, and implementation of ensemble learning enhances the performance of both DL and TL models. Following a comprehensive assessment, it is observed that the TL-ensemble model has delivered the optimal outcome, boasting an accuracy of 90.58% and a MCC of 0.7624. This study highlights the efficacy of our methodology. TL-ensemble models, especially, excelled, providing valuable insights into automatic UCSG systems\u2019 potential enhancement. Ensemble learning offers promise for enhancing accuracy and reliability in UCSG, with implications for future research in this field.","doi":"10.1007\/s40031-024-01099-8","cleaned_title":"severity grading of ulcerative colitis using endoscopy images: an ensembled deep learning and transfer learning approach","cleaned_abstract":"ulcerative colitis (uc) is a persistent condition necessitating prompt treatment to avert potential complications. detecting uc severity aids treatment decisions. the mayo-endoscopic subscore is a standard for uc severity grading (ucsg). deep learning (dl) and transfer learning (tl) have enhanced severity grading, but ensemble learning\u2019s impact remains unexplored. this study designed dl-ensemble and tl-ensemble models for ucsg. using the hyperkvasir dataset, we classified ucsg into two stages: initial and advanced. three deep convolutional neural networks were trained from scratch for dl, and three pre-trained networks were trained for tl. ucsg was conducted using a majority voting ensemble scheme. a detailed comparative analysis evaluated individual networks. it is observed that tl models perform better than the dl models, and implementation of ensemble learning enhances the performance of both dl and tl models. following a comprehensive assessment, it is observed that the tl-ensemble model has delivered the optimal outcome, boasting an accuracy of 90.58% and a mcc of 0.7624. this study highlights the efficacy of our methodology. tl-ensemble models, especially, excelled, providing valuable insights into automatic ucsg systems\u2019 potential enhancement. ensemble learning offers promise for enhancing accuracy and reliability in ucsg, with implications for future research in this field.","key_phrases":["ulcerative colitis","uc","a persistent condition","prompt treatment","potential complications","uc severity aids treatment decisions","the mayo-endoscopic","a standard","ucsg","deep learning","dl","learning","tl","severity grading","ensemble learning\u2019s impact","this study","dl","-ensemble and tl-ensemble models","ucsg","the hyperkvasir","we","ucsg","two stages","three deep convolutional neural networks","scratch","dl","three pre-trained networks","tl. ucsg","a majority voting ensemble scheme","a detailed comparative analysis","it","tl models","the dl models","implementation","the performance","both dl and tl models","a comprehensive assessment","it","the tl-ensemble model","the optimal outcome","an accuracy","90.58%","a mcc","this study","the efficacy","our methodology","tl-ensemble models","valuable insights","automatic ucsg systems\u2019 potential enhancement","ensemble learning","promise","accuracy","reliability","ucsg","implications","future research","this field","two","three","three","ucsg","90.58%","0.7624"]},{"title":"Learning and predicting the unknown class using evidential deep learning","authors":["Akihito Nagahama"],"abstract":"In practical deep-learning applications, such as medical image analysis, autonomous driving, and traffic simulation, the uncertainty of a classification model\u2019s output is critical. Evidential deep learning (EDL) can output this uncertainty for the prediction; however, its accuracy depends on a user-defined threshold, and it cannot handle training data with unknown classes that are unexpectedly contaminated or deliberately mixed for better classification of unknown class. To address these limitations, I propose a classification method called modified-EDL that extends classical EDL such that it outputs a prediction, i.e. an input belongs to a collective unknown class along with a probability. Although other methods handle unknown classes by creating new unknown classes and attempting to learn each class efficiently, the proposed m-EDL outputs, in a natural way, the \u201cuncertainty of the prediction\u201d of classical EDL and uses the output as the probability of an unknown class. Although classical EDL can also classify both known and unknown classes, experiments on three datasets from different domains demonstrated that m-EDL outperformed EDL on known classes when there were instances of unknown classes. Moreover, extensive experiments under different conditions established that m-EDL can predict unknown classes even when the unknown classes in the training and test data have different properties. If unknown class data are to be mixed intentionally during training to increase the discrimination accuracy of unknown classes, it is necessary to mix such data that the characteristics of the mixed data are as close as possible to those of known class data. This ability extends the range of practical applications that can benefit from deep learning-based classification and prediction models.","doi":"10.1038\/s41598-023-40649-w","cleaned_title":"learning and predicting the unknown class using evidential deep learning","cleaned_abstract":"in practical deep-learning applications, such as medical image analysis, autonomous driving, and traffic simulation, the uncertainty of a classification model\u2019s output is critical. evidential deep learning (edl) can output this uncertainty for the prediction; however, its accuracy depends on a user-defined threshold, and it cannot handle training data with unknown classes that are unexpectedly contaminated or deliberately mixed for better classification of unknown class. to address these limitations, i propose a classification method called modified-edl that extends classical edl such that it outputs a prediction, i.e. an input belongs to a collective unknown class along with a probability. although other methods handle unknown classes by creating new unknown classes and attempting to learn each class efficiently, the proposed m-edl outputs, in a natural way, the \u201cuncertainty of the prediction\u201d of classical edl and uses the output as the probability of an unknown class. although classical edl can also classify both known and unknown classes, experiments on three datasets from different domains demonstrated that m-edl outperformed edl on known classes when there were instances of unknown classes. moreover, extensive experiments under different conditions established that m-edl can predict unknown classes even when the unknown classes in the training and test data have different properties. if unknown class data are to be mixed intentionally during training to increase the discrimination accuracy of unknown classes, it is necessary to mix such data that the characteristics of the mixed data are as close as possible to those of known class data. this ability extends the range of practical applications that can benefit from deep learning-based classification and prediction models.","key_phrases":["practical deep-learning applications","medical image analysis","autonomous driving","traffic simulation","the uncertainty","a classification model\u2019s output","evidential deep learning","edl","this uncertainty","the prediction","its accuracy","a user-defined threshold","it","training data","unknown classes","that","better classification","unknown class","these limitations","i","a classification method","modified-edl","that","classical edl","it","a prediction","i.e. an input","a collective unknown class","a probability","other methods","unknown classes","new unknown classes","each class","the proposed m-edl outputs","a natural way","the prediction","classical edl","the output","the probability","an unknown class","classical edl","both known and unknown classes","experiments","three datasets","different domains","m-edl","edl","known classes","instances","unknown classes","extensive experiments","different conditions","m","edl","unknown classes","the unknown classes","the training","test data","different properties","unknown class data","training","the discrimination accuracy","unknown classes","it","such data","the characteristics","the mixed data","those","known class data","this ability","the range","practical applications","that","deep learning-based classification and prediction models","three"]},{"title":"Optimized model architectures for deep learning on genomic data","authors":["H\u00fcseyin Anil G\u00fcnd\u00fcz","Ren\u00e9 Mreches","Julia Moosbauer","Gary Robertson","Xiao-Yin To","Eric A. Franzosa","Curtis Huttenhower","Mina Rezaei","Alice C. McHardy","Bernd Bischl","Philipp C. M\u00fcnch","Martin Binder"],"abstract":"The success of deep learning in various applications depends on task-specific architecture design choices, including the types, hyperparameters, and number of layers. In computational biology, there is no consensus on the optimal architecture design, and decisions are often made using insights from more well-established fields such as computer vision. These may not consider the domain-specific characteristics of genome sequences, potentially limiting performance. Here, we present GenomeNet-Architect, a neural architecture design framework that automatically optimizes deep learning models for genome sequence data. It optimizes the overall layout of the architecture, with a search space specifically designed for genomics. Additionally, it optimizes hyperparameters of individual layers and the model training procedure. On a viral classification task, GenomeNet-Architect reduced the read-level misclassification rate by 19%, with 67% faster inference and 83% fewer parameters, and achieved similar contig-level accuracy with ~100 times fewer parameters compared to the best-performing deep learning baselines.","doi":"10.1038\/s42003-024-06161-1","cleaned_title":"optimized model architectures for deep learning on genomic data","cleaned_abstract":"the success of deep learning in various applications depends on task-specific architecture design choices, including the types, hyperparameters, and number of layers. in computational biology, there is no consensus on the optimal architecture design, and decisions are often made using insights from more well-established fields such as computer vision. these may not consider the domain-specific characteristics of genome sequences, potentially limiting performance. here, we present genomenet-architect, a neural architecture design framework that automatically optimizes deep learning models for genome sequence data. it optimizes the overall layout of the architecture, with a search space specifically designed for genomics. additionally, it optimizes hyperparameters of individual layers and the model training procedure. on a viral classification task, genomenet-architect reduced the read-level misclassification rate by 19%, with 67% faster inference and 83% fewer parameters, and achieved similar contig-level accuracy with ~100 times fewer parameters compared to the best-performing deep learning baselines.","key_phrases":["the success","deep learning","various applications","task-specific architecture design choices","the types","hyperparameters","number","layers","computational biology","no consensus","the optimal architecture design","decisions","insights","more well-established fields","computer vision","these","the domain-specific characteristics","genome sequences","performance","we","genomenet-architect","a neural architecture design framework","that","deep learning models","genome sequence data","it","the overall layout","the architecture","a search space","genomics","it","hyperparameters","individual layers","the model training procedure","a viral classification task","genomenet-architect","the read-level misclassification rate","19%","67% faster inference","83% fewer parameters","similar contig-level accuracy","~100 times fewer parameters","the best-performing deep learning baselines","19%","67%","83%"]},{"title":"Stroke detection in the brain using MRI and deep learning models","authors":["Subba Rao Polamuri"],"abstract":"When it comes to finding solutions to issues, deep learning models are pretty much everywhere. Medical image data is best analysed using models based on Convolutional Neural Networks (CNNs). Better methods for early detection are crucial due to the concerning increase in the number of people suffering from brain stroke. Among the several medical imaging modalities used for brain imaging, magnetic resonance imaging (MRI) stands out. When it comes to analysing medical photos, the deep learning models currently utilised with MRI have showed good outcomes. To improve the efficacy of brain stroke diagnosis, we suggested several upgrades to deep learning models in this work, including DenseNet121, ResNet50, and VGG16. Since these models are not purpose-built to solve any particular issue, they are modified according to the present situation involving the detection of brain strokes. To make use of all of these cutting-edge deep learning models in a pipeline, we proposed a strategy based on supervised learning. Results from the experiments showed that optimised models outperformed baseline models.","doi":"10.1007\/s11042-024-19318-1","cleaned_title":"stroke detection in the brain using mri and deep learning models","cleaned_abstract":"when it comes to finding solutions to issues, deep learning models are pretty much everywhere. medical image data is best analysed using models based on convolutional neural networks (cnns). better methods for early detection are crucial due to the concerning increase in the number of people suffering from brain stroke. among the several medical imaging modalities used for brain imaging, magnetic resonance imaging (mri) stands out. when it comes to analysing medical photos, the deep learning models currently utilised with mri have showed good outcomes. to improve the efficacy of brain stroke diagnosis, we suggested several upgrades to deep learning models in this work, including densenet121, resnet50, and vgg16. since these models are not purpose-built to solve any particular issue, they are modified according to the present situation involving the detection of brain strokes. to make use of all of these cutting-edge deep learning models in a pipeline, we proposed a strategy based on supervised learning. results from the experiments showed that optimised models outperformed baseline models.","key_phrases":["it","solutions","issues","deep learning models","medical image data","using models","convolutional neural networks","cnns","better methods","early detection","the concerning increase","the number","people","brain stroke","the several medical imaging modalities","brain imaging","magnetic resonance imaging","mri","it","medical photos","the deep learning models","mri","good outcomes","the efficacy","brain stroke diagnosis","we","several upgrades","deep learning models","this work","densenet121","resnet50","vgg16","these models","any particular issue","they","the present situation","the detection","brain strokes","use","all","these cutting-edge deep learning models","a pipeline","we","a strategy","supervised learning","results","the experiments","optimised models","baseline models","resnet50"]},{"title":"Combining Deep Learning with Good Old-Fashioned Machine Learning","authors":["Moshe Sipper"],"abstract":"We present a comprehensive, stacking-based framework for combining deep learning with good old-fashioned machine learning, called Deep GOld. Our framework involves ensemble selection from 51 retrained pretrained deep networks as first-level models, and 10 machine-learning algorithms as second-level models. Enabled by today\u2019s state-of-the-art software tools and hardware platforms, Deep GOld delivers consistent improvement when tested on four image-classification datasets: Fashion MNIST, CIFAR10, CIFAR100, and Tiny ImageNet. Of 120 experiments, in all but 10 Deep GOld improved the original networks\u2019 performance.","doi":"10.1007\/s42979-022-01505-2","cleaned_title":"combining deep learning with good old-fashioned machine learning","cleaned_abstract":"we present a comprehensive, stacking-based framework for combining deep learning with good old-fashioned machine learning, called deep gold. our framework involves ensemble selection from 51 retrained pretrained deep networks as first-level models, and 10 machine-learning algorithms as second-level models. enabled by today\u2019s state-of-the-art software tools and hardware platforms, deep gold delivers consistent improvement when tested on four image-classification datasets: fashion mnist, cifar10, cifar100, and tiny imagenet. of 120 experiments, in all but 10 deep gold improved the original networks\u2019 performance.","key_phrases":["we","a comprehensive, stacking-based framework","deep learning","good old-fashioned machine learning","deep gold","our framework","ensemble selection","deep networks","first-level models","10 machine-learning algorithms","second-level models","the-art","hardware platforms","deep gold","consistent improvement","four image-classification datasets","fashion mnist","cifar10","cifar100","tiny imagenet","120 experiments","all but 10 deep gold","the original networks\u2019 performance","51","first","10","second","today","four","cifar10","120","10"]},{"title":"Deep learning evaluation of echocardiograms to identify occult atrial fibrillation","authors":["Neal Yuan","Nathan R. Stein","Grant Duffy","Roopinder K. Sandhu","Sumeet S. Chugh","Peng-Sheng Chen","Carine Rosenberg","Christine M. Albert","Susan Cheng","Robert J. Siegel","David Ouyang"],"abstract":"Atrial fibrillation (AF) often escapes detection, given its frequent paroxysmal and asymptomatic presentation. Deep learning of transthoracic echocardiograms (TTEs), which have structural information, could help identify occult AF. We created a two-stage deep learning algorithm using a video-based convolutional neural network model that (1) distinguished whether TTEs were in sinus rhythm or AF and then (2) predicted which of the TTEs in sinus rhythm were in patients who had experienced AF within 90 days. Our model, trained on 111,319 TTE videos, distinguished TTEs in AF from those in sinus rhythm with high accuracy in a held-out test cohort (AUC 0.96 (0.95\u20130.96), AUPRC 0.91 (0.90\u20130.92)). Among TTEs in sinus rhythm, the model predicted the presence of concurrent paroxysmal AF (AUC 0.74 (0.71\u20130.77), AUPRC 0.19 (0.16\u20130.23)). Model discrimination remained similar in an external cohort of 10,203 TTEs (AUC of 0.69 (0.67\u20130.70), AUPRC 0.34 (0.31\u20130.36)). Performance held across patients who were women (AUC 0.76 (0.72\u20130.81)), older than 65 years (0.73 (0.69\u20130.76)), or had a CHA2DS2VASc \u22652 (0.73 (0.79\u20130.77)). The model performed better than using clinical risk factors (AUC 0.64 (0.62\u20130.67)), TTE measurements (0.64 (0.62\u20130.67)), left atrial size (0.63 (0.62\u20130.64)), or CHA2DS2VASc (0.61 (0.60\u20130.62)). An ensemble model in a cohort subset combining the TTE model with an electrocardiogram (ECGs) deep learning model performed better than using the ECG model alone (AUC 0.81 vs. 0.79, p\u2009=\u20090.01). Deep learning using TTEs can predict patients with active or occult AF and could be used for opportunistic AF screening that could lead to earlier treatment.","doi":"10.1038\/s41746-024-01090-z","cleaned_title":"deep learning evaluation of echocardiograms to identify occult atrial fibrillation","cleaned_abstract":"atrial fibrillation (af) often escapes detection, given its frequent paroxysmal and asymptomatic presentation. deep learning of transthoracic echocardiograms (ttes), which have structural information, could help identify occult af. we created a two-stage deep learning algorithm using a video-based convolutional neural network model that (1) distinguished whether ttes were in sinus rhythm or af and then (2) predicted which of the ttes in sinus rhythm were in patients who had experienced af within 90 days. our model, trained on 111,319 tte videos, distinguished ttes in af from those in sinus rhythm with high accuracy in a held-out test cohort (auc 0.96 (0.95\u20130.96), auprc 0.91 (0.90\u20130.92)). among ttes in sinus rhythm, the model predicted the presence of concurrent paroxysmal af (auc 0.74 (0.71\u20130.77), auprc 0.19 (0.16\u20130.23)). model discrimination remained similar in an external cohort of 10,203 ttes (auc of 0.69 (0.67\u20130.70), auprc 0.34 (0.31\u20130.36)). performance held across patients who were women (auc 0.76 (0.72\u20130.81)), older than 65 years (0.73 (0.69\u20130.76)), or had a cha2ds2vasc \u22652 (0.73 (0.79\u20130.77)). the model performed better than using clinical risk factors (auc 0.64 (0.62\u20130.67)), tte measurements (0.64 (0.62\u20130.67)), left atrial size (0.63 (0.62\u20130.64)), or cha2ds2vasc (0.61 (0.60\u20130.62)). an ensemble model in a cohort subset combining the tte model with an electrocardiogram (ecgs) deep learning model performed better than using the ecg model alone (auc 0.81 vs. 0.79, p = 0.01). deep learning using ttes can predict patients with active or occult af and could be used for opportunistic af screening that could lead to earlier treatment.","key_phrases":["atrial fibrillation","detection","its frequent paroxysmal and asymptomatic presentation","deep learning","transthoracic echocardiograms","ttes","which","structural information","occult","we","a two-stage deep learning algorithm","a video-based convolutional neural network model","ttes","sinus rhythm","which","the ttes","sinus rhythm","patients","who","90 days","our model","111,319 tte videos","ttes","af","those","sinus rhythm","high accuracy","a held-out test cohort","auc","(0.95\u20130.96","(0.90\u20130.92","ttes","sinus rhythm","the model","the presence","concurrent paroxysmal","auc","model discrimination","an external cohort","10,203 ttes","auc","performance","patients","who","women","auc","(0.72\u20130.81","older than 65 years","0.69\u20130.76","a cha2ds2vasc","the model","clinical risk factors","auc","tte measurements","(0.62\u20130.67","atrial size","cha2ds2vasc","an ensemble model","a cohort subset","the tte model","an electrocardiogram (ecgs","deep learning model","the ecg model","auc","=","deep learning","ttes","patients","occult","screening","that","earlier treatment","two","1","2","90 days","111,319","0.96","0.91","0.19","10,203","0.69","0.34","0.76","older than 65 years","0.73","0.73","0.64","0.64","0.61","0.81","0.79","0.01"]},{"title":"Relationship constraint deep metric learning","authors":["Yanbing Zhang","Ting Xiao","Zhe Wang","Xinru Wang","Wenyi Feng","Zhiling Fu","Hai Yang"],"abstract":"AbstractDeep metric learning\u00a0(DML) models aim to learn semantically meaningful representations in which similar samples are pulled together and dissimilar samples are pushed apart. However, the classification effect is limited due to the high time complexity of previous models and their poor performance in extracting data relationships. This paper presents a novel relationship constraint deep metric learning\u00a0(RCDML) approach, including proxy relationship constraint\u00a0(PRC) and sample relationship constraint\u00a0(SRC) for inter-class separability and intra-class compactness, to solve the above problems and improve the classification effect. The PRC combines the proxy-to-proxy relationship loss term with the proxy-to-sample relationship loss function to maximize the proxy features, hence enhancing inter-class separability by decreasing proxy similarity. Additionally, the SRC combines the sample-to-sample relationship loss term with the proxy-to-sample relationship loss function to maximize the sample features, which promotes intra-class compactness by increasing the similarity between the most different samples of the same class. Unlike existing proxy-based and pair-based methods, the relationship constraint framework uses a diverse range of proxy and sample data relationships. In addition, the proxy correction\u00a0(PC) module is used to optimize the proxy. Extensive tests conducted on the widely popular CUB-200-2011, CARS-196, and SOP datasets show that the framework is effective and attains state-of-the-art performance.Graphical abstract","doi":"10.1007\/s10489-024-05425-x","cleaned_title":"relationship constraint deep metric learning","cleaned_abstract":"abstractdeep metric learning (dml) models aim to learn semantically meaningful representations in which similar samples are pulled together and dissimilar samples are pushed apart. however, the classification effect is limited due to the high time complexity of previous models and their poor performance in extracting data relationships. this paper presents a novel relationship constraint deep metric learning (rcdml) approach, including proxy relationship constraint (prc) and sample relationship constraint (src) for inter-class separability and intra-class compactness, to solve the above problems and improve the classification effect. the prc combines the proxy-to-proxy relationship loss term with the proxy-to-sample relationship loss function to maximize the proxy features, hence enhancing inter-class separability by decreasing proxy similarity. additionally, the src combines the sample-to-sample relationship loss term with the proxy-to-sample relationship loss function to maximize the sample features, which promotes intra-class compactness by increasing the similarity between the most different samples of the same class. unlike existing proxy-based and pair-based methods, the relationship constraint framework uses a diverse range of proxy and sample data relationships. in addition, the proxy correction (pc) module is used to optimize the proxy. extensive tests conducted on the widely popular cub-200-2011, cars-196, and sop datasets show that the framework is effective and attains state-of-the-art performance.graphical abstract","key_phrases":["dml","semantically meaningful representations","which","similar samples","dissimilar samples","the classification effect","the high time complexity","previous models","their poor performance","data relationships","this paper","a novel relationship","deep metric learning (rcdml) approach","proxy relationship constraint","prc","sample relationship constraint","(src","inter-class separability","intra-class compactness","the above problems","the classification effect","the prc","proxy","sample","the proxy features","inter-class separability","proxy similarity","the src","sample","sample","the sample features","which","intra-class compactness","the similarity","the most different samples","the same class","existing proxy-based and pair-based methods","the relationship constraint framework","a diverse range","proxy and sample data relationships","addition","the proxy correction","(pc) module","the proxy","extensive tests","the widely popular cub-200","sop datasets","the framework","the-art","abstractdeep metric learning","prc","prc","cars-196"]},{"title":"Deep structure-level N-glycan identification using feature-induced structure diagnosis integrated with a deep learning model","authors":["Suideng Qin","Zhixin Tian"],"abstract":"Being a widely occurring protein post-translational modification, N-glycosylation features unique multi-dimensional structures including sequence and linkage isomers. There have been successful bioinformatics efforts in N-glycan structure identification using N-glycoproteomics data; however, symmetric \u201cmirror\u201d branch isomers and linkage isomers are largely unresolved. Here, we report deep structure-level N-glycan identification using feature-induced structure diagnosis (FISD) integrated with a deep learning model. A neural network model is integrated to conduct the identification of featured N-glycan motifs and boosts the process of structure diagnosis and distinction for linkage isomers. By adopting publicly available N-glycoproteomics datasets of five mouse tissues (17,136 intact N-glycopeptide spectrum matches) and a consideration of 23 motif features, a deep learning model integrated with a convolutional autoencoder and a multilayer perceptron was trained to be capable of predicting N-glycan featured motifs in the MS\/MS spectra with previously identified compositions. In the test of the trained model, a prediction accuracy of 0.8 and AUC value of 0.95 were achieved; 5701 previously unresolved N-glycan structures were assigned by matched structure-diagnostic ions; and by using an\u00a0explainable learning algorithm, two new fragmentation features of m\/z\u2009=\u2009674.25 and m\/z\u2009=\u2009835.28 were found to be significant to three N-glycan structure motifs with fucose, NeuAc, and NeuGc, proving the capability of FISD to discover new features in the MS\/MS spectra.Graphical Abstract","doi":"10.1007\/s00216-024-05505-4","cleaned_title":"deep structure-level n-glycan identification using feature-induced structure diagnosis integrated with a deep learning model","cleaned_abstract":"being a widely occurring protein post-translational modification, n-glycosylation features unique multi-dimensional structures including sequence and linkage isomers. there have been successful bioinformatics efforts in n-glycan structure identification using n-glycoproteomics data; however, symmetric \u201cmirror\u201d branch isomers and linkage isomers are largely unresolved. here, we report deep structure-level n-glycan identification using feature-induced structure diagnosis (fisd) integrated with a deep learning model. a neural network model is integrated to conduct the identification of featured n-glycan motifs and boosts the process of structure diagnosis and distinction for linkage isomers. by adopting publicly available n-glycoproteomics datasets of five mouse tissues (17,136 intact n-glycopeptide spectrum matches) and a consideration of 23 motif features, a deep learning model integrated with a convolutional autoencoder and a multilayer perceptron was trained to be capable of predicting n-glycan featured motifs in the ms\/ms spectra with previously identified compositions. in the test of the trained model, a prediction accuracy of 0.8 and auc value of 0.95 were achieved; 5701 previously unresolved n-glycan structures were assigned by matched structure-diagnostic ions; and by using an explainable learning algorithm, two new fragmentation features of m\/z = 674.25 and m\/z = 835.28 were found to be significant to three n-glycan structure motifs with fucose, neuac, and neugc, proving the capability of fisd to discover new features in the ms\/ms spectra.graphical abstract","key_phrases":["a widely occurring protein post-translational modification, n-glycosylation features","unique multi-dimensional structures","sequence and linkage isomers","successful bioinformatics efforts","n-glycan structure identification","n-glycoproteomics data","symmetric \u201cmirror\u201d branch isomers","linkage isomers","we","deep structure-level n-glycan identification","feature-induced structure diagnosis","(fisd","a deep learning model","a neural network model","the identification","featured n-glycan motifs","the process","structure diagnosis","distinction","linkage isomers","publicly available n-glycoproteomics datasets","five mouse tissues","17,136 intact n-glycopeptide spectrum matches","a consideration","23 motif features","a deep learning model","a convolutional autoencoder","a multilayer perceptron","n-glycan featured motifs","the ms\/ms spectra","previously identified compositions","the test","the trained model","a prediction accuracy","0.8 and auc value","5701 previously unresolved n-glycan structures","matched structure-diagnostic ions","an explainable learning algorithm","two new fragmentation features","m\/z","three n-glycan structure motifs","fucose","neuac","neugc","the capability","fisd","new features","the ms\/ms","spectra.graphical abstract","five","17,136","23","0.8","0.95","5701","two","674.25","835.28","three"]},{"title":"An empirical study of sentiment analysis utilizing machine learning and deep learning algorithms","authors":["Betul Erkantarci","Gokhan Bakal"],"abstract":"Among text-mining studies, one of the most studied topics is the text classification task applied in various domains, including medicine, social media, and academia. As a sub-problem in text classification, sentiment analysis has been widely investigated to classify often opinion-based textual elements. Specifically, user reviews and experiential feedback for products or services have been employed as fundamental data sources for sentiment analysis efforts. As a result of rapidly emerging technological advancements, social media platforms such as Twitter, Facebook, and Reddit, have become central opinion-sharing mediums since the early 2000s. In this sense, we build various machine-learning models to solve the sentiment analysis problem on the Reddit comments dataset in this work. The experimental models we constructed achieve F1 scores within intervals of 73\u201376%. Consequently, we present comparative performance scores obtained by traditional machine learning and deep learning models and discuss the results.","doi":"10.1007\/s42001-023-00236-5","cleaned_title":"an empirical study of sentiment analysis utilizing machine learning and deep learning algorithms","cleaned_abstract":"among text-mining studies, one of the most studied topics is the text classification task applied in various domains, including medicine, social media, and academia. as a sub-problem in text classification, sentiment analysis has been widely investigated to classify often opinion-based textual elements. specifically, user reviews and experiential feedback for products or services have been employed as fundamental data sources for sentiment analysis efforts. as a result of rapidly emerging technological advancements, social media platforms such as twitter, facebook, and reddit, have become central opinion-sharing mediums since the early 2000s. in this sense, we build various machine-learning models to solve the sentiment analysis problem on the reddit comments dataset in this work. the experimental models we constructed achieve f1 scores within intervals of 73\u201376%. consequently, we present comparative performance scores obtained by traditional machine learning and deep learning models and discuss the results.","key_phrases":["text-mining studies","the most studied topics","the text classification task","various domains","medicine","social media","academia","a sub","-","problem","text classification","sentiment analysis","opinion-based textual elements","user reviews","experiential feedback","products","services","fundamental data sources","sentiment analysis efforts","a result","rapidly emerging technological advancements","social media platforms","twitter","facebook","reddit","central opinion-sharing mediums","this sense","we","various machine-learning models","the sentiment analysis problem","the reddit comments","this work","the experimental models","we","f1 scores","intervals","73\u201376%","we","comparative performance scores","traditional machine learning","deep learning models","the results","one","the early 2000s","73\u201376%"]},{"title":"Deep learning in neglected vector-borne diseases: a systematic review","authors":["Atmika Mishra","Arya Pandey","Ruchika Malhotra"],"abstract":"This study explores the application of Deep Learning in combating neglected vector-borne Diseases, a significant global health concern, particularly in resource-limited areas. It examines areas where Deep Learning has proven effective, compares popular Deep Learning techniques, focuses on interdisciplinary approaches with translational impact, and finds untapped potential for deep learning application. Thorough searches across multiple databases yielded 64 pertinent studies, from which 16 were selected based on inclusion criteria and quality assessment. Deep Learning applications in disease transmission risk prediction, vector detection, parasite classification, and treatment procedure optimization were investigated and focused on diseases such as Schistosomiasis, Chagas disease, Leishmaniasis, Echinococcosis, and Trachoma. Convolutional neural networks, artificial neural networks, multilayer perceptrons, and AutoML algorithms surpassed traditional methods for disease prediction, species identification, and diagnosis. The interdisciplinary integration of Deep Learning with public health, entomology, and epidemiology provides prospects for improved disease control and understanding. Deep Learning models automate disease surveillance, simplify epidemiological data processing, and enable early detection, particularly in resource-constrained settings. Smartphone apps driven by deep learning allow for rapid disease diagnosis and identification, boosting healthcare accessibility and global health outcomes. Improved algorithms, broadening the scope of applications to areas such as one health approach, and community engagement, and expanding deep learning applications to diseases such as lymphatic filariasis, hydatidosis, and onchocerciasis hold promise for improving global health outcomes.","doi":"10.1007\/s13198-024-02380-1","cleaned_title":"deep learning in neglected vector-borne diseases: a systematic review","cleaned_abstract":"this study explores the application of deep learning in combating neglected vector-borne diseases, a significant global health concern, particularly in resource-limited areas. it examines areas where deep learning has proven effective, compares popular deep learning techniques, focuses on interdisciplinary approaches with translational impact, and finds untapped potential for deep learning application. thorough searches across multiple databases yielded 64 pertinent studies, from which 16 were selected based on inclusion criteria and quality assessment. deep learning applications in disease transmission risk prediction, vector detection, parasite classification, and treatment procedure optimization were investigated and focused on diseases such as schistosomiasis, chagas disease, leishmaniasis, echinococcosis, and trachoma. convolutional neural networks, artificial neural networks, multilayer perceptrons, and automl algorithms surpassed traditional methods for disease prediction, species identification, and diagnosis. the interdisciplinary integration of deep learning with public health, entomology, and epidemiology provides prospects for improved disease control and understanding. deep learning models automate disease surveillance, simplify epidemiological data processing, and enable early detection, particularly in resource-constrained settings. smartphone apps driven by deep learning allow for rapid disease diagnosis and identification, boosting healthcare accessibility and global health outcomes. improved algorithms, broadening the scope of applications to areas such as one health approach, and community engagement, and expanding deep learning applications to diseases such as lymphatic filariasis, hydatidosis, and onchocerciasis hold promise for improving global health outcomes.","key_phrases":["this study","the application","deep learning","neglected vector-borne diseases","a significant global health concern","resource-limited areas","it","areas","deep learning","popular deep learning techniques","interdisciplinary approaches","translational impact","untapped potential","deep learning application","thorough searches","multiple databases","64 pertinent studies","which","inclusion criteria","quality assessment","deep learning applications","disease transmission risk prediction","vector detection","classification","treatment procedure optimization","diseases","schistosomiasis","chagas disease","leishmaniasis","echinococcosis","trachoma","convolutional neural networks","artificial neural networks","multilayer perceptrons","automl algorithms","traditional methods","disease prediction","species identification","diagnosis","the interdisciplinary integration","deep learning","public health","entomology","epidemiology","prospects","improved disease control","understanding","deep learning models","automate disease surveillance","epidemiological data processing","early detection","resource-constrained settings","smartphone apps","deep learning","rapid disease diagnosis","identification","healthcare accessibility","global health outcomes","improved algorithms","the scope","applications","areas","one health approach","community engagement","deep learning applications","diseases","lymphatic filariasis","hydatidosis","onchocerciasis hold promise","global health outcomes","64","16","one"]},{"title":"Machine Learning and Deep Learning-Based Students\u2019 Grade Prediction","authors":["Adil Korchi","Fay\u00e7al Messaoudi","Ahmed Abatal","Youness Manzali"],"abstract":"Predicting student performance in a curriculum or program offers the prospect of improving academic outcomes. By using effective performance prediction methods, instructional leaders can allocate adequate resources and instruction more accurately. This paper aims to identify machine learning algorithm features for predicting student grades as an early intervention. Predictive models spot at-risk students early, allowing educators to provide timely support. Educators can customize teaching methods, and these models assess program success, helping institutions refine or expand them through data-driven decisions. But the problem definition of student grade prediction is to develop predictive models or algorithms that can forecast or estimate the future academic performance or grades of students based on various input features and historical data, and to do so, we utilized a student dataset comprising personal information and grades, employing various regression algorithms, including decision tree, random forest, linear regression, k-nearest neighbor, XGBoost, and deep neural network. We chose these algorithms for their suitability and distinct strengths. We assessed their performance using determination coefficient, mean average error, mean squared error, and root mean squared error. The results showed that the deep neural network outperformed others with a determination coefficient of 99.97%, confirming its reliability in predicting student grades and assessing performance, and this will certainly help to develop predictive models that can accurately forecast or estimate students\u2019 academic performance based on various input features and enable teaching staff to provide timely assistance in addressing these issues.","doi":"10.1007\/s43069-023-00267-8","cleaned_title":"machine learning and deep learning-based students\u2019 grade prediction","cleaned_abstract":"predicting student performance in a curriculum or program offers the prospect of improving academic outcomes. by using effective performance prediction methods, instructional leaders can allocate adequate resources and instruction more accurately. this paper aims to identify machine learning algorithm features for predicting student grades as an early intervention. predictive models spot at-risk students early, allowing educators to provide timely support. educators can customize teaching methods, and these models assess program success, helping institutions refine or expand them through data-driven decisions. but the problem definition of student grade prediction is to develop predictive models or algorithms that can forecast or estimate the future academic performance or grades of students based on various input features and historical data, and to do so, we utilized a student dataset comprising personal information and grades, employing various regression algorithms, including decision tree, random forest, linear regression, k-nearest neighbor, xgboost, and deep neural network. we chose these algorithms for their suitability and distinct strengths. we assessed their performance using determination coefficient, mean average error, mean squared error, and root mean squared error. the results showed that the deep neural network outperformed others with a determination coefficient of 99.97%, confirming its reliability in predicting student grades and assessing performance, and this will certainly help to develop predictive models that can accurately forecast or estimate students\u2019 academic performance based on various input features and enable teaching staff to provide timely assistance in addressing these issues.","key_phrases":["student performance","a curriculum","program","the prospect","academic outcomes","effective performance prediction methods","instructional leaders","adequate resources","instruction","this paper","machine learning","algorithm","student grades","an early intervention","predictive models","risk","educators","timely support","educators","teaching methods","these models","program success","institutions","them","data-driven decisions","the problem definition","student grade prediction","predictive models","algorithms","that","the future academic performance","grades","students","various input features","historical data","we","a student dataset","personal information","grades","various regression algorithms","decision tree","random forest","linear regression","k-nearest neighbor","xgboost","deep neural network","we","these algorithms","their suitability","distinct strengths","we","their performance","determination coefficient","average error","mean squared error","root mean squared error","the results","the deep neural network","others","a determination coefficient","99.97%","its reliability","student grades","performance","this","predictive models","that","students\u2019 academic performance","various input features","teaching staff","timely assistance","these issues","linear","99.97%"]},{"title":"Rf-based fingerprinting for indoor localization: deep transfer learning approach","authors":["Rokaya Safwat","Eman Shaaban","Shahinaz. M. Al-Tabbakh","Karim Emara"],"abstract":"Transfer Learning (TL) has emerged as a powerful approach for improving the performance of Deep Learning systems in various domains by leveraging pre-trained models. It was proven that features learned by deep learning can smoothly be reused across similar domains. Deep transfer learning schemes compensate for limited training data via transfer learning of a rich data environment. This paper investigates the effectiveness of applying TL schemes in indoor localization. It proposes four deep TL models where the knowledge is transferred from the rich-measurement data source domain to multiple target domains with limited data measurements. The architecture of the source domain is based on Convolutional Neural Network (CNN), where the four deep TL models for the target domain are: standalone feature extractor, integrated feature extractor, selective fine-tuning, and weight initialization. We employed a dataset of RF fingerprinting measurement signals representing common interior conditions, including extremely crowded, medium cluttered, low cluttered, and open environments, to test the effectiveness of the proposed TL models. We measured the accuracy and computation time of target-domain models trained, with varied percentages of restricted data sizes: 40%, 30%, 20%, 15%, 10%, 5%, and 2.5%. The experimental results show that all TL models are effective in achieving significant improvement in accuracy when compared to non-transferred models, even with minimal training data size. However, the proper determination of the TL model and the amount of training data profoundly influence the performance results.","doi":"10.1007\/s12652-024-04819-6","cleaned_title":"rf-based fingerprinting for indoor localization: deep transfer learning approach","cleaned_abstract":"transfer learning (tl) has emerged as a powerful approach for improving the performance of deep learning systems in various domains by leveraging pre-trained models. it was proven that features learned by deep learning can smoothly be reused across similar domains. deep transfer learning schemes compensate for limited training data via transfer learning of a rich data environment. this paper investigates the effectiveness of applying tl schemes in indoor localization. it proposes four deep tl models where the knowledge is transferred from the rich-measurement data source domain to multiple target domains with limited data measurements. the architecture of the source domain is based on convolutional neural network (cnn), where the four deep tl models for the target domain are: standalone feature extractor, integrated feature extractor, selective fine-tuning, and weight initialization. we employed a dataset of rf fingerprinting measurement signals representing common interior conditions, including extremely crowded, medium cluttered, low cluttered, and open environments, to test the effectiveness of the proposed tl models. we measured the accuracy and computation time of target-domain models trained, with varied percentages of restricted data sizes: 40%, 30%, 20%, 15%, 10%, 5%, and 2.5%. the experimental results show that all tl models are effective in achieving significant improvement in accuracy when compared to non-transferred models, even with minimal training data size. however, the proper determination of the tl model and the amount of training data profoundly influence the performance results.","key_phrases":["tl","a powerful approach","the performance","deep learning systems","various domains","pre-trained models","it","features","deep learning","similar domains","deep transfer learning schemes","limited training data","transfer learning","a rich data environment","this paper","the effectiveness","tl schemes","indoor localization","it","four deep tl models","the knowledge","the rich-measurement data source domain","multiple target domains","limited data measurements","the architecture","the source domain","convolutional neural network","cnn","the four deep tl models","the target domain","standalone feature extractor","integrated feature extractor","selective fine-tuning, and weight initialization","we","a dataset","rf fingerprinting measurement signals","common interior conditions","extremely crowded, medium cluttered, low cluttered, and open environments","the effectiveness","the proposed tl models","we","the accuracy","computation time","target-domain models","varied percentages","restricted data sizes","40%","30%","20%","15%","10%","5%","2.5%","the experimental results","all tl models","significant improvement","accuracy","non-transferred models","minimal training data size","the proper determination","the tl model","the amount","training data","the performance results","four","cnn","four","40%","30%","20%","15%","10%","5%","2.5%"]},{"title":"Satellite image classification using deep learning approach","authors":["Divakar Yadav","Kritarth Kapoor","Arun Kumar Yadav","Mohit Kumar","Arti Jain","Jorge Morato"],"abstract":"Our planet Earth comprises distinguished topologies based on temperature, location, latitude, longitude, and altitude, which can be captured using Remote Sensing Satellites. In this paper, the classification of satellite images is performed based on their topologies and geographical features. Researchers have worked on several machine learning and deep learning methods like support vector machine, k-nearest neighbor, maximum likelihood, deep belief network, etc. that can be used to solve satellite image classification tasks. All strategies give promising results. Recent trends show that a Convolutional Neural Network (CNN) is an excellent deep learning model for classification purposes, which is used in this paper. The open-source EuroSAT dataset is used for classifying the remote images which contain 27,000 images distributed among ten classes. The 3 baseline CNN models are pre-trained, namely- ResNet50, ResNet101, and GoogleNet models. They have other sequence layers added to them with respect to CNN, and data is pre-processed using LAB channel operations. The highest accuracy of 99.68%, precision of 99.42%, recall of 99.51%, and F- Score of 99.45% are achieved using GoogleNet over the pre-processed dataset. The proposed work is compared with the state-of-art methods and it is observed that more layers in CNN do not necessarily provide a better outcome for a medium-sized dataset. The GoogleNet, a 22-layer CNN, performs faster and better than the 50 layers CNN- ResNet50, and 101 layers CNN- ResNet101.","doi":"10.1007\/s12145-024-01301-x","cleaned_title":"satellite image classification using deep learning approach","cleaned_abstract":"our planet earth comprises distinguished topologies based on temperature, location, latitude, longitude, and altitude, which can be captured using remote sensing satellites. in this paper, the classification of satellite images is performed based on their topologies and geographical features. researchers have worked on several machine learning and deep learning methods like support vector machine, k-nearest neighbor, maximum likelihood, deep belief network, etc. that can be used to solve satellite image classification tasks. all strategies give promising results. recent trends show that a convolutional neural network (cnn) is an excellent deep learning model for classification purposes, which is used in this paper. the open-source eurosat dataset is used for classifying the remote images which contain 27,000 images distributed among ten classes. the 3 baseline cnn models are pre-trained, namely- resnet50, resnet101, and googlenet models. they have other sequence layers added to them with respect to cnn, and data is pre-processed using lab channel operations. the highest accuracy of 99.68%, precision of 99.42%, recall of 99.51%, and f- score of 99.45% are achieved using googlenet over the pre-processed dataset. the proposed work is compared with the state-of-art methods and it is observed that more layers in cnn do not necessarily provide a better outcome for a medium-sized dataset. the googlenet, a 22-layer cnn, performs faster and better than the 50 layers cnn- resnet50, and 101 layers cnn- resnet101.","key_phrases":["our planet earth","distinguished topologies","temperature","location","latitude","altitude","which","remote sensing satellites","this paper","the classification","satellite images","their topologies","geographical features","researchers","several machine learning","deep learning methods","support vector machine","k-nearest neighbor","maximum likelihood","deep belief network","that","satellite image classification tasks","all strategies","promising results","recent trends","a convolutional neural network","cnn","an excellent deep learning model","classification purposes","which","this paper","the open-source eurosat dataset","the remote images","which","27,000 images","ten classes","the 3 baseline cnn models","-trained, namely- resnet50","resnet101","googlenet models","they","other sequence layers","them","respect","cnn","data","-processed using lab channel operations","the highest accuracy","99.68%","precision","99.42%","recall","99.51%","f- score","99.45%","googlenet","the pre-processed dataset","the proposed work","art","it","more layers","cnn","a better outcome","a medium-sized dataset","the googlenet","a 22-layer cnn","the 50 layers cnn- resnet50","101 layers","resnet101","cnn","27,000","ten","3","cnn","cnn","99.68%","99.42%","99.51%","99.45%","cnn","22","cnn","50","resnet50","101"]},{"title":"Amplifying document categorization with advanced features and deep learning","authors":["M. Kavitha","K. Akila"],"abstract":"The field of natural language processing (NLP) plays a pivotal role in discerning unstructured data from diverse origins. This study employs advanced techniques rooted in machine learning and deep learning to effectively categorize news articles. Notably, deep learning models have demonstrated superior performance over traditional machine learning algorithms, rendering them a popular choice for a range of NLP tasks. The research employs feature extraction techniques to identify multiword tokens, negation words, and out-of-vocabulary words and replace them. Additionally, convolutional neural network models leverage embedding, convolutional layers, and max pooling layers to capture intricate features. For tasks requiring an understanding of dependencies among long phrases, long short-term memory models come into play. The evaluation of the proposed model hinges on training it with datasets like AG News, BBC, and 20 Newsgroup, gauging its efficacy. The study delves into the myriad challenges inherent to text classification. These challenges are thoughtfully discussed, shedding light on the intricacies of the process. Furthermore, the research furnishes comprehensive test outcomes for both conventional machine learning and deep learning models. The significance of this proposed model is that it uses a multiword expression lexicon, wordnet synset, and word embedding techniques for feature extraction. The performance of the models is increased when using these feature extraction techniques.","doi":"10.1007\/s11042-024-18483-7","cleaned_title":"amplifying document categorization with advanced features and deep learning","cleaned_abstract":"the field of natural language processing (nlp) plays a pivotal role in discerning unstructured data from diverse origins. this study employs advanced techniques rooted in machine learning and deep learning to effectively categorize news articles. notably, deep learning models have demonstrated superior performance over traditional machine learning algorithms, rendering them a popular choice for a range of nlp tasks. the research employs feature extraction techniques to identify multiword tokens, negation words, and out-of-vocabulary words and replace them. additionally, convolutional neural network models leverage embedding, convolutional layers, and max pooling layers to capture intricate features. for tasks requiring an understanding of dependencies among long phrases, long short-term memory models come into play. the evaluation of the proposed model hinges on training it with datasets like ag news, bbc, and 20 newsgroup, gauging its efficacy. the study delves into the myriad challenges inherent to text classification. these challenges are thoughtfully discussed, shedding light on the intricacies of the process. furthermore, the research furnishes comprehensive test outcomes for both conventional machine learning and deep learning models. the significance of this proposed model is that it uses a multiword expression lexicon, wordnet synset, and word embedding techniques for feature extraction. the performance of the models is increased when using these feature extraction techniques.","key_phrases":["the field","natural language processing","nlp","a pivotal role","unstructured data","diverse origins","this study","advanced techniques","machine learning","deep learning","news articles","deep learning models","superior performance","traditional machine learning algorithms","them","a popular choice","a range","nlp tasks","the research","feature extraction techniques","multiword tokens","negation words","out-of-vocabulary words","them","convolutional neural network models","max pooling layers","intricate features","tasks","an understanding","dependencies","long phrases","long short-term memory models","play","the evaluation","the proposed model hinges","it","datasets","ag news","20 newsgroup","its efficacy","the study","the myriad challenges","classification","these challenges","light","the intricacies","the process","the research","comprehensive test outcomes","both conventional machine learning","deep learning models","the significance","this proposed model","it","a multiword expression lexicon","wordnet synset","techniques","feature extraction","the performance","the models","these feature extraction techniques","multiword tokens","max","ag news","bbc","20","multiword"]},{"title":"Image steganalysis using deep learning models","authors":["Alexandr Kuznetsov","Nicolas Luhanko","Emanuele Frontoni","Luca Romeo","Riccardo Rosati"],"abstract":"In the domain of digital steganography, the problem of efficient and accurate steganalysis is of utmost importance. Steganalysis seeks to detect the presence of hidden data within digital media, a task that is continually evolving due to advancements in steganographic techniques. This study undertakes a detailed exploration of the SRNet model, a prominent deep learning model for steganalysis. We aim to evaluate the impact of various factors, including choice of deep learning framework, model initialization and optimization parameters, and architectural modifications on the model's steganalysis performance. Three separate implementations of the SRNet model are examined in this study: our custom implementation, an implementation using TensorFlow, and another utilizing PyTorch. Each model is evaluated on its ability to detect different payloads, or bytes of hidden data per pixel, in digital images. This investigation includes a thorough comparative analysis of different performance metrics including accuracy, recall, precision, and F1-score. Our findings indicate that the choice of deep learning framework and the parameters utilized for model initialization and optimization play significant roles in influencing the model's steganalysis effectiveness. Notably, the TensorFlow implementation, enhanced with an additional dense layer, outperforms all other models. In contrast, our custom SRNet implementation, trained with fewer epochs, offers a balance between computational cost and steganalysis performance. This study thus provides valuable insights into the adaptability and potential of the SRNet model for steganalysis, illustrating the model's performance under different configurations and implementations. It underscores the importance of continued exploration and optimization in the field of steganalysis, offering guidance for future research in this evolving domain.","doi":"10.1007\/s11042-023-17591-0","cleaned_title":"image steganalysis using deep learning models","cleaned_abstract":"in the domain of digital steganography, the problem of efficient and accurate steganalysis is of utmost importance. steganalysis seeks to detect the presence of hidden data within digital media, a task that is continually evolving due to advancements in steganographic techniques. this study undertakes a detailed exploration of the srnet model, a prominent deep learning model for steganalysis. we aim to evaluate the impact of various factors, including choice of deep learning framework, model initialization and optimization parameters, and architectural modifications on the model's steganalysis performance. three separate implementations of the srnet model are examined in this study: our custom implementation, an implementation using tensorflow, and another utilizing pytorch. each model is evaluated on its ability to detect different payloads, or bytes of hidden data per pixel, in digital images. this investigation includes a thorough comparative analysis of different performance metrics including accuracy, recall, precision, and f1-score. our findings indicate that the choice of deep learning framework and the parameters utilized for model initialization and optimization play significant roles in influencing the model's steganalysis effectiveness. notably, the tensorflow implementation, enhanced with an additional dense layer, outperforms all other models. in contrast, our custom srnet implementation, trained with fewer epochs, offers a balance between computational cost and steganalysis performance. this study thus provides valuable insights into the adaptability and potential of the srnet model for steganalysis, illustrating the model's performance under different configurations and implementations. it underscores the importance of continued exploration and optimization in the field of steganalysis, offering guidance for future research in this evolving domain.","key_phrases":["the domain","digital steganography","the problem","efficient and accurate steganalysis","utmost importance","steganalysis","the presence","hidden data","digital media","a task","that","advancements","steganographic techniques","this study","a detailed exploration","the srnet model","a prominent deep learning model","steganalysis","we","the impact","various factors","choice","deep learning framework","model initialization","optimization parameters","architectural modifications","the model's steganalysis performance","three separate implementations","the srnet model","this study","our custom implementation","an implementation","tensorflow","another utilizing pytorch","each model","its ability","different payloads","bytes","hidden data","pixel","digital images","this investigation","a thorough comparative analysis","different performance metrics","accuracy","recall","precision","f1-score","our findings","the choice","deep learning framework","the parameters","model initialization","optimization","significant roles","the model's steganalysis effectiveness","the tensorflow implementation","an additional dense layer","all other models","contrast","our custom srnet implementation","fewer epochs","a balance","computational cost","steganalysis performance","this study","valuable insights","the adaptability","potential","the srnet model","steganalysis","the model's performance","different configurations","implementations","it","the importance","continued exploration","optimization","the field","steganalysis","guidance","future research","this evolving domain","three"]},{"title":"Control learning rate for autism facial detection via deep transfer learning","authors":["Abdelkrim El Mouatasim","Mohamed Ikermane"],"abstract":"Autism spectrum disorder (ASD) is a complex neurodevelopmental disorder that affects social interaction and communication. Early detection of ASD can significantly improve outcomes for individuals with the disorder, and there has been increasing interest in using machine learning techniques to aid in the diagnosis of ASD. One promising approach is the use of deep learning techniques, particularly convolutional neural networks (CNNs), to classify facial images as indicative of ASD or not. However, choosing a learning rate for optimizing the performance of these deep CNNs can be tedious and may not always result in optimal convergence. In this paper, we propose a novel approach called the control subgradient algorithm (CSA) for tackling ASD diagnosis based on facial images using deep CNNs. CSA is a variation of the subgradient method in which the learning rate is updated by a control step in each iteration of each epoch. We apply CSA to the popular DensNet-121 CNN model and evaluate its performance on a publicly available facial ASD dataset. Our results show that CSA is faster than the baseline method and improves the classification accuracy and loss compared to the baseline. We also demonstrate the effectiveness of using CSA with \\(L_1\\)-regularization to further improve the performance of our deep CNN model.","doi":"10.1007\/s11760-023-02598-9","cleaned_title":"control learning rate for autism facial detection via deep transfer learning","cleaned_abstract":"autism spectrum disorder (asd) is a complex neurodevelopmental disorder that affects social interaction and communication. early detection of asd can significantly improve outcomes for individuals with the disorder, and there has been increasing interest in using machine learning techniques to aid in the diagnosis of asd. one promising approach is the use of deep learning techniques, particularly convolutional neural networks (cnns), to classify facial images as indicative of asd or not. however, choosing a learning rate for optimizing the performance of these deep cnns can be tedious and may not always result in optimal convergence. in this paper, we propose a novel approach called the control subgradient algorithm (csa) for tackling asd diagnosis based on facial images using deep cnns. csa is a variation of the subgradient method in which the learning rate is updated by a control step in each iteration of each epoch. we apply csa to the popular densnet-121 cnn model and evaluate its performance on a publicly available facial asd dataset. our results show that csa is faster than the baseline method and improves the classification accuracy and loss compared to the baseline. we also demonstrate the effectiveness of using csa with \\(l_1\\)-regularization to further improve the performance of our deep cnn model.","key_phrases":["autism spectrum disorder","asd","a complex neurodevelopmental disorder","that","social interaction","communication","early detection","asd","outcomes","individuals","the disorder","interest","machine learning techniques","the diagnosis","asd","one promising approach","the use","deep learning techniques","particularly convolutional neural networks","cnns","facial images","asd","a learning rate","the performance","these deep cnns","optimal convergence","this paper","we","a novel approach","the control subgradient","algorithm","csa","asd diagnosis","facial images","deep cnns","csa","a variation","the subgradient method","which","the learning rate","a control step","each iteration","each epoch","we","the popular densnet-121 cnn model","its performance","a publicly available facial asd dataset","our results","csa","the baseline method","the classification accuracy","loss","the baseline","we","the effectiveness","\\(l_1\\)-regularization","the performance","our deep cnn model","one","cnn"]},{"title":"Optimized model architectures for deep learning on genomic data","authors":["H\u00fcseyin Anil G\u00fcnd\u00fcz","Ren\u00e9 Mreches","Julia Moosbauer","Gary Robertson","Xiao-Yin To","Eric A. Franzosa","Curtis Huttenhower","Mina Rezaei","Alice C. McHardy","Bernd Bischl","Philipp C. M\u00fcnch","Martin Binder"],"abstract":"The success of deep learning in various applications depends on task-specific architecture design choices, including the types, hyperparameters, and number of layers. In computational biology, there is no consensus on the optimal architecture design, and decisions are often made using insights from more well-established fields such as computer vision. These may not consider the domain-specific characteristics of genome sequences, potentially limiting performance. Here, we present GenomeNet-Architect, a neural architecture design framework that automatically optimizes deep learning models for genome sequence data. It optimizes the overall layout of the architecture, with a search space specifically designed for genomics. Additionally, it optimizes hyperparameters of individual layers and the model training procedure. On a viral classification task, GenomeNet-Architect reduced the read-level misclassification rate by 19%, with 67% faster inference and 83% fewer parameters, and achieved similar contig-level accuracy with ~100 times fewer parameters compared to the best-performing deep learning baselines.","doi":"10.1038\/s42003-024-06161-1","cleaned_title":"optimized model architectures for deep learning on genomic data","cleaned_abstract":"the success of deep learning in various applications depends on task-specific architecture design choices, including the types, hyperparameters, and number of layers. in computational biology, there is no consensus on the optimal architecture design, and decisions are often made using insights from more well-established fields such as computer vision. these may not consider the domain-specific characteristics of genome sequences, potentially limiting performance. here, we present genomenet-architect, a neural architecture design framework that automatically optimizes deep learning models for genome sequence data. it optimizes the overall layout of the architecture, with a search space specifically designed for genomics. additionally, it optimizes hyperparameters of individual layers and the model training procedure. on a viral classification task, genomenet-architect reduced the read-level misclassification rate by 19%, with 67% faster inference and 83% fewer parameters, and achieved similar contig-level accuracy with ~100 times fewer parameters compared to the best-performing deep learning baselines.","key_phrases":["the success","deep learning","various applications","task-specific architecture design choices","the types","hyperparameters","number","layers","computational biology","no consensus","the optimal architecture design","decisions","insights","more well-established fields","computer vision","these","the domain-specific characteristics","genome sequences","performance","we","genomenet-architect","a neural architecture design framework","that","deep learning models","genome sequence data","it","the overall layout","the architecture","a search space","genomics","it","hyperparameters","individual layers","the model training procedure","a viral classification task","genomenet-architect","the read-level misclassification rate","19%","67% faster inference","83% fewer parameters","similar contig-level accuracy","~100 times fewer parameters","the best-performing deep learning baselines","19%","67%","83%"]},{"title":"Virtual reality-empowered deep-learning analysis of brain cells","authors":["Doris Kaltenecker","Rami Al-Maskari","Moritz Negwer","Luciano Hoeher","Florian Kofler","Shan Zhao","Mihail Todorov","Zhouyi Rong","Johannes Christian Paetzold","Benedikt Wiestler","Marie Piraud","Daniel Rueckert","Julia Geppert","Pauline Morigny","Maria Rohm","Bjoern H. Menze","Stephan Herzig","Mauricio Berriel Diaz","Ali Ert\u00fcrk"],"abstract":"Automated detection of specific cells in three-dimensional datasets such as whole-brain light-sheet image stacks is challenging. Here, we present DELiVR, a virtual reality-trained deep-learning pipeline for detecting c-Fos+ cells as markers for neuronal activity in cleared mouse brains. Virtual reality annotation substantially accelerated training data generation, enabling DELiVR to outperform state-of-the-art cell-segmenting approaches. Our pipeline is available in a user-friendly Docker container that runs with a standalone Fiji plugin. DELiVR features a comprehensive toolkit for data visualization and can be customized to other cell types of interest, as we did here for microglia somata, using Fiji for dataset-specific training. We applied DELiVR to investigate cancer-related brain activity, unveiling an activation pattern that distinguishes weight-stable cancer from cancers associated with weight loss. Overall, DELiVR is a robust deep-learning tool that does not require advanced coding skills to analyze whole-brain imaging data in health and disease.","doi":"10.1038\/s41592-024-02245-2","cleaned_title":"virtual reality-empowered deep-learning analysis of brain cells","cleaned_abstract":"automated detection of specific cells in three-dimensional datasets such as whole-brain light-sheet image stacks is challenging. here, we present delivr, a virtual reality-trained deep-learning pipeline for detecting c-fos+ cells as markers for neuronal activity in cleared mouse brains. virtual reality annotation substantially accelerated training data generation, enabling delivr to outperform state-of-the-art cell-segmenting approaches. our pipeline is available in a user-friendly docker container that runs with a standalone fiji plugin. delivr features a comprehensive toolkit for data visualization and can be customized to other cell types of interest, as we did here for microglia somata, using fiji for dataset-specific training. we applied delivr to investigate cancer-related brain activity, unveiling an activation pattern that distinguishes weight-stable cancer from cancers associated with weight loss. overall, delivr is a robust deep-learning tool that does not require advanced coding skills to analyze whole-brain imaging data in health and disease.","key_phrases":["automated detection","specific cells","three-dimensional datasets","whole-brain light-sheet image stacks","we","delivr","a virtual reality-trained deep-learning pipeline","c-fos+ cells","markers","neuronal activity","cleared mouse brains","virtual reality annotation","training data generation","delivr","the-art","our pipeline","a user-friendly docker container","that","a standalone fiji plugin","delivr","a comprehensive toolkit","data visualization","other cell types","interest","we","microglia somata","fiji","dataset-specific training","we","delivr","cancer-related brain activity","an activation pattern","that","weight-stable cancer","cancers","weight loss","delivr","a robust deep-learning tool","that","advanced coding skills","whole-brain imaging data","health","disease","three","delivr","delivr"]},{"title":"Fast flow field prediction of pollutant leakage diffusion based on deep learning","authors":["Wan YunBo","Zhao Zhong","Liu Jie","Zuo KuiJun","Zhang Yong"],"abstract":"AbstractPredicting pollutant leakage and diffusion processes is crucial for ensuring people\u2019s safety. While the deep learning method offers high simulation efficiency and superior generalization, there is currently a lack of research on predicting pollutant leakage and diffusion flow field using deep learning. Therefore, it is necessary to conduct further studies in this area. This paper introduces a two-level network method to model the flow characteristics of pollutant diffusion. The proposed method in this study demonstrates a significant enhancement in flow field prediction accuracy compared to traditional deep learning methods. Moreover, it improves computational efficiency by over 800 times compared to traditional computational fluid dynamics (CFD) methods. Unlike conventional CFD methods that require grid expansion to calculate all operation conditions, the deep learning method is not confined by grid limitations. While deep learning methods may not entirely replace CFD methods, they can serve as a valuable supplementary tool, expanding the versatility of CFD methods. The findings of this research establish a robust foundation for incorporating deep learning methods in addressing pollutant leakage and diffusion challenges.Graphical abstractA deep learning method for simulating pollutant leakage diffusion.","doi":"10.1007\/s11356-024-34462-9","cleaned_title":"fast flow field prediction of pollutant leakage diffusion based on deep learning","cleaned_abstract":"abstractpredicting pollutant leakage and diffusion processes is crucial for ensuring people\u2019s safety. while the deep learning method offers high simulation efficiency and superior generalization, there is currently a lack of research on predicting pollutant leakage and diffusion flow field using deep learning. therefore, it is necessary to conduct further studies in this area. this paper introduces a two-level network method to model the flow characteristics of pollutant diffusion. the proposed method in this study demonstrates a significant enhancement in flow field prediction accuracy compared to traditional deep learning methods. moreover, it improves computational efficiency by over 800 times compared to traditional computational fluid dynamics (cfd) methods. unlike conventional cfd methods that require grid expansion to calculate all operation conditions, the deep learning method is not confined by grid limitations. while deep learning methods may not entirely replace cfd methods, they can serve as a valuable supplementary tool, expanding the versatility of cfd methods. the findings of this research establish a robust foundation for incorporating deep learning methods in addressing pollutant leakage and diffusion challenges.graphical abstracta deep learning method for simulating pollutant leakage diffusion.","key_phrases":["pollutant leakage","diffusion processes","people\u2019s safety","the deep learning method","high simulation efficiency","superior generalization","a lack","research","pollutant leakage","diffusion flow field","deep learning","it","further studies","this area","this paper","a two-level network method","the flow characteristics","pollutant diffusion","the proposed method","this study","a significant enhancement","flow field prediction accuracy","traditional deep learning methods","it","computational efficiency","over 800 times","traditional computational fluid dynamics","(cfd) methods","conventional cfd methods","that","grid expansion","all operation conditions","the deep learning method","grid limitations","deep learning methods","cfd methods","they","a valuable supplementary tool","the versatility","cfd methods","the findings","this research","a robust foundation","deep learning methods","pollutant leakage","diffusion challenges.graphical abstracta","deep learning method","pollutant leakage diffusion","two"]},{"title":"Stroke detection in the brain using MRI and deep learning models","authors":["Subba Rao Polamuri"],"abstract":"When it comes to finding solutions to issues, deep learning models are pretty much everywhere. Medical image data is best analysed using models based on Convolutional Neural Networks (CNNs). Better methods for early detection are crucial due to the concerning increase in the number of people suffering from brain stroke. Among the several medical imaging modalities used for brain imaging, magnetic resonance imaging (MRI) stands out. When it comes to analysing medical photos, the deep learning models currently utilised with MRI have showed good outcomes. To improve the efficacy of brain stroke diagnosis, we suggested several upgrades to deep learning models in this work, including DenseNet121, ResNet50, and VGG16. Since these models are not purpose-built to solve any particular issue, they are modified according to the present situation involving the detection of brain strokes. To make use of all of these cutting-edge deep learning models in a pipeline, we proposed a strategy based on supervised learning. Results from the experiments showed that optimised models outperformed baseline models.","doi":"10.1007\/s11042-024-19318-1","cleaned_title":"stroke detection in the brain using mri and deep learning models","cleaned_abstract":"when it comes to finding solutions to issues, deep learning models are pretty much everywhere. medical image data is best analysed using models based on convolutional neural networks (cnns). better methods for early detection are crucial due to the concerning increase in the number of people suffering from brain stroke. among the several medical imaging modalities used for brain imaging, magnetic resonance imaging (mri) stands out. when it comes to analysing medical photos, the deep learning models currently utilised with mri have showed good outcomes. to improve the efficacy of brain stroke diagnosis, we suggested several upgrades to deep learning models in this work, including densenet121, resnet50, and vgg16. since these models are not purpose-built to solve any particular issue, they are modified according to the present situation involving the detection of brain strokes. to make use of all of these cutting-edge deep learning models in a pipeline, we proposed a strategy based on supervised learning. results from the experiments showed that optimised models outperformed baseline models.","key_phrases":["it","solutions","issues","deep learning models","medical image data","using models","convolutional neural networks","cnns","better methods","early detection","the concerning increase","the number","people","brain stroke","the several medical imaging modalities","brain imaging","magnetic resonance imaging","mri","it","medical photos","the deep learning models","mri","good outcomes","the efficacy","brain stroke diagnosis","we","several upgrades","deep learning models","this work","densenet121","resnet50","vgg16","these models","any particular issue","they","the present situation","the detection","brain strokes","use","all","these cutting-edge deep learning models","a pipeline","we","a strategy","supervised learning","results","the experiments","optimised models","baseline models","resnet50"]},{"title":"Combining Deep Learning with Good Old-Fashioned Machine Learning","authors":["Moshe Sipper"],"abstract":"We present a comprehensive, stacking-based framework for combining deep learning with good old-fashioned machine learning, called Deep GOld. Our framework involves ensemble selection from 51 retrained pretrained deep networks as first-level models, and 10 machine-learning algorithms as second-level models. Enabled by today\u2019s state-of-the-art software tools and hardware platforms, Deep GOld delivers consistent improvement when tested on four image-classification datasets: Fashion MNIST, CIFAR10, CIFAR100, and Tiny ImageNet. Of 120 experiments, in all but 10 Deep GOld improved the original networks\u2019 performance.","doi":"10.1007\/s42979-022-01505-2","cleaned_title":"combining deep learning with good old-fashioned machine learning","cleaned_abstract":"we present a comprehensive, stacking-based framework for combining deep learning with good old-fashioned machine learning, called deep gold. our framework involves ensemble selection from 51 retrained pretrained deep networks as first-level models, and 10 machine-learning algorithms as second-level models. enabled by today\u2019s state-of-the-art software tools and hardware platforms, deep gold delivers consistent improvement when tested on four image-classification datasets: fashion mnist, cifar10, cifar100, and tiny imagenet. of 120 experiments, in all but 10 deep gold improved the original networks\u2019 performance.","key_phrases":["we","a comprehensive, stacking-based framework","deep learning","good old-fashioned machine learning","deep gold","our framework","ensemble selection","deep networks","first-level models","10 machine-learning algorithms","second-level models","the-art","hardware platforms","deep gold","consistent improvement","four image-classification datasets","fashion mnist","cifar10","cifar100","tiny imagenet","120 experiments","all but 10 deep gold","the original networks\u2019 performance","51","first","10","second","today","four","cifar10","120","10"]},{"title":"Image classification of intracranial tumor using deep residual learning technique","authors":["G. Vidya Sagar","M. Ravi Kumar","Sk. Hasane Ahammad","Chella Santhosh"],"abstract":"Classifying brain tumours is essential for diagnosing tumour progression and planning effective treatments. Different imaging modalities are used to diagnose brain tumours. The opposite is true for magnetic resonance imaging (MRI), which has gained widespread use due to its superior image quality and the fact that it does not require ionizing radiation. Image classification of intracranial tumors using deep residual learning technique is an application of deep learning in the field of medical imaging analysis. It involves using convolutional neural networks (CNNs) to automatically classify brain images into different categories based on the presence or absence of tumors. ResNet is a deep neural network that addresses the problem of vanishing gradients during training of very deep networks.The deep learning subfield of machine learning has recently shown remarkable success, especially in classification and segmentation. We trained a deep residual network using picture datasets to distinguish between several brain cancers. The information generated by MRI scans is extensive. A radiologist analyses these images. The three most common brain tumours are meningioma, glioma, and pituitary tumour. Brain tumours are complex diseases, and a manual examination may be fraught with error. Experimental outcomes based on various techniques under augmentation with image-based datasets are presented. The accuracy with no augmentation is about 98% and under augmentation is approximately 99.08%.By leveraging deep residual learning techniques, image classification of intracranial tumors can benefit from the ability of deep neural networks to automatically learn complex representations from raw image data.Classification methods that use machine learning to automate the process have proven superior to human curation. Thus, we present a system that can identify and classify utilizing deep CNN-based residual networks.","doi":"10.1007\/s11042-023-17712-9","cleaned_title":"image classification of intracranial tumor using deep residual learning technique","cleaned_abstract":"classifying brain tumours is essential for diagnosing tumour progression and planning effective treatments. different imaging modalities are used to diagnose brain tumours. the opposite is true for magnetic resonance imaging (mri), which has gained widespread use due to its superior image quality and the fact that it does not require ionizing radiation. image classification of intracranial tumors using deep residual learning technique is an application of deep learning in the field of medical imaging analysis. it involves using convolutional neural networks (cnns) to automatically classify brain images into different categories based on the presence or absence of tumors. resnet is a deep neural network that addresses the problem of vanishing gradients during training of very deep networks.the deep learning subfield of machine learning has recently shown remarkable success, especially in classification and segmentation. we trained a deep residual network using picture datasets to distinguish between several brain cancers. the information generated by mri scans is extensive. a radiologist analyses these images. the three most common brain tumours are meningioma, glioma, and pituitary tumour. brain tumours are complex diseases, and a manual examination may be fraught with error. experimental outcomes based on various techniques under augmentation with image-based datasets are presented. the accuracy with no augmentation is about 98% and under augmentation is approximately 99.08%.by leveraging deep residual learning techniques, image classification of intracranial tumors can benefit from the ability of deep neural networks to automatically learn complex representations from raw image data.classification methods that use machine learning to automate the process have proven superior to human curation. thus, we present a system that can identify and classify utilizing deep cnn-based residual networks.","key_phrases":["brain tumours","tumour progression","effective treatments","different imaging modalities","brain tumours","magnetic resonance imaging","mri","which","widespread use","its superior image quality","the fact","it","ionizing radiation","image classification","intracranial tumors","deep residual learning technique","an application","deep learning","the field","medical imaging analysis","it","convolutional neural networks","cnns","brain images","different categories","the presence","absence","tumors","resnet","a deep neural network","that","the problem","vanishing gradients","training","very deep networks.the deep learning subfield","machine learning","remarkable success","classification","segmentation","we","a deep residual network","picture datasets","several brain cancers","the information","mri scans","a radiologist","these images","the three most common brain tumours","brain tumours","complex diseases","a manual examination","error","experimental outcomes","various techniques","augmentation","image-based datasets","the accuracy","no augmentation","about 98%","augmentation","approximately 99.08%.by","deep residual learning techniques","image classification","intracranial tumors","the ability","deep neural networks","complex representations","raw image data.classification methods","that","the process","human curation","we","a system","that","deep cnn-based residual networks","three","about 98%","cnn"]},{"title":"Attractor Inspired Deep Learning for Modelling Chaotic Systems","authors":["Anurag Dutta","John Harshith","A. Ramamoorthy","K. Lakshmanan"],"abstract":"Predicting and understanding the behavior of dynamic systems have driven advancements in various approaches, including physics-based models and data-driven techniques like deep neural networks. Chaotic systems, with their stochastic nature and unpredictable behavior, pose challenges for accurate modeling and forecasting, especially during extreme events. In this paper, we propose a novel deep learning framework called Attractor-Inspired Deep Learning (AiDL), which seamlessly integrates actual statistics and mathematical models of system kinetics. AiDL combines the strengths of physics-informed machine learning and data-driven methods, offering a promising solution for modeling nonlinear systems. By leveraging the intricate dynamics of attractors, AiDL bridges the gap between physics-based models and deep neural networks. We demonstrate the effectiveness of AiDL using real-world data from various domains, including catastrophic weather mechanics, El Ni\u00f1o cycles, and disease transmission. Our empirical results showcase AiDL\u2019s ability to substantially enhance the modeling of extreme events. The proposed AiDL paradigm holds promise for advancing research in Time Series Prediction of Extreme Events and has applications in real-world chaotic system transformations.","doi":"10.1007\/s44230-023-00045-z","cleaned_title":"attractor inspired deep learning for modelling chaotic systems","cleaned_abstract":"predicting and understanding the behavior of dynamic systems have driven advancements in various approaches, including physics-based models and data-driven techniques like deep neural networks. chaotic systems, with their stochastic nature and unpredictable behavior, pose challenges for accurate modeling and forecasting, especially during extreme events. in this paper, we propose a novel deep learning framework called attractor-inspired deep learning (aidl), which seamlessly integrates actual statistics and mathematical models of system kinetics. aidl combines the strengths of physics-informed machine learning and data-driven methods, offering a promising solution for modeling nonlinear systems. by leveraging the intricate dynamics of attractors, aidl bridges the gap between physics-based models and deep neural networks. we demonstrate the effectiveness of aidl using real-world data from various domains, including catastrophic weather mechanics, el ni\u00f1o cycles, and disease transmission. our empirical results showcase aidl\u2019s ability to substantially enhance the modeling of extreme events. the proposed aidl paradigm holds promise for advancing research in time series prediction of extreme events and has applications in real-world chaotic system transformations.","key_phrases":["the behavior","dynamic systems","advancements","various approaches","physics-based models","data-driven techniques","deep neural networks","chaotic systems","their stochastic nature","unpredictable behavior","challenges","accurate modeling","forecasting","extreme events","this paper","we","a novel deep learning framework","attractor-inspired deep learning","aidl","which","actual statistics","mathematical models","system kinetics","aidl","the strengths","physics-informed machine learning","data-driven methods","a promising solution","nonlinear systems","the intricate dynamics","attractors","aidl","the gap","physics-based models","deep neural networks","we","the effectiveness","aidl","real-world data","various domains","catastrophic weather mechanics","el ni\u00f1o cycles","disease transmission","our empirical results","showcase","the modeling","extreme events","the proposed aidl paradigm","promise","research","time series prediction","extreme events","applications","real-world chaotic system transformations","el ni\u00f1o"]},{"title":"Deep learning for high-resolution seismic imaging","authors":["Liyun Ma","Liguo Han","Qiang Feng"],"abstract":"Seismic imaging techniques play a crucial role in interpreting subsurface geological structures by analyzing the propagation and reflection of seismic waves. However, traditional methods face challenges in achieving high resolution due to theoretical constraints and computational costs. Leveraging recent advancements in deep learning, this study introduces a neural network framework that integrates Transformer and Convolutional Neural Network (CNN) architectures, enhanced through Adaptive Spatial Feature Fusion (ASFF), to achieve high-resolution seismic imaging. Our approach directly maps seismic data to reflection models, eliminating the need for post-processing low-resolution results. Through extensive numerical experiments, we demonstrate the outstanding ability of this method to accurately infer subsurface structures. Evaluation metrics including Root Mean Square Error (RMSE), Correlation Coefficient (CC), and Structural Similarity Index (SSIM) emphasize the model's capacity to faithfully reconstruct subsurface features. Furthermore, noise injection experiments showcase the reliability of this efficient seismic imaging method, further underscoring the potential of deep learning in seismic imaging.","doi":"10.1038\/s41598-024-61251-8","cleaned_title":"deep learning for high-resolution seismic imaging","cleaned_abstract":"seismic imaging techniques play a crucial role in interpreting subsurface geological structures by analyzing the propagation and reflection of seismic waves. however, traditional methods face challenges in achieving high resolution due to theoretical constraints and computational costs. leveraging recent advancements in deep learning, this study introduces a neural network framework that integrates transformer and convolutional neural network (cnn) architectures, enhanced through adaptive spatial feature fusion (asff), to achieve high-resolution seismic imaging. our approach directly maps seismic data to reflection models, eliminating the need for post-processing low-resolution results. through extensive numerical experiments, we demonstrate the outstanding ability of this method to accurately infer subsurface structures. evaluation metrics including root mean square error (rmse), correlation coefficient (cc), and structural similarity index (ssim) emphasize the model's capacity to faithfully reconstruct subsurface features. furthermore, noise injection experiments showcase the reliability of this efficient seismic imaging method, further underscoring the potential of deep learning in seismic imaging.","key_phrases":["seismic imaging techniques","a crucial role","subsurface geological structures","the propagation","reflection","seismic waves","traditional methods","challenges","high resolution","theoretical constraints","computational costs","recent advancements","deep learning","this study","a neural network framework","that","transformer","cnn","adaptive spatial feature fusion","(asff","high-resolution seismic imaging","our approach","seismic data","reflection models","the need","post-processing low-resolution results","extensive numerical experiments","we","the outstanding ability","this method","subsurface structures","evaluation metrics","root mean square error","rmse","correlation","structural similarity index","ssim","the model's capacity","subsurface features","noise injection experiments","the reliability","this efficient seismic imaging method","the potential","deep learning","seismic imaging","cnn"]},{"title":"Predicting equilibrium distributions for molecular systems with deep learning","authors":["Shuxin Zheng","Jiyan He","Chang Liu","Yu Shi","Ziheng Lu","Weitao Feng","Fusong Ju","Jiaxi Wang","Jianwei Zhu","Yaosen Min","He Zhang","Shidi Tang","Hongxia Hao","Peiran Jin","Chi Chen","Frank No\u00e9","Haiguang Liu","Tie-Yan Liu"],"abstract":"Advances in deep learning have greatly improved structure prediction of molecules. However, many macroscopic observations that are important for real-world applications are not functions of a single molecular structure but rather determined from the equilibrium distribution of structures. Conventional methods for obtaining these distributions, such as molecular dynamics simulation, are computationally expensive and often intractable. Here we introduce a deep learning framework, called Distributional Graphormer (DiG), in an attempt to predict the equilibrium distribution of molecular systems. Inspired by the annealing process in thermodynamics, DiG uses deep neural networks to transform a simple distribution towards the equilibrium distribution, conditioned on a descriptor of a molecular system such as a chemical graph or a protein sequence. This framework enables the efficient generation of diverse conformations and provides estimations of state densities, orders of magnitude faster than conventional methods. We demonstrate applications of DiG on several molecular tasks, including protein conformation sampling, ligand structure sampling, catalyst\u2013adsorbate sampling and property-guided structure generation. DiG presents a substantial advancement in methodology for statistically understanding molecular systems, opening up new research opportunities in the molecular sciences.","doi":"10.1038\/s42256-024-00837-3","cleaned_title":"predicting equilibrium distributions for molecular systems with deep learning","cleaned_abstract":"advances in deep learning have greatly improved structure prediction of molecules. however, many macroscopic observations that are important for real-world applications are not functions of a single molecular structure but rather determined from the equilibrium distribution of structures. conventional methods for obtaining these distributions, such as molecular dynamics simulation, are computationally expensive and often intractable. here we introduce a deep learning framework, called distributional graphormer (dig), in an attempt to predict the equilibrium distribution of molecular systems. inspired by the annealing process in thermodynamics, dig uses deep neural networks to transform a simple distribution towards the equilibrium distribution, conditioned on a descriptor of a molecular system such as a chemical graph or a protein sequence. this framework enables the efficient generation of diverse conformations and provides estimations of state densities, orders of magnitude faster than conventional methods. we demonstrate applications of dig on several molecular tasks, including protein conformation sampling, ligand structure sampling, catalyst\u2013adsorbate sampling and property-guided structure generation. dig presents a substantial advancement in methodology for statistically understanding molecular systems, opening up new research opportunities in the molecular sciences.","key_phrases":["advances","deep learning","structure prediction","molecules","many macroscopic observations","that","real-world applications","functions","a single molecular structure","the equilibrium distribution","structures","conventional methods","these distributions","molecular dynamics simulation","we","a deep learning framework","distributional graphormer","dig","an attempt","the equilibrium distribution","molecular systems","the annealing process","thermodynamics","dig","deep neural networks","a simple distribution","the equilibrium distribution","a descriptor","a molecular system","a chemical graph","a protein sequence","this framework","the efficient generation","diverse conformations","estimations","state densities","orders","magnitude","conventional methods","we","applications","dig","several molecular tasks","protein conformation sampling","ligand structure sampling","catalyst","sampling","property-guided structure generation","dig","a substantial advancement","methodology","molecular systems","new research opportunities","the molecular sciences"]},{"title":"Privacy-Preserving Classification on Deep Learning with Exponential Mechanism","authors":["Quan Ju","Rongqing Xia","Shuhong Li","Xiaojian Zhang"],"abstract":"How to protect the privacy of training data in deep learning has been the subject of increasing amounts of related research in recent years. Private Aggregation of Teacher Ensembles (PATE) uses transfer learning and differential privacy methods to provide a broadly applicable data privacy framework in deep learning. PATE combines the Laplacian mechanism and the voting method to achieve deep learning privacy classification. However, the Laplacian mechanism may greatly distort the histogram vote counts of each class. This paper proposes a novel exponential mechanism with PATE to ensure the privacy protection. This proposed method improves the protection effect and accuracy through the screening algorithm and uses the differential privacy combination theorems to reduce the total privacy budget. The data-dependent analysis demonstrates that the exponential mechanism outperforms the original Laplace mechanism. Experimental results show that the proposed method can train models with improved accuracy while requiring a smaller privacy budget when compared to the original Pate framework.","doi":"10.1007\/s44196-024-00422-x","cleaned_title":"privacy-preserving classification on deep learning with exponential mechanism","cleaned_abstract":"how to protect the privacy of training data in deep learning has been the subject of increasing amounts of related research in recent years. private aggregation of teacher ensembles (pate) uses transfer learning and differential privacy methods to provide a broadly applicable data privacy framework in deep learning. pate combines the laplacian mechanism and the voting method to achieve deep learning privacy classification. however, the laplacian mechanism may greatly distort the histogram vote counts of each class. this paper proposes a novel exponential mechanism with pate to ensure the privacy protection. this proposed method improves the protection effect and accuracy through the screening algorithm and uses the differential privacy combination theorems to reduce the total privacy budget. the data-dependent analysis demonstrates that the exponential mechanism outperforms the original laplace mechanism. experimental results show that the proposed method can train models with improved accuracy while requiring a smaller privacy budget when compared to the original pate framework.","key_phrases":["the privacy","training data","deep learning","the subject","increasing amounts","related research","recent years","private aggregation","teacher ensembles","transfer learning","differential privacy methods","a broadly applicable data privacy framework","deep learning","the laplacian mechanism","the voting method","deep learning privacy classification","the laplacian mechanism","the histogram vote counts","each class","this paper","a novel exponential mechanism","pate","the privacy protection","this proposed method","the protection effect","accuracy","the screening algorithm","the differential privacy combination","the total privacy budget","the data-dependent analysis","the exponential mechanism","the original laplace mechanism","experimental results","the proposed method","models","improved accuracy","a smaller privacy budget","the original pate framework","recent years"]},{"title":"An empirical study of sentiment analysis utilizing machine learning and deep learning algorithms","authors":["Betul Erkantarci","Gokhan Bakal"],"abstract":"Among text-mining studies, one of the most studied topics is the text classification task applied in various domains, including medicine, social media, and academia. As a sub-problem in text classification, sentiment analysis has been widely investigated to classify often opinion-based textual elements. Specifically, user reviews and experiential feedback for products or services have been employed as fundamental data sources for sentiment analysis efforts. As a result of rapidly emerging technological advancements, social media platforms such as Twitter, Facebook, and Reddit, have become central opinion-sharing mediums since the early 2000s. In this sense, we build various machine-learning models to solve the sentiment analysis problem on the Reddit comments dataset in this work. The experimental models we constructed achieve F1 scores within intervals of 73\u201376%. Consequently, we present comparative performance scores obtained by traditional machine learning and deep learning models and discuss the results.","doi":"10.1007\/s42001-023-00236-5","cleaned_title":"an empirical study of sentiment analysis utilizing machine learning and deep learning algorithms","cleaned_abstract":"among text-mining studies, one of the most studied topics is the text classification task applied in various domains, including medicine, social media, and academia. as a sub-problem in text classification, sentiment analysis has been widely investigated to classify often opinion-based textual elements. specifically, user reviews and experiential feedback for products or services have been employed as fundamental data sources for sentiment analysis efforts. as a result of rapidly emerging technological advancements, social media platforms such as twitter, facebook, and reddit, have become central opinion-sharing mediums since the early 2000s. in this sense, we build various machine-learning models to solve the sentiment analysis problem on the reddit comments dataset in this work. the experimental models we constructed achieve f1 scores within intervals of 73\u201376%. consequently, we present comparative performance scores obtained by traditional machine learning and deep learning models and discuss the results.","key_phrases":["text-mining studies","the most studied topics","the text classification task","various domains","medicine","social media","academia","a sub","-","problem","text classification","sentiment analysis","opinion-based textual elements","user reviews","experiential feedback","products","services","fundamental data sources","sentiment analysis efforts","a result","rapidly emerging technological advancements","social media platforms","twitter","facebook","reddit","central opinion-sharing mediums","this sense","we","various machine-learning models","the sentiment analysis problem","the reddit comments","this work","the experimental models","we","f1 scores","intervals","73\u201376%","we","comparative performance scores","traditional machine learning","deep learning models","the results","one","the early 2000s","73\u201376%"]},{"title":"Application of deep learning in isolated tooth identification","authors":["Meng-Xun Li","Zhi-Wei Wang","Xin-Ran Chen","Gui-Song Xia","Yong Zheng","Cui Huang","Zhi Li"],"abstract":"BackgroundTeeth identification has a pivotal role in the dental curriculum and provides one of the important foundations of clinical practice. Accurately identifying teeth is a vital aspect of dental education and clinical practice, but can be challenging due to the anatomical similarities between categories. In this study, we aim to explore the possibility of using a deep learning model to classify isolated tooth by a set of photographs.MethodsA collection of 5,100 photographs from 850 isolated human tooth specimens were assembled to serve as the dataset for this study. Each tooth was carefully labeled during the data collection phase through direct observation. We developed a deep learning model that incorporates the state-of-the-art feature extractor and attention mechanism to classify each tooth based on a set of 6 photographs captured from multiple angles. To increase the validity of model evaluation, a voting-based strategy was applied to refine the test set to generate a more reliable label, and the model was evaluated under different types of classification granularities.ResultsThis deep learning model achieved top-3 accuracies of over 90% in all classification types, with an average AUC of 0.95. The Cohen\u2019s Kappa demonstrated good agreement between model prediction and the test set.ConclusionsThis deep learning model can achieve performance comparable to that of human experts and has the potential to become a valuable tool for dental education and various applications in accurately identifying isolated tooth.","doi":"10.1186\/s12903-024-04274-x","cleaned_title":"application of deep learning in isolated tooth identification","cleaned_abstract":"backgroundteeth identification has a pivotal role in the dental curriculum and provides one of the important foundations of clinical practice. accurately identifying teeth is a vital aspect of dental education and clinical practice, but can be challenging due to the anatomical similarities between categories. in this study, we aim to explore the possibility of using a deep learning model to classify isolated tooth by a set of photographs.methodsa collection of 5,100 photographs from 850 isolated human tooth specimens were assembled to serve as the dataset for this study. each tooth was carefully labeled during the data collection phase through direct observation. we developed a deep learning model that incorporates the state-of-the-art feature extractor and attention mechanism to classify each tooth based on a set of 6 photographs captured from multiple angles. to increase the validity of model evaluation, a voting-based strategy was applied to refine the test set to generate a more reliable label, and the model was evaluated under different types of classification granularities.resultsthis deep learning model achieved top-3 accuracies of over 90% in all classification types, with an average auc of 0.95. the cohen\u2019s kappa demonstrated good agreement between model prediction and the test set.conclusionsthis deep learning model can achieve performance comparable to that of human experts and has the potential to become a valuable tool for dental education and various applications in accurately identifying isolated tooth.","key_phrases":["backgroundteeth identification","a pivotal role","the dental curriculum","the important foundations","clinical practice","teeth","a vital aspect","dental education","clinical practice","the anatomical similarities","categories","this study","we","the possibility","a deep learning model","isolated tooth","a set","photographs.methodsa collection","5,100 photographs","850 isolated human tooth specimens","the dataset","this study","each tooth","the data collection phase","direct observation","we","a deep learning model","that","the-art","attention mechanism","each tooth","a set","6 photographs","multiple angles","the validity","model evaluation","a voting-based strategy","the test","a more reliable label","the model","different types","classification","granularities.resultsthis deep learning model","top-3 accuracies","over 90%","all classification types","an average auc","the cohen\u2019s kappa","good agreement","model prediction","the test","deep learning model","performance","that","human experts","the potential","a valuable tool","dental education","various applications","isolated tooth","5,100","850","6","over 90%","0.95"]},{"title":"Deep learning in neglected vector-borne diseases: a systematic review","authors":["Atmika Mishra","Arya Pandey","Ruchika Malhotra"],"abstract":"This study explores the application of Deep Learning in combating neglected vector-borne Diseases, a significant global health concern, particularly in resource-limited areas. It examines areas where Deep Learning has proven effective, compares popular Deep Learning techniques, focuses on interdisciplinary approaches with translational impact, and finds untapped potential for deep learning application. Thorough searches across multiple databases yielded 64 pertinent studies, from which 16 were selected based on inclusion criteria and quality assessment. Deep Learning applications in disease transmission risk prediction, vector detection, parasite classification, and treatment procedure optimization were investigated and focused on diseases such as Schistosomiasis, Chagas disease, Leishmaniasis, Echinococcosis, and Trachoma. Convolutional neural networks, artificial neural networks, multilayer perceptrons, and AutoML algorithms surpassed traditional methods for disease prediction, species identification, and diagnosis. The interdisciplinary integration of Deep Learning with public health, entomology, and epidemiology provides prospects for improved disease control and understanding. Deep Learning models automate disease surveillance, simplify epidemiological data processing, and enable early detection, particularly in resource-constrained settings. Smartphone apps driven by deep learning allow for rapid disease diagnosis and identification, boosting healthcare accessibility and global health outcomes. Improved algorithms, broadening the scope of applications to areas such as one health approach, and community engagement, and expanding deep learning applications to diseases such as lymphatic filariasis, hydatidosis, and onchocerciasis hold promise for improving global health outcomes.","doi":"10.1007\/s13198-024-02380-1","cleaned_title":"deep learning in neglected vector-borne diseases: a systematic review","cleaned_abstract":"this study explores the application of deep learning in combating neglected vector-borne diseases, a significant global health concern, particularly in resource-limited areas. it examines areas where deep learning has proven effective, compares popular deep learning techniques, focuses on interdisciplinary approaches with translational impact, and finds untapped potential for deep learning application. thorough searches across multiple databases yielded 64 pertinent studies, from which 16 were selected based on inclusion criteria and quality assessment. deep learning applications in disease transmission risk prediction, vector detection, parasite classification, and treatment procedure optimization were investigated and focused on diseases such as schistosomiasis, chagas disease, leishmaniasis, echinococcosis, and trachoma. convolutional neural networks, artificial neural networks, multilayer perceptrons, and automl algorithms surpassed traditional methods for disease prediction, species identification, and diagnosis. the interdisciplinary integration of deep learning with public health, entomology, and epidemiology provides prospects for improved disease control and understanding. deep learning models automate disease surveillance, simplify epidemiological data processing, and enable early detection, particularly in resource-constrained settings. smartphone apps driven by deep learning allow for rapid disease diagnosis and identification, boosting healthcare accessibility and global health outcomes. improved algorithms, broadening the scope of applications to areas such as one health approach, and community engagement, and expanding deep learning applications to diseases such as lymphatic filariasis, hydatidosis, and onchocerciasis hold promise for improving global health outcomes.","key_phrases":["this study","the application","deep learning","neglected vector-borne diseases","a significant global health concern","resource-limited areas","it","areas","deep learning","popular deep learning techniques","interdisciplinary approaches","translational impact","untapped potential","deep learning application","thorough searches","multiple databases","64 pertinent studies","which","inclusion criteria","quality assessment","deep learning applications","disease transmission risk prediction","vector detection","classification","treatment procedure optimization","diseases","schistosomiasis","chagas disease","leishmaniasis","echinococcosis","trachoma","convolutional neural networks","artificial neural networks","multilayer perceptrons","automl algorithms","traditional methods","disease prediction","species identification","diagnosis","the interdisciplinary integration","deep learning","public health","entomology","epidemiology","prospects","improved disease control","understanding","deep learning models","automate disease surveillance","epidemiological data processing","early detection","resource-constrained settings","smartphone apps","deep learning","rapid disease diagnosis","identification","healthcare accessibility","global health outcomes","improved algorithms","the scope","applications","areas","one health approach","community engagement","deep learning applications","diseases","lymphatic filariasis","hydatidosis","onchocerciasis hold promise","global health outcomes","64","16","one"]},{"title":"State of the Art on Deep Learning-enhanced Rendering Methods","authors":["Qi Wang","Zhihua Zhong","Yuchi Huo","Hujun Bao","Rui Wang"],"abstract":"Photorealistic rendering of the virtual world is an important and classic problem in the field of computer graphics. With the development of GPU hardware and continuous research on computer graphics, representing and rendering virtual scenes has become easier and more efficient. However, there are still unresolved challenges in efficiently rendering global illumination effects. At the same time, machine learning and computer vision provide real-world image analysis and synthesis methods, which can be exploited by computer graphics rendering pipelines. Deep learning-enhanced rendering combines techniques from deep learning and computer vision into the traditional graphics rendering pipeline to enhance existing rasterization or Monte Carlo integration renderers. This state-of-the-art report summarizes recent studies of deep learning-enhanced rendering in the computer graphics community. Specifically, we focus on works of renderers represented using neural networks, whether the scene is represented by neural networks or traditional scene files. These works are either for general scenes or specific scenes, which are differentiated by the need to retrain the network for new scenes.","doi":"10.1007\/s11633-022-1400-x","cleaned_title":"state of the art on deep learning-enhanced rendering methods","cleaned_abstract":"photorealistic rendering of the virtual world is an important and classic problem in the field of computer graphics. with the development of gpu hardware and continuous research on computer graphics, representing and rendering virtual scenes has become easier and more efficient. however, there are still unresolved challenges in efficiently rendering global illumination effects. at the same time, machine learning and computer vision provide real-world image analysis and synthesis methods, which can be exploited by computer graphics rendering pipelines. deep learning-enhanced rendering combines techniques from deep learning and computer vision into the traditional graphics rendering pipeline to enhance existing rasterization or monte carlo integration renderers. this state-of-the-art report summarizes recent studies of deep learning-enhanced rendering in the computer graphics community. specifically, we focus on works of renderers represented using neural networks, whether the scene is represented by neural networks or traditional scene files. these works are either for general scenes or specific scenes, which are differentiated by the need to retrain the network for new scenes.","key_phrases":["photorealistic rendering","the virtual world","an important and classic problem","the field","computer graphics","the development","gpu hardware","continuous research","computer graphics","virtual scenes","unresolved challenges","global illumination effects","the same time","machine learning","computer vision","real-world image analysis","synthesis methods","which","computer graphics","pipelines","deep learning-enhanced rendering","techniques","deep learning and computer vision","the traditional graphics","pipeline","existing rasterization","monte carlo integration renderers","the-art","recent studies","deep learning-enhanced rendering","the computer graphics community","we","works","renderers","neural networks","the scene","neural networks","traditional scene files","these works","general scenes","specific scenes","which","the need","the network","new scenes"]},{"title":"Deep neural networks watermark via universal deep hiding and metric learning","authors":["Zhicheng Ye","Xinpeng Zhang","Guorui Feng"],"abstract":"With the rising costs of model training, it is urgent to safeguard the intellectual property of deep neural networks. To achieve this, researchers have proposed various model watermarking techniques. Existing methods utilize visible trigger patterns, which are vulnerable to being detected by humans or detectors. Moreover, these approaches fail to establish active protection mechanisms that link the model with the user\u2019s identity. In this study, we present an innovative imperceptible model watermarking approach that utilizes deep hiding to encode the user\u2019s copyright verification information. This process superimposes a trigger pattern onto clean images, resulting in watermark trigger images. These watermark trigger images closely mimic the original images, achieving excellent stealthiness while enabling the retrieval of the user\u2019s copyright verification information, thus definitively asserting ownership rights. Slight alterations made to the images to maintain stealthiness can weaken the triggering of the watermark pattern. We first leverage the triple loss in metric learning to tackle this challenge of training watermark samples. Using watermark trigger images as anchor samples and selecting appropriate positive and negative samples, we enhance the model\u2019s capability to discern the watermark trigger. Experimental results on CIFAR-10, GTSRB, and Tiny-ImageNet confirm the defender\u2019s capability to embed watermark successfully. The average watermark accuracy exceeds 90%, while the average performance loss is less than 0.05% points. It is also robust to existing watermark removal attacks and backdoor detection methods.","doi":"10.1007\/s00521-024-09469-5","cleaned_title":"deep neural networks watermark via universal deep hiding and metric learning","cleaned_abstract":"with the rising costs of model training, it is urgent to safeguard the intellectual property of deep neural networks. to achieve this, researchers have proposed various model watermarking techniques. existing methods utilize visible trigger patterns, which are vulnerable to being detected by humans or detectors. moreover, these approaches fail to establish active protection mechanisms that link the model with the user\u2019s identity. in this study, we present an innovative imperceptible model watermarking approach that utilizes deep hiding to encode the user\u2019s copyright verification information. this process superimposes a trigger pattern onto clean images, resulting in watermark trigger images. these watermark trigger images closely mimic the original images, achieving excellent stealthiness while enabling the retrieval of the user\u2019s copyright verification information, thus definitively asserting ownership rights. slight alterations made to the images to maintain stealthiness can weaken the triggering of the watermark pattern. we first leverage the triple loss in metric learning to tackle this challenge of training watermark samples. using watermark trigger images as anchor samples and selecting appropriate positive and negative samples, we enhance the model\u2019s capability to discern the watermark trigger. experimental results on cifar-10, gtsrb, and tiny-imagenet confirm the defender\u2019s capability to embed watermark successfully. the average watermark accuracy exceeds 90%, while the average performance loss is less than 0.05% points. it is also robust to existing watermark removal attacks and backdoor detection methods.","key_phrases":["the rising costs","model training","it","the intellectual property","deep neural networks","this","researchers","various model watermarking techniques","existing methods","visible trigger patterns","which","humans","detectors","these approaches","active protection mechanisms","that","the model","the user\u2019s identity","this study","we","an innovative imperceptible model watermarking approach","that","deep hiding","the user\u2019s copyright verification information","this process","a trigger pattern","clean images","watermark trigger images","these watermark trigger images","the original images","excellent stealthiness","the retrieval","the user\u2019s copyright verification information","ownership rights","slight alterations","the images","stealthiness","the triggering","the watermark pattern","we","the triple loss","metric learning","this challenge","training watermark samples","watermark trigger images","anchor samples","appropriate positive and negative samples","we","the model\u2019s capability","the watermark trigger","experimental results","cifar-10","gtsrb","tiny-imagenet","the defender\u2019s capability","the average watermark accuracy","90%","the average performance loss","less than 0.05% points","it","existing watermark removal attacks","backdoor detection methods","first","cifar-10","90%","less than","0.05%"]},{"title":"Deep learning for higher-order nonparametric spatial autoregressive model","authors":["Zitong Li","Yunquan Song","Ling Jian"],"abstract":"Deep learning technology has been successfully applied in more and more fields. In this paper, the application of deep neural networks in higher-order nonparametric spatial autoregressive models is studied. For spatial model, we propose the higher-order nonparametric spatial autoregressive neural network (HNSARNN) to fit the model. This method offers both good interpretability and prediction performance, and solves the black box problem in deep learning models to some degree. In various scenarios of spatial data distribution, the proposed method demonstrates superior performance compared to traditional approaches for handling nonparametric functions (such as the B-spline method). Simulation results show the effectiveness of the proposed model.","doi":"10.1007\/s10489-024-05541-8","cleaned_title":"deep learning for higher-order nonparametric spatial autoregressive model","cleaned_abstract":"deep learning technology has been successfully applied in more and more fields. in this paper, the application of deep neural networks in higher-order nonparametric spatial autoregressive models is studied. for spatial model, we propose the higher-order nonparametric spatial autoregressive neural network (hnsarnn) to fit the model. this method offers both good interpretability and prediction performance, and solves the black box problem in deep learning models to some degree. in various scenarios of spatial data distribution, the proposed method demonstrates superior performance compared to traditional approaches for handling nonparametric functions (such as the b-spline method). simulation results show the effectiveness of the proposed model.","key_phrases":["deep learning technology","more and more fields","this paper","the application","deep neural networks","higher-order nonparametric spatial autoregressive models","spatial model","we","the higher-order nonparametric spatial autoregressive neural network","hnsarnn","the model","this method","both good interpretability","prediction performance","the black box problem","deep learning models","some degree","various scenarios","spatial data distribution","the proposed method","superior performance","traditional approaches","nonparametric functions","the b-spline method","simulation results","the effectiveness","the proposed model"]},{"title":"A meta-analysis on diabetic retinopathy and deep learning applications","authors":["Abd\u00fcssamed Erciyas","Necaattin Bari\u015f\u00e7i"],"abstract":"Diabetic retinopathy is one of the negative effects of diabetes on the eye. Early diagnosis of this disease, which can progress to blindness, is very important in this sense. There are many studies that detect and classify diabetic retinopathy, especially Machine Learning and Deep Learning methods. It is known that Deep Learning has been used more and more on disease detection and classification in recent years. There are three important reasons why deep learning is more successful in disease detection than methods such as image processing or machine learning. The first of these is that it achieves higher accuracies. Secondly, there is no need to develop an algorithm for each disease, that is, the algorithm learns the disease itself. Thirdly, faster results can be achieved with GPU (Graphics Processing Unit) support. For these reasons, in this study, articles written between 2015 and 2022 on the classification of diabetic retinopathy with deep learning were examined, and meta and statistical analysis was performed. Considering the work in the last two years the combined SEN value is 0.97 [95% CI, 0.92, 0.98], and the SPE value is 0.99 [95% CI, 0.98, 1.00]. The results obtained show how effective and necessary deep learning is in the early diagnosis of diabetic retinopathy.","doi":"10.1007\/s11042-023-17784-7","cleaned_title":"a meta-analysis on diabetic retinopathy and deep learning applications","cleaned_abstract":"diabetic retinopathy is one of the negative effects of diabetes on the eye. early diagnosis of this disease, which can progress to blindness, is very important in this sense. there are many studies that detect and classify diabetic retinopathy, especially machine learning and deep learning methods. it is known that deep learning has been used more and more on disease detection and classification in recent years. there are three important reasons why deep learning is more successful in disease detection than methods such as image processing or machine learning. the first of these is that it achieves higher accuracies. secondly, there is no need to develop an algorithm for each disease, that is, the algorithm learns the disease itself. thirdly, faster results can be achieved with gpu (graphics processing unit) support. for these reasons, in this study, articles written between 2015 and 2022 on the classification of diabetic retinopathy with deep learning were examined, and meta and statistical analysis was performed. considering the work in the last two years the combined sen value is 0.97 [95% ci, 0.92, 0.98], and the spe value is 0.99 [95% ci, 0.98, 1.00]. the results obtained show how effective and necessary deep learning is in the early diagnosis of diabetic retinopathy.","key_phrases":["diabetic retinopathy","the negative effects","diabetes","the eye","early diagnosis","this disease","which","blindness","this sense","many studies","that","diabetic retinopathy","especially machine learning","deep learning methods","it","deep learning","disease detection","classification","recent years","three important reasons","deep learning","disease detection","methods","image processing","machine learning","these","it","higher accuracies","no need","an algorithm","each disease","the algorithm","the disease","itself","faster results","gpu","graphics processing unit","support","these reasons","this study","articles","the classification","diabetic retinopathy","deep learning","meta and statistical analysis","the work","the last two years","the combined sen value","[95% ci","the spe value","[95% ci","the results","how effective and necessary deep learning","the early diagnosis","diabetic retinopathy","recent years","three","first","secondly","thirdly","gpu","between 2015 and 2022","the last two years","0.97","95%","0.92","0.98","0.99","95%","0.98","1.00"]},{"title":"Naturalistic Scene Modelling: Deep Learning with Insights from Biology","authors":["Kofi Appiah","Zhiyong Jin","Lei Shi","Sze Chai Kwok"],"abstract":"Advances in machine learning coupled with the abundances of training data has facilitated the deep learning era, which has demonstrated its ability and effectiveness in solving complex detection and recognition problems. In general application areas with elements of machine learning have seen exponential growth with promising new and sophisticated solutions to complex learning problems. In computer vision, the challenge related to the detection of known objects in a scene is a thing of the past. With the tremendous increase in detection accuracies, some close to that of human detection, there are several areas still lagging in computer vision and machine learning where improvements may call for more architectural designs. In this paper, we propose a physiologically inspired model for scene understanding that encodes three key components: object location, size and category. Our aim is to develop an energy efficient artificial intelligent model for naturalistic scene understanding capable of deploying on a low power neuromorphic hardware. We have reviewed recent advances in deep learning architecture that have taken inspiration from human or primate learning systems and provided direct to future advancement on deep learning with inspiration from physiological experiments. Upon a review of areas that have benefitted from deep learning, we provide recommendations for enhancing those areas that might have stalled or grinded to a halt with little or no significant improvement.","doi":"10.1007\/s11265-023-01894-4","cleaned_title":"naturalistic scene modelling: deep learning with insights from biology","cleaned_abstract":"advances in machine learning coupled with the abundances of training data has facilitated the deep learning era, which has demonstrated its ability and effectiveness in solving complex detection and recognition problems. in general application areas with elements of machine learning have seen exponential growth with promising new and sophisticated solutions to complex learning problems. in computer vision, the challenge related to the detection of known objects in a scene is a thing of the past. with the tremendous increase in detection accuracies, some close to that of human detection, there are several areas still lagging in computer vision and machine learning where improvements may call for more architectural designs. in this paper, we propose a physiologically inspired model for scene understanding that encodes three key components: object location, size and category. our aim is to develop an energy efficient artificial intelligent model for naturalistic scene understanding capable of deploying on a low power neuromorphic hardware. we have reviewed recent advances in deep learning architecture that have taken inspiration from human or primate learning systems and provided direct to future advancement on deep learning with inspiration from physiological experiments. upon a review of areas that have benefitted from deep learning, we provide recommendations for enhancing those areas that might have stalled or grinded to a halt with little or no significant improvement.","key_phrases":["advances","machine learning","the abundances","training data","the deep learning era","which","its ability","effectiveness","complex detection and recognition problems","general application areas","elements","machine learning","exponential growth","new and sophisticated solutions","complex learning problems","computer vision","the challenge","the detection","known objects","a scene","a thing","the past","the tremendous increase","detection accuracies","that","human detection","several areas","computer vision","machine learning","improvements","more architectural designs","this paper","we","a physiologically inspired model","three key components","object location","size","category","our aim","an energy efficient artificial intelligent model","naturalistic scene","a low power neuromorphic hardware","we","recent advances","deep learning architecture","that","inspiration","learning systems","future advancement","deep learning","inspiration","physiological experiments","a review","areas","that","deep learning","we","recommendations","those areas","that","a halt","little or no significant improvement","three"]},{"title":"Needle tracking in low-resolution ultrasound volumes using deep learning","authors":["Sarah Grube","Sarah Latus","Finn Behrendt","Oleksandra Riabova","Maximilian Neidhardt","Alexander Schlaefer"],"abstract":"PurposeClinical needle insertion into tissue, commonly assisted by 2D ultrasound imaging for real-time navigation, faces the challenge of precise needle and probe alignment to reduce out-of-plane movement. Recent studies investigate 3D ultrasound imaging together with deep learning to overcome this problem, focusing on acquiring high-resolution images to create optimal conditions for needle tip detection. However, high-resolution also requires a lot of time for image acquisition and processing, which limits the real-time capability. Therefore, we aim to maximize the US volume rate with the trade-off of low image resolution. We propose a deep learning approach to directly extract the 3D needle tip position from sparsely sampled US volumes.MethodsWe design an experimental setup with a robot inserting a needle into water and chicken liver tissue. In contrast to manual annotation, we assess the needle tip position from the known robot pose. During insertion, we acquire a large data set of low-resolution volumes using a 16\u00a0\\(\\times \\)\u00a016 element matrix transducer with a volume rate of 4\u00a0Hz. We compare the performance of our deep learning approach with conventional needle segmentation.ResultsOur experiments in water and liver show that deep learning outperforms the conventional approach while achieving sub-millimeter accuracy. We achieve mean position errors of 0.54\u00a0mm in water and 1.54\u00a0mm in liver for deep learning.ConclusionOur study underlines the strengths of deep learning to predict the 3D needle positions from low-resolution ultrasound volumes. This is an important milestone for real-time needle navigation, simplifying the alignment of needle and ultrasound probe and enabling a 3D motion analysis.","doi":"10.1007\/s11548-024-03234-8","cleaned_title":"needle tracking in low-resolution ultrasound volumes using deep learning","cleaned_abstract":"purposeclinical needle insertion into tissue, commonly assisted by 2d ultrasound imaging for real-time navigation, faces the challenge of precise needle and probe alignment to reduce out-of-plane movement. recent studies investigate 3d ultrasound imaging together with deep learning to overcome this problem, focusing on acquiring high-resolution images to create optimal conditions for needle tip detection. however, high-resolution also requires a lot of time for image acquisition and processing, which limits the real-time capability. therefore, we aim to maximize the us volume rate with the trade-off of low image resolution. we propose a deep learning approach to directly extract the 3d needle tip position from sparsely sampled us volumes.methodswe design an experimental setup with a robot inserting a needle into water and chicken liver tissue. in contrast to manual annotation, we assess the needle tip position from the known robot pose. during insertion, we acquire a large data set of low-resolution volumes using a 16 \\(\\times \\) 16 element matrix transducer with a volume rate of 4 hz. we compare the performance of our deep learning approach with conventional needle segmentation.resultsour experiments in water and liver show that deep learning outperforms the conventional approach while achieving sub-millimeter accuracy. we achieve mean position errors of 0.54 mm in water and 1.54 mm in liver for deep learning.conclusionour study underlines the strengths of deep learning to predict the 3d needle positions from low-resolution ultrasound volumes. this is an important milestone for real-time needle navigation, simplifying the alignment of needle and ultrasound probe and enabling a 3d motion analysis.","key_phrases":["purposeclinical needle insertion","tissue","2d ultrasound imaging","real-time navigation","the challenge","precise needle","probe alignment","plane","recent studies","3d ultrasound","deep learning","this problem","high-resolution images","optimal conditions","needle tip detection","high-resolution","a lot","time","image acquisition","processing","which","the real-time capability","we","the us volume rate","the trade-off","low image resolution","we","a deep learning approach","the 3d needle tip position","us volumes.methodswe design","a robot","a needle","water and chicken liver tissue","contrast","manual annotation","we","the needle tip position","the known robot","insertion","we","a large data","low-resolution volumes","a 16 \\(\\times","16 element matrix transducer","a volume rate","4 hz","we","the performance","our deep learning approach","conventional needle","segmentation.resultsour experiments","water","liver","deep learning","the conventional approach","sub-millimeter accuracy","we","mean position errors","0.54 mm","water","1.54 mm","liver","deep learning.conclusionour study","the strengths","deep learning","the 3d needle positions","low-resolution ultrasound volumes","this","an important milestone","real-time needle navigation","the alignment","needle and ultrasound probe","a 3d motion analysis","2d","3d","us","3d","16 \\(\\times \\","16","4","segmentation.resultsour","0.54 mm","1.54 mm","3d","3d"]},{"title":"Deep learning, textual sentiment, and financial market","authors":["Fuwei Jiang","Yumin Liu","Lingchao Meng","Huajing Zhang"],"abstract":"In this paper, we apply the BERT model, a cut-edging deep learning model, to construct a novel textual sentiment index in the Chinese stock market. By introducing the stock market returns as sentiment labels, our BERT model effectively extracts textual sentiment-related information useful for asset pricing. We find that the BERT-based sentiment has much greater predictive power for stock market returns than the traditional dictionary method as well as the Baker\u2013Wurgler investor sentiment index both in and out of sample. The BERT-based sentiment shows strong predictive power during economic downturns and can significantly predict future macroeconomic conditions. Overall, our BERT model offers a better measure of textual investor sentiment, highlighting the potentially significant value of deep learning, AI, and FinTech in financial market.","doi":"10.1007\/s10799-024-00428-z","cleaned_title":"deep learning, textual sentiment, and financial market","cleaned_abstract":"in this paper, we apply the bert model, a cut-edging deep learning model, to construct a novel textual sentiment index in the chinese stock market. by introducing the stock market returns as sentiment labels, our bert model effectively extracts textual sentiment-related information useful for asset pricing. we find that the bert-based sentiment has much greater predictive power for stock market returns than the traditional dictionary method as well as the baker\u2013wurgler investor sentiment index both in and out of sample. the bert-based sentiment shows strong predictive power during economic downturns and can significantly predict future macroeconomic conditions. overall, our bert model offers a better measure of textual investor sentiment, highlighting the potentially significant value of deep learning, ai, and fintech in financial market.","key_phrases":["this paper","we","the bert model","a cut-edging deep learning model","a novel textual sentiment index","the chinese stock market","the stock market returns","sentiment labels","our bert model","textual sentiment-related information","asset pricing","we","the bert-based sentiment","much greater predictive power","stock market returns","the traditional dictionary method","the baker","wurgler investor sentiment index","sample","the bert-based sentiment","strong predictive power","economic downturns","future macroeconomic conditions","our bert model","a better measure","textual investor sentiment","the potentially significant value","deep learning","financial market","chinese"]},{"title":"Research progress on intelligent monitoring of tool condition based on deep learning","authors":["Dahu Cao","Wei Liu","Jimin Ge","Shishuai Du","Wang Liu","Zhaohui Deng","Jia Chen"],"abstract":"Intelligent monitoring of tool condition is the key to ensuring workshop manufacturing efficiency, product machining quality, and accuracy, and is also an indispensable part of intelligent processing. In the face of complex and massive, multi-source heterogeneous, and low-value density machining process data, the monitoring method based on traditional machine learning is challenging to meet the development needs of intelligent manufacturing. In contrast, with its powerful data processing and automatic feature extraction capabilities, deep learning shows broad application prospects in tool condition intelligent monitoring. Given this, this paper first systematically introduces the components of tool condition intelligent monitoring framework based on deep learning. Subsequently, the basic principles, modeling process, and application status of the four most widely used deep learning models (deep belief network, stacked auto-encoder network, convolutional neural network, and recurrent neural network) in the field of tool condition monitoring are detailed, and the advantages and disadvantages of different models are comparably discussed. Finally, the challenges and prospects of the current tool condition intelligent monitoring based on deep learning are summarized.","doi":"10.1007\/s00170-024-14273-5","cleaned_title":"research progress on intelligent monitoring of tool condition based on deep learning","cleaned_abstract":"intelligent monitoring of tool condition is the key to ensuring workshop manufacturing efficiency, product machining quality, and accuracy, and is also an indispensable part of intelligent processing. in the face of complex and massive, multi-source heterogeneous, and low-value density machining process data, the monitoring method based on traditional machine learning is challenging to meet the development needs of intelligent manufacturing. in contrast, with its powerful data processing and automatic feature extraction capabilities, deep learning shows broad application prospects in tool condition intelligent monitoring. given this, this paper first systematically introduces the components of tool condition intelligent monitoring framework based on deep learning. subsequently, the basic principles, modeling process, and application status of the four most widely used deep learning models (deep belief network, stacked auto-encoder network, convolutional neural network, and recurrent neural network) in the field of tool condition monitoring are detailed, and the advantages and disadvantages of different models are comparably discussed. finally, the challenges and prospects of the current tool condition intelligent monitoring based on deep learning are summarized.","key_phrases":["intelligent monitoring","tool condition","the key","workshop manufacturing efficiency","product machining quality","accuracy","an indispensable part","intelligent processing","the face","low-value density machining process data","the monitoring method","traditional machine learning","the development needs","intelligent manufacturing","contrast","its powerful data processing","automatic feature extraction capabilities","deep learning","broad application prospects","tool condition intelligent monitoring","this","this paper","the components","tool condition intelligent monitoring framework","deep learning","the basic principles","modeling process","application status","the four most widely used deep learning models","deep belief network","stacked auto-encoder network","convolutional neural network","recurrent neural network","the field","tool condition monitoring","the advantages","disadvantages","different models","the challenges","prospects","the current tool condition intelligent monitoring","deep learning","first","four"]},{"title":"Prospective de novo drug design with deep interactome learning","authors":["Kenneth Atz","Leandro Cotos","Clemens Isert","Maria H\u00e5kansson","Dorota Focht","Mattis Hilleke","David F. Nippa","Michael Iff","Jann Ledergerber","Carl C. G. Schiebroek","Valentina Romeo","Jan A. Hiss","Daniel Merk","Petra Schneider","Bernd Kuhn","Uwe Grether","Gisbert Schneider"],"abstract":"De novo drug design aims to generate molecules from scratch that possess specific chemical and pharmacological properties. We present a computational approach utilizing interactome-based deep learning for ligand- and structure-based generation of drug-like molecules. This method capitalizes on the unique strengths of both graph neural networks and chemical language models, offering an alternative to the need for application-specific reinforcement, transfer, or few-shot learning. It enables the \u201czero-shot\" construction of compound libraries tailored to possess specific bioactivity, synthesizability, and structural novelty. In order to proactively evaluate the deep interactome learning framework for protein\u00a0structure-based drug design, potential new ligands targeting the binding site of the human peroxisome proliferator-activated receptor (PPAR) subtype gamma are generated. The top-ranking designs are chemically synthesized and computationally, biophysically, and biochemically characterized. Potent PPAR partial agonists are identified, demonstrating favorable activity and the desired selectivity profiles for both nuclear receptors and off-target interactions. Crystal structure determination of the ligand-receptor complex confirms the anticipated binding mode. This successful outcome positively advocates interactome-based de novo design for application in bioorganic and medicinal chemistry, enabling the creation of innovative bioactive molecules.","doi":"10.1038\/s41467-024-47613-w","cleaned_title":"prospective de novo drug design with deep interactome learning","cleaned_abstract":"de novo drug design aims to generate molecules from scratch that possess specific chemical and pharmacological properties. we present a computational approach utilizing interactome-based deep learning for ligand- and structure-based generation of drug-like molecules. this method capitalizes on the unique strengths of both graph neural networks and chemical language models, offering an alternative to the need for application-specific reinforcement, transfer, or few-shot learning. it enables the \u201czero-shot\" construction of compound libraries tailored to possess specific bioactivity, synthesizability, and structural novelty. in order to proactively evaluate the deep interactome learning framework for protein structure-based drug design, potential new ligands targeting the binding site of the human peroxisome proliferator-activated receptor (ppar) subtype gamma are generated. the top-ranking designs are chemically synthesized and computationally, biophysically, and biochemically characterized. potent ppar partial agonists are identified, demonstrating favorable activity and the desired selectivity profiles for both nuclear receptors and off-target interactions. crystal structure determination of the ligand-receptor complex confirms the anticipated binding mode. this successful outcome positively advocates interactome-based de novo design for application in bioorganic and medicinal chemistry, enabling the creation of innovative bioactive molecules.","key_phrases":["de novo drug design","molecules","scratch","that","specific chemical and pharmacological properties","we","a computational approach","interactome-based deep learning","ligand-","structure-based generation","drug-like molecules","this method capitalizes","the unique strengths","both graph neural networks","chemical language models","an alternative","the need","application-specific reinforcement","transfer","few-shot learning","it","the \u201czero-shot\" construction","compound libraries","specific bioactivity","synthesizability","structural novelty","order","the deep interactome learning framework","protein","structure-based drug design","potential new ligands","the binding site","the human peroxisome proliferator-activated receptor","ppar) subtype gamma","the top-ranking designs","potent ppar partial agonists","favorable activity","the desired selectivity profiles","both nuclear receptors","off-target interactions","crystal structure determination","the ligand-receptor complex","the anticipated binding mode","this successful outcome","interactome-based de novo design","application","bioorganic and medicinal chemistry","the creation","innovative bioactive molecules","zero","de novo"]},{"title":"Handloomed fabrics recognition with deep learning","authors":["Lipi B. Mahanta","Deva Raj Mahanta","Taibur Rahman","Chandan Chakraborty"],"abstract":"Every nation treasures its handloom heritage, and in India, the handloom industry safeguards cultural traditions, sustains millions of artisans, and preserves ancient weaving techniques. To protect this legacy, a critical need arises to distinguish genuine handloom products, exemplified by the renowned \u201cgamucha\u201d from India\u2019s northeast, from counterfeit powerloom imitations. Our study\u2019s objective is to create an AI tool for effortless detection of authentic handloom items amidst a sea of fakes. Six deep learning architectures\u2014VGG16, VGG19, ResNet50, InceptionV3, InceptionResNetV2, and DenseNet201\u2014were trained on annotated image repositories of handloom and powerloom towels (17,484 images in total, with 14,020 for training and 3464 for validation). A novel deep learning model was also proposed. Despite respectable training accuracies, the pre-trained models exhibited lower performance on the validation dataset compared to our novel model. The proposed model outperformed pre-trained models, demonstrating superior validation accuracy, lower validation loss, computational efficiency, and adaptability to the specific classification problem. Notably, the existing models showed challenges in generalizing to unseen data and raised concerns about practical deployment due to computational expenses. This study pioneers a computer-assisted approach for automated differentiation between authentic handwoven \u201cgamucha\u201ds and counterfeit powerloom imitations\u2014a groundbreaking recognition method. The methodology presented not only holds scalability potential and opportunities for accuracy improvement but also suggests broader applications across diverse fabric products.","doi":"10.1038\/s41598-024-58750-z","cleaned_title":"handloomed fabrics recognition with deep learning","cleaned_abstract":"every nation treasures its handloom heritage, and in india, the handloom industry safeguards cultural traditions, sustains millions of artisans, and preserves ancient weaving techniques. to protect this legacy, a critical need arises to distinguish genuine handloom products, exemplified by the renowned \u201cgamucha\u201d from india\u2019s northeast, from counterfeit powerloom imitations. our study\u2019s objective is to create an ai tool for effortless detection of authentic handloom items amidst a sea of fakes. six deep learning architectures\u2014vgg16, vgg19, resnet50, inceptionv3, inceptionresnetv2, and densenet201\u2014were trained on annotated image repositories of handloom and powerloom towels (17,484 images in total, with 14,020 for training and 3464 for validation). a novel deep learning model was also proposed. despite respectable training accuracies, the pre-trained models exhibited lower performance on the validation dataset compared to our novel model. the proposed model outperformed pre-trained models, demonstrating superior validation accuracy, lower validation loss, computational efficiency, and adaptability to the specific classification problem. notably, the existing models showed challenges in generalizing to unseen data and raised concerns about practical deployment due to computational expenses. this study pioneers a computer-assisted approach for automated differentiation between authentic handwoven \u201cgamucha\u201ds and counterfeit powerloom imitations\u2014a groundbreaking recognition method. the methodology presented not only holds scalability potential and opportunities for accuracy improvement but also suggests broader applications across diverse fabric products.","key_phrases":["every nation","its handloom heritage","india","cultural traditions","millions","artisans","ancient weaving techniques","this legacy","a critical need","genuine handloom products","the renowned \u201cgamucha","india\u2019s northeast","counterfeit powerloom imitations","our study\u2019s objective","an ai tool","effortless detection","authentic handloom items","a sea","fakes","six deep learning architectures","vgg16","resnet50","inceptionv3","densenet201","annotated image repositories","handloom","powerloom towels","17,484 images","total","training","validation","a novel deep learning model","respectable training accuracies","the pre-trained models","lower performance","the validation dataset","our novel model","the proposed model","pre-trained models","superior validation accuracy","the specific classification problem","the existing models","challenges","unseen data","concerns","practical deployment","computational expenses","this study","a computer-assisted approach","automated differentiation","authentic handwoven \u201cgamucha\u201ds","counterfeit powerloom imitations","a groundbreaking recognition method","the methodology","scalability potential","opportunities","accuracy improvement","broader applications","diverse fabric products","india","millions","india","six","resnet50","inceptionv3","inceptionresnetv2","17,484","14,020","3464"]},{"title":"Deep fake detection using an optimal deep learning model with multi head attention-based feature extraction scheme","authors":["R. Raja Sekar","T. Dhiliphan Rajkumar","Koteswara Rao Anne"],"abstract":"Face forgery, or deep fake, is a frequently used method to produce fake face images, network pornography, blackmail, and other illegal activities. Researchers developed several detection approaches based on the changing traces presented by deep forgery to limit the damage caused by deep fake methods. They obtain limited performance when evaluating cross-datum scenarios. This paper proposes an optimal deep learning approach with an attention-based feature learning scheme to perform DFD more accurately. The proposed system mainly comprises \u20185\u2019 phases: face detection, preprocessing, texture feature extraction, spatial feature extraction, and classification. The face regions are initially detected from the collected data using the Viola\u2013Jones (VJ) algorithm. Then, preprocessing is carried out, which resizes and normalizes the detected face regions to improve their quality for detection purposes. Next, texture features are learned using the Butterfly Optimized Gabor Filter to get information about the local features of objects in an image. Then, the spatial features are extracted using Residual Network-50 with Multi Head Attention (RN50MHA) to represent the data globally. Finally, classification is done using the Optimal Long Short-Term Memory (OLSTM), which classifies the data as fake or real, in which optimization of network is done using Enhanced Archimedes Optimization Algorithm. The proposed system is evaluated on four benchmark datasets such as Face Forensics +\u2009\u2009+\u2009(FF +\u2009+), Deepfake Detection Challenge, Celebrity Deepfake (CDF), and Wild Deepfake. The experimental results show that DFD using OLSTM and RN50MHA achieves a higher inter and intra-dataset detection rate than existing state-of-the-art methods.","doi":"10.1007\/s00371-024-03567-0","cleaned_title":"deep fake detection using an optimal deep learning model with multi head attention-based feature extraction scheme","cleaned_abstract":"face forgery, or deep fake, is a frequently used method to produce fake face images, network pornography, blackmail, and other illegal activities. researchers developed several detection approaches based on the changing traces presented by deep forgery to limit the damage caused by deep fake methods. they obtain limited performance when evaluating cross-datum scenarios. this paper proposes an optimal deep learning approach with an attention-based feature learning scheme to perform dfd more accurately. the proposed system mainly comprises \u20185\u2019 phases: face detection, preprocessing, texture feature extraction, spatial feature extraction, and classification. the face regions are initially detected from the collected data using the viola\u2013jones (vj) algorithm. then, preprocessing is carried out, which resizes and normalizes the detected face regions to improve their quality for detection purposes. next, texture features are learned using the butterfly optimized gabor filter to get information about the local features of objects in an image. then, the spatial features are extracted using residual network-50 with multi head attention (rn50mha) to represent the data globally. finally, classification is done using the optimal long short-term memory (olstm), which classifies the data as fake or real, in which optimization of network is done using enhanced archimedes optimization algorithm. the proposed system is evaluated on four benchmark datasets such as face forensics + + (ff + +), deepfake detection challenge, celebrity deepfake (cdf), and wild deepfake. the experimental results show that dfd using olstm and rn50mha achieves a higher inter and intra-dataset detection rate than existing state-of-the-art methods.","key_phrases":["face forgery","a frequently used method","fake face images","researchers","several detection approaches","the changing traces","deep forgery","the damage","deep fake methods","they","limited performance","cross-datum scenarios","this paper","an optimal deep learning approach","an attention-based feature learning scheme","the proposed system","5\u2019 phases","face","detection","preprocessing","texture feature extraction","spatial feature extraction","classification","the face regions","the collected data","the viola","jones","vj","algorithm","preprocessing","which","the detected face regions","their quality","detection purposes","texture features","the butterfly optimized gabor filter","information","the local features","objects","an image","the spatial features","residual network-50","multi head attention","rn50mha","the data","classification","the optimal long short-term memory","olstm","which","the data","which optimization","network","enhanced archimedes optimization algorithm","the proposed system","four benchmark datasets","face forensics","+ (ff + +), deepfake detection challenge","celebrity deepfake","cdf","wild deepfake","the experimental results","olstm","rn50mha","a higher inter","intra-dataset detection rate","the-art","5","network-50","four","olstm"]},{"title":"Imaging-based deep learning in kidney diseases: recent progress and future prospects","authors":["Meng Zhang","Zheng Ye","Enyu Yuan","Xinyang Lv","Yiteng Zhang","Yuqi Tan","Chunchao Xia","Jing Tang","Jin Huang","Zhenlin Li"],"abstract":"Kidney diseases result from various causes, which can generally be divided into neoplastic and non-neoplastic diseases. Deep learning based on medical imaging is an established methodology for further data mining and an evolving field of expertise, which provides the possibility for precise management of kidney diseases. Recently, imaging-based deep learning has been widely applied to many clinical scenarios of kidney diseases including organ segmentation, lesion detection, differential diagnosis, surgical planning, and prognosis prediction, which can provide support for disease diagnosis and management. In this review, we will introduce the basic methodology of imaging-based deep learning and its recent clinical applications in neoplastic and non-neoplastic kidney diseases. Additionally, we further discuss its current challenges and future prospects and conclude that achieving data balance, addressing heterogeneity, and managing data size remain challenges for imaging-based deep learning. Meanwhile, the interpretability of algorithms, ethical risks, and barriers of bias assessment are also issues that require consideration in future development. We hope to provide urologists, nephrologists, and radiologists with clear ideas about imaging-based deep learning and reveal its great potential in clinical practice.Critical relevance statement The wide clinical applications of imaging-based deep learning in kidney diseases can help doctors to diagnose, treat, and manage patients with neoplastic or non-neoplastic renal diseases.Key points\u2022 Imaging-based deep learning is widely applied to neoplastic and non-neoplastic renal diseases.\u2022 Imaging-based deep learning improves the accuracy of the delineation, diagnosis, and evaluation of kidney diseases.\u2022 The small dataset, various lesion sizes, and so on are still challenges for deep learning.Graphical Abstract","doi":"10.1186\/s13244-024-01636-5","cleaned_title":"imaging-based deep learning in kidney diseases: recent progress and future prospects","cleaned_abstract":"kidney diseases result from various causes, which can generally be divided into neoplastic and non-neoplastic diseases. deep learning based on medical imaging is an established methodology for further data mining and an evolving field of expertise, which provides the possibility for precise management of kidney diseases. recently, imaging-based deep learning has been widely applied to many clinical scenarios of kidney diseases including organ segmentation, lesion detection, differential diagnosis, surgical planning, and prognosis prediction, which can provide support for disease diagnosis and management. in this review, we will introduce the basic methodology of imaging-based deep learning and its recent clinical applications in neoplastic and non-neoplastic kidney diseases. additionally, we further discuss its current challenges and future prospects and conclude that achieving data balance, addressing heterogeneity, and managing data size remain challenges for imaging-based deep learning. meanwhile, the interpretability of algorithms, ethical risks, and barriers of bias assessment are also issues that require consideration in future development. we hope to provide urologists, nephrologists, and radiologists with clear ideas about imaging-based deep learning and reveal its great potential in clinical practice.critical relevance statement the wide clinical applications of imaging-based deep learning in kidney diseases can help doctors to diagnose, treat, and manage patients with neoplastic or non-neoplastic renal diseases.key points\u2022 imaging-based deep learning is widely applied to neoplastic and non-neoplastic renal diseases.\u2022 imaging-based deep learning improves the accuracy of the delineation, diagnosis, and evaluation of kidney diseases.\u2022 the small dataset, various lesion sizes, and so on are still challenges for deep learning.graphical abstract","key_phrases":["kidney diseases","various causes","which","neoplastic and non-neoplastic diseases","deep learning","medical imaging","an established methodology","further data mining","an evolving field","expertise","which","the possibility","precise management","kidney diseases","imaging-based deep learning","many clinical scenarios","kidney diseases","organ segmentation","lesion detection","differential diagnosis","surgical planning","prognosis prediction","which","support","disease diagnosis","management","this review","we","the basic methodology","imaging-based deep learning","its recent clinical applications","neoplastic and non-neoplastic kidney diseases","we","its current challenges","future prospects","data balance","heterogeneity","managing data size","challenges","imaging-based deep learning","the interpretability","algorithms","ethical risks","barriers","bias assessment","issues","that","consideration","future development","we","urologists","nephrologists","radiologists","clear ideas","imaging-based deep learning","its great potential","clinical practice.critical relevance statement","the wide clinical applications","imaging-based deep learning","kidney diseases","doctors","patients","neoplastic or non-neoplastic renal diseases.key","imaging-based deep learning","neoplastic and non-neoplastic renal diseases.\u2022 imaging-based deep learning","the accuracy","the delineation","diagnosis","evaluation","kidney","the small dataset",", various lesion sizes","challenges","deep learning.graphical abstract"]},{"title":"Genotype imputation methods for whole and complex genomic regions utilizing deep learning technology","authors":["Tatsuhiko Naito","Yukinori Okada"],"abstract":"The imputation of unmeasured genotypes is essential in human genetic research, particularly in enhancing the power of genome-wide association studies and conducting subsequent fine-mapping. Recently, several deep learning-based genotype imputation methods for genome-wide variants with the capability of learning complex linkage disequilibrium patterns have been developed. Additionally, deep learning-based imputation has been applied to a distinct genomic region known as the major histocompatibility complex, referred to as HLA imputation. Despite their various advantages, the current deep learning-based genotype imputation methods do have certain limitations and have not yet become standard. These limitations include the modest accuracy improvement over statistical and conventional machine learning-based methods. However, their benefits include other aspects, such as their \u201creference-free\u201d nature, which ensures complete privacy protection, and their higher computational efficiency. Furthermore, the continuing evolution of deep learning technologies is expected to contribute to further improvements in prediction accuracy and usability in the future.","doi":"10.1038\/s10038-023-01213-6","cleaned_title":"genotype imputation methods for whole and complex genomic regions utilizing deep learning technology","cleaned_abstract":"the imputation of unmeasured genotypes is essential in human genetic research, particularly in enhancing the power of genome-wide association studies and conducting subsequent fine-mapping. recently, several deep learning-based genotype imputation methods for genome-wide variants with the capability of learning complex linkage disequilibrium patterns have been developed. additionally, deep learning-based imputation has been applied to a distinct genomic region known as the major histocompatibility complex, referred to as hla imputation. despite their various advantages, the current deep learning-based genotype imputation methods do have certain limitations and have not yet become standard. these limitations include the modest accuracy improvement over statistical and conventional machine learning-based methods. however, their benefits include other aspects, such as their \u201creference-free\u201d nature, which ensures complete privacy protection, and their higher computational efficiency. furthermore, the continuing evolution of deep learning technologies is expected to contribute to further improvements in prediction accuracy and usability in the future.","key_phrases":["the imputation","unmeasured genotypes","human genetic research","the power","genome-wide association studies","subsequent fine-mapping","several deep learning-based genotype imputation methods","genome-wide variants","the capability","complex linkage disequilibrium patterns","deep learning-based imputation","a distinct genomic region","the major histocompatibility complex","hla imputation","their various advantages","the current deep learning-based genotype imputation methods","certain limitations","these limitations","the modest accuracy improvement","statistical and conventional machine learning-based methods","their benefits","other aspects","their \u201creference-free\u201d nature","which","complete privacy protection","their higher computational efficiency","the continuing evolution","deep learning technologies","further improvements","prediction accuracy","usability","the future"]},{"title":"Deep learning-assisted medical image compression challenges and opportunities: systematic review","authors":["Nour El Houda Bourai","Hayet Farida Merouani","Akila Djebbar"],"abstract":"Over the preceding decade, there has been a discernible surge in the prominence of artificial intelligence, marked by the development of various methodologies, among which deep learning emerges as a particularly auspicious technique. The captivating attribute of deep learning, characterised by its capacity to glean intricate feature representations from data, has served as a catalyst for pioneering approaches and methodologies spanning a multitude of domains. In the face of the burgeoning exponential growth in digital medical image data, the exigency for adept image compression methodologies has become increasingly pronounced. These methodologies are designed to preserve bandwidth and storage resources, thereby ensuring the seamless and efficient transmission of data within medical applications. The critical nature of medical image compression accentuates the imperative to confront the challenges precipitated by the escalating deluge of medical image data. This review paper undertakes a comprehensive examination of medical image compression, with a predominant focus on sophisticated, research-driven deep learning techniques. It delves into a spectrum of approaches, encompassing the amalgamation of deep learning with conventional compression algorithms and the application of deep learning to enhance compression quality. Additionally, the review endeavours to explicate these fundamental concepts, elucidating their inherent characteristics, merits, and limitations.","doi":"10.1007\/s00521-024-09660-8","cleaned_title":"deep learning-assisted medical image compression challenges and opportunities: systematic review","cleaned_abstract":"over the preceding decade, there has been a discernible surge in the prominence of artificial intelligence, marked by the development of various methodologies, among which deep learning emerges as a particularly auspicious technique. the captivating attribute of deep learning, characterised by its capacity to glean intricate feature representations from data, has served as a catalyst for pioneering approaches and methodologies spanning a multitude of domains. in the face of the burgeoning exponential growth in digital medical image data, the exigency for adept image compression methodologies has become increasingly pronounced. these methodologies are designed to preserve bandwidth and storage resources, thereby ensuring the seamless and efficient transmission of data within medical applications. the critical nature of medical image compression accentuates the imperative to confront the challenges precipitated by the escalating deluge of medical image data. this review paper undertakes a comprehensive examination of medical image compression, with a predominant focus on sophisticated, research-driven deep learning techniques. it delves into a spectrum of approaches, encompassing the amalgamation of deep learning with conventional compression algorithms and the application of deep learning to enhance compression quality. additionally, the review endeavours to explicate these fundamental concepts, elucidating their inherent characteristics, merits, and limitations.","key_phrases":["the preceding decade","a discernible surge","the prominence","artificial intelligence","the development","various methodologies","which","deep learning","a particularly auspicious technique","the captivating attribute","deep learning","its capacity","intricate feature representations","data","a catalyst","approaches","methodologies","a multitude","domains","the face","the burgeoning exponential growth","digital medical image data","the exigency","adept image compression methodologies","these methodologies","bandwidth and storage resources","the seamless and efficient transmission","data","medical applications","the critical nature","medical image compression","the imperative","the challenges","the escalating deluge","medical image data","this review paper","a comprehensive examination","medical image compression","a predominant focus","sophisticated, research-driven deep learning techniques","it","a spectrum","approaches","the amalgamation","deep learning","conventional compression algorithms","the application","deep learning","compression quality","the review","these fundamental concepts","their inherent characteristics","merits","limitations","the preceding decade"]},{"title":"Open and reusable deep learning for pathology with WSInfer and QuPath","authors":["Jakub R. Kaczmarzyk","Alan O\u2019Callaghan","Fiona Inglis","Swarad Gat","Tahsin Kurc","Rajarsi Gupta","Erich Bremer","Peter Bankhead","Joel H. Saltz"],"abstract":"Digital pathology has seen a proliferation of deep learning models in recent years, but many models are not readily reusable. To address this challenge, we developed WSInfer: an open-source software ecosystem designed to streamline the sharing and reuse of deep learning models for digital pathology. The increased access to trained models can augment research on the diagnostic, prognostic, and predictive capabilities of digital pathology.","doi":"10.1038\/s41698-024-00499-9","cleaned_title":"open and reusable deep learning for pathology with wsinfer and qupath","cleaned_abstract":"digital pathology has seen a proliferation of deep learning models in recent years, but many models are not readily reusable. to address this challenge, we developed wsinfer: an open-source software ecosystem designed to streamline the sharing and reuse of deep learning models for digital pathology. the increased access to trained models can augment research on the diagnostic, prognostic, and predictive capabilities of digital pathology.","key_phrases":["digital pathology","a proliferation","deep learning models","recent years","many models","this challenge","we","an open-source software ecosystem","the sharing","reuse","deep learning models","digital pathology","the increased access","trained models","research","predictive capabilities","digital pathology","digital pathology","recent years"]},{"title":"Adversarial robustness of deep reinforcement learning-based intrusion detection","authors":["Mohamed Amine Merzouk","Christopher Neal","Jos\u00e9phine Delas","Reda Yaich","Nora Boulahia-Cuppens","Fr\u00e9d\u00e9ric Cuppens"],"abstract":"Machine learning techniques, including Deep Reinforcement Learning (DRL), enhance intrusion detection systems by adapting to new threats. However, DRL\u2019s reliance on vulnerable deep neural networks leads to susceptibility to adversarial examples-perturbations designed to evade detection. While adversarial examples are well-studied in deep learning, their impact on DRL-based intrusion detection remains underexplored, particularly in critical domains. This article conducts a thorough analysis of DRL-based intrusion detection\u2019s vulnerability to adversarial examples. It systematically evaluates key hyperparameters such as DRL algorithms, neural network depth, and width, impacting agents\u2019 robustness. The study extends to black-box attacks, demonstrating adversarial transferability across DRL algorithms. Findings emphasize neural network architecture\u2019s critical role in DRL agent robustness, addressing underfitting and overfitting challenges. Practical implications include insights for optimizing DRL-based intrusion detection agents to enhance performance and resilience. Experiments encompass multiple DRL algorithms tested on three datasets: NSL-KDD, UNSW-NB15, and CICIoV2024, against gradient-based adversarial attacks, with publicly available implementation code.","doi":"10.1007\/s10207-024-00903-2","cleaned_title":"adversarial robustness of deep reinforcement learning-based intrusion detection","cleaned_abstract":"machine learning techniques, including deep reinforcement learning (drl), enhance intrusion detection systems by adapting to new threats. however, drl\u2019s reliance on vulnerable deep neural networks leads to susceptibility to adversarial examples-perturbations designed to evade detection. while adversarial examples are well-studied in deep learning, their impact on drl-based intrusion detection remains underexplored, particularly in critical domains. this article conducts a thorough analysis of drl-based intrusion detection\u2019s vulnerability to adversarial examples. it systematically evaluates key hyperparameters such as drl algorithms, neural network depth, and width, impacting agents\u2019 robustness. the study extends to black-box attacks, demonstrating adversarial transferability across drl algorithms. findings emphasize neural network architecture\u2019s critical role in drl agent robustness, addressing underfitting and overfitting challenges. practical implications include insights for optimizing drl-based intrusion detection agents to enhance performance and resilience. experiments encompass multiple drl algorithms tested on three datasets: nsl-kdd, unsw-nb15, and ciciov2024, against gradient-based adversarial attacks, with publicly available implementation code.","key_phrases":["machine learning techniques","deep reinforcement learning","drl","intrusion detection systems","new threats","drl\u2019s reliance","vulnerable deep neural networks","susceptibility","adversarial examples-perturbations","detection","adversarial examples","deep learning","their impact","drl-based intrusion detection","critical domains","this article","a thorough analysis","drl-based intrusion detection\u2019s vulnerability","adversarial examples","it","key hyperparameters","drl algorithms","neural network depth","width","impacting agents\u2019 robustness","the study","black-box attacks","adversarial transferability","drl algorithms","findings","neural network architecture\u2019s critical role","drl agent robustness","underfitting and overfitting challenges","practical implications","insights","drl-based intrusion detection agents","performance","resilience","experiments","multiple drl algorithms","three datasets","nsl-kdd, unsw-nb15","ciciov2024","gradient-based adversarial attacks","publicly available implementation code","drl\u2019s","three","ciciov2024"]},{"title":"Yoga with Deep Learning: Linking Mind and Machine","authors":["Sakshi","Sandeep Saini"],"abstract":"Health and fitness play a crucial role in every aspect of an individual\u2019s life. In an era where well-being is an absolute target, Yoga is one of the most straightforward ways to remain healthy and achieve mental focus, breath control and body balance without undergoing useless expenditures. The motive of this study is to combine the knowledge of Yoga with the continuously developing power of deep learning. In this study, we observe yoga pose estimation along with breath control and focus instructions for 10 different poses with the help of deep learning and Mediapipe. The proposed work predicts pose and instructions through model matching with landmark extraction done by obtaining keypoints on the pose by initializing the Mediapipe pose model. Here, a hybrid model is created using ANN with existing state-of-the-art models to achieve the pose estimation. The maximum accuracy achieved during the model implementation was around 93.43%. Later, when the model was deployed on Raspberry Pi-4 hardware, the maximum accuracy achieved was around 90.9%. The user also gets guidance on breath and focus control during a particular pose. Integration with deep learning and data science is seen to have created a whole new personalized Yoga experience with on-demand posture monitoring, customized feedback, and global access while creating future scopes for risk analysis and injury prevention.","doi":"10.1007\/s42979-024-02784-7","cleaned_title":"yoga with deep learning: linking mind and machine","cleaned_abstract":"health and fitness play a crucial role in every aspect of an individual\u2019s life. in an era where well-being is an absolute target, yoga is one of the most straightforward ways to remain healthy and achieve mental focus, breath control and body balance without undergoing useless expenditures. the motive of this study is to combine the knowledge of yoga with the continuously developing power of deep learning. in this study, we observe yoga pose estimation along with breath control and focus instructions for 10 different poses with the help of deep learning and mediapipe. the proposed work predicts pose and instructions through model matching with landmark extraction done by obtaining keypoints on the pose by initializing the mediapipe pose model. here, a hybrid model is created using ann with existing state-of-the-art models to achieve the pose estimation. the maximum accuracy achieved during the model implementation was around 93.43%. later, when the model was deployed on raspberry pi-4 hardware, the maximum accuracy achieved was around 90.9%. the user also gets guidance on breath and focus control during a particular pose. integration with deep learning and data science is seen to have created a whole new personalized yoga experience with on-demand posture monitoring, customized feedback, and global access while creating future scopes for risk analysis and injury prevention.","key_phrases":["health","fitness","a crucial role","every aspect","an individual\u2019s life","an era","well-being","an absolute target","yoga","the most straightforward ways","mental focus","breath control","body balance","useless expenditures","the motive","this study","the knowledge","yoga","the continuously developing power","deep learning","this study","we","yoga pose estimation","breath control","instructions","10 different poses","the help","deep learning","mediapipe","the proposed work predicts","instructions","model","landmark extraction","keypoints","the pose","the mediapipe pose model","a hybrid model","ann","the-art","the pose estimation","the maximum accuracy","the model implementation","93.43%","the model","raspberry pi-4 hardware","the maximum accuracy","around 90.9%","the user","guidance","breath","control","a particular pose","integration","deep learning and data science","a whole new personalized yoga experience","demand","customized feedback","global access","future scopes","risk analysis","injury prevention","one","10","around 93.43%","around 90.9%"]},{"title":"A review of deep learning and machine learning techniques for hydrological inflow forecasting","authors":["Sarmad Dashti Latif","Ali Najah Ahmed"],"abstract":"Conventional machine learning models have been widely used for reservoir inflow and rainfall prediction. Nowadays, researchers focus on a new computing architecture in the area of AI, namely, deep learning for hydrological forecasting parameters. This review paper tends to broadcast more of the intriguing interest in reservoir inflow prediction utilizing deep learning and machine learning algorithms. The AI models utilized for different hydrology sectors, as well as the most prevalent machine learning techniques, will be explored in this thorough study, which divides AI techniques into two primary categories: deep learning and machine learning. In this study, we look at the long short-term memory deep learning method as well as three traditional machine learning algorithms: support vector machine, random forest, and boosted regression tree. Under each part, a summary of the findings is provided. For convenience of reference, some of the benefits and drawbacks discovered through literature reviews have been listed. Finally, future recommendations and overall conclusions based on research findings are given. This review focuses on papers from high-impact factor periodicals published over a 4\u00a0years period beginning in 2018 onwards.","doi":"10.1007\/s10668-023-03131-1","cleaned_title":"a review of deep learning and machine learning techniques for hydrological inflow forecasting","cleaned_abstract":"conventional machine learning models have been widely used for reservoir inflow and rainfall prediction. nowadays, researchers focus on a new computing architecture in the area of ai, namely, deep learning for hydrological forecasting parameters. this review paper tends to broadcast more of the intriguing interest in reservoir inflow prediction utilizing deep learning and machine learning algorithms. the ai models utilized for different hydrology sectors, as well as the most prevalent machine learning techniques, will be explored in this thorough study, which divides ai techniques into two primary categories: deep learning and machine learning. in this study, we look at the long short-term memory deep learning method as well as three traditional machine learning algorithms: support vector machine, random forest, and boosted regression tree. under each part, a summary of the findings is provided. for convenience of reference, some of the benefits and drawbacks discovered through literature reviews have been listed. finally, future recommendations and overall conclusions based on research findings are given. this review focuses on papers from high-impact factor periodicals published over a 4 years period beginning in 2018 onwards.","key_phrases":["conventional machine learning models","reservoir inflow","rainfall prediction","researchers","a new computing architecture","the area","hydrological forecasting parameters","this review paper","the intriguing interest","reservoir inflow prediction","deep learning and machine learning algorithms","the ai models","different hydrology sectors","the most prevalent machine learning techniques","this thorough study","which","techniques","two primary categories","deep learning","machine learning","this study","we","the long short-term memory deep learning method","three traditional machine learning algorithms","vector machine","random forest","regression tree","each part","a summary","the findings","convenience","reference","some","the benefits","drawbacks","literature reviews","future recommendations","overall conclusions","research findings","this review","papers","high-impact factor periodicals","a 4 years period","two","three","4 years","2018"]},{"title":"Deep Metric Learning: Loss Functions Comparison","authors":["R. L. Vasilev","A. G. D\u2019yakonov"],"abstract":"AbstractAn overview of deep metric learning methods is presented. Although they have appeared in recent years, these methods were compared only with their predecessors, with neural networks of outdated architectures used for representation learning (representations on which the metric is calculated). The described methods were compared on different datasets from several domains, using pre-trained neural networks comparable in performance to SotA (state of the art): ConvNeXt for images and DistilBERT for texts. Labeled datasets were used, divided into two parts (train and test) so that the classes did not overlap (i.e., for each class its objects are fully in train or fully in test). Such a large-scale honest comparison was made for the first time and led to unexpected conclusions, viz. some \u201cold\u201d methods, for example, Tuplet Margin Loss, are superior in performance to their modern modifications and methods proposed in very recent works.","doi":"10.1134\/S1064562423701053","cleaned_title":"deep metric learning: loss functions comparison","cleaned_abstract":"abstractan overview of deep metric learning methods is presented. although they have appeared in recent years, these methods were compared only with their predecessors, with neural networks of outdated architectures used for representation learning (representations on which the metric is calculated). the described methods were compared on different datasets from several domains, using pre-trained neural networks comparable in performance to sota (state of the art): convnext for images and distilbert for texts. labeled datasets were used, divided into two parts (train and test) so that the classes did not overlap (i.e., for each class its objects are fully in train or fully in test). such a large-scale honest comparison was made for the first time and led to unexpected conclusions, viz. some \u201cold\u201d methods, for example, tuplet margin loss, are superior in performance to their modern modifications and methods proposed in very recent works.","key_phrases":["abstractan overview","deep metric learning methods","they","recent years","these methods","their predecessors","neural networks","outdated architectures","representation learning","representations","which","the metric","the described methods","different datasets","several domains","pre-trained neural networks","performance","sota","(state","the art","convnext","images","texts","labeled datasets","two parts","train","test","the classes","each class","its objects","train","test","such a large-scale honest comparison","the first time","unexpected conclusions","some \u201cold\u201d methods","example","margin loss","performance","their modern modifications","methods","very recent works","abstractan","recent years","two","first"]},{"title":"Comparing Machine Learning and Deep Learning Techniques for Text Analytics: Detecting the Severity of Hate Comments Online","authors":["Alaa Marshan","Farah Nasreen Mohamed Nizar","Athina Ioannou","Konstantina Spanaki"],"abstract":"Social media platforms have become an increasingly popular tool for individuals to share their thoughts and opinions with other people. However, very often people tend to misuse social media posting abusive comments. Abusive and harassing behaviours can have adverse effects on people's lives. This study takes a novel approach to combat harassment in online platforms by detecting the severity of abusive comments, that has not been investigated before. The study compares the performance of machine learning models such as Na\u00efve Bayes, Random Forest, and Support Vector Machine, with deep learning models such as Convolutional Neural Network (CNN) and Bi-directional Long Short-Term Memory (Bi-LSTM). Moreover, in this work we investigate the effect of text pre-processing on the performance of the machine and deep learning models, the feature set for the abusive comments was made using unigrams and bigrams for the machine learning models and word embeddings for the deep learning models. The comparison of the models\u2019 performances showed that the Random Forest with bigrams achieved the best overall performance with an accuracy of (0.94), a precision of (0.91), a recall of (0.94), and an F1 score of (0.92). The study develops an efficient model to detect severity of abusive language in online platforms, offering important implications both to theory and practice.","doi":"10.1007\/s10796-023-10446-x","cleaned_title":"comparing machine learning and deep learning techniques for text analytics: detecting the severity of hate comments online","cleaned_abstract":"social media platforms have become an increasingly popular tool for individuals to share their thoughts and opinions with other people. however, very often people tend to misuse social media posting abusive comments. abusive and harassing behaviours can have adverse effects on people's lives. this study takes a novel approach to combat harassment in online platforms by detecting the severity of abusive comments, that has not been investigated before. the study compares the performance of machine learning models such as na\u00efve bayes, random forest, and support vector machine, with deep learning models such as convolutional neural network (cnn) and bi-directional long short-term memory (bi-lstm). moreover, in this work we investigate the effect of text pre-processing on the performance of the machine and deep learning models, the feature set for the abusive comments was made using unigrams and bigrams for the machine learning models and word embeddings for the deep learning models. the comparison of the models\u2019 performances showed that the random forest with bigrams achieved the best overall performance with an accuracy of (0.94), a precision of (0.91), a recall of (0.94), and an f1 score of (0.92). the study develops an efficient model to detect severity of abusive language in online platforms, offering important implications both to theory and practice.","key_phrases":["social media platforms","an increasingly popular tool","individuals","their thoughts","opinions","other people","people","social media","abusive comments","abusive and harassing behaviours","adverse effects","people's lives","this study","a novel approach","combat harassment","online platforms","the severity","abusive comments","that","the study","the performance","machine learning models","na\u00efve bayes","random forest","vector machine","deep learning models","convolutional neural network","cnn","bi-directional long short-term memory","bi","-","lstm","this work","we","the effect","text pre","the performance","the machine","deep learning models","the feature","the abusive comments","unigrams","bigrams","the machine learning models","word embeddings","the deep learning models","the comparison","the models\u2019 performances","the random forest","bigrams","the best overall performance","an accuracy","the study","an efficient model","severity","abusive language","online platforms","important implications","theory","practice","cnn","0.94","0.91","0.94","0.92"]},{"title":"Interpretable predictions of chaotic dynamical systems using dynamical system deep learning","authors":["Mingyu Wang","Jianping Li"],"abstract":"Making accurate predictions of chaotic dynamical systems is an essential but challenging task with many practical applications in various disciplines. However, the current dynamical methods can only provide short-term precise predictions, while prevailing deep learning techniques with better performances always suffer from model complexity and interpretability. Here, we propose a new dynamic-based deep learning method, namely the dynamical system deep learning (DSDL), to achieve interpretable long-term precise predictions by the combination of nonlinear dynamics theory and deep learning methods. As validated by four chaotic dynamical systems with different complexities, the DSDL framework significantly outperforms other dynamical and deep learning methods. Furthermore, the DSDL also reduces the model complexity and realizes the model transparency to make it more interpretable. We firmly believe that the DSDL framework is a promising and effective method for comprehending and predicting chaotic dynamical systems.","doi":"10.1038\/s41598-024-53169-y","cleaned_title":"interpretable predictions of chaotic dynamical systems using dynamical system deep learning","cleaned_abstract":"making accurate predictions of chaotic dynamical systems is an essential but challenging task with many practical applications in various disciplines. however, the current dynamical methods can only provide short-term precise predictions, while prevailing deep learning techniques with better performances always suffer from model complexity and interpretability. here, we propose a new dynamic-based deep learning method, namely the dynamical system deep learning (dsdl), to achieve interpretable long-term precise predictions by the combination of nonlinear dynamics theory and deep learning methods. as validated by four chaotic dynamical systems with different complexities, the dsdl framework significantly outperforms other dynamical and deep learning methods. furthermore, the dsdl also reduces the model complexity and realizes the model transparency to make it more interpretable. we firmly believe that the dsdl framework is a promising and effective method for comprehending and predicting chaotic dynamical systems.","key_phrases":["accurate predictions","chaotic dynamical systems","an essential but challenging task","many practical applications","various disciplines","the current dynamical methods","short-term precise predictions","deep learning techniques","better performances","model complexity","interpretability","we","a new dynamic-based deep learning method","namely the dynamical system","deep learning","dsdl","interpretable long-term precise predictions","the combination","nonlinear dynamics theory","deep learning methods","four chaotic dynamical systems","different complexities","the dsdl framework","other dynamical and deep learning methods","the dsdl","the model complexity","the model transparency","it","we","the dsdl framework","a promising and effective method","chaotic dynamical systems","four"]},{"title":"A comprehensive review of COVID-19 detection with machine learning and deep learning techniques","authors":["Sreeparna Das","Ishan Ayus","Deepak Gupta"],"abstract":"PurposeThe first transmission of coronavirus to humans started in Wuhan city of China, took the shape of a pandemic called Corona Virus Disease 2019 (COVID-19), and posed a principal threat to the entire world. The researchers are trying to inculcate artificial intelligence (Machine learning or deep learning models) for the efficient detection of COVID-19. This research explores all the existing machine learning (ML) or deep learning (DL) models, used for COVID-19 detection which may help the researcher to explore in different directions. The main purpose of this review article is to present a compact overview of the application of artificial intelligence to the research experts, helping them to explore the future scopes of improvement.MethodsThe researchers have used various machine learning, deep learning, and a combination of machine and deep learning models for extracting significant features and classifying various health conditions in COVID-19 patients. For this purpose, the researchers have utilized different image modalities such as CT-Scan, X-Ray, etc. This study has collected over 200 research papers from various repositories like Google Scholar, PubMed, Web of Science, etc. These research papers were passed through various levels of scrutiny and finally, 50 research articles were selected.ResultsIn those listed articles, the ML \/ DL models showed an accuracy of 99% and above while performing the classification of COVID-19. This study has also presented various clinical applications of various research. This study specifies the importance of various machine and deep learning models in the field of medical diagnosis and research.ConclusionIn conclusion, it is evident that ML\/DL models have made significant progress in recent years, but there are still limitations that need to be addressed. Overfitting is one such limitation that can lead to incorrect predictions and overburdening of the models. The research community must continue to work towards finding ways to overcome these limitations and make machine and deep learning models even more effective and efficient. Through this ongoing research and development, we can expect even greater advances in the future.","doi":"10.1007\/s12553-023-00757-z","cleaned_title":"a comprehensive review of covid-19 detection with machine learning and deep learning techniques","cleaned_abstract":"purposethe first transmission of coronavirus to humans started in wuhan city of china, took the shape of a pandemic called corona virus disease 2019 (covid-19), and posed a principal threat to the entire world. the researchers are trying to inculcate artificial intelligence (machine learning or deep learning models) for the efficient detection of covid-19. this research explores all the existing machine learning (ml) or deep learning (dl) models, used for covid-19 detection which may help the researcher to explore in different directions. the main purpose of this review article is to present a compact overview of the application of artificial intelligence to the research experts, helping them to explore the future scopes of improvement.methodsthe researchers have used various machine learning, deep learning, and a combination of machine and deep learning models for extracting significant features and classifying various health conditions in covid-19 patients. for this purpose, the researchers have utilized different image modalities such as ct-scan, x-ray, etc. this study has collected over 200 research papers from various repositories like google scholar, pubmed, web of science, etc. these research papers were passed through various levels of scrutiny and finally, 50 research articles were selected.resultsin those listed articles, the ml \/ dl models showed an accuracy of 99% and above while performing the classification of covid-19. this study has also presented various clinical applications of various research. this study specifies the importance of various machine and deep learning models in the field of medical diagnosis and research.conclusionin conclusion, it is evident that ml\/dl models have made significant progress in recent years, but there are still limitations that need to be addressed. overfitting is one such limitation that can lead to incorrect predictions and overburdening of the models. the research community must continue to work towards finding ways to overcome these limitations and make machine and deep learning models even more effective and efficient. through this ongoing research and development, we can expect even greater advances in the future.","key_phrases":["purposethe first transmission","coronavirus","humans","wuhan city","china","the shape","a pandemic called corona virus disease","covid-19","a principal threat","the entire world","the researchers","artificial intelligence","machine learning","deep learning models","the efficient detection","covid-19","this research","all the existing machine learning","ml","deep learning","(dl) models","covid-19 detection","which","the researcher","different directions","the main purpose","this review article","a compact overview","the application","artificial intelligence","the research experts","them","the future scopes","improvement.methodsthe researchers","various machine learning","deep learning","a combination","machine","deep learning models","significant features","various health conditions","covid-19 patients","this purpose","the researchers","different image modalities","ct-scan","x","-","ray","this study","over 200 research papers","various repositories","google scholar","pubmed","web","science","these research papers","various levels","scrutiny","50 research articles","those listed articles","the ml \/ dl models","an accuracy","99%","the classification","covid-19","this study","various clinical applications","various research","this study","the importance","various machine","deep learning models","the field","medical diagnosis","research.conclusionin conclusion","it","ml\/dl models","significant progress","recent years","limitations","that","one such limitation","that","incorrect predictions","the models","the research community","ways","these limitations","machine","deep learning models","this ongoing research","development","we","even greater advances","the future","first","wuhan city","china","2019","covid-19","covid-19","covid-19","covid-19","200","50","99%","covid-19","recent years","one"]},{"title":"Continual learning, deep reinforcement learning, and microcircuits: a novel method for clever game playing","authors":["Oscar Chang","Leo Ramos","Manuel Eugenio Morocho-Cayamcela","Rolando Armas","Luis Zhinin-Vera"],"abstract":"Contemporary neural networks frequently encounter the challenge of catastrophic forgetting, wherein newly acquired learning can overwrite and erase previously learned information. The paradigm of continual learning offers a promising solution by enabling intelligent systems to retain and build upon their acquired knowledge over time. This paper introduces a novel approach within the continual learning framework, employing deep reinforcement learning agents that process unprocessed pixel data and interact with microcircuit-like components. These agents autonomously advance through a series of learning stages, culminating in the development of a sophisticated neural network system optimized for predictive performance in the game of tic-tac-toe. Structured to operate in sequential order, each agent is tasked with achieving forward-looking objectives based on Bellman\u2019s principles of reinforcement learning. Knowledge retention is facilitated through the integration of specific microcircuits, which securely store the insights gained by each agent. During the training phase, these microcircuits work in concert, employing high-energy, sparse encoding techniques to enhance learning efficiency and effectiveness. The core contribution of this paper is the establishment of an artificial neural network system capable of accurately predicting tic-tac-toe moves, akin to the observational strategies employed by humans. Our experimental results demonstrate that after approximately 5000 cycles of backpropagation, the system significantly reduced the training loss to \\(L_{DQN}<0.1\\), thereby increasing the expected cumulative reward. This advancement in training efficiency translates into superior predictive capabilities, enabling the system to secure consistent victories by anticipating up to four moves ahead.","doi":"10.1007\/s11042-024-18925-2","cleaned_title":"continual learning, deep reinforcement learning, and microcircuits: a novel method for clever game playing","cleaned_abstract":"contemporary neural networks frequently encounter the challenge of catastrophic forgetting, wherein newly acquired learning can overwrite and erase previously learned information. the paradigm of continual learning offers a promising solution by enabling intelligent systems to retain and build upon their acquired knowledge over time. this paper introduces a novel approach within the continual learning framework, employing deep reinforcement learning agents that process unprocessed pixel data and interact with microcircuit-like components. these agents autonomously advance through a series of learning stages, culminating in the development of a sophisticated neural network system optimized for predictive performance in the game of tic-tac-toe. structured to operate in sequential order, each agent is tasked with achieving forward-looking objectives based on bellman\u2019s principles of reinforcement learning. knowledge retention is facilitated through the integration of specific microcircuits, which securely store the insights gained by each agent. during the training phase, these microcircuits work in concert, employing high-energy, sparse encoding techniques to enhance learning efficiency and effectiveness. the core contribution of this paper is the establishment of an artificial neural network system capable of accurately predicting tic-tac-toe moves, akin to the observational strategies employed by humans. our experimental results demonstrate that after approximately 5000 cycles of backpropagation, the system significantly reduced the training loss to \\(l_{dqn}<0.1\\), thereby increasing the expected cumulative reward. this advancement in training efficiency translates into superior predictive capabilities, enabling the system to secure consistent victories by anticipating up to four moves ahead.","key_phrases":["contemporary neural networks","the challenge","catastrophic forgetting","newly acquired learning","erase","information","the paradigm","continual learning","a promising solution","intelligent systems","their acquired knowledge","time","this paper","a novel approach","the continual learning framework","deep reinforcement learning agents","that","unprocessed pixel data","interact","microcircuit-like components","these agents","a series","learning stages","the development","a sophisticated neural network system","predictive performance","the game","tic-tac-toe","sequential order","each agent","forward-looking objectives","bellman\u2019s principles","reinforcement learning","knowledge retention","the integration","specific microcircuits","which","the insights","each agent","the training phase","these microcircuits","concert","high-energy","techniques","efficiency","effectiveness","the core contribution","this paper","the establishment","an artificial neural network system","tic-tac-toe moves","the observational strategies","humans","our experimental results","approximately 5000 cycles","backpropagation","the system","the training loss","\\(l_{dqn}<0.1\\","the expected cumulative reward","this advancement","training efficiency","superior predictive capabilities","the system","consistent victories","up to four moves","contemporary neural networks","approximately 5000","\\(l_{dqn}<0.1\\","up to four"]},{"title":"Application of deep learning for characterizing microstructures in SBS modified asphalt","authors":["Enhao Zhang","Liyan Shan","Yapeng Guo","Shuang Liu"],"abstract":"Microstructures in asphalt, often resembling bee structures, are pivotal in influencing asphalt performance and, by extension, sustainable fuel production. This study employs deep learning techniques to investigate the impact of different Styrene\u2013Butadiene\u2013Styrene (SBS) modifiers on asphalt microstructures, akin to bee structures. The employed deep learning model was trained on a diverse dataset comprising 200 of images sourced from testing. The dataset was carefully curated to address specific challenges in data labeling precision. This involved individualized labeling sessions and adjustments in the number of targets per image, contributing to enhanced precision and increased dataset size. The research begins with the development of a deep learning model trained on a dataset comprising images featuring bee-like structures within asphalt. The model excels in accurately identifying and segmenting these structures. Subsequently, the deep learning approach is compared with existing methods for bee structure segmentation to establish its precision and superiority. Employing frequency distribution histograms, the distribution patterns of bee structures within various types of SBS-modified asphalt is analyzed, quantitatively assessing the influence of diverse modifier types on these microstructural attributes. The findings in this study underscore the deep learning model's efficacy in recognizing and segmenting bee structures with introduced metrics effectively capturing the distinctive characteristics of various asphalt microstructures. This study paves the way for comprehensive analyses of microstructural metrics, including parameters such as perimeter, area, quantity, and related indicators, thus contributing to the development of fundamental asphalt structural units suitable for processes like molecular simulation and finite element analysis. Moreover, it propels the application of deep learning methodologies in the realm of road materials research, opening new avenues for innovative explorations that can ultimately benefit sustainable fuel production.","doi":"10.1617\/s11527-024-02341-x","cleaned_title":"application of deep learning for characterizing microstructures in sbs modified asphalt","cleaned_abstract":"microstructures in asphalt, often resembling bee structures, are pivotal in influencing asphalt performance and, by extension, sustainable fuel production. this study employs deep learning techniques to investigate the impact of different styrene\u2013butadiene\u2013styrene (sbs) modifiers on asphalt microstructures, akin to bee structures. the employed deep learning model was trained on a diverse dataset comprising 200 of images sourced from testing. the dataset was carefully curated to address specific challenges in data labeling precision. this involved individualized labeling sessions and adjustments in the number of targets per image, contributing to enhanced precision and increased dataset size. the research begins with the development of a deep learning model trained on a dataset comprising images featuring bee-like structures within asphalt. the model excels in accurately identifying and segmenting these structures. subsequently, the deep learning approach is compared with existing methods for bee structure segmentation to establish its precision and superiority. employing frequency distribution histograms, the distribution patterns of bee structures within various types of sbs-modified asphalt is analyzed, quantitatively assessing the influence of diverse modifier types on these microstructural attributes. the findings in this study underscore the deep learning model's efficacy in recognizing and segmenting bee structures with introduced metrics effectively capturing the distinctive characteristics of various asphalt microstructures. this study paves the way for comprehensive analyses of microstructural metrics, including parameters such as perimeter, area, quantity, and related indicators, thus contributing to the development of fundamental asphalt structural units suitable for processes like molecular simulation and finite element analysis. moreover, it propels the application of deep learning methodologies in the realm of road materials research, opening new avenues for innovative explorations that can ultimately benefit sustainable fuel production.","key_phrases":["microstructures","asphalt","bee structures","asphalt performance","extension","sustainable fuel production","this study","deep learning techniques","the impact","different styrene","butadiene\u2013styrene (sbs) modifiers","asphalt microstructures","bee structures","the employed deep learning model","a diverse dataset","images","the dataset","specific challenges","data labeling precision","this involved individualized labeling sessions","adjustments","the number","targets","image","enhanced precision","the research","the development","a deep learning model","a dataset comprising images","bee-like structures","asphalt","the model excels","these structures","the deep learning approach","existing methods","bee structure segmentation","its precision","superiority","frequency distribution histograms","the distribution patterns","bee structures","various types","sbs-modified asphalt","the influence","diverse modifier types","these microstructural attributes","the findings","this study","the deep learning model's efficacy","segmenting bee structures","introduced metrics","the distinctive characteristics","various asphalt microstructures","this study","the way","comprehensive analyses","microstructural metrics","parameters","perimeter","area","quantity","related indicators","the development","fundamental asphalt structural units","processes","molecular simulation","finite element analysis","it","the application","deep learning methodologies","the realm","road materials research","new avenues","innovative explorations","that","sustainable fuel production","200"]},{"title":"Deep Machine Learning in Optimization of Scientific Research Activities","authors":["E. V. Melnikova"],"abstract":"Abstract\u2014This article provides a general overview of machine learning, a subdomain of artificial intelligence. The substance of the deep learning process is explained, and key features of deep learning as a high-level artificial intelligence technology are outlined. Differences between deep and conventional machine learning are analyzed. The architecture of deep learning models is considered. Issues with using deep learning in neural networks are outlined, and key processes of the functioning of neural networks are described. The importance of deep learning neural networks for processing big data is noted. Specific examples of application of deep learning algorithms in various research fields, specifically, scientometrics, bibliometrics, medicine, geoseismic research, and others, are provided. It is shown that deep learning plays an important role in optimizing research activities and improving research productivity.","doi":"10.3103\/S0147688223010082","cleaned_title":"deep machine learning in optimization of scientific research activities","cleaned_abstract":"abstract\u2014this article provides a general overview of machine learning, a subdomain of artificial intelligence. the substance of the deep learning process is explained, and key features of deep learning as a high-level artificial intelligence technology are outlined. differences between deep and conventional machine learning are analyzed. the architecture of deep learning models is considered. issues with using deep learning in neural networks are outlined, and key processes of the functioning of neural networks are described. the importance of deep learning neural networks for processing big data is noted. specific examples of application of deep learning algorithms in various research fields, specifically, scientometrics, bibliometrics, medicine, geoseismic research, and others, are provided. it is shown that deep learning plays an important role in optimizing research activities and improving research productivity.","key_phrases":["this article","a general overview","machine learning","a subdomain","artificial intelligence","the substance","the deep learning process","key features","deep learning","a high-level artificial intelligence technology","differences","deep and conventional machine learning","the architecture","deep learning models","issues","deep learning","neural networks","key processes","the functioning","neural networks","the importance","deep learning neural networks","big data","specific examples","application","deep learning algorithms","various research fields","specifically, scientometrics","bibliometrics","medicine","geoseismic research","others","it","deep learning","an important role","research activities","research productivity"]},{"title":"Detecting and Mitigating Encoded Bias in Deep Learning-Based Stealth Assessment Models for Reflection-Enriched Game-Based Learning Environments","authors":["Anisha Gupta","Dan Carpenter","Wookhee Min","Jonathan Rowe","Roger Azevedo","James Lester"],"abstract":"Reflection plays a critical role in learning. Game-based learning environments have significant potential to elicit and support student reflection by prompting learners to think critically about their own learning processes and performance. Stealth assessment models, used for unobtrusively assessing student competencies from evidence of game interaction data and facilitating learning through adaptive feedback, can be enhanced by incorporating evidence from students\u2019 written reflections. We present a deep learning-based stealth assessment framework that predicts depth of student reflections and science content post-test scores during game-based learning. With the increasing adoption of AI techniques in decision-making processes, it is important to evaluate the fairness of these models. To address this concern, we investigate encoded bias in our stealth assessment model with respect to student gender and prior game-playing experience in deep learning-based stealth assessment models and examine the impact of debiasing on the models\u2019 predictive performance. We evaluate the predictive performance of the deep learning-based stealth assessment models and measure encoded bias with the Absolute Between-ROC Area (ABROCA) statistic using gameplay data from 119 students collected in a series of classroom studies with a reflection-enriched game-based learning environment for middle school microbiology, Crystal Island. The results demonstrate the effectiveness of deep learning-based stealth assessment models and multiple debiasing techniques for deriving algorithmically fair stealth assessment models.","doi":"10.1007\/s40593-023-00379-6","cleaned_title":"detecting and mitigating encoded bias in deep learning-based stealth assessment models for reflection-enriched game-based learning environments","cleaned_abstract":"reflection plays a critical role in learning. game-based learning environments have significant potential to elicit and support student reflection by prompting learners to think critically about their own learning processes and performance. stealth assessment models, used for unobtrusively assessing student competencies from evidence of game interaction data and facilitating learning through adaptive feedback, can be enhanced by incorporating evidence from students\u2019 written reflections. we present a deep learning-based stealth assessment framework that predicts depth of student reflections and science content post-test scores during game-based learning. with the increasing adoption of ai techniques in decision-making processes, it is important to evaluate the fairness of these models. to address this concern, we investigate encoded bias in our stealth assessment model with respect to student gender and prior game-playing experience in deep learning-based stealth assessment models and examine the impact of debiasing on the models\u2019 predictive performance. we evaluate the predictive performance of the deep learning-based stealth assessment models and measure encoded bias with the absolute between-roc area (abroca) statistic using gameplay data from 119 students collected in a series of classroom studies with a reflection-enriched game-based learning environment for middle school microbiology, crystal island. the results demonstrate the effectiveness of deep learning-based stealth assessment models and multiple debiasing techniques for deriving algorithmically fair stealth assessment models.","key_phrases":["reflection","a critical role","game-based learning environments","significant potential","student reflection","learners","their own learning processes","performance","stealth assessment models","student competencies","evidence","game interaction data","adaptive feedback","evidence","students\u2019 written reflections","we","a deep learning-based stealth assessment framework","that","depth","student reflections","science content post-test scores","game-based learning","the increasing adoption","ai techniques","decision-making processes","it","the fairness","these models","this concern","we","encoded bias","our stealth assessment model","respect","student gender","prior game-playing experience","deep learning-based stealth assessment models","the impact","the models\u2019 predictive performance","we","the predictive performance","the deep learning-based stealth assessment models","measure","bias","-roc","gameplay data","119 students","a series","classroom studies","a reflection-enriched game-based learning environment","middle school microbiology","crystal island","the results","the effectiveness","deep learning-based stealth assessment models","multiple debiasing techniques","algorithmically fair stealth assessment models","119","crystal island"]},{"title":"Reading Between the Lines: Machine Learning Ensemble and Deep Learning for Implied Threat Detection in Textual Data","authors":["Muhammad Owais Raza","Areej Fatemah Meghji","Naeem Ahmed Mahoto","Mana Saleh Al Reshan","Hamad Ali Abosaq","Adel Sulaiman","Asadullah Shaikh"],"abstract":"With the increase in the generation and spread of textual content on social media, natural language processing (NLP) has become an important area of research for detecting underlying threats, racial abuse, violence, and implied warnings in the content. The subtlety and ambiguity of language make the development of effective models for detecting threats in text a challenging task. This task is further complicated when the threat is not explicitly conveyed. This study focuses on the task of implied threat detection using an explicitly designed machine-generated dataset with both linguistic and lexical features. We evaluated the performance of different machine learning algorithms on these features including Support Vector Machines, Logistic Regression, Naive Bayes, Decision Tree, and K-nearest neighbors. The ensembling approaches of Adaboost, Random Forest, and Gradient Boosting were also explored. Deep learning modeling was performed using Long Short-Term Memory, Deep Neural Networks (DNN), and Bidirectional Long Short-Term Memory (BiLSTM). Based on the evaluation, it was observed that classical and ensemble models overfit while working with linguistic features. The performance of these models improved when working with lexical features. The model based on logistic regression exhibited superior performance with an F1 score of 77.13%. While experimenting with deep learning models, DNN achieved an F1 score of 91.49% while the BiLSTM achieved an F1 score of 91.61% while working with lexical features. The current study provides a baseline for future research in the domain of implied threat detection.","doi":"10.1007\/s44196-024-00580-y","cleaned_title":"reading between the lines: machine learning ensemble and deep learning for implied threat detection in textual data","cleaned_abstract":"with the increase in the generation and spread of textual content on social media, natural language processing (nlp) has become an important area of research for detecting underlying threats, racial abuse, violence, and implied warnings in the content. the subtlety and ambiguity of language make the development of effective models for detecting threats in text a challenging task. this task is further complicated when the threat is not explicitly conveyed. this study focuses on the task of implied threat detection using an explicitly designed machine-generated dataset with both linguistic and lexical features. we evaluated the performance of different machine learning algorithms on these features including support vector machines, logistic regression, naive bayes, decision tree, and k-nearest neighbors. the ensembling approaches of adaboost, random forest, and gradient boosting were also explored. deep learning modeling was performed using long short-term memory, deep neural networks (dnn), and bidirectional long short-term memory (bilstm). based on the evaluation, it was observed that classical and ensemble models overfit while working with linguistic features. the performance of these models improved when working with lexical features. the model based on logistic regression exhibited superior performance with an f1 score of 77.13%. while experimenting with deep learning models, dnn achieved an f1 score of 91.49% while the bilstm achieved an f1 score of 91.61% while working with lexical features. the current study provides a baseline for future research in the domain of implied threat detection.","key_phrases":["the increase","the generation","spread","textual content","social media","natural language processing","nlp","an important area","research","underlying threats","racial abuse","violence","implied warnings","the content","the subtlety","ambiguity","language","the development","effective models","threats","text","a challenging task","this task","the threat","this study","the task","implied threat detection","an explicitly designed machine-generated dataset","both linguistic and lexical features","we","the performance","different machine learning algorithms","these features","support vector machines","logistic regression","naive bayes","decision tree","k-nearest neighbors","the ensembling approaches","adaboost, random forest","gradient boosting","deep learning modeling","long short-term memory","deep neural networks","dnn","bidirectional long short-term memory","bilstm","the evaluation","it","classical and ensemble models","linguistic features","the performance","these models","lexical features","the model","logistic regression","superior performance","an f1 score","77.13%","deep learning models","dnn","an f1 score","91.49%","the bilstm","an f1 score","91.61%","lexical features","the current study","a baseline","future research","the domain","implied threat detection","77.13%","91.49%","91.61%"]},{"title":"Deep learning-based phenotyping reclassifies combined hepatocellular-cholangiocarcinoma","authors":["Julien Calderaro","Narmin Ghaffari Laleh","Qinghe Zeng","Pascale Maille","Loetitia Favre","Ana\u00efs Pujals","Christophe Klein","C\u00e9line Bazille","Lara R. Heij","Arnaud Uguen","Tom Luedde","Luca Di Tommaso","Aur\u00e9lie Beaufr\u00e8re","Augustin Chatain","Delphine Gastineau","Cong Trung Nguyen","Hiep Nguyen-Canh","Khuyen Nguyen Thi","Viviane Gnemmi","Rondell P. Graham","Fr\u00e9d\u00e9ric Charlotte","Dominique Wendum","Mukul Vij","Daniela S. Allende","Federico Aucejo","Alba Diaz","Benjamin Rivi\u00e8re","Astrid Herrero","Katja Evert","Diego Francesco Calvisi","J\u00e9r\u00e9my Augustin","Wei Qiang Leow","Howard Ho Wai Leung","Emmanuel Boleslawski","Mohamed Rela","Arnaud Fran\u00e7ois","Anthony Wing-Hung Cha","Alejandro Forner","Maria Reig","Manon Allaire","Olivier Scatton","Denis Chatelain","Camille Boulagnon-Rombi","Nathalie Sturm","Benjamin Menahem","Eric Frouin","David Tougeron","Christophe Tournigand","Emmanuelle Kempf","Haeryoung Kim","Massih Ningarhari","Sophie Michalak-Provost","Purva Gopal","Raffaele Brustia","Eric Vibert","Kornelius Schulze","Darius F. R\u00fcther","S\u00f6ren A. Weidemann","Rami Rhaiem","Jean-Michel Pawlotsky","Xuchen Zhang","Alain Luciani","S\u00e9bastien Mul\u00e9","Alexis Laurent","Giuliana Amaddeo","H\u00e9l\u00e8ne Regnault","Eleonora De Martin","Christine Sempoux","Pooja Navale","Maria Westerhoff","Regina Cheuk-Lam Lo","Jan Bednarsch","Annette Gouw","Catherine Guettier","Marie Lequoy","Kenichi Harada","Pimsiri Sripongpun","Poowadon Wetwittayaklang","Nicolas Lom\u00e9nie","Jarukit Tantipisit","Apichat Kaewdech","Jeanne Shen","Val\u00e9rie Paradis","Stefano Caruso","Jakob Nikolas Kather"],"abstract":"Primary liver cancer arises either from hepatocytic or biliary lineage cells, giving rise to hepatocellular carcinoma (HCC) or intrahepatic cholangiocarcinoma (ICCA). Combined hepatocellular- cholangiocarcinomas (cHCC-CCA) exhibit equivocal or mixed features of both, causing diagnostic uncertainty and difficulty in determining proper management. Here, we perform a comprehensive deep learning-based phenotyping of multiple cohorts of patients. We show that deep learning can reproduce the diagnosis of HCC vs. CCA with a high performance. We analyze a series of 405 cHCC-CCA patients and demonstrate that the model can reclassify the tumors as HCC or ICCA, and that the predictions are consistent with clinical outcomes, genetic alterations and in situ spatial gene expression profiling. This type of approach could improve treatment decisions and ultimately clinical outcome for patients with rare and biphenotypic cancers such as cHCC-CCA.","doi":"10.1038\/s41467-023-43749-3","cleaned_title":"deep learning-based phenotyping reclassifies combined hepatocellular-cholangiocarcinoma","cleaned_abstract":"primary liver cancer arises either from hepatocytic or biliary lineage cells, giving rise to hepatocellular carcinoma (hcc) or intrahepatic cholangiocarcinoma (icca). combined hepatocellular- cholangiocarcinomas (chcc-cca) exhibit equivocal or mixed features of both, causing diagnostic uncertainty and difficulty in determining proper management. here, we perform a comprehensive deep learning-based phenotyping of multiple cohorts of patients. we show that deep learning can reproduce the diagnosis of hcc vs. cca with a high performance. we analyze a series of 405 chcc-cca patients and demonstrate that the model can reclassify the tumors as hcc or icca, and that the predictions are consistent with clinical outcomes, genetic alterations and in situ spatial gene expression profiling. this type of approach could improve treatment decisions and ultimately clinical outcome for patients with rare and biphenotypic cancers such as chcc-cca.","key_phrases":["primary liver cancer","hepatocytic or biliary lineage cells","rise","hepatocellular carcinoma","hcc","intrahepatic cholangiocarcinoma","icca","combined hepatocellular- cholangiocarcinomas","chcc-cca","equivocal or mixed features","both","diagnostic uncertainty","difficulty","proper management","we","a comprehensive deep learning-based phenotyping","multiple cohorts","patients","we","deep learning","the diagnosis","hcc","cca","a high performance","we","a series","405 chcc-cca patients","the model","the tumors","hcc","icca","the predictions","clinical outcomes","genetic alterations","situ spatial gene expression profiling","this type","approach","treatment decisions","ultimately clinical outcome","patients","rare and biphenotypic cancers","chcc-cca","405"]},{"title":"Advancing Naturalistic Affective Science with Deep Learning","authors":["Chujun Lin","Landry S. Bulls","Lindsey J. Tepfer","Amisha D. Vyas","Mark A. Thornton"],"abstract":"People express their own emotions and perceive others\u2019 emotions via a variety of channels, including facial movements, body gestures, vocal prosody, and language. Studying these channels of affective behavior offers insight into both the experience and perception of emotion. Prior research has predominantly focused on studying individual channels of affective behavior in isolation using tightly controlled, non-naturalistic experiments. This approach limits our understanding of emotion in more naturalistic contexts where different channels of information tend to interact. Traditional methods struggle to address this limitation: manually annotating behavior is time-consuming, making it infeasible to do at large scale; manually selecting and manipulating stimuli based on hypotheses may neglect unanticipated features, potentially generating biased conclusions; and common linear modeling approaches cannot fully capture the complex, nonlinear, and interactive nature of real-life affective processes. In this methodology review, we describe how deep learning can be applied to address these challenges to advance a more naturalistic affective science. First, we describe current practices in affective research and explain why existing methods face challenges in revealing a more naturalistic understanding of emotion. Second, we introduce deep learning approaches and explain how they can be applied to tackle three main challenges: quantifying naturalistic behaviors, selecting and manipulating naturalistic stimuli, and modeling naturalistic affective processes. Finally, we describe the limitations of these deep learning methods, and how these limitations might be avoided or mitigated. By detailing the promise and the peril of deep learning, this review aims to pave the way for a more naturalistic affective science.","doi":"10.1007\/s42761-023-00215-z","cleaned_title":"advancing naturalistic affective science with deep learning","cleaned_abstract":"people express their own emotions and perceive others\u2019 emotions via a variety of channels, including facial movements, body gestures, vocal prosody, and language. studying these channels of affective behavior offers insight into both the experience and perception of emotion. prior research has predominantly focused on studying individual channels of affective behavior in isolation using tightly controlled, non-naturalistic experiments. this approach limits our understanding of emotion in more naturalistic contexts where different channels of information tend to interact. traditional methods struggle to address this limitation: manually annotating behavior is time-consuming, making it infeasible to do at large scale; manually selecting and manipulating stimuli based on hypotheses may neglect unanticipated features, potentially generating biased conclusions; and common linear modeling approaches cannot fully capture the complex, nonlinear, and interactive nature of real-life affective processes. in this methodology review, we describe how deep learning can be applied to address these challenges to advance a more naturalistic affective science. first, we describe current practices in affective research and explain why existing methods face challenges in revealing a more naturalistic understanding of emotion. second, we introduce deep learning approaches and explain how they can be applied to tackle three main challenges: quantifying naturalistic behaviors, selecting and manipulating naturalistic stimuli, and modeling naturalistic affective processes. finally, we describe the limitations of these deep learning methods, and how these limitations might be avoided or mitigated. by detailing the promise and the peril of deep learning, this review aims to pave the way for a more naturalistic affective science.","key_phrases":["people","their own emotions","others\u2019 emotions","a variety","channels","facial movements","body gestures","vocal prosody","language","these channels","affective behavior","insight","both the experience","perception","emotion","prior research","individual channels","affective behavior","isolation","tightly controlled, non-naturalistic experiments","this approach","our understanding","emotion","more naturalistic contexts","different channels","information","traditional methods","this limitation","behavior","it","large scale","stimuli","hypotheses","unanticipated features","biased conclusions","common linear modeling approaches","the complex, nonlinear, and interactive nature","real-life affective processes","this methodology review","we","how deep learning","these challenges","a more naturalistic affective science","we","current practices","affective research","existing methods","challenges","a more naturalistic understanding","emotion","we","deep learning approaches","they","three main challenges","naturalistic behaviors","naturalistic stimuli","naturalistic affective processes","we","the limitations","these deep learning methods","these limitations","the promise","the peril","deep learning","this review","the way","a more naturalistic affective science","linear","first","second","three"]},{"title":"Bearing Fault Diagnosis Based on Artificial Intelligence Methods: Machine Learning and Deep Learning","authors":["Ahmed Ghorbel","Sarra Eddai","Bouthayna Limam","Nabih Feki","Mohamed Haddar"],"abstract":"This paper presents a comprehensive study on the application of Artificial Intelligence (AI) methods, specifically machine learning and deep learning, for the diagnosis of bearing faults. The study explores both data preprocessing-dependent methods (Support Vector Machine, Nearest Neighbor, and Decision Tree) and a preprocessing-independent method (1D Convolutional Neural Network). The experiment setup utilizes the Case Western Reserve University dataset for signal acquisition. A detailed strategy for data processing is developed, encompassing initialization, data loading, signal filtration, decomposition, feature extraction in both time- and frequency-domains, and feature selection. Indeed, the study involves working with four datasets, selected based on the distribution curves of the indicators as a function of the number of observations. The results demonstrate remarkable performance of the AI methods in bearing fault diagnosis. The 1D-CNN model, in particular, shows high robustness and accuracy, even in the presence of load variations. The findings of this study shed light on the significant potential of AI methods in improving the accuracy and efficiency of bearing fault diagnosis.","doi":"10.1007\/s13369-024-09488-3","cleaned_title":"bearing fault diagnosis based on artificial intelligence methods: machine learning and deep learning","cleaned_abstract":"this paper presents a comprehensive study on the application of artificial intelligence (ai) methods, specifically machine learning and deep learning, for the diagnosis of bearing faults. the study explores both data preprocessing-dependent methods (support vector machine, nearest neighbor, and decision tree) and a preprocessing-independent method (1d convolutional neural network). the experiment setup utilizes the case western reserve university dataset for signal acquisition. a detailed strategy for data processing is developed, encompassing initialization, data loading, signal filtration, decomposition, feature extraction in both time- and frequency-domains, and feature selection. indeed, the study involves working with four datasets, selected based on the distribution curves of the indicators as a function of the number of observations. the results demonstrate remarkable performance of the ai methods in bearing fault diagnosis. the 1d-cnn model, in particular, shows high robustness and accuracy, even in the presence of load variations. the findings of this study shed light on the significant potential of ai methods in improving the accuracy and efficiency of bearing fault diagnosis.","key_phrases":["this paper","a comprehensive study","the application","artificial intelligence (ai) methods","specifically machine learning","deep learning","the diagnosis","faults","the study","both data preprocessing-dependent methods","vector machine","neighbor","decision tree","a preprocessing-independent method","1d convolutional neural network","the experiment setup","the case","western reserve university dataset","signal acquisition","a detailed strategy","data processing","initialization","data loading","signal filtration","decomposition","extraction","both time- and frequency-domains","the study","four datasets","the distribution curves","the indicators","a function","the number","observations","the results","remarkable performance","the ai methods","fault diagnosis","the 1d-cnn model","high robustness","accuracy","the presence","load variations","the findings","this study","light","the significant potential","ai methods","the accuracy","efficiency","fault diagnosis","1d","western reserve university","four","1d-cnn"]},{"title":"Exploring the effects of digital technology on deep learning: a meta-analysis","authors":["Xiu-Yi Wu"],"abstract":"The impact of digital technology on learning outcomes, specifically deep learning, has been a subject of considerable debate and scrutiny in educational settings. This study aims to provide clarity by conducting a meta-analysis of empirical publications that examine students' deep learning outcomes in relation to digital technology. A comprehensive search of databases and a thorough literature review yielded 60 high-quality, peer-reviewed journal articles that met the inclusion criteria. Using Review Manager 5.4.1 software, a meta-analysis was conducted to assess the overall effectiveness of digital technology. The calculated effect size indicates a positive influence of digital technology on students' deep learning outcomes. Furthermore, a moderator variable analysis revealed several significant findings: 1. Different categories of digital technology tools have a favorable impact on deep learning outcomes; 2. The duration of digital technology treatment does not significantly affect deep learning outcomes; 3. Digital technology demonstrates a highly positive influence on deep learning within the humanities and social sciences disciplines; 4. Combining online and offline utilization of digital technology in education leads to a substantially greater enhancement in deep learning compared to relying solely on online methods; 5. The effectiveness of digital technology on deep learning is enhanced when accompanied by appropriate instructional guidance; 6. Utilizing digital technology in a systematic manner produces different outcomes compared to fragmented approaches, highlighting the importance of a cohesive implementation; 7. Integrating digital technology with collaborative learning has a more pronounced effect on deep learning compared to independent learning. These findings contribute to our understanding of the impact of digital technology on deep learning outcomes and underscore the importance of thoughtful integration and instructional support in educational contexts.","doi":"10.1007\/s10639-023-12307-1","cleaned_title":"exploring the effects of digital technology on deep learning: a meta-analysis","cleaned_abstract":"the impact of digital technology on learning outcomes, specifically deep learning, has been a subject of considerable debate and scrutiny in educational settings. this study aims to provide clarity by conducting a meta-analysis of empirical publications that examine students' deep learning outcomes in relation to digital technology. a comprehensive search of databases and a thorough literature review yielded 60 high-quality, peer-reviewed journal articles that met the inclusion criteria. using review manager 5.4.1 software, a meta-analysis was conducted to assess the overall effectiveness of digital technology. the calculated effect size indicates a positive influence of digital technology on students' deep learning outcomes. furthermore, a moderator variable analysis revealed several significant findings: 1. different categories of digital technology tools have a favorable impact on deep learning outcomes; 2. the duration of digital technology treatment does not significantly affect deep learning outcomes; 3. digital technology demonstrates a highly positive influence on deep learning within the humanities and social sciences disciplines; 4. combining online and offline utilization of digital technology in education leads to a substantially greater enhancement in deep learning compared to relying solely on online methods; 5. the effectiveness of digital technology on deep learning is enhanced when accompanied by appropriate instructional guidance; 6. utilizing digital technology in a systematic manner produces different outcomes compared to fragmented approaches, highlighting the importance of a cohesive implementation; 7. integrating digital technology with collaborative learning has a more pronounced effect on deep learning compared to independent learning. these findings contribute to our understanding of the impact of digital technology on deep learning outcomes and underscore the importance of thoughtful integration and instructional support in educational contexts.","key_phrases":["the impact","digital technology","outcomes","specifically deep learning","a subject","considerable debate","scrutiny","educational settings","this study","clarity","a meta-analysis","empirical publications","that","outcomes","relation","digital technology","a comprehensive search","databases","a thorough literature review","60 high-quality, peer-reviewed journal articles","that","the inclusion criteria","review manager 5.4.1 software","a meta-analysis","the overall effectiveness","digital technology","the calculated effect size","a positive influence","digital technology","students' deep learning outcomes","a moderator variable analysis","several significant findings","1. different categories","digital technology tools","a favorable impact","deep learning outcomes","the duration","digital technology treatment","deep learning outcomes","3. digital technology","a highly positive influence","deep learning","the humanities","social sciences disciplines","online and offline utilization","digital technology","education","a substantially greater enhancement","deep learning","online methods","the effectiveness","digital technology","deep learning","appropriate instructional guidance","digital technology","a systematic manner","different outcomes","fragmented approaches","the importance","a cohesive implementation","digital technology","collaborative learning","a more pronounced effect","deep learning","independent learning","these findings","our understanding","the impact","digital technology","deep learning outcomes","the importance","thoughtful integration","instructional support","educational contexts","60","5.4.1","1","2","3","4","5","6","7"]},{"title":"Abstraction, mimesis and the evolution of deep learning","authors":["Jon Ekl\u00f6f","Thomas Hamelryck","Cadell Last","Alexander Grima","Ulrika Lundh Snis"],"abstract":"Deep learning developers typically rely on deep learning software frameworks (DLSFs)\u2014simply described as pre-packaged libraries of programming components that provide high-level access to deep learning functionality. New DLSFs progressively encapsulate mathematical, statistical and computational complexity. Such higher levels of abstraction subsequently make it easier for deep learning methodology to spread through mimesis (i.e., imitation of models perceived as successful). In this study, we quantify this increase in abstraction and discuss its implications. Analyzing publicly available code from Github, we found that the introduction of DLSFs correlates both with significant increases in the number of deep learning projects and substantial reductions in the number of lines of code used. We subsequently discuss and argue the importance of abstraction in deep learning with respect to ephemeralization, technological advancement, democratization, adopting timely levels of abstraction, the emergence of mimetic deadlocks, issues related to the use of black box methods including privacy and fairness, and the concentration of technological power. Finally, we also discuss abstraction as a symptom of an ongoing technological metatransition.","doi":"10.1007\/s00146-023-01688-z","cleaned_title":"abstraction, mimesis and the evolution of deep learning","cleaned_abstract":"deep learning developers typically rely on deep learning software frameworks (dlsfs)\u2014simply described as pre-packaged libraries of programming components that provide high-level access to deep learning functionality. new dlsfs progressively encapsulate mathematical, statistical and computational complexity. such higher levels of abstraction subsequently make it easier for deep learning methodology to spread through mimesis (i.e., imitation of models perceived as successful). in this study, we quantify this increase in abstraction and discuss its implications. analyzing publicly available code from github, we found that the introduction of dlsfs correlates both with significant increases in the number of deep learning projects and substantial reductions in the number of lines of code used. we subsequently discuss and argue the importance of abstraction in deep learning with respect to ephemeralization, technological advancement, democratization, adopting timely levels of abstraction, the emergence of mimetic deadlocks, issues related to the use of black box methods including privacy and fairness, and the concentration of technological power. finally, we also discuss abstraction as a symptom of an ongoing technological metatransition.","key_phrases":["deep learning developers","deep learning software frameworks","pre-packaged libraries","programming components","that","high-level access","deep learning functionality","mathematical, statistical and computational complexity","such higher levels","abstraction","it","deep learning methodology","mimesis","i.e., imitation","models","this study","we","this increase","abstraction","its implications","publicly available code","github","we","the introduction","significant increases","the number","deep learning projects","substantial reductions","the number","lines","code","we","the importance","abstraction","deep learning","respect","ephemeralization","technological advancement","democratization","timely levels","abstraction","the emergence","mimetic deadlocks","issues","the use","black box methods","privacy","fairness","the concentration","technological power","we","abstraction","a symptom","an ongoing technological metatransition"]},{"title":"COVID-19 classification based on a deep learning and machine learning fusion technique using chest CT images","authors":["Gerges M. Salama","Asmaa Mohamed","Mahmoud Khaled Abd-Ellah"],"abstract":"Coronavirus disease (COVID-19), impacted by SARS-CoV-2, is one of the greatest challenges of the twenty-first century. COVID-19 broke out in the world over the last 2\u00a0years and has caused many injuries and killed persons. Computer-aided diagnosis has become a necessary tool to prevent the spreading of this virus. Detecting COVID-19 at an early stage is essential to reduce the mortality risk of patients. Researchers seek to find rapid solutions based on techniques of Machine Learning and Deep Learning. In this paper, we introduced a hybrid model for COVID-19 detection based on machine learning and deep learning models. We used 10 different deep CNN network models to extract features from CT images. We extract features from different layers in each network and find the optimum layer that gives the best-extracted features for each CNN network. Then, for classifying these features, we used five different classifiers based on machine learning. The dataset consists of 2481 CT images divided into COVID-19 and non-COVID-19 categories. Three folds are extracted with a different size between testing and training. Through experiments, we define the best layer for all used CNN networks, the best network, and the best-used classifier. The measured performance shows the superiority of the proposed system over the literature with a highest accuracy of 99.39%. Our models are tested with the three folds that gained maximum average accuracy. The result is 98.69%.","doi":"10.1007\/s00521-023-09346-7","cleaned_title":"covid-19 classification based on a deep learning and machine learning fusion technique using chest ct images","cleaned_abstract":"coronavirus disease (covid-19), impacted by sars-cov-2, is one of the greatest challenges of the twenty-first century. covid-19 broke out in the world over the last 2 years and has caused many injuries and killed persons. computer-aided diagnosis has become a necessary tool to prevent the spreading of this virus. detecting covid-19 at an early stage is essential to reduce the mortality risk of patients. researchers seek to find rapid solutions based on techniques of machine learning and deep learning. in this paper, we introduced a hybrid model for covid-19 detection based on machine learning and deep learning models. we used 10 different deep cnn network models to extract features from ct images. we extract features from different layers in each network and find the optimum layer that gives the best-extracted features for each cnn network. then, for classifying these features, we used five different classifiers based on machine learning. the dataset consists of 2481 ct images divided into covid-19 and non-covid-19 categories. three folds are extracted with a different size between testing and training. through experiments, we define the best layer for all used cnn networks, the best network, and the best-used classifier. the measured performance shows the superiority of the proposed system over the literature with a highest accuracy of 99.39%. our models are tested with the three folds that gained maximum average accuracy. the result is 98.69%.","key_phrases":["coronavirus disease","covid-19","sars","the greatest challenges","the twenty-first century","covid-19","the world","the last 2 years","many injuries","persons","computer-aided diagnosis","a necessary tool","the spreading","this virus","covid-19","an early stage","the mortality risk","patients","researchers","rapid solutions","techniques","machine learning","deep learning","this paper","we","a hybrid model","covid-19 detection","machine learning","deep learning models","we","10 different deep cnn network models","features","ct images","we","features","different layers","each network","the optimum layer","that","the best-extracted features","each cnn network","these features","we","five different classifiers","machine learning","the dataset","2481 ct images","covid-19 and non-covid-19 categories","three folds","a different size","testing","training","experiments","we","the best layer","all used cnn networks","the best network","the best-used classifier","the measured performance","the superiority","the proposed system","the literature","a highest accuracy","99.39%","our models","the three folds","that","maximum average accuracy","the result","98.69%","covid-19","the twenty-first century","covid-19","the last 2 years","covid-19","covid-19","10","cnn","cnn","five","2481","covid-19","non-covid-19","three","cnn","99.39%","three","98.69%"]},{"title":"A framework for training larger networks for deep Reinforcement learning","authors":["Kei Ota","Devesh K. Jha","Asako Kanezaki"],"abstract":"The success of deep learning in computer vision and natural language processing communities can be attributed to the training of very deep neural networks with millions or billions of parameters, which can then be trained with massive amounts of data. However, a similar trend has largely eluded the training of deep reinforcement learning (RL) algorithms where larger networks do not lead to performance improvement. Previous work has shown that this is mostly due to instability during the training of deep RL agents when using larger networks. In this paper, we make an attempt to understand and address the training of larger networks for deep RL. We first show that naively increasing network capacity does not improve performance. Then, we propose a novel method that consists of (1) wider networks with DenseNet connection, (2) decoupling representation learning from the training of RL, and (3) a distributed training method to mitigate overfitting problems. Using this three-fold technique, we show that we can train very large networks that result in significant performance gains. We present several ablation studies to demonstrate the efficacy of the proposed method and some intuitive understanding of the reasons for performance gain. We show that our proposed method outperforms other baseline algorithms on several challenging locomotion tasks.","doi":"10.1007\/s10994-024-06547-6","cleaned_title":"a framework for training larger networks for deep reinforcement learning","cleaned_abstract":"the success of deep learning in computer vision and natural language processing communities can be attributed to the training of very deep neural networks with millions or billions of parameters, which can then be trained with massive amounts of data. however, a similar trend has largely eluded the training of deep reinforcement learning (rl) algorithms where larger networks do not lead to performance improvement. previous work has shown that this is mostly due to instability during the training of deep rl agents when using larger networks. in this paper, we make an attempt to understand and address the training of larger networks for deep rl. we first show that naively increasing network capacity does not improve performance. then, we propose a novel method that consists of (1) wider networks with densenet connection, (2) decoupling representation learning from the training of rl, and (3) a distributed training method to mitigate overfitting problems. using this three-fold technique, we show that we can train very large networks that result in significant performance gains. we present several ablation studies to demonstrate the efficacy of the proposed method and some intuitive understanding of the reasons for performance gain. we show that our proposed method outperforms other baseline algorithms on several challenging locomotion tasks.","key_phrases":["the success","deep learning","computer vision","natural language processing communities","the training","very deep neural networks","millions","billions","parameters","which","massive amounts","data","a similar trend","the training","deep reinforcement learning","rl) algorithms","larger networks","performance improvement","previous work","this","instability","the training","deep rl agents","larger networks","this paper","we","an attempt","the training","larger networks","deep rl","we","naively increasing network capacity","performance","we","a novel method","that","wider networks","densenet connection","representation","the training","rl","overfitting problems","this three-fold technique","we","we","very large networks","that","significant performance gains","we","several ablation studies","the efficacy","the proposed method","some intuitive understanding","the reasons","performance gain","we","our proposed method","other baseline algorithms","several challenging locomotion tasks","millions or billions","first","1","2","3","three-fold"]},{"title":"Deep Learning-Based Watermarking Techniques Challenges: A Review of Current and Future Trends","authors":["Saoussen Ben Jabra","Mohamed Ben Farah"],"abstract":"The digital revolution places great emphasis on digital media watermarking due to the increased vulnerability of multimedia content to unauthorized alterations. Recently, in the digital boom in the technology of hiding data, research has been tending to perform watermarking with numerous architectures of deep learning, which has explored a variety of problems since its inception. Several watermarking approaches based on deep learning have been proposed, and they have proven their efficiency compared to traditional methods. This paper summarizes recent developments in conventional and deep learning image and video watermarking techniques. It shows that although there are many conventional techniques focused on video watermarking, there are yet to be any deep learning models focusing on this area; however, for image watermarking, different deep learning-based techniques where efficiency in invisibility and robustness depends on the used network architecture are observed. This study has been concluded by discussing possible research directions in deep learning-based video watermarking.","doi":"10.1007\/s00034-024-02651-z","cleaned_title":"deep learning-based watermarking techniques challenges: a review of current and future trends","cleaned_abstract":"the digital revolution places great emphasis on digital media watermarking due to the increased vulnerability of multimedia content to unauthorized alterations. recently, in the digital boom in the technology of hiding data, research has been tending to perform watermarking with numerous architectures of deep learning, which has explored a variety of problems since its inception. several watermarking approaches based on deep learning have been proposed, and they have proven their efficiency compared to traditional methods. this paper summarizes recent developments in conventional and deep learning image and video watermarking techniques. it shows that although there are many conventional techniques focused on video watermarking, there are yet to be any deep learning models focusing on this area; however, for image watermarking, different deep learning-based techniques where efficiency in invisibility and robustness depends on the used network architecture are observed. this study has been concluded by discussing possible research directions in deep learning-based video watermarking.","key_phrases":["the digital revolution","great emphasis","digital media watermarking","the increased vulnerability","multimedia content","unauthorized alterations","the digital boom","the technology","hiding data","research","numerous architectures","deep learning","which","a variety","problems","its inception","several watermarking approaches","deep learning","they","their efficiency","traditional methods","this paper","recent developments","conventional and deep learning image","video watermarking techniques","it","many conventional techniques","video watermarking","any deep learning models","this area","image watermarking",", different deep learning-based techniques","efficiency","invisibility","robustness","the used network architecture","this study","possible research directions","deep learning-based video watermarking"]},{"title":"Advancing plant biology through deep learning-powered natural language processing","authors":["Shuang Peng","Lo\u00efc Rajjou"],"abstract":"The application of deep learning methods, specifically the utilization of Large Language Models (LLMs), in the field of plant biology holds significant promise for generating novel knowledge on plant cell systems. The LLM framework exhibits exceptional potential, particularly with the development of Protein Language Models (PLMs), allowing for in-depth analyses of nucleic acid and protein sequences. This analytical capacity facilitates the discernment of intricate patterns and relationships within biological data, encompassing multi-scale information within DNA or protein sequences. The contribution of PLMs extends beyond mere sequence patterns and structure\u2013\u2013function recognition; it also supports advancements in genetic improvements for agriculture. The integration of deep learning approaches into the domain of plant sciences offers opportunities for major breakthroughs in basic research across multi-scale plant traits. Consequently, the strategic application of deep learning methodologies, particularly leveraging the potential of LLMs, will undoubtedly play a pivotal role in advancing plant sciences, plant production, plant uses and propelling the trajectory toward sustainable agroecological and agro-food transitions.","doi":"10.1007\/s00299-024-03294-9","cleaned_title":"advancing plant biology through deep learning-powered natural language processing","cleaned_abstract":"the application of deep learning methods, specifically the utilization of large language models (llms), in the field of plant biology holds significant promise for generating novel knowledge on plant cell systems. the llm framework exhibits exceptional potential, particularly with the development of protein language models (plms), allowing for in-depth analyses of nucleic acid and protein sequences. this analytical capacity facilitates the discernment of intricate patterns and relationships within biological data, encompassing multi-scale information within dna or protein sequences. the contribution of plms extends beyond mere sequence patterns and structure\u2013\u2013function recognition; it also supports advancements in genetic improvements for agriculture. the integration of deep learning approaches into the domain of plant sciences offers opportunities for major breakthroughs in basic research across multi-scale plant traits. consequently, the strategic application of deep learning methodologies, particularly leveraging the potential of llms, will undoubtedly play a pivotal role in advancing plant sciences, plant production, plant uses and propelling the trajectory toward sustainable agroecological and agro-food transitions.","key_phrases":["the application","deep learning methods","specifically the utilization","large language models","llms","the field","plant biology","significant promise","novel knowledge","plant cell systems","the llm framework","exceptional potential","the development","protein language models","plms","in-depth analyses","nucleic acid and protein sequences","this analytical capacity","the discernment","intricate patterns","relationships","biological data","multi-scale information","dna","protein sequences","the contribution","plms","mere sequence patterns","structure\u2013\u2013function recognition","it","advancements","genetic improvements","agriculture","the integration","deep learning approaches","the domain","plant sciences","opportunities","major breakthroughs","basic research","multi-scale plant traits","the strategic application","deep learning methodologies","the potential","llms","a pivotal role","plant sciences","plant production","the trajectory","sustainable agroecological and agro-food transitions"]},{"title":"Machine learning and deep learning techniques for detecting malicious android applications: An empirical analysis","authors":["Parnika Bhat","Sunny Behal","Kamlesh Dutta"],"abstract":"The open system architecture of android makes it vulnerable to a variety of cyberattacks. Cybercriminals use android applications to intrude into the system and steal confidential data. This situation poses a threat to user privacy and integrity of the system. This paper proposes a static analysis approach to detect malicious and benign Android applications using various machine learning and deep learning algorithms. The proposed work has been validated using a bench marked dataset comprising 11,449 benign and malicious Android applications. The proposed approach applies a wrapper-based feature selection method to filter irrelevant features. The results clearly show that the deep learning algorithms of DBN and MLP outperformed machine learning algorithms in detecting malicious Android applications.","doi":"10.1007\/s43538-023-00182-w","cleaned_title":"machine learning and deep learning techniques for detecting malicious android applications: an empirical analysis","cleaned_abstract":"the open system architecture of android makes it vulnerable to a variety of cyberattacks. cybercriminals use android applications to intrude into the system and steal confidential data. this situation poses a threat to user privacy and integrity of the system. this paper proposes a static analysis approach to detect malicious and benign android applications using various machine learning and deep learning algorithms. the proposed work has been validated using a bench marked dataset comprising 11,449 benign and malicious android applications. the proposed approach applies a wrapper-based feature selection method to filter irrelevant features. the results clearly show that the deep learning algorithms of dbn and mlp outperformed machine learning algorithms in detecting malicious android applications.","key_phrases":["the open system architecture","android","it","a variety","cyberattacks","cybercriminals","android applications","the system","confidential data","this situation","a threat","user privacy","integrity","the system","this paper","a static analysis approach","malicious and benign android applications","various machine learning","deep learning algorithms","the proposed work","a bench","marked dataset","11,449 benign and malicious android applications","the proposed approach","a wrapper-based feature selection method","irrelevant features","the results","the deep learning algorithms","dbn","mlp","algorithms","malicious android applications","11,449"]},{"title":"Deep reinforcement learning-based scheduling in distributed systems: a critical review","authors":["Zahra Jalali Khalil Abadi","Najme Mansouri","Mohammad Masoud Javidi"],"abstract":"Many fields of research use parallelized and distributed computing environments, including astronomy, earth science, and bioinformatics. Due to an increase in client requests, service providers face various challenges, such as task scheduling, security, resource management, and virtual machine migration. NP-hard scheduling problems require a long time to implement an optimal or suboptimal solution due to their large solution space. With recent advances in artificial intelligence, deep reinforcement learning (DRL) can be used to solve scheduling problems. The DRL approach combines the strength of deep learning and neural networks with reinforcement learning\u2019s feedback-based learning. This paper provides a comprehensive overview of DRL-based scheduling algorithms in distributed systems by categorizing algorithms and applications. As a result, several articles are assessed based on their main objectives, quality of service and scheduling parameters, as well as evaluation environments (i.e., simulation tools, real-world environment). The literature review indicates that algorithms based on RL, such as Q-learning, are effective for learning scaling and scheduling policies in a cloud environment. Additionally, the challenges and directions for further research on deep reinforcement learning to address scheduling problems were summarized (e.g., edge intelligence, ideal dynamic task scheduling framework, human\u2013machine interaction, resource-hungry artificial intelligence (AI) and sustainability).","doi":"10.1007\/s10115-024-02167-7","cleaned_title":"deep reinforcement learning-based scheduling in distributed systems: a critical review","cleaned_abstract":"many fields of research use parallelized and distributed computing environments, including astronomy, earth science, and bioinformatics. due to an increase in client requests, service providers face various challenges, such as task scheduling, security, resource management, and virtual machine migration. np-hard scheduling problems require a long time to implement an optimal or suboptimal solution due to their large solution space. with recent advances in artificial intelligence, deep reinforcement learning (drl) can be used to solve scheduling problems. the drl approach combines the strength of deep learning and neural networks with reinforcement learning\u2019s feedback-based learning. this paper provides a comprehensive overview of drl-based scheduling algorithms in distributed systems by categorizing algorithms and applications. as a result, several articles are assessed based on their main objectives, quality of service and scheduling parameters, as well as evaluation environments (i.e., simulation tools, real-world environment). the literature review indicates that algorithms based on rl, such as q-learning, are effective for learning scaling and scheduling policies in a cloud environment. additionally, the challenges and directions for further research on deep reinforcement learning to address scheduling problems were summarized (e.g., edge intelligence, ideal dynamic task scheduling framework, human\u2013machine interaction, resource-hungry artificial intelligence (ai) and sustainability).","key_phrases":["many fields","research use","astronomy","earth science","bioinformatics","an increase","client requests","service providers","various challenges","task scheduling","security","resource management","virtual machine migration","np-hard scheduling problems","a long time","an optimal or suboptimal solution","their large solution space","recent advances","artificial intelligence","deep reinforcement learning","drl","scheduling problems","the drl approach","the strength","deep learning","neural networks","reinforcement learning\u2019s feedback-based learning","this paper","a comprehensive overview","drl-based scheduling algorithms","distributed systems","algorithms","applications","a result","several articles","their main objectives","quality","service and scheduling parameters","evaluation environments","i.e., simulation tools","real-world environment","the literature review","algorithms","rl","q-learning","scaling and scheduling policies","a cloud environment","the challenges","directions","further research","deep reinforcement learning","scheduling problems","edge intelligence","ideal dynamic task scheduling framework","human\u2013machine interaction","resource-hungry artificial intelligence","ai","sustainability"]},{"title":"Deep-WET: a deep learning-based approach for predicting DNA-binding proteins using word embedding techniques with weighted features","authors":["S. M. Hasan Mahmud","Kah Ong Michael Goh","Md. Faruk Hosen","Dip Nandi","Watshara Shoombuatong"],"abstract":"DNA-binding proteins (DBPs) play a significant role in all phases of genetic processes, including DNA recombination, repair, and modification. They are often utilized in drug discovery as fundamental elements of steroids, antibiotics, and anticancer drugs. Predicting them poses the most challenging task in proteomics research. Conventional experimental methods for DBP identification are costly and sometimes biased toward prediction. Therefore, developing powerful computational methods that can accurately and rapidly identify DBPs from sequence information is an urgent need. In this study, we propose a novel deep learning-based method called Deep-WET to accurately identify DBPs from primary sequence information. In Deep-WET, we employed three powerful feature encoding schemes containing Global Vectors, Word2Vec, and fastText to encode the protein sequence. Subsequently, these three features were sequentially combined and weighted using the weights obtained from the elements learned through the differential evolution (DE) algorithm. To enhance the predictive performance of Deep-WET, we applied the SHapley Additive exPlanations approach to remove irrelevant features. Finally, the optimal feature subset was input into convolutional neural networks to construct the Deep-WET predictor. Both cross-validation and independent tests indicated that Deep-WET achieved superior predictive performance compared to conventional machine learning classifiers. In addition, in extensive independent test, Deep-WET was effective and outperformed than several state-of-the-art methods for DBP prediction, with accuracy of 78.08%, MCC of 0.559, and AUC of 0.805. This superior performance shows that Deep-WET has a tremendous predictive capacity to predict DBPs. The web server of Deep-WET and curated datasets in this study are available at https:\/\/deepwet-dna.monarcatechnical.com\/. The proposed Deep-WET is anticipated to serve the community-wide effort for large-scale identification of potential DBPs.","doi":"10.1038\/s41598-024-52653-9","cleaned_title":"deep-wet: a deep learning-based approach for predicting dna-binding proteins using word embedding techniques with weighted features","cleaned_abstract":"dna-binding proteins (dbps) play a significant role in all phases of genetic processes, including dna recombination, repair, and modification. they are often utilized in drug discovery as fundamental elements of steroids, antibiotics, and anticancer drugs. predicting them poses the most challenging task in proteomics research. conventional experimental methods for dbp identification are costly and sometimes biased toward prediction. therefore, developing powerful computational methods that can accurately and rapidly identify dbps from sequence information is an urgent need. in this study, we propose a novel deep learning-based method called deep-wet to accurately identify dbps from primary sequence information. in deep-wet, we employed three powerful feature encoding schemes containing global vectors, word2vec, and fasttext to encode the protein sequence. subsequently, these three features were sequentially combined and weighted using the weights obtained from the elements learned through the differential evolution (de) algorithm. to enhance the predictive performance of deep-wet, we applied the shapley additive explanations approach to remove irrelevant features. finally, the optimal feature subset was input into convolutional neural networks to construct the deep-wet predictor. both cross-validation and independent tests indicated that deep-wet achieved superior predictive performance compared to conventional machine learning classifiers. in addition, in extensive independent test, deep-wet was effective and outperformed than several state-of-the-art methods for dbp prediction, with accuracy of 78.08%, mcc of 0.559, and auc of 0.805. this superior performance shows that deep-wet has a tremendous predictive capacity to predict dbps. the web server of deep-wet and curated datasets in this study are available at https:\/\/deepwet-dna.monarcatechnical.com\/. the proposed deep-wet is anticipated to serve the community-wide effort for large-scale identification of potential dbps.","key_phrases":["dna-binding proteins","dbps","a significant role","all phases","genetic processes","dna recombination","repair","modification","they","drug discovery","fundamental elements","steroids","antibiotics","anticancer drugs","them","the most challenging task","proteomics research","conventional experimental methods","dbp identification","prediction","powerful computational methods","that","dbps","sequence information","an urgent need","this study","we","a novel deep learning-based method","dbps","primary sequence information","we","three powerful feature encoding schemes","global vectors","word2vec","fasttext","the protein sequence","these three features","the weights","the elements","the differential evolution","de","algorithm","the predictive performance","we","the shapley additive explanations approach","irrelevant features","the optimal feature subset","input","convolutional neural networks","the deep-wet predictor","both cross-validation and independent tests","superior predictive performance","conventional machine learning classifiers","addition","extensive independent test","the-art","dbp prediction","accuracy","78.08%","mcc","auc","this superior performance","a tremendous predictive capacity","dbps","the web server","datasets","this study","https:\/\/deepwet-dna.monarcatechnical.com\/.","the community-wide effort","large-scale identification","potential dbps","three","three","78.08%","0.559","0.805"]},{"title":"DDoS attack traffic classification in SDN using deep learning","authors":["Nisha Ahuja","Debajyoti Mukhopadhyay","Gaurav Singal"],"abstract":"Software-defined networking will be a critical component of the networking domain as it transitions from a standard networking design to an automation network. To meet the needs of the current scenario, this architecture redesign becomes mandatory. Besides, machine learning (ML) and deep learning (DL) techniques provide a significant solution in network attack detection, traffic classification, etc. The DDoS attack is still wreaking havoc. Previous work for DDoS attack detection in SDN has not yielded significant results, so the author has used the most recent deep learning technique to detect the attacks. In this paper, we aim to classify the network traffic into normal and malicious classes based on features in the available dataset by using various deep learning techniques. TCP, UDP, and ICMP traffic are considered normal; however, malicious traffic includes TCP Syn Attack, UDP Flood, and ICMP Flood, all of which are DDoS attack traffic. The major contribution of this paper is the identification of novel features for DDoS attack detection. Novel features are logged into the CSV file to create the dataset, and machine learning algorithms are trained on the created SDN dataset. Various work which has already been done for DDoS attack detection either used a non-SDN dataset or the research data is not made public. A novel hybrid machine learning model is utilized to perform the classification. The dataset used by the ML\/DL algorithms is a collection of public datasets on DDoS attacks as well as an experimental DDoS dataset generated by us and publicly available on the Mendeley Data repository. A Python application performs the classification of traffic into one of the classes. From the various classifiers used, the accuracy score of 99.75% is achieved with Stacked Auto-Encoder Multi-layer Perceptron (SAE-MLP). To measure the effectiveness of the SDN-DDoS dataset, the other publicly available datasets are also evaluated against the same deep learning algorithms, and traffic classification accuracy is found to be significantly higher with the SDN-DDoS dataset. The attack detection time of 216.39\u00a0s also serve as experimental evidence.","doi":"10.1007\/s00779-023-01785-2","cleaned_title":"ddos attack traffic classification in sdn using deep learning","cleaned_abstract":"software-defined networking will be a critical component of the networking domain as it transitions from a standard networking design to an automation network. to meet the needs of the current scenario, this architecture redesign becomes mandatory. besides, machine learning (ml) and deep learning (dl) techniques provide a significant solution in network attack detection, traffic classification, etc. the ddos attack is still wreaking havoc. previous work for ddos attack detection in sdn has not yielded significant results, so the author has used the most recent deep learning technique to detect the attacks. in this paper, we aim to classify the network traffic into normal and malicious classes based on features in the available dataset by using various deep learning techniques. tcp, udp, and icmp traffic are considered normal; however, malicious traffic includes tcp syn attack, udp flood, and icmp flood, all of which are ddos attack traffic. the major contribution of this paper is the identification of novel features for ddos attack detection. novel features are logged into the csv file to create the dataset, and machine learning algorithms are trained on the created sdn dataset. various work which has already been done for ddos attack detection either used a non-sdn dataset or the research data is not made public. a novel hybrid machine learning model is utilized to perform the classification. the dataset used by the ml\/dl algorithms is a collection of public datasets on ddos attacks as well as an experimental ddos dataset generated by us and publicly available on the mendeley data repository. a python application performs the classification of traffic into one of the classes. from the various classifiers used, the accuracy score of 99.75% is achieved with stacked auto-encoder multi-layer perceptron (sae-mlp). to measure the effectiveness of the sdn-ddos dataset, the other publicly available datasets are also evaluated against the same deep learning algorithms, and traffic classification accuracy is found to be significantly higher with the sdn-ddos dataset. the attack detection time of 216.39 s also serve as experimental evidence.","key_phrases":["software-defined networking","a critical component","the networking domain","it","a standard networking design","an automation network","the needs","the current scenario","this architecture redesign","machine learning","ml","deep learning","(dl) techniques","a significant solution","network attack detection","traffic classification","the ddos attack","havoc","previous work","ddos attack detection","sdn","significant results","the author","the most recent deep learning technique","the attacks","this paper","we","the network traffic","normal and malicious classes","features","the available dataset","various deep learning techniques","tcp","udp","icmp traffic","malicious traffic","tcp syn attack","udp flood","icmp flood","all","which","ddos attack traffic","the major contribution","this paper","the identification","novel features","ddos attack detection","novel features","the csv file","the dataset","machine learning algorithms","the created sdn dataset","various work","which","ddos attack detection","a non-sdn dataset","the research data","a novel hybrid machine learning model","the classification","the dataset","the ml\/dl algorithms","a collection","public datasets","ddos attacks","an experimental ddos dataset","us","the mendeley data repository","a python application","the classification","traffic","the classes","the various classifiers","the accuracy score","99.75%","stacked auto-encoder multi-layer perceptron","sae-mlp","the effectiveness","the sdn-ddos dataset","the other publicly available datasets","the same deep learning algorithms","traffic classification accuracy","the sdn-ddos dataset","the attack detection time","216.39 s","experimental evidence","icmp","mendeley","99.75%","216.39"]},{"title":"A Deep Learning-Based Object Representation Algorithm for Smart Retail Management","authors":["Bin Liu"],"abstract":"This study underscores the vital role of object representation and detection in smart retail management systems for optimizing customer experiences and operational efficiency. The literature review reveals a preference for deep learning techniques, citing their superior accuracy compared to traditional methods. While acknowledging the challenges of achieving high accuracy and low computation costs simultaneously in deep learning-based object representation, the paper proposes a solution using the YOLOv7 framework. In order to navigate the ever-changing landscape of smart retail technologies, the study clarifies the potential scalability and flexibility of deep learning approaches. The method employs a custom dataset, and experimental results demonstrate the model\u2019s efficacy, showcasing accurate results and enhanced performance in various experiments and analyses.","doi":"10.1007\/s40031-024-01051-w","cleaned_title":"a deep learning-based object representation algorithm for smart retail management","cleaned_abstract":"this study underscores the vital role of object representation and detection in smart retail management systems for optimizing customer experiences and operational efficiency. the literature review reveals a preference for deep learning techniques, citing their superior accuracy compared to traditional methods. while acknowledging the challenges of achieving high accuracy and low computation costs simultaneously in deep learning-based object representation, the paper proposes a solution using the yolov7 framework. in order to navigate the ever-changing landscape of smart retail technologies, the study clarifies the potential scalability and flexibility of deep learning approaches. the method employs a custom dataset, and experimental results demonstrate the model\u2019s efficacy, showcasing accurate results and enhanced performance in various experiments and analyses.","key_phrases":["this study","the vital role","object representation","detection","smart retail management systems","customer experiences","operational efficiency","the literature review","a preference","deep learning techniques","their superior accuracy","traditional methods","the challenges","high accuracy","low computation costs","deep learning-based object representation","the paper","a solution","the yolov7 framework","order","the ever-changing landscape","smart retail technologies","the study","the potential scalability","flexibility","deep learning approaches","the method","a custom dataset","experimental results","the model\u2019s efficacy","accurate results","enhanced performance","various experiments","analyses"]},{"title":"Shallow and deep learning classifiers in medical image analysis","authors":["Francesco Prinzi","Tiziana Currieri","Salvatore Gaglio","Salvatore Vitabile"],"abstract":"An increasingly strong connection between artificial intelligence and medicine has enabled the development of predictive models capable of supporting physicians\u2019 decision-making. Artificial intelligence encompasses much more than machine learning, which nevertheless is its most cited and used sub-branch in the last decade. Since most clinical problems can be modeled through machine learning classifiers, it is essential to discuss their main elements. This review aims to give primary educational insights on the most accessible and widely employed classifiers in radiology field, distinguishing between \u201cshallow\u201d learning (i.e., traditional machine learning) algorithms, including support vector machines, random forest and XGBoost, and \u201cdeep\u201d learning architectures including convolutional neural networks and vision transformers. In addition, the paper outlines the key steps for classifiers training and highlights the differences between the most common algorithms and architectures. Although the choice of an algorithm depends on the task and dataset dealing with, general guidelines for classifier selection are proposed in relation to task analysis, dataset size, explainability requirements, and available computing resources. Considering the enormous interest in these innovative models and architectures, the problem of machine learning algorithms interpretability is finally discussed, providing a future perspective on trustworthy artificial intelligence.Relevance statement The growing synergy between artificial intelligence and medicine fosters predictive models aiding physicians. Machine learning classifiers, from shallow learning to deep learning, are offering crucial insights for the development of clinical decision support systems in healthcare. Explainability is a key feature of models that leads systems toward integration into clinical practice.\nKey points\n\u2022 Training a shallow classifier requires extracting disease-related features from region of interests (e.g., radiomics).\u2022 Deep classifiers implement automatic feature extraction and classification.\u2022 The classifier selection\u00a0is based on data and computational resources availability, task, and explanation needs.Graphical Abstract","doi":"10.1186\/s41747-024-00428-2","cleaned_title":"shallow and deep learning classifiers in medical image analysis","cleaned_abstract":"an increasingly strong connection between artificial intelligence and medicine has enabled the development of predictive models capable of supporting physicians\u2019 decision-making. artificial intelligence encompasses much more than machine learning, which nevertheless is its most cited and used sub-branch in the last decade. since most clinical problems can be modeled through machine learning classifiers, it is essential to discuss their main elements. this review aims to give primary educational insights on the most accessible and widely employed classifiers in radiology field, distinguishing between \u201cshallow\u201d learning (i.e., traditional machine learning) algorithms, including support vector machines, random forest and xgboost, and \u201cdeep\u201d learning architectures including convolutional neural networks and vision transformers. in addition, the paper outlines the key steps for classifiers training and highlights the differences between the most common algorithms and architectures. although the choice of an algorithm depends on the task and dataset dealing with, general guidelines for classifier selection are proposed in relation to task analysis, dataset size, explainability requirements, and available computing resources. considering the enormous interest in these innovative models and architectures, the problem of machine learning algorithms interpretability is finally discussed, providing a future perspective on trustworthy artificial intelligence.relevance statement the growing synergy between artificial intelligence and medicine fosters predictive models aiding physicians. machine learning classifiers, from shallow learning to deep learning, are offering crucial insights for the development of clinical decision support systems in healthcare. explainability is a key feature of models that leads systems toward integration into clinical practice. key points \u2022 training a shallow classifier requires extracting disease-related features from region of interests (e.g., radiomics).\u2022 deep classifiers implement automatic feature extraction and classification.\u2022 the classifier selection is based on data and computational resources availability, task, and explanation needs.graphical abstract","key_phrases":["an increasingly strong connection","artificial intelligence","medicine","the development","predictive models","physicians\u2019 decision-making","artificial intelligence","machine learning","which","sub","the last decade","most clinical problems","machine learning classifiers","it","their main elements","this review","primary educational insights","the most accessible and widely employed classifiers","radiology field","\u201cshallow\u201d learning","(i.e., traditional machine learning","algorithms","support vector machines","random forest","xgboost","\u201cdeep\u201d learning architectures","convolutional neural networks","vision transformers","addition","the paper","the key steps","classifiers training","the differences","the most common algorithms","architectures","the choice","an algorithm","the task","general guidelines","classifier selection","relation","task analysis","dataset size","explainability requirements","available computing resources","the enormous interest","these innovative models","architectures","the problem","machine learning algorithms interpretability","a future perspective","trustworthy artificial intelligence.relevance statement","the growing synergy","artificial intelligence and medicine fosters predictive models","physicians","machine learning classifiers","shallow learning","deep learning","crucial insights","the development","clinical decision support systems","healthcare","explainability","a key feature","models","that","systems","integration","clinical practice","key points \u2022","a shallow classifier","disease-related features","region","interests","deep classifiers","automatic feature extraction","classification.\u2022","the classifier selection","data","computational resources availability","task","the last decade","\u2022"]},{"title":"Survey of Research on Application of Deep Learning in Modulation Recognition","authors":["Yongjun Sun","Wanting Wu"],"abstract":"Modulation recognition is an important research branch in the field of communication, which is widely used in civil and military fields. The classic methods depend on decision theory, signal feature and the choice of classifier, while the deep learning network can get the signal feature directly from the data, and its recognition accuracy is higher than the classic methods. This paper summarized the application of deep learning in modulation recognition. Firstly, the basic concept of deep learning and the common network structure in modulation recognition were introduced. Secondly, the common signal forms and signal preprocessing technologies of input deep learning network were given, and the characteristics and performance of different deep learning networks were summarized and analyzed. Finally, the challenges and future research directions in this field were discussed.","doi":"10.1007\/s11277-023-10826-1","cleaned_title":"survey of research on application of deep learning in modulation recognition","cleaned_abstract":"modulation recognition is an important research branch in the field of communication, which is widely used in civil and military fields. the classic methods depend on decision theory, signal feature and the choice of classifier, while the deep learning network can get the signal feature directly from the data, and its recognition accuracy is higher than the classic methods. this paper summarized the application of deep learning in modulation recognition. firstly, the basic concept of deep learning and the common network structure in modulation recognition were introduced. secondly, the common signal forms and signal preprocessing technologies of input deep learning network were given, and the characteristics and performance of different deep learning networks were summarized and analyzed. finally, the challenges and future research directions in this field were discussed.","key_phrases":["modulation recognition","an important research branch","the field","communication","which","civil and military fields","the classic methods","decision theory","classifier","the deep learning network","the signal feature","the data","its recognition accuracy","the classic methods","this paper","the application","deep learning","modulation recognition","the basic concept","deep learning","the common network structure","modulation recognition","the common signal forms","signal","preprocessing technologies","input deep learning network","the characteristics","performance","different deep learning networks","the challenges","future research directions","this field","firstly","secondly"]},{"title":"Ocular biomarkers: useful incidental findings by deep learning algorithms in fundus photographs","authors":["Eve Martin","Angus G. Cook","Shaun M. Frost","Angus W. Turner","Fred K. Chen","Ian L. McAllister","Janis M. Nolde","Markus P. Schlaich"],"abstract":"Background\/ObjectivesArtificial intelligence can assist with ocular image analysis for screening and diagnosis, but it is not yet capable of autonomous full-spectrum screening. Hypothetically, false-positive results may have unrealized screening potential arising from signals persisting despite training and\/or ambiguous signals such as from biomarker overlap or high comorbidity. The study aimed to explore the potential to detect clinically useful incidental ocular biomarkers by screening fundus photographs of hypertensive adults using diabetic deep learning algorithms.Subjects\/MethodsPatients referred for treatment-resistant hypertension were imaged at a hospital unit in Perth, Australia, between 2016 and 2022. The same 45\u00b0 colour fundus photograph selected for each of the 433 participants imaged was processed by three deep learning algorithms. Two expert retinal specialists graded all false-positive results for diabetic retinopathy in non-diabetic participants.ResultsOf the 29 non-diabetic participants misclassified as positive for diabetic retinopathy, 28 (97%) had clinically useful retinal biomarkers. The models designed to screen for fewer diseases captured more incidental disease. All three algorithms showed a positive correlation between severity of hypertensive retinopathy and misclassified diabetic retinopathy.ConclusionsThe results suggest that diabetic deep learning models may be responsive to hypertensive and other clinically useful retinal biomarkers within an at-risk, hypertensive cohort. Observing that models trained for fewer diseases captured more incidental pathology increases confidence in signalling hypotheses aligned with using self-supervised learning to develop autonomous comprehensive screening. Meanwhile, non-referable and false-positive outputs of other deep learning screening models could be explored for immediate clinical use in other populations.","doi":"10.1038\/s41433-024-03085-2","cleaned_title":"ocular biomarkers: useful incidental findings by deep learning algorithms in fundus photographs","cleaned_abstract":"background\/objectivesartificial intelligence can assist with ocular image analysis for screening and diagnosis, but it is not yet capable of autonomous full-spectrum screening. hypothetically, false-positive results may have unrealized screening potential arising from signals persisting despite training and\/or ambiguous signals such as from biomarker overlap or high comorbidity. the study aimed to explore the potential to detect clinically useful incidental ocular biomarkers by screening fundus photographs of hypertensive adults using diabetic deep learning algorithms.subjects\/methodspatients referred for treatment-resistant hypertension were imaged at a hospital unit in perth, australia, between 2016 and 2022. the same 45\u00b0 colour fundus photograph selected for each of the 433 participants imaged was processed by three deep learning algorithms. two expert retinal specialists graded all false-positive results for diabetic retinopathy in non-diabetic participants.resultsof the 29 non-diabetic participants misclassified as positive for diabetic retinopathy, 28 (97%) had clinically useful retinal biomarkers. the models designed to screen for fewer diseases captured more incidental disease. all three algorithms showed a positive correlation between severity of hypertensive retinopathy and misclassified diabetic retinopathy.conclusionsthe results suggest that diabetic deep learning models may be responsive to hypertensive and other clinically useful retinal biomarkers within an at-risk, hypertensive cohort. observing that models trained for fewer diseases captured more incidental pathology increases confidence in signalling hypotheses aligned with using self-supervised learning to develop autonomous comprehensive screening. meanwhile, non-referable and false-positive outputs of other deep learning screening models could be explored for immediate clinical use in other populations.","key_phrases":["background\/objectivesartificial intelligence","ocular image analysis","screening","diagnosis","it","autonomous full-spectrum screening","false-positive results","screening potential","signals","training and\/or ambiguous signals","biomarker overlap","high comorbidity","the study","the potential","clinically useful incidental ocular biomarkers","fundus photographs","hypertensive adults","diabetic deep learning algorithms.subjects\/methodspatients","treatment-resistant hypertension","a hospital unit","perth","australia","the same 45\u00b0 colour fundus photograph","each","the 433 participants","three deep learning algorithms","two expert retinal specialists","all false-positive results","diabetic retinopathy","non-diabetic participants.resultsof","the 29 non-diabetic participants","diabetic retinopathy","28 (97%","clinically useful retinal biomarkers","the models","fewer diseases","more incidental disease","all three algorithms","a positive correlation","severity","hypertensive retinopathy","retinopathy.conclusionsthe results","diabetic deep learning models","hypertensive and other clinically useful retinal biomarkers","an at-risk, hypertensive cohort","that models","fewer diseases","more incidental pathology","confidence","hypotheses","self-supervised learning","autonomous comprehensive screening","non-referable and false-positive outputs","other deep learning screening models","immediate clinical use","other populations","perth","australia","between 2016 and 2022","the same 45","433","three","two","29","28","97%","three"]},{"title":"Deep residual learning with Anscombe transformation for low-dose digital tomosynthesis","authors":["Youngjin Lee","Seungwan Lee","Chanrok Park"],"abstract":"Deep learning-based convolutional neural networks (CNNs) have been proposed for enhancing the quality of digital tomosynthesis (DTS) images. However, the direct applications of the conventional CNNs for low-dose DTS imaging are limited to provide acceptable image quality due to the inaccurate recognition of complex texture patterns. In this study, a deep residual learning network combined with the Anscombe transformation was proposed for simplifying the complex texture and restoring the low-dose DTS image quality. The proposed network consisted of convolution layers, max-pooling layers, up-sampling layers, and skip connections. The network training was performed to learn the residual images between the ground-truth and low-dose projections, which were converted using the Anscombe transformation. As a result, the proposed network enhanced the quantitative accuracy and noise characteristic of DTS images by 1.01\u20131.27 and 1.14\u20131.71 times, respectively, in comparison to low-dose DTS images and other deep learning networks. The spatial resolution of the DTS image restored using the proposed network was 1.12 times higher than that obtained using a deep image learning network. In conclusion, the proposed network can restore the low-dose DTS image quality and provide an optimal model for low-dose DTS imaging.","doi":"10.1007\/s40042-024-01117-4","cleaned_title":"deep residual learning with anscombe transformation for low-dose digital tomosynthesis","cleaned_abstract":"deep learning-based convolutional neural networks (cnns) have been proposed for enhancing the quality of digital tomosynthesis (dts) images. however, the direct applications of the conventional cnns for low-dose dts imaging are limited to provide acceptable image quality due to the inaccurate recognition of complex texture patterns. in this study, a deep residual learning network combined with the anscombe transformation was proposed for simplifying the complex texture and restoring the low-dose dts image quality. the proposed network consisted of convolution layers, max-pooling layers, up-sampling layers, and skip connections. the network training was performed to learn the residual images between the ground-truth and low-dose projections, which were converted using the anscombe transformation. as a result, the proposed network enhanced the quantitative accuracy and noise characteristic of dts images by 1.01\u20131.27 and 1.14\u20131.71 times, respectively, in comparison to low-dose dts images and other deep learning networks. the spatial resolution of the dts image restored using the proposed network was 1.12 times higher than that obtained using a deep image learning network. in conclusion, the proposed network can restore the low-dose dts image quality and provide an optimal model for low-dose dts imaging.","key_phrases":["deep learning-based convolutional neural networks","cnns","the quality","digital tomosynthesis (dts) images","the direct applications","the conventional cnns","low-dose dts imaging","acceptable image quality","the inaccurate recognition","complex texture patterns","this study","a deep residual learning network","the anscombe transformation","the complex texture","the low-dose dts image quality","the proposed network","convolution layers","max-pooling layers","up-sampling layers","skip connections","the network training","the residual images","the ground-truth and low-dose projections","which","the anscombe transformation","a result","the proposed network","the quantitative accuracy","dts images","1.01\u20131.27 and 1.14\u20131.71 times","comparison","low-dose dts images","other deep learning networks","the spatial resolution","the dts image","the proposed network","that","a deep image learning network","conclusion","the proposed network","the low-dose dts image quality","an optimal model","low-dose dts imaging","max","1.01\u20131.27","1.14\u20131.71","1.12"]},{"title":"Colon cancer diagnosis by means of explainable deep learning","authors":["Marcello Di Giammarco","Fabio Martinelli","Antonella Santone","Mario Cesarelli","Francesco Mercaldo"],"abstract":"Early detection of the adenocarcinoma cancer in colon tissue by means of explainable deep learning, by classifying histological images and providing visual explainability on model prediction. Considering that in recent years, deep learning techniques have emerged as powerful techniques in medical image analysis, offering unprecedented accuracy and efficiency, in this paper we propose a method to automatically detect the presence of cancerous cells in colon tissue images. Various deep learning architectures are considered, with the aim of considering the best one in terms of quantitative and qualitative results. As a matter of fact, we consider qualitative results by taking into account the so-called prediction explainability, by providing a way to highlight on the tissue images the areas that from the model point of view are related to the presence of colon cancer. The experimental analysis, performed on 10,000 colon issue images, showed the effectiveness of the proposed method by obtaining an accuracy equal to 0.99. The experimental analysis shows that the proposed method can be successfully exploited for colon cancer detection and localisation from tissue images.","doi":"10.1038\/s41598-024-63659-8","cleaned_title":"colon cancer diagnosis by means of explainable deep learning","cleaned_abstract":"early detection of the adenocarcinoma cancer in colon tissue by means of explainable deep learning, by classifying histological images and providing visual explainability on model prediction. considering that in recent years, deep learning techniques have emerged as powerful techniques in medical image analysis, offering unprecedented accuracy and efficiency, in this paper we propose a method to automatically detect the presence of cancerous cells in colon tissue images. various deep learning architectures are considered, with the aim of considering the best one in terms of quantitative and qualitative results. as a matter of fact, we consider qualitative results by taking into account the so-called prediction explainability, by providing a way to highlight on the tissue images the areas that from the model point of view are related to the presence of colon cancer. the experimental analysis, performed on 10,000 colon issue images, showed the effectiveness of the proposed method by obtaining an accuracy equal to 0.99. the experimental analysis shows that the proposed method can be successfully exploited for colon cancer detection and localisation from tissue images.","key_phrases":["early detection","the adenocarcinoma cancer","colon tissue","means","explainable deep learning","histological images","visual explainability","model prediction","recent years","deep learning techniques","powerful techniques","medical image analysis","unprecedented accuracy","efficiency","this paper","we","a method","the presence","cancerous cells","colon tissue images","various deep learning architectures","the aim","terms","quantitative and qualitative results","a matter","fact","we","qualitative results","account","the so-called prediction explainability","a way","the tissue","the areas","the model point","view","the presence","colon cancer","the experimental analysis","10,000 colon issue images","the effectiveness","the proposed method","an accuracy","the experimental analysis","the proposed method","colon cancer detection","localisation","tissue images","recent years","10,000","0.99"]},{"title":"Deep learning based multiclass classification for citrus anomaly detection in agriculture","authors":["Ebru Erg\u00fcn"],"abstract":"In regions where citrus crops are threatened by diseases caused by fungi, bacteria, pests and viruses, growers are actively seeking automated technologies that can accurately detect citrus anomalies to minimize economic losses. Recent advances in deep learning techniques have shown potential in automating and improving the accuracy of citrus anomaly categorization. This research explores the use of deep learning methods, specifically DenseNet, to construct robust models capable of accurately distinguishing between different types of citrus anomalies. The dataset consists of high-resolution images of different orange leaves of the species Citrus sinensis osbeck, collected from orange groves in the states of Tamaulipas and San Luis Potosi in northeastern Mexico was used in study. Experimental results demonstrated the effectiveness of the proposed deep learning models in simultaneously identifying 12 different classes of citrus anomalies. Evaluation metrics, including accuracy, recall, precision and the confusion matrix, underscore the discriminative power of the models. Among the convolutional neural network architectures used, DenseNet achieved the highest classification accuracy at 99.50%. The study concluded by highlighting the potential for scalable and effective citrus anomaly classification and management using deep learning-based systems.","doi":"10.1007\/s11760-024-03452-2","cleaned_title":"deep learning based multiclass classification for citrus anomaly detection in agriculture","cleaned_abstract":"in regions where citrus crops are threatened by diseases caused by fungi, bacteria, pests and viruses, growers are actively seeking automated technologies that can accurately detect citrus anomalies to minimize economic losses. recent advances in deep learning techniques have shown potential in automating and improving the accuracy of citrus anomaly categorization. this research explores the use of deep learning methods, specifically densenet, to construct robust models capable of accurately distinguishing between different types of citrus anomalies. the dataset consists of high-resolution images of different orange leaves of the species citrus sinensis osbeck, collected from orange groves in the states of tamaulipas and san luis potosi in northeastern mexico was used in study. experimental results demonstrated the effectiveness of the proposed deep learning models in simultaneously identifying 12 different classes of citrus anomalies. evaluation metrics, including accuracy, recall, precision and the confusion matrix, underscore the discriminative power of the models. among the convolutional neural network architectures used, densenet achieved the highest classification accuracy at 99.50%. the study concluded by highlighting the potential for scalable and effective citrus anomaly classification and management using deep learning-based systems.","key_phrases":["regions","citrus crops","diseases","fungi","bacteria","pests","viruses","growers","automated technologies","that","citrus anomalies","economic losses","recent advances","deep learning techniques","potential","the accuracy","citrus anomaly categorization","this research","the use","deep learning methods","specifically densenet","robust models","different types","citrus anomalies","the dataset","high-resolution images","different orange leaves","the species citrus sinensis osbeck","orange groves","the states","tamaulipas","san luis potosi","northeastern mexico","study","experimental results","the effectiveness","the proposed deep learning models","12 different classes","citrus anomalies","evaluation metrics","accuracy","recall","precision","the confusion matrix","the discriminative power","the models","the convolutional neural network architectures","densenet","the highest classification accuracy","99.50%","the study","the potential","scalable and effective citrus anomaly classification","management","deep learning-based systems","san luis","mexico","12","99.50%"]},{"title":"Ensemble of deep learning and machine learning approach for classification of handwritten Hindi numerals","authors":["Danveer Rajpal","Akhil Ranjan Garg"],"abstract":"Given the vast range of factors, including shape, size, skew, and orientation of handwritten numerals, their machine-based recognition is a difficult challenge for researchers in the pattern recognition field. Due to the abundance of curves and resembling shapes of the symbols, the recognition of Devnagari numerals can leverage the difficulty level of the recognition. The suggested low-classification-cost method for obtaining fine features from given numeral images used benchmark deep learning models, VGG-16Net, VGG-19Net, ResNet-50, and Inception-v3, to address these issues. Principal component analysis, a powerful dimensionality reduction method, was used to efficiently reduce the number of dimensions in the information that pre-trained deep convolutional neural network models provided. The method for improving recognition accuracy by fusing features was provided in the scheme. A machine learning algorithm: support vector machine was employed for the recognition task due to its capacity to distinguish between patterns belonging to distinct classes. The system was able to obtain a recognition accuracy of 99.72% and was effective in demonstrating the importance of ensemble machine learning and deep learning approaches.","doi":"10.1186\/s44147-023-00252-2","cleaned_title":"ensemble of deep learning and machine learning approach for classification of handwritten hindi numerals","cleaned_abstract":"given the vast range of factors, including shape, size, skew, and orientation of handwritten numerals, their machine-based recognition is a difficult challenge for researchers in the pattern recognition field. due to the abundance of curves and resembling shapes of the symbols, the recognition of devnagari numerals can leverage the difficulty level of the recognition. the suggested low-classification-cost method for obtaining fine features from given numeral images used benchmark deep learning models, vgg-16net, vgg-19net, resnet-50, and inception-v3, to address these issues. principal component analysis, a powerful dimensionality reduction method, was used to efficiently reduce the number of dimensions in the information that pre-trained deep convolutional neural network models provided. the method for improving recognition accuracy by fusing features was provided in the scheme. a machine learning algorithm: support vector machine was employed for the recognition task due to its capacity to distinguish between patterns belonging to distinct classes. the system was able to obtain a recognition accuracy of 99.72% and was effective in demonstrating the importance of ensemble machine learning and deep learning approaches.","key_phrases":["the vast range","factors","shape","size","skew","orientation","handwritten numerals","their machine-based recognition","a difficult challenge","researchers","the pattern recognition field","the abundance","curves","shapes","the symbols","the recognition","devnagari numerals","the difficulty level","the recognition","the suggested low-classification-cost method","fine features","numeral images","benchmark deep learning models","vgg-16net","vgg-19net","resnet-50","inception-v3","these issues","principal component analysis","a powerful dimensionality reduction method","the number","dimensions","the information","that","pre-trained deep convolutional neural network models","the method","recognition accuracy","fusing features","the scheme","a machine learning","algorithm","support vector machine","the recognition task","its capacity","patterns","distinct classes","the system","a recognition accuracy","99.72%","the importance","ensemble machine learning","deep learning approaches","resnet-50","99.72%"]},{"title":"Strengthening KMS Security with Advanced Cryptography, Machine Learning, Deep Learning, and IoT Technologies","authors":["Justin Onyarin Ogala","Shahnawaz Ahmad","Iman Shakeel","Javed Ahmad","Shabana Mehfuz"],"abstract":"This paper presents an innovative approach to strengthening Key Management Systems (KMS) against the escalating landscape of cyber threats by integrating advanced cryptographic technologies, machine learning, deep learning, and the Internet of Things (IoT). As digital reliance and cyber-attacks surge, strengthening KMS security becomes paramount. Our research provides a comprehensive overview of the state-of-the-art in cloud data security, identifying key vulnerabilities in existing KMS. The paper also outlines a distinctive framework based on the combined application of advanced cryptography, machine learning, deep learning, and IoT, which represents a novel approach in the quest for robust KMS security. Our experimental results substantiate the efficacy of this unique blend of technologies, providing solid empirical evidence that such a fusion can successfully strengthen KMS against potential threats. As technologies and threat landscapes continue to evolve, our framework can serve as a benchmark for future research and practical implementations. It highlights the potential of integrated technological solutions to counter complex cybersecurity issues. Moreover, the approach we've developed can be adapted and expanded to cater to the specific needs of different sectors, such as finance, healthcare, and e-commerce, which are particularly vulnerable to cyber threats. The novelty of our work lies in the amalgamation of the four technologies and the creation of an empirically backed, robust framework, marking a significant stride in KMS security.","doi":"10.1007\/s42979-023-02073-9","cleaned_title":"strengthening kms security with advanced cryptography, machine learning, deep learning, and iot technologies","cleaned_abstract":"this paper presents an innovative approach to strengthening key management systems (kms) against the escalating landscape of cyber threats by integrating advanced cryptographic technologies, machine learning, deep learning, and the internet of things (iot). as digital reliance and cyber-attacks surge, strengthening kms security becomes paramount. our research provides a comprehensive overview of the state-of-the-art in cloud data security, identifying key vulnerabilities in existing kms. the paper also outlines a distinctive framework based on the combined application of advanced cryptography, machine learning, deep learning, and iot, which represents a novel approach in the quest for robust kms security. our experimental results substantiate the efficacy of this unique blend of technologies, providing solid empirical evidence that such a fusion can successfully strengthen kms against potential threats. as technologies and threat landscapes continue to evolve, our framework can serve as a benchmark for future research and practical implementations. it highlights the potential of integrated technological solutions to counter complex cybersecurity issues. moreover, the approach we've developed can be adapted and expanded to cater to the specific needs of different sectors, such as finance, healthcare, and e-commerce, which are particularly vulnerable to cyber threats. the novelty of our work lies in the amalgamation of the four technologies and the creation of an empirically backed, robust framework, marking a significant stride in kms security.","key_phrases":["this paper","an innovative approach","key management systems","kms","the escalating landscape","cyber threats","advanced cryptographic technologies","machine learning","deep learning","the internet","things","iot","digital reliance","cyber-attacks","kms security","our research","a comprehensive overview","the state","the-art","cloud data security","key vulnerabilities","existing kms","the paper","a distinctive framework","the combined application","advanced cryptography","machine learning","deep learning","iot","which","a novel approach","the quest","robust kms security","our experimental results","the efficacy","this unique blend","technologies","solid empirical evidence","such a fusion","kms","potential threats","technologies","threat landscapes","our framework","a benchmark","future research and practical implementations","it","the potential","integrated technological solutions","complex cybersecurity issues","the approach","we","the specific needs","different sectors","finance","healthcare","e","-","commerce","which","cyber threats","the novelty","our work","the amalgamation","the four technologies","the creation","an empirically backed, robust framework","a significant stride","kms security","four"]},{"title":"Skin cancer detection using ensemble of machine learning and deep learning techniques","authors":["Jitendra V. Tembhurne","Nachiketa Hebbar","Hemprasad Y. Patil","Tausif Diwan"],"abstract":"Skin cancer is one of the most common forms of cancer, which makes it pertinent to be able to diagnose it accurately. In particular, melanoma is a form of skin cancer that is fatal and accounts for 6 of every 7-skin cancer related death. Moreover, in hospitals where dermatologists have to diagnose multiple cases of skin cancer, there are high possibilities of false negatives in diagnosis. To avoid such incidents, there here has been exhaustive research conducted by the research community all over the world to build highly accurate automated tools for skin cancer detection. In this paper, we introduce a novel approach of combining machine learning and deep learning techniques to solve the problem of skin cancer detection. The deep learning model uses state-of-the-art neural networks to extract features from images whereas the machine learning model processes image features which are obtained after performing the techniques such as Contourlet Transform and Local Binary Pattern Histogram. Meaningful feature extraction is crucial for any image classification roblem. As a result, by combining the manual and automated features, our designed model achieves a higher accuracy of 93% with an individual recall score of 99.7% and 86% for the benign and malignant forms of cancer, respectively. We benchmarked the model on publicly available Kaggle dataset containing processed images from ISIC Archive dataset. The proposed ensemble outperforms both expert dermatologists as well as other state-of-the-art deep learning and machine learning methods. Thus, this novel method can be of high assistance to dermatologists to help prevent any misdiagnosis.","doi":"10.1007\/s11042-023-14697-3","cleaned_title":"skin cancer detection using ensemble of machine learning and deep learning techniques","cleaned_abstract":"skin cancer is one of the most common forms of cancer, which makes it pertinent to be able to diagnose it accurately. in particular, melanoma is a form of skin cancer that is fatal and accounts for 6 of every 7-skin cancer related death. moreover, in hospitals where dermatologists have to diagnose multiple cases of skin cancer, there are high possibilities of false negatives in diagnosis. to avoid such incidents, there here has been exhaustive research conducted by the research community all over the world to build highly accurate automated tools for skin cancer detection. in this paper, we introduce a novel approach of combining machine learning and deep learning techniques to solve the problem of skin cancer detection. the deep learning model uses state-of-the-art neural networks to extract features from images whereas the machine learning model processes image features which are obtained after performing the techniques such as contourlet transform and local binary pattern histogram. meaningful feature extraction is crucial for any image classification roblem. as a result, by combining the manual and automated features, our designed model achieves a higher accuracy of 93% with an individual recall score of 99.7% and 86% for the benign and malignant forms of cancer, respectively. we benchmarked the model on publicly available kaggle dataset containing processed images from isic archive dataset. the proposed ensemble outperforms both expert dermatologists as well as other state-of-the-art deep learning and machine learning methods. thus, this novel method can be of high assistance to dermatologists to help prevent any misdiagnosis.","key_phrases":["skin cancer","the most common forms","cancer","which","it","it","melanoma","a form","skin cancer","that","every 7-skin cancer related death","hospitals","dermatologists","multiple cases","skin cancer","high possibilities","false negatives","diagnosis","such incidents","exhaustive research","the research community","the world","highly accurate automated tools","skin cancer detection","this paper","we","a novel approach","machine learning","deep learning techniques","the problem","skin cancer detection","the deep learning model","the-art","features","images","the machine learning model","image features","which","the techniques","contourlet transform","local binary pattern histogram","meaningful feature extraction","any image classification roblem","a result","the manual and automated features","our designed model","a higher accuracy","93%","an individual recall score","99.7%","86%","the benign and malignant forms","cancer","we","the model","publicly available kaggle dataset","processed images","isic archive dataset","the proposed ensemble outperforms","both expert dermatologists","the-art","this novel method","high assistance","dermatologists","any misdiagnosis","one","6","7","93%","99.7% and","86%"]},{"title":"Deep transfer learning for tool condition monitoring under different processing conditions","authors":["Yongqing Wang","Mengmeng Niu","Kuo Liu","Haibo Liu","Bo Qin","Yiming Cui"],"abstract":"Deep learning methods have developed rapidly in the field of tool condition monitoring, but due to the complexity and diversity of working conditions, it is difficult to ensure the high accuracy, strong generalization performance, and wide applicability of monitoring models. Therefore, this article proposes a deep transfer learning method with center loss for tool condition monitoring under different processing conditions. Firstly, Deep Extreme Learning Machine (DELM) is used to extract sample features and tool condition monitoring, and deep coral is integrated into the last feature extraction layer of DELM for domain adaptation. Secondly, the center loss is introduced into the transfer learning model to improve the intra-class compactness by minimizing the center loss, thereby obtaining a broader decision boundary. A tool wear experiment was conducted on a milling machine. Research shows that the proposed method can achieve the tool condition monitoring under different processing conditions. The introduction of central loss is beneficial for promoting the separation of samples from different categories in the target domain and effectively improves the model\u2019s applicability. Compared with other domain adaptation methods, this method has better accuracy and generalization ability.","doi":"10.1007\/s00170-024-13713-6","cleaned_title":"deep transfer learning for tool condition monitoring under different processing conditions","cleaned_abstract":"deep learning methods have developed rapidly in the field of tool condition monitoring, but due to the complexity and diversity of working conditions, it is difficult to ensure the high accuracy, strong generalization performance, and wide applicability of monitoring models. therefore, this article proposes a deep transfer learning method with center loss for tool condition monitoring under different processing conditions. firstly, deep extreme learning machine (delm) is used to extract sample features and tool condition monitoring, and deep coral is integrated into the last feature extraction layer of delm for domain adaptation. secondly, the center loss is introduced into the transfer learning model to improve the intra-class compactness by minimizing the center loss, thereby obtaining a broader decision boundary. a tool wear experiment was conducted on a milling machine. research shows that the proposed method can achieve the tool condition monitoring under different processing conditions. the introduction of central loss is beneficial for promoting the separation of samples from different categories in the target domain and effectively improves the model\u2019s applicability. compared with other domain adaptation methods, this method has better accuracy and generalization ability.","key_phrases":["deep learning methods","the field","tool condition monitoring","the complexity","diversity","working conditions","it","the high accuracy, strong generalization performance","wide applicability","monitoring models","this article","a deep transfer learning method","center loss","tool condition monitoring","different processing conditions","deep extreme learning machine","(delm","sample features","tool condition monitoring","deep coral","the last feature extraction layer","delm","domain adaptation","the center loss","the transfer learning model","the intra-class compactness","the center loss","a broader decision boundary","a tool wear experiment","a milling machine","research","the proposed method","the tool condition monitoring","different processing conditions","the introduction","central loss","the separation","samples","different categories","the target domain","the model\u2019s applicability","other domain adaptation methods","this method","better accuracy","generalization ability","firstly","secondly"]},{"title":"Breast Mammograms Diagnosis Using Deep Learning: State of Art Tutorial Review","authors":["Osama Bin Naeem","Yasir Saleem","M. Usman Ghani Khan","Amjad Rehman Khan","Tanzila Saba","Saeed Ali Bahaj","Noor Ayesha"],"abstract":"Usually, screening (mostly mammography) is used by radiologists to manually detect breast cancer. The likelihood of identifying suspected cases as false positives or false negatives is significant, contingent on the experience of the radiologist and the kind of imaging screening device\/method utilized. The confirmation of the type of tumour seen by the radiologist is sent for histological investigation (microscopic analysis) through a biopsy, where the tumor's grade and stage, which are used in the latter stages of treatment, are ascertained through biopsy. However, a secondary issue with the cancer detection process is that only 15 to 30% of instances that are referred for biopsy result in malignant findings. Since deep learning demonstrated remarkable performance in visual recognition challenges, it has been widely applied to a variety of tasks. Similar examples include deep learning applications in healthcare, which are gaining a lot of interest from the research community. Deep learning is used to identify, categories tumours, and breast cancer is a significant global health concern. The medical sciences could now make more accurate diagnoses and detections due to recent advancements in machine learning techniques. Hence due to systems potential accuracy, it could offer optimistic outcomes when used to read malignant images. In imaging domains, deep learning-based methods have achieved remarkable success in constituent segmentation (UNet), localization (DenseNet), and classification (VGG-19). This study examines, how deep learning methods are assisting in the highly accurate diagnosis of benign or malignant tumours based on screened images. In contrast to a mammogram, which is covered in detail, this paper briefly discusses imaging methods for cancer detection. Early detection and cost effectiveness are two main benefits of applying machine learning and deep learning techniques to mammograms.","doi":"10.1007\/s11831-023-10052-9","cleaned_title":"breast mammograms diagnosis using deep learning: state of art tutorial review","cleaned_abstract":"usually, screening (mostly mammography) is used by radiologists to manually detect breast cancer. the likelihood of identifying suspected cases as false positives or false negatives is significant, contingent on the experience of the radiologist and the kind of imaging screening device\/method utilized. the confirmation of the type of tumour seen by the radiologist is sent for histological investigation (microscopic analysis) through a biopsy, where the tumor's grade and stage, which are used in the latter stages of treatment, are ascertained through biopsy. however, a secondary issue with the cancer detection process is that only 15 to 30% of instances that are referred for biopsy result in malignant findings. since deep learning demonstrated remarkable performance in visual recognition challenges, it has been widely applied to a variety of tasks. similar examples include deep learning applications in healthcare, which are gaining a lot of interest from the research community. deep learning is used to identify, categories tumours, and breast cancer is a significant global health concern. the medical sciences could now make more accurate diagnoses and detections due to recent advancements in machine learning techniques. hence due to systems potential accuracy, it could offer optimistic outcomes when used to read malignant images. in imaging domains, deep learning-based methods have achieved remarkable success in constituent segmentation (unet), localization (densenet), and classification (vgg-19). this study examines, how deep learning methods are assisting in the highly accurate diagnosis of benign or malignant tumours based on screened images. in contrast to a mammogram, which is covered in detail, this paper briefly discusses imaging methods for cancer detection. early detection and cost effectiveness are two main benefits of applying machine learning and deep learning techniques to mammograms.","key_phrases":["(mostly mammography","radiologists","breast cancer","the likelihood","suspected cases","false positives","false negatives","the experience","the radiologist","the kind","device\/method","the confirmation","the type","tumour","the radiologist","histological investigation","microscopic analysis","a biopsy","the tumor's grade","stage","which","the latter stages","treatment","biopsy","a secondary issue","the cancer detection process","only 15 to 30%","instances","that","biopsy result","malignant findings","deep learning","remarkable performance","visual recognition challenges","it","a variety","tasks","similar examples","deep learning applications","healthcare","which","a lot","interest","the research community","deep learning","categories tumours","breast cancer","a significant global health concern","the medical sciences","more accurate diagnoses","detections","recent advancements","machine learning techniques","systems potential accuracy","it","optimistic outcomes","malignant images","imaging","domains","deep learning-based methods","remarkable success","constituent segmentation","unet","localization","densenet","classification","vgg-19","this study examines","deep learning methods","the highly accurate diagnosis","benign or malignant tumours","screened images","contrast","a mammogram","which","detail","this paper","imaging methods","cancer detection","early detection and cost effectiveness","two main benefits","machine learning","deep learning techniques","mammograms","only 15 to","two"]},{"title":"A Survey on ensemble learning under the era of deep learning","authors":["Yongquan Yang","Haijun Lv","Ning Chen"],"abstract":"Due to the dominant position of deep learning (mostly deep neural networks) in various artificial intelligence applications, recently, ensemble learning based on deep neural networks (ensemble deep learning) has shown significant performances in improving the generalization of learning system. However, since modern deep neural networks usually have millions to billions of parameters, the time and space overheads for training multiple base deep learners and testing with the ensemble deep learner are far greater than that of traditional ensemble learning. Though several algorithms of fast ensemble deep learning have been proposed to promote the deployment of ensemble deep learning in some applications, further advances still need to be made for many applications in specific fields, where the developing time and computing resources are usually restricted or the data to be processed is of large dimensionality. An urgent problem needs to be solved is how to take the significant advantages of ensemble deep learning while reduce the required expenses so that many more applications in specific fields can benefit from it. For the alleviation of this problem, it is essential to know about how ensemble learning has developed under the era of deep learning. Thus, in this article, we present discussions focusing on data analyses of published works, methodologies, recent advances and unattainability of traditional ensemble learning and ensemble deep learning. We hope this article will be helpful to realize the intrinsic problems and technical challenges faced by future developments of ensemble learning under the era of deep learning.","doi":"10.1007\/s10462-022-10283-5","cleaned_title":"a survey on ensemble learning under the era of deep learning","cleaned_abstract":"due to the dominant position of deep learning (mostly deep neural networks) in various artificial intelligence applications, recently, ensemble learning based on deep neural networks (ensemble deep learning) has shown significant performances in improving the generalization of learning system. however, since modern deep neural networks usually have millions to billions of parameters, the time and space overheads for training multiple base deep learners and testing with the ensemble deep learner are far greater than that of traditional ensemble learning. though several algorithms of fast ensemble deep learning have been proposed to promote the deployment of ensemble deep learning in some applications, further advances still need to be made for many applications in specific fields, where the developing time and computing resources are usually restricted or the data to be processed is of large dimensionality. an urgent problem needs to be solved is how to take the significant advantages of ensemble deep learning while reduce the required expenses so that many more applications in specific fields can benefit from it. for the alleviation of this problem, it is essential to know about how ensemble learning has developed under the era of deep learning. thus, in this article, we present discussions focusing on data analyses of published works, methodologies, recent advances and unattainability of traditional ensemble learning and ensemble deep learning. we hope this article will be helpful to realize the intrinsic problems and technical challenges faced by future developments of ensemble learning under the era of deep learning.","key_phrases":["the dominant position","deep learning","mostly deep neural networks","various artificial intelligence applications",", recently, ensemble learning","deep neural networks","ensemble deep learning","significant performances","the generalization","learning system","modern deep neural networks","millions","billions","parameters","the time and space overheads","multiple base deep learners","the ensemble deep learner","that","traditional ensemble learning","several algorithms","fast ensemble deep learning","the deployment","ensemble deep learning","some applications","further advances","many applications","specific fields","the developing time","computing resources","the data","large dimensionality","an urgent problem","the significant advantages","ensemble deep learning","the required expenses","many more applications","specific fields","it","the alleviation","this problem","it","how ensemble learning","the era","deep learning","this article","we","discussions","data analyses","published works","methodologies","recent advances","unattainability","traditional ensemble learning","ensemble deep learning","we","this article","the intrinsic problems","technical challenges","future developments","ensemble learning","the era","deep learning","millions to billions"]},{"title":"Realistic fault detection of li-ion battery via dynamical deep learning","authors":["Jingzhao Zhang","Yanan Wang","Benben Jiang","Haowei He","Shaobo Huang","Chen Wang","Yang Zhang","Xuebing Han","Dongxu Guo","Guannan He","Minggao Ouyang"],"abstract":"Accurate evaluation of Li-ion battery (LiB) safety conditions can reduce unexpected cell failures, facilitate battery deployment, and promote low-carbon economies. Despite the recent progress in artificial intelligence, anomaly detection methods are not customized for or validated in realistic battery settings due to the complex failure mechanisms and the lack of real-world testing frameworks with large-scale datasets. Here, we develop a realistic deep-learning framework for electric vehicle (EV) LiB anomaly detection. It features a dynamical autoencoder tailored for dynamical systems and configured by social and financial factors. We test our detection algorithm on released datasets comprising over 690,000 LiB charging snippets from 347 EVs. Our model overcomes the limitations of state-of-the-art fault detection models, including deep learning ones. Moreover, it reduces the expected direct EV battery fault and inspection costs. Our work highlights the potential of deep learning in improving LiB safety and the significance of social and financial information in designing deep learning models.","doi":"10.1038\/s41467-023-41226-5","cleaned_title":"realistic fault detection of li-ion battery via dynamical deep learning","cleaned_abstract":"accurate evaluation of li-ion battery (lib) safety conditions can reduce unexpected cell failures, facilitate battery deployment, and promote low-carbon economies. despite the recent progress in artificial intelligence, anomaly detection methods are not customized for or validated in realistic battery settings due to the complex failure mechanisms and the lack of real-world testing frameworks with large-scale datasets. here, we develop a realistic deep-learning framework for electric vehicle (ev) lib anomaly detection. it features a dynamical autoencoder tailored for dynamical systems and configured by social and financial factors. we test our detection algorithm on released datasets comprising over 690,000 lib charging snippets from 347 evs. our model overcomes the limitations of state-of-the-art fault detection models, including deep learning ones. moreover, it reduces the expected direct ev battery fault and inspection costs. our work highlights the potential of deep learning in improving lib safety and the significance of social and financial information in designing deep learning models.","key_phrases":["accurate evaluation","li-ion battery (lib) safety conditions","unexpected cell failures","facilitate battery deployment","low-carbon economies","the recent progress","artificial intelligence","anomaly detection methods","realistic battery settings","the complex failure mechanisms","the lack","real-world testing frameworks","large-scale datasets","we","a realistic deep-learning framework","electric vehicle (ev) lib anomaly detection","it","a dynamical autoencoder","dynamical systems","social and financial factors","we","our detection algorithm","released datasets","over 690,000 lib","snippets","347 evs","our model","the limitations","the-art","deep learning ones","it","the expected direct ev battery fault and inspection costs","our work","the potential","deep learning","lib safety","the significance","social and financial information","deep learning models","690,000","347"]},{"title":"A Deep Learning-Based Object Representation Algorithm for Smart Retail Management","authors":["Bin Liu"],"abstract":"This study underscores the vital role of object representation and detection in smart retail management systems for optimizing customer experiences and operational efficiency. The literature review reveals a preference for deep learning techniques, citing their superior accuracy compared to traditional methods. While acknowledging the challenges of achieving high accuracy and low computation costs simultaneously in deep learning-based object representation, the paper proposes a solution using the YOLOv7 framework. In order to navigate the ever-changing landscape of smart retail technologies, the study clarifies the potential scalability and flexibility of deep learning approaches. The method employs a custom dataset, and experimental results demonstrate the model\u2019s efficacy, showcasing accurate results and enhanced performance in various experiments and analyses.","doi":"10.1007\/s40031-024-01051-w","cleaned_title":"a deep learning-based object representation algorithm for smart retail management","cleaned_abstract":"this study underscores the vital role of object representation and detection in smart retail management systems for optimizing customer experiences and operational efficiency. the literature review reveals a preference for deep learning techniques, citing their superior accuracy compared to traditional methods. while acknowledging the challenges of achieving high accuracy and low computation costs simultaneously in deep learning-based object representation, the paper proposes a solution using the yolov7 framework. in order to navigate the ever-changing landscape of smart retail technologies, the study clarifies the potential scalability and flexibility of deep learning approaches. the method employs a custom dataset, and experimental results demonstrate the model\u2019s efficacy, showcasing accurate results and enhanced performance in various experiments and analyses.","key_phrases":["this study","the vital role","object representation","detection","smart retail management systems","customer experiences","operational efficiency","the literature review","a preference","deep learning techniques","their superior accuracy","traditional methods","the challenges","high accuracy","low computation costs","deep learning-based object representation","the paper","a solution","the yolov7 framework","order","the ever-changing landscape","smart retail technologies","the study","the potential scalability","flexibility","deep learning approaches","the method","a custom dataset","experimental results","the model\u2019s efficacy","accurate results","enhanced performance","various experiments","analyses"]},{"title":"Deep transfer learning for tool condition monitoring under different processing conditions","authors":["Yongqing Wang","Mengmeng Niu","Kuo Liu","Haibo Liu","Bo Qin","Yiming Cui"],"abstract":"Deep learning methods have developed rapidly in the field of tool condition monitoring, but due to the complexity and diversity of working conditions, it is difficult to ensure the high accuracy, strong generalization performance, and wide applicability of monitoring models. Therefore, this article proposes a deep transfer learning method with center loss for tool condition monitoring under different processing conditions. Firstly, Deep Extreme Learning Machine (DELM) is used to extract sample features and tool condition monitoring, and deep coral is integrated into the last feature extraction layer of DELM for domain adaptation. Secondly, the center loss is introduced into the transfer learning model to improve the intra-class compactness by minimizing the center loss, thereby obtaining a broader decision boundary. A tool wear experiment was conducted on a milling machine. Research shows that the proposed method can achieve the tool condition monitoring under different processing conditions. The introduction of central loss is beneficial for promoting the separation of samples from different categories in the target domain and effectively improves the model\u2019s applicability. Compared with other domain adaptation methods, this method has better accuracy and generalization ability.","doi":"10.1007\/s00170-024-13713-6","cleaned_title":"deep transfer learning for tool condition monitoring under different processing conditions","cleaned_abstract":"deep learning methods have developed rapidly in the field of tool condition monitoring, but due to the complexity and diversity of working conditions, it is difficult to ensure the high accuracy, strong generalization performance, and wide applicability of monitoring models. therefore, this article proposes a deep transfer learning method with center loss for tool condition monitoring under different processing conditions. firstly, deep extreme learning machine (delm) is used to extract sample features and tool condition monitoring, and deep coral is integrated into the last feature extraction layer of delm for domain adaptation. secondly, the center loss is introduced into the transfer learning model to improve the intra-class compactness by minimizing the center loss, thereby obtaining a broader decision boundary. a tool wear experiment was conducted on a milling machine. research shows that the proposed method can achieve the tool condition monitoring under different processing conditions. the introduction of central loss is beneficial for promoting the separation of samples from different categories in the target domain and effectively improves the model\u2019s applicability. compared with other domain adaptation methods, this method has better accuracy and generalization ability.","key_phrases":["deep learning methods","the field","tool condition monitoring","the complexity","diversity","working conditions","it","the high accuracy, strong generalization performance","wide applicability","monitoring models","this article","a deep transfer learning method","center loss","tool condition monitoring","different processing conditions","deep extreme learning machine","(delm","sample features","tool condition monitoring","deep coral","the last feature extraction layer","delm","domain adaptation","the center loss","the transfer learning model","the intra-class compactness","the center loss","a broader decision boundary","a tool wear experiment","a milling machine","research","the proposed method","the tool condition monitoring","different processing conditions","the introduction","central loss","the separation","samples","different categories","the target domain","the model\u2019s applicability","other domain adaptation methods","this method","better accuracy","generalization ability","firstly","secondly"]},{"title":"Breast Mammograms Diagnosis Using Deep Learning: State of Art Tutorial Review","authors":["Osama Bin Naeem","Yasir Saleem","M. Usman Ghani Khan","Amjad Rehman Khan","Tanzila Saba","Saeed Ali Bahaj","Noor Ayesha"],"abstract":"Usually, screening (mostly mammography) is used by radiologists to manually detect breast cancer. The likelihood of identifying suspected cases as false positives or false negatives is significant, contingent on the experience of the radiologist and the kind of imaging screening device\/method utilized. The confirmation of the type of tumour seen by the radiologist is sent for histological investigation (microscopic analysis) through a biopsy, where the tumor's grade and stage, which are used in the latter stages of treatment, are ascertained through biopsy. However, a secondary issue with the cancer detection process is that only 15 to 30% of instances that are referred for biopsy result in malignant findings. Since deep learning demonstrated remarkable performance in visual recognition challenges, it has been widely applied to a variety of tasks. Similar examples include deep learning applications in healthcare, which are gaining a lot of interest from the research community. Deep learning is used to identify, categories tumours, and breast cancer is a significant global health concern. The medical sciences could now make more accurate diagnoses and detections due to recent advancements in machine learning techniques. Hence due to systems potential accuracy, it could offer optimistic outcomes when used to read malignant images. In imaging domains, deep learning-based methods have achieved remarkable success in constituent segmentation (UNet), localization (DenseNet), and classification (VGG-19). This study examines, how deep learning methods are assisting in the highly accurate diagnosis of benign or malignant tumours based on screened images. In contrast to a mammogram, which is covered in detail, this paper briefly discusses imaging methods for cancer detection. Early detection and cost effectiveness are two main benefits of applying machine learning and deep learning techniques to mammograms.","doi":"10.1007\/s11831-023-10052-9","cleaned_title":"breast mammograms diagnosis using deep learning: state of art tutorial review","cleaned_abstract":"usually, screening (mostly mammography) is used by radiologists to manually detect breast cancer. the likelihood of identifying suspected cases as false positives or false negatives is significant, contingent on the experience of the radiologist and the kind of imaging screening device\/method utilized. the confirmation of the type of tumour seen by the radiologist is sent for histological investigation (microscopic analysis) through a biopsy, where the tumor's grade and stage, which are used in the latter stages of treatment, are ascertained through biopsy. however, a secondary issue with the cancer detection process is that only 15 to 30% of instances that are referred for biopsy result in malignant findings. since deep learning demonstrated remarkable performance in visual recognition challenges, it has been widely applied to a variety of tasks. similar examples include deep learning applications in healthcare, which are gaining a lot of interest from the research community. deep learning is used to identify, categories tumours, and breast cancer is a significant global health concern. the medical sciences could now make more accurate diagnoses and detections due to recent advancements in machine learning techniques. hence due to systems potential accuracy, it could offer optimistic outcomes when used to read malignant images. in imaging domains, deep learning-based methods have achieved remarkable success in constituent segmentation (unet), localization (densenet), and classification (vgg-19). this study examines, how deep learning methods are assisting in the highly accurate diagnosis of benign or malignant tumours based on screened images. in contrast to a mammogram, which is covered in detail, this paper briefly discusses imaging methods for cancer detection. early detection and cost effectiveness are two main benefits of applying machine learning and deep learning techniques to mammograms.","key_phrases":["(mostly mammography","radiologists","breast cancer","the likelihood","suspected cases","false positives","false negatives","the experience","the radiologist","the kind","device\/method","the confirmation","the type","tumour","the radiologist","histological investigation","microscopic analysis","a biopsy","the tumor's grade","stage","which","the latter stages","treatment","biopsy","a secondary issue","the cancer detection process","only 15 to 30%","instances","that","biopsy result","malignant findings","deep learning","remarkable performance","visual recognition challenges","it","a variety","tasks","similar examples","deep learning applications","healthcare","which","a lot","interest","the research community","deep learning","categories tumours","breast cancer","a significant global health concern","the medical sciences","more accurate diagnoses","detections","recent advancements","machine learning techniques","systems potential accuracy","it","optimistic outcomes","malignant images","imaging","domains","deep learning-based methods","remarkable success","constituent segmentation","unet","localization","densenet","classification","vgg-19","this study examines","deep learning methods","the highly accurate diagnosis","benign or malignant tumours","screened images","contrast","a mammogram","which","detail","this paper","imaging methods","cancer detection","early detection and cost effectiveness","two main benefits","machine learning","deep learning techniques","mammograms","only 15 to","two"]},{"title":"Optimization of Fed-Batch Baker\u2019s Yeast Fermentation Using Deep Reinforcement Learning","authors":["Wan Ying Chai","Min Keng Tan","Kenneth Tze Kin Teo","Heng Jin Tham"],"abstract":"Fermentation is widely used in chemical industries to produce valuable products. It consumes less energy and has a lesser environmental impact compared to conventional chemical processes. However, the inherent nonlinearity of the fermentation process and limited comprehension of its metabolic mechanisms present challenges for control and optimization, particularly in minimizing the formation of by-products. Reinforcement learning is a machine learning method where an agent learns through exploration and experience. By receiving feedback in the form of rewards, it computes an optimal policy which produces maximum cumulative reward. The integration of deep learning with reinforcement learning has further improved the efficiency of classical reinforcement learning, particularly in continuous control. This paper focuses on optimizing substrate feeding rate in a simulated fed-batch baker\u2019s yeast fermentation using deep reinforcement learning. Artificial neural network (ANN) was applied as the function approximator to estimate the state-action function for determining the substrate feeding rate within a large state space, which includes substrate concentration, yeast concentration, and the change in ethanol concentration. The deep reinforcement learning algorithm was formulated based on the optimization objective of maximizing yeast production while minimizing ethanol formation. The performance of the feeding strategy proposed using deep reinforcement learning was compared to a commonly used pre-determined exponential feeding profile. The results show that the proposed feeding strategy outperformed the exponential feeding strategy with the yeast yield increased by 25.66% with negligible ethanol production. In addition, the proposed algorithm exhibits effective handling of various initial conditions compared to the exponential feeding approach.","doi":"10.1007\/s41660-024-00406-6","cleaned_title":"optimization of fed-batch baker\u2019s yeast fermentation using deep reinforcement learning","cleaned_abstract":"fermentation is widely used in chemical industries to produce valuable products. it consumes less energy and has a lesser environmental impact compared to conventional chemical processes. however, the inherent nonlinearity of the fermentation process and limited comprehension of its metabolic mechanisms present challenges for control and optimization, particularly in minimizing the formation of by-products. reinforcement learning is a machine learning method where an agent learns through exploration and experience. by receiving feedback in the form of rewards, it computes an optimal policy which produces maximum cumulative reward. the integration of deep learning with reinforcement learning has further improved the efficiency of classical reinforcement learning, particularly in continuous control. this paper focuses on optimizing substrate feeding rate in a simulated fed-batch baker\u2019s yeast fermentation using deep reinforcement learning. artificial neural network (ann) was applied as the function approximator to estimate the state-action function for determining the substrate feeding rate within a large state space, which includes substrate concentration, yeast concentration, and the change in ethanol concentration. the deep reinforcement learning algorithm was formulated based on the optimization objective of maximizing yeast production while minimizing ethanol formation. the performance of the feeding strategy proposed using deep reinforcement learning was compared to a commonly used pre-determined exponential feeding profile. the results show that the proposed feeding strategy outperformed the exponential feeding strategy with the yeast yield increased by 25.66% with negligible ethanol production. in addition, the proposed algorithm exhibits effective handling of various initial conditions compared to the exponential feeding approach.","key_phrases":["fermentation","chemical industries","valuable products","it","less energy","a lesser environmental impact","conventional chemical processes","however, the inherent nonlinearity","the fermentation process","limited comprehension","its metabolic mechanisms","present challenges","control","optimization","the formation","products","reinforcement learning","a machine learning method","an agent","exploration","experience","feedback","the form","rewards","it","an optimal policy","which","maximum cumulative reward","the integration","deep learning","reinforcement learning","the efficiency","classical reinforcement learning","continuous control","this paper","substrate feeding rate","a simulated fed-batch baker\u2019s yeast fermentation","deep reinforcement learning","artificial neural network","ann","the function approximator","the state-action function","the substrate feeding rate","a large state space","which","substrate concentration","yeast concentration","the change","ethanol concentration","the deep reinforcement learning algorithm","the optimization objective","yeast production","ethanol formation","the performance","the feeding strategy","deep reinforcement learning","a commonly used pre-determined exponential feeding profile","the results","the proposed feeding strategy","the exponential feeding strategy","the yeast yield","25.66%","negligible ethanol production","addition","the proposed algorithm","effective handling","various initial conditions","the exponential feeding approach","baker","25.66%"]},{"title":"Deep residual learning with Anscombe transformation for low-dose digital tomosynthesis","authors":["Youngjin Lee","Seungwan Lee","Chanrok Park"],"abstract":"Deep learning-based convolutional neural networks (CNNs) have been proposed for enhancing the quality of digital tomosynthesis (DTS) images. However, the direct applications of the conventional CNNs for low-dose DTS imaging are limited to provide acceptable image quality due to the inaccurate recognition of complex texture patterns. In this study, a deep residual learning network combined with the Anscombe transformation was proposed for simplifying the complex texture and restoring the low-dose DTS image quality. The proposed network consisted of convolution layers, max-pooling layers, up-sampling layers, and skip connections. The network training was performed to learn the residual images between the ground-truth and low-dose projections, which were converted using the Anscombe transformation. As a result, the proposed network enhanced the quantitative accuracy and noise characteristic of DTS images by 1.01\u20131.27 and 1.14\u20131.71 times, respectively, in comparison to low-dose DTS images and other deep learning networks. The spatial resolution of the DTS image restored using the proposed network was 1.12 times higher than that obtained using a deep image learning network. In conclusion, the proposed network can restore the low-dose DTS image quality and provide an optimal model for low-dose DTS imaging.","doi":"10.1007\/s40042-024-01117-4","cleaned_title":"deep residual learning with anscombe transformation for low-dose digital tomosynthesis","cleaned_abstract":"deep learning-based convolutional neural networks (cnns) have been proposed for enhancing the quality of digital tomosynthesis (dts) images. however, the direct applications of the conventional cnns for low-dose dts imaging are limited to provide acceptable image quality due to the inaccurate recognition of complex texture patterns. in this study, a deep residual learning network combined with the anscombe transformation was proposed for simplifying the complex texture and restoring the low-dose dts image quality. the proposed network consisted of convolution layers, max-pooling layers, up-sampling layers, and skip connections. the network training was performed to learn the residual images between the ground-truth and low-dose projections, which were converted using the anscombe transformation. as a result, the proposed network enhanced the quantitative accuracy and noise characteristic of dts images by 1.01\u20131.27 and 1.14\u20131.71 times, respectively, in comparison to low-dose dts images and other deep learning networks. the spatial resolution of the dts image restored using the proposed network was 1.12 times higher than that obtained using a deep image learning network. in conclusion, the proposed network can restore the low-dose dts image quality and provide an optimal model for low-dose dts imaging.","key_phrases":["deep learning-based convolutional neural networks","cnns","the quality","digital tomosynthesis (dts) images","the direct applications","the conventional cnns","low-dose dts imaging","acceptable image quality","the inaccurate recognition","complex texture patterns","this study","a deep residual learning network","the anscombe transformation","the complex texture","the low-dose dts image quality","the proposed network","convolution layers","max-pooling layers","up-sampling layers","skip connections","the network training","the residual images","the ground-truth and low-dose projections","which","the anscombe transformation","a result","the proposed network","the quantitative accuracy","dts images","1.01\u20131.27 and 1.14\u20131.71 times","comparison","low-dose dts images","other deep learning networks","the spatial resolution","the dts image","the proposed network","that","a deep image learning network","conclusion","the proposed network","the low-dose dts image quality","an optimal model","low-dose dts imaging","max","1.01\u20131.27","1.14\u20131.71","1.12"]},{"title":"An integrated deep-learning model for smart waste classification","authors":["Shivendu Mishra","Ritika Yaduvanshi","Prince Rajpoot","Sharad Verma","Amit Kumar Pandey","Digvijay Pandey"],"abstract":"Efficient waste management is essential for human well-being and environmental health, as neglecting proper disposal practices can lead to financial losses and the depletion of natural resources. Given the rapid urbanization and population growth, developing an automated, innovative waste classification model becomes imperative. To address this need, our paper introduces a novel and robust solution \u2014 a smart waste classification model that leverages a hybrid deep learning model (Optimized DenseNet-121 + SVM) to categorize waste items using the TrashNet datasets. Our proposed approach uses the advanced deep learning model DenseNet-121, optimized for superior performance, to extract meaningful features from an expanded TrashNet dataset. These features are subsequently fed into a support vector machine (SVM) for precise classification. Employing data augmentation techniques further enhances classification accuracy while mitigating the risk of overfitting, especially when working with limited TrashNet data. The results of our experimental evaluation of this hybrid deep learning model are highly promising, with an impressive accuracy rate of 99.84%. This accuracy surpasses similar existing models, affirming the efficacy and potential of our approach to revolutionizing waste classification for a sustainable and cleaner future.","doi":"10.1007\/s10661-024-12410-x","cleaned_title":"an integrated deep-learning model for smart waste classification","cleaned_abstract":"efficient waste management is essential for human well-being and environmental health, as neglecting proper disposal practices can lead to financial losses and the depletion of natural resources. given the rapid urbanization and population growth, developing an automated, innovative waste classification model becomes imperative. to address this need, our paper introduces a novel and robust solution \u2014 a smart waste classification model that leverages a hybrid deep learning model (optimized densenet-121 + svm) to categorize waste items using the trashnet datasets. our proposed approach uses the advanced deep learning model densenet-121, optimized for superior performance, to extract meaningful features from an expanded trashnet dataset. these features are subsequently fed into a support vector machine (svm) for precise classification. employing data augmentation techniques further enhances classification accuracy while mitigating the risk of overfitting, especially when working with limited trashnet data. the results of our experimental evaluation of this hybrid deep learning model are highly promising, with an impressive accuracy rate of 99.84%. this accuracy surpasses similar existing models, affirming the efficacy and potential of our approach to revolutionizing waste classification for a sustainable and cleaner future.","key_phrases":["efficient waste management","human well-being and environmental health","proper disposal practices","financial losses","the depletion","natural resources","the rapid urbanization","population growth","an automated, innovative waste classification model","this need","our paper","a novel and robust solution","a smart waste classification model","that","a hybrid deep learning model","svm","waste items","the trashnet datasets","our proposed approach","the advanced deep learning model densenet-121","superior performance","meaningful features","an expanded trashnet dataset","these features","a support vector machine","svm","precise classification","data augmentation","classification accuracy","the risk","overfitting","limited trashnet data","the results","our experimental evaluation","this hybrid deep learning model","an impressive accuracy rate","99.84%","this accuracy","similar existing models","the efficacy","potential","our approach","waste classification","a sustainable and cleaner future","99.84%"]},{"title":"Deep learning based multiclass classification for citrus anomaly detection in agriculture","authors":["Ebru Erg\u00fcn"],"abstract":"In regions where citrus crops are threatened by diseases caused by fungi, bacteria, pests and viruses, growers are actively seeking automated technologies that can accurately detect citrus anomalies to minimize economic losses. Recent advances in deep learning techniques have shown potential in automating and improving the accuracy of citrus anomaly categorization. This research explores the use of deep learning methods, specifically DenseNet, to construct robust models capable of accurately distinguishing between different types of citrus anomalies. The dataset consists of high-resolution images of different orange leaves of the species Citrus sinensis osbeck, collected from orange groves in the states of Tamaulipas and San Luis Potosi in northeastern Mexico was used in study. Experimental results demonstrated the effectiveness of the proposed deep learning models in simultaneously identifying 12 different classes of citrus anomalies. Evaluation metrics, including accuracy, recall, precision and the confusion matrix, underscore the discriminative power of the models. Among the convolutional neural network architectures used, DenseNet achieved the highest classification accuracy at 99.50%. The study concluded by highlighting the potential for scalable and effective citrus anomaly classification and management using deep learning-based systems.","doi":"10.1007\/s11760-024-03452-2","cleaned_title":"deep learning based multiclass classification for citrus anomaly detection in agriculture","cleaned_abstract":"in regions where citrus crops are threatened by diseases caused by fungi, bacteria, pests and viruses, growers are actively seeking automated technologies that can accurately detect citrus anomalies to minimize economic losses. recent advances in deep learning techniques have shown potential in automating and improving the accuracy of citrus anomaly categorization. this research explores the use of deep learning methods, specifically densenet, to construct robust models capable of accurately distinguishing between different types of citrus anomalies. the dataset consists of high-resolution images of different orange leaves of the species citrus sinensis osbeck, collected from orange groves in the states of tamaulipas and san luis potosi in northeastern mexico was used in study. experimental results demonstrated the effectiveness of the proposed deep learning models in simultaneously identifying 12 different classes of citrus anomalies. evaluation metrics, including accuracy, recall, precision and the confusion matrix, underscore the discriminative power of the models. among the convolutional neural network architectures used, densenet achieved the highest classification accuracy at 99.50%. the study concluded by highlighting the potential for scalable and effective citrus anomaly classification and management using deep learning-based systems.","key_phrases":["regions","citrus crops","diseases","fungi","bacteria","pests","viruses","growers","automated technologies","that","citrus anomalies","economic losses","recent advances","deep learning techniques","potential","the accuracy","citrus anomaly categorization","this research","the use","deep learning methods","specifically densenet","robust models","different types","citrus anomalies","the dataset","high-resolution images","different orange leaves","the species citrus sinensis osbeck","orange groves","the states","tamaulipas","san luis potosi","northeastern mexico","study","experimental results","the effectiveness","the proposed deep learning models","12 different classes","citrus anomalies","evaluation metrics","accuracy","recall","precision","the confusion matrix","the discriminative power","the models","the convolutional neural network architectures","densenet","the highest classification accuracy","99.50%","the study","the potential","scalable and effective citrus anomaly classification","management","deep learning-based systems","san luis","mexico","12","99.50%"]},{"title":"Skin cancer detection using ensemble of machine learning and deep learning techniques","authors":["Jitendra V. Tembhurne","Nachiketa Hebbar","Hemprasad Y. Patil","Tausif Diwan"],"abstract":"Skin cancer is one of the most common forms of cancer, which makes it pertinent to be able to diagnose it accurately. In particular, melanoma is a form of skin cancer that is fatal and accounts for 6 of every 7-skin cancer related death. Moreover, in hospitals where dermatologists have to diagnose multiple cases of skin cancer, there are high possibilities of false negatives in diagnosis. To avoid such incidents, there here has been exhaustive research conducted by the research community all over the world to build highly accurate automated tools for skin cancer detection. In this paper, we introduce a novel approach of combining machine learning and deep learning techniques to solve the problem of skin cancer detection. The deep learning model uses state-of-the-art neural networks to extract features from images whereas the machine learning model processes image features which are obtained after performing the techniques such as Contourlet Transform and Local Binary Pattern Histogram. Meaningful feature extraction is crucial for any image classification roblem. As a result, by combining the manual and automated features, our designed model achieves a higher accuracy of 93% with an individual recall score of 99.7% and 86% for the benign and malignant forms of cancer, respectively. We benchmarked the model on publicly available Kaggle dataset containing processed images from ISIC Archive dataset. The proposed ensemble outperforms both expert dermatologists as well as other state-of-the-art deep learning and machine learning methods. Thus, this novel method can be of high assistance to dermatologists to help prevent any misdiagnosis.","doi":"10.1007\/s11042-023-14697-3","cleaned_title":"skin cancer detection using ensemble of machine learning and deep learning techniques","cleaned_abstract":"skin cancer is one of the most common forms of cancer, which makes it pertinent to be able to diagnose it accurately. in particular, melanoma is a form of skin cancer that is fatal and accounts for 6 of every 7-skin cancer related death. moreover, in hospitals where dermatologists have to diagnose multiple cases of skin cancer, there are high possibilities of false negatives in diagnosis. to avoid such incidents, there here has been exhaustive research conducted by the research community all over the world to build highly accurate automated tools for skin cancer detection. in this paper, we introduce a novel approach of combining machine learning and deep learning techniques to solve the problem of skin cancer detection. the deep learning model uses state-of-the-art neural networks to extract features from images whereas the machine learning model processes image features which are obtained after performing the techniques such as contourlet transform and local binary pattern histogram. meaningful feature extraction is crucial for any image classification roblem. as a result, by combining the manual and automated features, our designed model achieves a higher accuracy of 93% with an individual recall score of 99.7% and 86% for the benign and malignant forms of cancer, respectively. we benchmarked the model on publicly available kaggle dataset containing processed images from isic archive dataset. the proposed ensemble outperforms both expert dermatologists as well as other state-of-the-art deep learning and machine learning methods. thus, this novel method can be of high assistance to dermatologists to help prevent any misdiagnosis.","key_phrases":["skin cancer","the most common forms","cancer","which","it","it","melanoma","a form","skin cancer","that","every 7-skin cancer related death","hospitals","dermatologists","multiple cases","skin cancer","high possibilities","false negatives","diagnosis","such incidents","exhaustive research","the research community","the world","highly accurate automated tools","skin cancer detection","this paper","we","a novel approach","machine learning","deep learning techniques","the problem","skin cancer detection","the deep learning model","the-art","features","images","the machine learning model","image features","which","the techniques","contourlet transform","local binary pattern histogram","meaningful feature extraction","any image classification roblem","a result","the manual and automated features","our designed model","a higher accuracy","93%","an individual recall score","99.7%","86%","the benign and malignant forms","cancer","we","the model","publicly available kaggle dataset","processed images","isic archive dataset","the proposed ensemble outperforms","both expert dermatologists","the-art","this novel method","high assistance","dermatologists","any misdiagnosis","one","6","7","93%","99.7% and","86%"]},{"title":"Prediction of fiber Rayleigh scattering responses based on deep learning","authors":["Yongxin Liang","Jianhui Sun","Jialei Zhang","Yuyao Wang","Anchi Wan","Shibo Zhang","Zhenyu Ye","Shengtao Lin","Zinan Wang"],"abstract":"Distributed acoustic sensing (DAS) is a fiber sensing technology based on Rayleigh scattering, which transforms optical fiber into a series of sensing units. It has become an indispensable part in the field of seismic monitoring, vehicle tracking, and pipeline monitoring. Fiber Rayleigh scattering responses lay at the core of DAS. However, there are few in-depth studies on the purpose of acquiring fiber Rayleigh scattering responses. In this paper, we establish a deep learning framework based on the bidirectional gated recurrent unit, which is the first time to predict the fiber Rayleigh scattering responses, to the best of our knowledge. The deep learning framework is trained with a numerical simulation dataset only, but it can process experimental data successfully. Moreover, since the responses could have a wider effective bandwidth than the experimental probing pulses, a finer spatial resolution could be obtained after demodulation. This work indicates that the deep learning framework can capture the characteristics of the fiber Rayleigh scattering responses effectively, which paves the way for intelligent DAS.","doi":"10.1007\/s11432-022-3734-0","cleaned_title":"prediction of fiber rayleigh scattering responses based on deep learning","cleaned_abstract":"distributed acoustic sensing (das) is a fiber sensing technology based on rayleigh scattering, which transforms optical fiber into a series of sensing units. it has become an indispensable part in the field of seismic monitoring, vehicle tracking, and pipeline monitoring. fiber rayleigh scattering responses lay at the core of das. however, there are few in-depth studies on the purpose of acquiring fiber rayleigh scattering responses. in this paper, we establish a deep learning framework based on the bidirectional gated recurrent unit, which is the first time to predict the fiber rayleigh scattering responses, to the best of our knowledge. the deep learning framework is trained with a numerical simulation dataset only, but it can process experimental data successfully. moreover, since the responses could have a wider effective bandwidth than the experimental probing pulses, a finer spatial resolution could be obtained after demodulation. this work indicates that the deep learning framework can capture the characteristics of the fiber rayleigh scattering responses effectively, which paves the way for intelligent das.","key_phrases":["acoustic sensing","das","a fiber sensing technology","rayleigh scattering","which","optical fiber","a series","sensing units","it","an indispensable part","the field","seismic monitoring","vehicle tracking","pipeline monitoring","fiber rayleigh scattering responses","the core","das","few in-depth studies","the purpose","fiber rayleigh scattering responses","this paper","we","a deep learning framework","the bidirectional gated recurrent unit","which","the first time","the fiber rayleigh scattering responses","our knowledge","the deep learning framework","a numerical simulation","it","experimental data","the responses","the experimental probing pulses","a finer spatial resolution","demodulation","this work","the deep learning framework","the characteristics","the fiber rayleigh scattering responses","which","the way","intelligent das","das","das","first"]},{"title":"Machine learning, deep learning and hernia surgery. Are we pushing the limits of abdominal core health? A qualitative systematic review","authors":["D. L. Lima","J. Kasakewitch","D. Q. Nguyen","R. Nogueira","L. T. Cavazzola","B. T. Heniford","F. Malcher"],"abstract":"IntroductionThis systematic review aims to evaluate the use of machine learning and artificial intelligence in hernia surgery.MethodsThe PRISMA guidelines were followed throughout this systematic review. The ROBINS\u2014I and Rob 2 tools were used to perform qualitative assessment of all studies included in this review. Recommendations were then summarized for the following pre-defined key items: protocol, research question, search strategy, study eligibility, data extraction, study design, risk of bias, publication bias, and statistical analysis.ResultsA total of 13 articles were ultimately included for this review, describing the use of machine learning and deep learning for hernia surgery. All studies were published from 2020 to 2023. Articles varied regarding the population studied, type of machine learning or Deep Learning Model (DLM) used, and hernia type. Of the thirteen included studies, all included either inguinal, ventral, or incisional hernias. Four studies evaluated recognition of surgical steps during inguinal hernia repair videos. Two studies predicted outcomes using image-based DMLs. Seven studies developed and validated deep learning algorithms to predict outcomes and identify factors associated with postoperative complications.ConclusionThe use of ML for abdominal wall reconstruction has been shown to be a promising tool for predicting outcomes and identifying factors that could lead to postoperative complications.","doi":"10.1007\/s10029-024-03069-x","cleaned_title":"machine learning, deep learning and hernia surgery. are we pushing the limits of abdominal core health? a qualitative systematic review","cleaned_abstract":"introductionthis systematic review aims to evaluate the use of machine learning and artificial intelligence in hernia surgery.methodsthe prisma guidelines were followed throughout this systematic review. the robins\u2014i and rob 2 tools were used to perform qualitative assessment of all studies included in this review. recommendations were then summarized for the following pre-defined key items: protocol, research question, search strategy, study eligibility, data extraction, study design, risk of bias, publication bias, and statistical analysis.resultsa total of 13 articles were ultimately included for this review, describing the use of machine learning and deep learning for hernia surgery. all studies were published from 2020 to 2023. articles varied regarding the population studied, type of machine learning or deep learning model (dlm) used, and hernia type. of the thirteen included studies, all included either inguinal, ventral, or incisional hernias. four studies evaluated recognition of surgical steps during inguinal hernia repair videos. two studies predicted outcomes using image-based dmls. seven studies developed and validated deep learning algorithms to predict outcomes and identify factors associated with postoperative complications.conclusionthe use of ml for abdominal wall reconstruction has been shown to be a promising tool for predicting outcomes and identifying factors that could lead to postoperative complications.","key_phrases":["introductionthis systematic review","the use","machine learning","artificial intelligence","hernia surgery.methodsthe prisma guidelines","this systematic review","the robins","i","rob","2 tools","qualitative assessment","all studies","this review","recommendations","the following pre-defined key items","protocol","research question","search strategy","study eligibility","data extraction","study design","risk","bias","publication bias","statistical analysis.resultsa total","13 articles","this review","the use","machine learning","deep learning","hernia surgery","all studies","articles","the population","type","machine learning","deep learning model","dlm","hernia type","the thirteen included studies","all","either inguinal, ventral, or incisional hernias","four studies","recognition","surgical steps","inguinal hernia repair videos","two studies","outcomes","image-based dmls","seven studies","deep learning algorithms","outcomes","factors","postoperative complications.conclusionthe use","ml","abdominal wall reconstruction","a promising tool","outcomes","factors","that","postoperative complications","hernia","2","13","2020","thirteen","four","two","seven"]},{"title":"Big data and deep learning for RNA biology","authors":["Hyeonseo Hwang","Hyeonseong Jeon","Nagyeong Yeo","Daehyun Baek"],"abstract":"The exponential growth of big data in RNA biology (RB) has led to the development of deep learning (DL) models that have driven crucial discoveries. As constantly evidenced by DL studies in other fields, the successful implementation of DL in RB depends heavily on the effective utilization of large-scale datasets from public databases. In achieving this goal, data encoding methods, learning algorithms, and techniques that align well with biological domain knowledge have played pivotal roles. In this review, we provide guiding principles for applying these DL concepts to various problems in RB by demonstrating successful examples and associated methodologies. We also discuss the remaining challenges in developing DL models for RB and suggest strategies to overcome these challenges. Overall, this review aims to illuminate the compelling potential of DL for RB and ways to apply this powerful technology to investigate the intriguing biology of RNA more effectively.","doi":"10.1038\/s12276-024-01243-w","cleaned_title":"big data and deep learning for rna biology","cleaned_abstract":"the exponential growth of big data in rna biology (rb) has led to the development of deep learning (dl) models that have driven crucial discoveries. as constantly evidenced by dl studies in other fields, the successful implementation of dl in rb depends heavily on the effective utilization of large-scale datasets from public databases. in achieving this goal, data encoding methods, learning algorithms, and techniques that align well with biological domain knowledge have played pivotal roles. in this review, we provide guiding principles for applying these dl concepts to various problems in rb by demonstrating successful examples and associated methodologies. we also discuss the remaining challenges in developing dl models for rb and suggest strategies to overcome these challenges. overall, this review aims to illuminate the compelling potential of dl for rb and ways to apply this powerful technology to investigate the intriguing biology of rna more effectively.","key_phrases":["the exponential growth","big data","rna biology","rb","the development","deep learning (dl) models","that","crucial discoveries","dl studies","other fields","the successful implementation","dl","the effective utilization","large-scale datasets","public databases","this goal","data encoding methods","algorithms","techniques","that","biological domain knowledge","pivotal roles","this review","we","guiding principles","these dl concepts","various problems","rb","successful examples","associated methodologies","we","the remaining challenges","dl models","rb","strategies","these challenges","this review","the compelling potential","dl","ways","this powerful technology","the intriguing biology","rna"]},{"title":"Classification of hazelnut varieties based on bigtransfer deep learning model","authors":["Emrah D\u00f6nmez","Serhat K\u0131l\u0131\u00e7arslan","Aykut Diker"],"abstract":"Hazelnut is an agricultural product that contributes greatly to the economy of the countries where it is grown. The human factor plays a major role in hazelnut classification. The typical approach involves manual inspection of each sample by experts, a process that is both labor-intensive and time-consuming, and often suffers from limited sensitivity. The deep learning techniques are extremely important in the classification and detection of agricultural products. Deep learning has great potential in the agricultural sector. This technology can improve product quality, increase productivity, and offer farmers the ability to classify and detect their produce more effectively. This is important for sustainability and efficiency in the agricultural industry. In this paper aims to the application of deep learning algorithms to streamline hazelnut classification, reducing the need for manual labor, time, and cost in the sorting process. The study utilized hazelnut images from three different varieties: Giresun, Ordu, and Van, comprising a dataset of 1165 images for Giresun, 1324 for Ordu, and 1138 for Van hazelnuts. This dataset is an open-access dataset. In the study, experiments were carried out on the determination of hazelnut varieties with BigTransfer (BiT)-M R50\u2009\u00d7\u20091, BiT-M R101\u2009\u00d7\u20093 and BiT-M R152\u2009\u00d7\u20094 models. Deep learning models, including big transfer was employed for classification. The classification task involved 3627 nut images and resulted in a remarkable accuracy of 99.49% with the BiT-M R152\u2009\u00d7\u20094 model. These innovative methods can also lead to patentable products and devices in various industries, thereby boosting the economic value of the country.","doi":"10.1007\/s00217-024-04468-1","cleaned_title":"classification of hazelnut varieties based on bigtransfer deep learning model","cleaned_abstract":"hazelnut is an agricultural product that contributes greatly to the economy of the countries where it is grown. the human factor plays a major role in hazelnut classification. the typical approach involves manual inspection of each sample by experts, a process that is both labor-intensive and time-consuming, and often suffers from limited sensitivity. the deep learning techniques are extremely important in the classification and detection of agricultural products. deep learning has great potential in the agricultural sector. this technology can improve product quality, increase productivity, and offer farmers the ability to classify and detect their produce more effectively. this is important for sustainability and efficiency in the agricultural industry. in this paper aims to the application of deep learning algorithms to streamline hazelnut classification, reducing the need for manual labor, time, and cost in the sorting process. the study utilized hazelnut images from three different varieties: giresun, ordu, and van, comprising a dataset of 1165 images for giresun, 1324 for ordu, and 1138 for van hazelnuts. this dataset is an open-access dataset. in the study, experiments were carried out on the determination of hazelnut varieties with bigtransfer (bit)-m r50 \u00d7 1, bit-m r101 \u00d7 3 and bit-m r152 \u00d7 4 models. deep learning models, including big transfer was employed for classification. the classification task involved 3627 nut images and resulted in a remarkable accuracy of 99.49% with the bit-m r152 \u00d7 4 model. these innovative methods can also lead to patentable products and devices in various industries, thereby boosting the economic value of the country.","key_phrases":["hazelnut","an agricultural product","that","the economy","the countries","it","the human factor","a major role","hazelnut classification","the typical approach","manual inspection","each sample","experts","a process","that","limited sensitivity","the deep learning techniques","the classification","detection","agricultural products","deep learning","great potential","the agricultural sector","this technology","product quality","productivity","farmers","the ability","their produce","this","sustainability","efficiency","the agricultural industry","this paper","the application","deep learning algorithms","hazelnut classification","the need","manual labor","time","cost","the sorting process","the study","hazelnut images","three different varieties","giresun","ordu","van","a dataset","1165 images","giresun","ordu","van hazelnuts","this dataset","an open-access dataset","the study","experiments","the determination","hazelnut varieties","bigtransfer","r101","r152 \u00d7 4 models","deep learning models","big transfer","classification","the classification task","3627 nut images","a remarkable accuracy","99.49%","the bit-m","r152","4 model","these innovative methods","patentable products","devices","various industries","the economic value","the country","three","van","1165","1324","1138","bit)-m","\u00d7 4","3627","99.49%","\u00d7 4"]},{"title":"Deep learning solutions for smart city challenges in urban development","authors":["Pengjun Wu","Zhanzhi Zhang","Xueyi Peng","Ran Wang"],"abstract":"In the realm of urban planning, the integration of deep learning technologies has emerged as a transformative force, promising to revolutionize the way cities are designed, managed, and optimized. This research embarks on a multifaceted exploration that combines the power of deep learning with Bayesian regularization techniques to enhance the performance and reliability of neural networks tailored for urban planning applications. Deep learning, characterized by its ability to extract complex patterns from vast urban datasets, has the potential to offer unprecedented insights into urban dynamics, transportation networks, and environmental sustainability. However, the complexity of these models often leads to challenges such as overfitting and limited interpretability. To address these issues, Bayesian regularization methods are employed to imbue neural networks with a principled framework that enhances generalization while quantifying predictive uncertainty. This research unfolds with the practical implementation of Bayesian regularization within neural networks, focusing on applications ranging from traffic prediction, urban infrastructure, data privacy, safety and security. By integrating Bayesian regularization, the aim is to, not only improve model performance in terms of accuracy and reliability but also to provide planners and decision-makers with probabilistic insights into the outcomes of various urban interventions. In tandem with quantitative assessments, graphical analysis is wielded as a crucial tool to visualize the inner workings of deep learning models in the context of urban planning. Through graphical representations, network visualizations, and decision boundary analysis, we uncover how Bayesian regularization influences neural network architecture and enhances interpretability.","doi":"10.1038\/s41598-024-55928-3","cleaned_title":"deep learning solutions for smart city challenges in urban development","cleaned_abstract":"in the realm of urban planning, the integration of deep learning technologies has emerged as a transformative force, promising to revolutionize the way cities are designed, managed, and optimized. this research embarks on a multifaceted exploration that combines the power of deep learning with bayesian regularization techniques to enhance the performance and reliability of neural networks tailored for urban planning applications. deep learning, characterized by its ability to extract complex patterns from vast urban datasets, has the potential to offer unprecedented insights into urban dynamics, transportation networks, and environmental sustainability. however, the complexity of these models often leads to challenges such as overfitting and limited interpretability. to address these issues, bayesian regularization methods are employed to imbue neural networks with a principled framework that enhances generalization while quantifying predictive uncertainty. this research unfolds with the practical implementation of bayesian regularization within neural networks, focusing on applications ranging from traffic prediction, urban infrastructure, data privacy, safety and security. by integrating bayesian regularization, the aim is to, not only improve model performance in terms of accuracy and reliability but also to provide planners and decision-makers with probabilistic insights into the outcomes of various urban interventions. in tandem with quantitative assessments, graphical analysis is wielded as a crucial tool to visualize the inner workings of deep learning models in the context of urban planning. through graphical representations, network visualizations, and decision boundary analysis, we uncover how bayesian regularization influences neural network architecture and enhances interpretability.","key_phrases":["the realm","urban planning","the integration","deep learning technologies","a transformative force","the way","cities","this research embarks","a multifaceted exploration","that","the power","deep learning","bayesian regularization techniques","the performance","reliability","neural networks","urban planning applications","deep learning","its ability","complex patterns","vast urban datasets","the potential","unprecedented insights","urban dynamics","transportation networks","environmental sustainability","the complexity","these models","challenges","limited interpretability","these issues","bayesian regularization methods","imbue neural networks","a principled framework","that","generalization","predictive uncertainty","this research","the practical implementation","bayesian regularization","neural networks","applications","traffic prediction","urban infrastructure","data privacy","safety","security","bayesian regularization","the aim","model performance","terms","accuracy","reliability","planners","decision-makers","probabilistic insights","the outcomes","various urban interventions","tandem","quantitative assessments","graphical analysis","a crucial tool","the inner workings","deep learning models","the context","urban planning","graphical representations","network visualizations","decision boundary analysis","we","bayesian regularization","neural network architecture","interpretability","bayesian"]},{"title":"Age transformation based on deep learning: a survey","authors":["Yingchun Guo","Xin Su","Gang Yan","Ye Zhu","Xueqi Lv"],"abstract":"Age transformation aims to preserve personalized facial information while altering a given face to appear at a target age. This technique finds extensive applications in fields such as face recognition, movie special effects, and social entertainment, among others. With the advancement of deep learning, particularly Generative Adversarial Networks (GANs), research on age transformation has made significant progress, leading to the emergence of a diverse range of deep learning-based methods. However, a comprehensive and systematic literature review of these methods is currently lacking. In this survey, we provide an all-encompassing review of deep learning methods for facial aging. Firstly, we summarize the key aspects of feature preservation during the age transformation process. Subsequently, we present a comprehensive overview of facial age transformation techniques, categorized according to various deep learning network architectures. Additionally, we conduct an analysis and comparison of commonly used face image datasets, offering recommendations for dataset selection. Furthermore, we consolidate the qualitative and quantitative evaluation metrics commonly employed in age transformation methodologies through experimental assessment. Finally, we address potential areas of future research in age transformation methods, based on the current challenges and limitations.","doi":"10.1007\/s00521-023-09376-1","cleaned_title":"age transformation based on deep learning: a survey","cleaned_abstract":"age transformation aims to preserve personalized facial information while altering a given face to appear at a target age. this technique finds extensive applications in fields such as face recognition, movie special effects, and social entertainment, among others. with the advancement of deep learning, particularly generative adversarial networks (gans), research on age transformation has made significant progress, leading to the emergence of a diverse range of deep learning-based methods. however, a comprehensive and systematic literature review of these methods is currently lacking. in this survey, we provide an all-encompassing review of deep learning methods for facial aging. firstly, we summarize the key aspects of feature preservation during the age transformation process. subsequently, we present a comprehensive overview of facial age transformation techniques, categorized according to various deep learning network architectures. additionally, we conduct an analysis and comparison of commonly used face image datasets, offering recommendations for dataset selection. furthermore, we consolidate the qualitative and quantitative evaluation metrics commonly employed in age transformation methodologies through experimental assessment. finally, we address potential areas of future research in age transformation methods, based on the current challenges and limitations.","key_phrases":["age transformation","personalized facial information","a given face","a target age","this technique","extensive applications","fields","face recognition","movie special effects","social entertainment","others","the advancement","deep learning","particularly generative adversarial networks","gans","age transformation","significant progress","the emergence","a diverse range","deep learning-based methods","a comprehensive and systematic literature review","these methods","this survey","we","an all-encompassing review","deep learning methods","facial aging","we","the key aspects","feature preservation","the age transformation process","we","a comprehensive overview","facial age transformation techniques","various deep learning network architectures","we","an analysis","comparison","commonly used face image datasets","recommendations","dataset selection","we","the qualitative and quantitative evaluation metrics","age transformation methodologies","experimental assessment","we","potential areas","future research","age transformation methods","the current challenges","limitations","firstly"]},{"title":"Advances of Pipeline Model Parallelism for Deep Learning Training: An Overview","authors":["Lei Guan\u00a0\n            (\u5173\u3000\u78ca)","Dong-Sheng Li\u00a0\n            (\u674e\u4e1c\u5347)","Ji-Ye Liang\u00a0\n            (\u6881\u5409\u4e1a)","Wen-Jian Wang\u00a0\n            (\u738b\u6587\u5251)","Ke-Shi Ge\u00a0\n            (\u845b\u53ef\u9002)","Xi-Cheng Lu\u00a0\n            (\u5362\u9521\u57ce)"],"abstract":"Deep learning has become the cornerstone of artificial intelligence, playing an increasingly important role in human production and lifestyle. However, as the complexity of problem-solving increases, deep learning models become increasingly intricate, resulting in a proliferation of large language models with an astonishing number of parameters. Pipeline model parallelism (PMP) has emerged as one of the mainstream approaches to addressing the significant challenge of training \u201cbig models\u201d. This paper presents a comprehensive review of PMP. It covers the basic concepts and main challenges of PMP. It also comprehensively compares synchronous and asynchronous pipeline schedules for PMP approaches, and discusses the main techniques to achieve load balance for both intra-node and inter-node training. Furthermore, the main techniques to optimize computation, storage, and communication are presented, with potential research directions being discussed.","doi":"10.1007\/s11390-024-3872-3","cleaned_title":"advances of pipeline model parallelism for deep learning training: an overview","cleaned_abstract":"deep learning has become the cornerstone of artificial intelligence, playing an increasingly important role in human production and lifestyle. however, as the complexity of problem-solving increases, deep learning models become increasingly intricate, resulting in a proliferation of large language models with an astonishing number of parameters. pipeline model parallelism (pmp) has emerged as one of the mainstream approaches to addressing the significant challenge of training \u201cbig models\u201d. this paper presents a comprehensive review of pmp. it covers the basic concepts and main challenges of pmp. it also comprehensively compares synchronous and asynchronous pipeline schedules for pmp approaches, and discusses the main techniques to achieve load balance for both intra-node and inter-node training. furthermore, the main techniques to optimize computation, storage, and communication are presented, with potential research directions being discussed.","key_phrases":["deep learning","the cornerstone","artificial intelligence","an increasingly important role","human production","lifestyle","the complexity","problem-solving increases","deep learning models","a proliferation","large language models","an astonishing number","parameters","pipeline model parallelism","the mainstream approaches","the significant challenge","\u201cbig models","this paper","a comprehensive review","pmp","it","the basic concepts","main challenges","pmp","it","synchronous and asynchronous pipeline schedules","pmp approaches","the main techniques","load balance","both intra-node and inter-node training","the main techniques","computation","storage","communication","potential research directions"]},{"title":"Deep learning-based channel estimation for wireless ultraviolet MIMO communication systems","authors":["Taifei Zhao","Yuxin Sun","Xinzhe L\u00fc","Shuang Zhang"],"abstract":"To solve the problems of pulse broadening and channel fading caused by atmospheric scattering and turbulence, multiple-input multiple-output (MIMO) technology is a valid way. A wireless ultraviolet (UV) MIMO channel estimation approach based on deep learning is provided in this paper. The deep learning is used to convert the channel estimation into the image processing. By combining convolutional neural network (CNN) and attention mechanism (AM), the learning model is designed to extract the depth features of channel state information (CSI). The simulation results show that the approach proposed in this paper can perform channel estimation effectively for UV MIMO communication and can better suppress the fading caused by scattering and turbulence in the MIMO scattering channel.","doi":"10.1007\/s11801-024-3069-6","cleaned_title":"deep learning-based channel estimation for wireless ultraviolet mimo communication systems","cleaned_abstract":"to solve the problems of pulse broadening and channel fading caused by atmospheric scattering and turbulence, multiple-input multiple-output (mimo) technology is a valid way. a wireless ultraviolet (uv) mimo channel estimation approach based on deep learning is provided in this paper. the deep learning is used to convert the channel estimation into the image processing. by combining convolutional neural network (cnn) and attention mechanism (am), the learning model is designed to extract the depth features of channel state information (csi). the simulation results show that the approach proposed in this paper can perform channel estimation effectively for uv mimo communication and can better suppress the fading caused by scattering and turbulence in the mimo scattering channel.","key_phrases":["the problems","pulse","channel fading","atmospheric scattering","turbulence","multiple-input multiple-output (mimo) technology","a valid way","a wireless ultraviolet (uv) mimo channel estimation approach","deep learning","this paper","the deep learning","the channel estimation","the image processing","convolutional neural network","cnn","attention mechanism","am","the learning model","the depth features","channel state information","csi","the simulation results","the approach","this paper","channel estimation","uv mimo communication","the fading","scattering","turbulence","the mimo scattering channel","mimo channel estimation","cnn"]},{"title":"Imbalcbl: addressing deep learning challenges with small and imbalanced datasets","authors":["Saqib ul Sabha","Assif Assad","Sadaf Shafi","Nusrat Mohi Ud Din","Rayees Ahmad Dar","Muzafar Rasool Bhat"],"abstract":"Deep learning, while transformative for computer vision, frequently falters when confronted with small and imbalanced datasets. Despite substantial progress in this domain, prevailing models often underachieve under these constraints. Addressing this, we introduce an innovative contrast-based learning strategy for small and imbalanced data that significantly bolsters the proficiency of deep learning architectures on these challenging datasets. By ingeniously concatenating training images, the effective training dataset expands from n to \\(n^2\\), affording richer data for model training, even when n is very small. Remarkably, our solution remains indifferent to specific loss functions or network architectures, endorsing its adaptability for diverse classification scenarios. Rigorously benchmarked against four benchmark datasets, our approach was juxtaposed with state-of-the-art oversampling paradigms. The empirical evidence underscores our method\u2019s superior efficacy, outshining contemporaries across metrics like Balanced accuracy, F1 score, and Geometric mean. Noteworthy increments include 7\u201316% on the Covid-19 dataset, 4\u201320% for Honey bees, 1\u20136% on CIFAR-10, and 1\u20139% on FashionMNIST. In essence, our proposed method offers a potent remedy for the perennial issues stemming from scanty and skewed data in deep learning.","doi":"10.1007\/s13198-024-02346-3","cleaned_title":"imbalcbl: addressing deep learning challenges with small and imbalanced datasets","cleaned_abstract":"deep learning, while transformative for computer vision, frequently falters when confronted with small and imbalanced datasets. despite substantial progress in this domain, prevailing models often underachieve under these constraints. addressing this, we introduce an innovative contrast-based learning strategy for small and imbalanced data that significantly bolsters the proficiency of deep learning architectures on these challenging datasets. by ingeniously concatenating training images, the effective training dataset expands from n to \\(n^2\\), affording richer data for model training, even when n is very small. remarkably, our solution remains indifferent to specific loss functions or network architectures, endorsing its adaptability for diverse classification scenarios. rigorously benchmarked against four benchmark datasets, our approach was juxtaposed with state-of-the-art oversampling paradigms. the empirical evidence underscores our method\u2019s superior efficacy, outshining contemporaries across metrics like balanced accuracy, f1 score, and geometric mean. noteworthy increments include 7\u201316% on the covid-19 dataset, 4\u201320% for honey bees, 1\u20136% on cifar-10, and 1\u20139% on fashionmnist. in essence, our proposed method offers a potent remedy for the perennial issues stemming from scanty and skewed data in deep learning.","key_phrases":["deep learning","computer vision","small and imbalanced datasets","substantial progress","this domain","prevailing models","these constraints","this","we","an innovative contrast-based learning strategy","small and imbalanced data","that","the proficiency","deep learning architectures","these challenging datasets","training images","the effective training dataset","n","\\(n^2\\","richer data","model training","n","our solution","specific loss functions","network architectures","its adaptability","diverse classification scenarios","four benchmark datasets","our approach","the-art","the empirical evidence","our method\u2019s superior efficacy","contemporaries","metrics","balanced accuracy","f1 score","geometric mean","noteworthy increments","7\u201316%","the covid-19 dataset","4\u201320%","honey bees","1\u20136%","cifar-10","1\u20139%","fashionmnist","essence","our proposed method","a potent remedy","the perennial issues","scanty","skewed data","deep learning","four","noteworthy","7\u201316%","covid-19","4\u201320%","1\u20136%","cifar-10","1\u20139%"]},{"title":"ConvDepthTransEnsembleNet: An Improved Deep Learning Approach for Rice Crop Leaf Disease Classification","authors":["Kavita Bathe","Nita Patil","Sanjay Patil","Devanand Bathe","Kuldeep Kumar"],"abstract":"Agricultural sector, in association with its allied sectors, plays pivotal role in the progress of a country. In spite of notable contribution, Agricultural sector suffers from several challenges. Crop loss due to diseases is considered as one of the major challenges. Rice is one of the major cereal crops in India. Its production is highly impacted due to crop diseases. Early and accurate detection of rice crop leaf diseases is essential for aforementioned issue. Over the years, conventional methods of crop leaf disease detection are used but with its own limitations. Recent technological advancement in computer vision, deep learning has created new pathways in agricultural sector. Deep learning models require huge data which is quite a great challenge. In such case, building up of a robust deep learning model could work with limited and unbalanced dataset, with good generalization performances. In this study, weighted deep ensemble learning approach is used for performance improvement of rice crop leaf disease classification task. The ensemble method ConvDepthTransEnsembleNet is proposed. To show the effectiveness of proposed work, experiments are conducted on diverse datasets. ConvDepthTransEnsembleNet is a lightweight model which has achieved the accuracy of 96.88% on limited and unbalanced dataset. The experimental results show that proposed model outperforms the individual classifiers based on conventional methods and transfer learning approach in terms of significant reduction in parameters and improved upon generalization performance. The proposed model is highly useful for implementing deep learning models with resources constrained devices.","doi":"10.1007\/s42979-024-02783-8","cleaned_title":"convdepthtransensemblenet: an improved deep learning approach for rice crop leaf disease classification","cleaned_abstract":"agricultural sector, in association with its allied sectors, plays pivotal role in the progress of a country. in spite of notable contribution, agricultural sector suffers from several challenges. crop loss due to diseases is considered as one of the major challenges. rice is one of the major cereal crops in india. its production is highly impacted due to crop diseases. early and accurate detection of rice crop leaf diseases is essential for aforementioned issue. over the years, conventional methods of crop leaf disease detection are used but with its own limitations. recent technological advancement in computer vision, deep learning has created new pathways in agricultural sector. deep learning models require huge data which is quite a great challenge. in such case, building up of a robust deep learning model could work with limited and unbalanced dataset, with good generalization performances. in this study, weighted deep ensemble learning approach is used for performance improvement of rice crop leaf disease classification task. the ensemble method convdepthtransensemblenet is proposed. to show the effectiveness of proposed work, experiments are conducted on diverse datasets. convdepthtransensemblenet is a lightweight model which has achieved the accuracy of 96.88% on limited and unbalanced dataset. the experimental results show that proposed model outperforms the individual classifiers based on conventional methods and transfer learning approach in terms of significant reduction in parameters and improved upon generalization performance. the proposed model is highly useful for implementing deep learning models with resources constrained devices.","key_phrases":["agricultural sector","association","its allied sectors","pivotal role","the progress","a country","spite","notable contribution","agricultural sector","several challenges","crop loss","diseases","the major challenges","rice","the major cereal crops","india","its production","crop diseases","early and accurate detection","rice crop leaf diseases","aforementioned issue","the years","conventional methods","crop leaf disease detection","its own limitations","recent technological advancement","computer vision","deep learning","new pathways","agricultural sector","deep learning models","huge data","which","quite a great challenge","such case","a robust deep learning model","limited and unbalanced dataset","good generalization performances","this study","deep ensemble learning approach","performance improvement","rice crop leaf disease classification task","the ensemble method convdepthtransensemblenet","the effectiveness","proposed work","experiments","diverse datasets","convdepthtransensemblenet","a lightweight model","which","the accuracy","96.88%","limited and unbalanced dataset","the experimental results","proposed model","the individual classifiers","conventional methods","learning approach","terms","significant reduction","parameters","generalization performance","the proposed model","deep learning models","resources constrained devices","rice","one","india","the years","deep","96.88%"]},{"title":"Transfer learning based cascaded deep learning network and mask recognition for COVID-19","authors":["Fengyin Li","Xiaojiao Wang","Yuhong Sun","Tao Li","Junrong Ge"],"abstract":"The COVID-19 is still spreading today, and it has caused great harm to human beings. The system at the entrance of public places such as shopping malls and stations should check whether pedestrians are wearing masks. However, pedestrians often pass the system inspection by wearing cotton masks, scarves, etc. Therefore, the detection system not only needs to check whether pedestrians are wearing masks, but also needs to detect the type of masks. Based on the lightweight network architecture MobilenetV3, this paper proposes a cascaded deep learning network based on transfer learning, and then designs a mask recognition system based on the cascaded deep learning network. By modifying the activation function of the MobilenetV3 output layer and the structure of the model, two MobilenetV3 networks suitable for cascading are obtained. By introducing transfer learning into the training process of two modified MobilenetV3 networks and a multi-task convolutional neural network, the ImagNet underlying parameters of the network models are obtained in advance, which reduces the computational load of the models. The cascaded deep learning network consists of a multi-task convolutional neural network cascaded with these two modified MobilenetV3 networks. A multi-task convolutional neural network is used to detect faces in images, and two modified MobilenetV3 networks are used as the backbone network to extract the features of masks. After comparing with the classification results of the modified MobilenetV3 neural network before cascading, the classification accuracy of the cascading learning network is improved by 7%, and the excellent performance of the cascading network can be seen.","doi":"10.1007\/s11280-023-01149-z","cleaned_title":"transfer learning based cascaded deep learning network and mask recognition for covid-19","cleaned_abstract":"the covid-19 is still spreading today, and it has caused great harm to human beings. the system at the entrance of public places such as shopping malls and stations should check whether pedestrians are wearing masks. however, pedestrians often pass the system inspection by wearing cotton masks, scarves, etc. therefore, the detection system not only needs to check whether pedestrians are wearing masks, but also needs to detect the type of masks. based on the lightweight network architecture mobilenetv3, this paper proposes a cascaded deep learning network based on transfer learning, and then designs a mask recognition system based on the cascaded deep learning network. by modifying the activation function of the mobilenetv3 output layer and the structure of the model, two mobilenetv3 networks suitable for cascading are obtained. by introducing transfer learning into the training process of two modified mobilenetv3 networks and a multi-task convolutional neural network, the imagnet underlying parameters of the network models are obtained in advance, which reduces the computational load of the models. the cascaded deep learning network consists of a multi-task convolutional neural network cascaded with these two modified mobilenetv3 networks. a multi-task convolutional neural network is used to detect faces in images, and two modified mobilenetv3 networks are used as the backbone network to extract the features of masks. after comparing with the classification results of the modified mobilenetv3 neural network before cascading, the classification accuracy of the cascading learning network is improved by 7%, and the excellent performance of the cascading network can be seen.","key_phrases":["the covid-19","it","great harm","human beings","the system","the entrance","public places","shopping malls","stations","pedestrians","masks","pedestrians","the system inspection","cotton masks","scarves","the detection system","pedestrians","masks","the type","masks","the lightweight network architecture","this paper","a cascaded deep learning network","transfer learning","a mask recognition system","the cascaded deep learning network","the activation function","the mobilenetv3 output layer","the structure","the model","two mobilenetv3 networks","transfer","the training process","two modified mobilenetv3 networks","a multi-task convolutional neural network","the imagnet underlying parameters","the network models","advance","which","the computational load","the models","the cascaded deep learning network","a multi-task convolutional neural network","these two modified mobilenetv3 networks","a multi-task convolutional neural network","faces","images","two modified mobilenetv3 networks","the backbone network","the features","masks","the classification results","the modified mobilenetv3 neural network","the classification accuracy","the cascading learning network","7%","the excellent performance","the cascading network","covid-19","today","mobilenetv3","mobilenetv3","two","mobilenetv3","two","mobilenetv3","two","mobilenetv3","two","mobilenetv3","mobilenetv3","7%"]},{"title":"Efficient screening framework for organic solar cells with deep learning and ensemble learning","authors":["Hongshuai Wang","Jie Feng","Zhihao Dong","Lujie Jin","Miaomiao Li","Jianyu Yuan","Youyong Li"],"abstract":"Organic photovoltaics have attracted worldwide interest due to their unique advantages in developing low-cost, lightweight, and flexible power sources. Functional molecular design and synthesis have been put forward to accelerate the discovery of ideal organic semiconductors. However, it is extremely expensive to conduct experimental screening of the wide organic compound space. Here we develop a framework by combining a deep learning model (graph neural network) and an ensemble learning model (Light Gradient Boosting Machine), which enables rapid and accurate screening of organic photovoltaic molecules. This framework establishes the relationship between molecular structure, molecular properties, and device efficiency. Our framework evaluates the chemical structure of the organic photovoltaic molecules directly and accurately. Since it does not involve density functional theory calculations, it makes fast predictions. The reliability of our framework is verified with data from previous reports and our newly synthesized organic molecules. Our work provides an efficient method for developing new organic optoelectronic materials.","doi":"10.1038\/s41524-023-01155-9","cleaned_title":"efficient screening framework for organic solar cells with deep learning and ensemble learning","cleaned_abstract":"organic photovoltaics have attracted worldwide interest due to their unique advantages in developing low-cost, lightweight, and flexible power sources. functional molecular design and synthesis have been put forward to accelerate the discovery of ideal organic semiconductors. however, it is extremely expensive to conduct experimental screening of the wide organic compound space. here we develop a framework by combining a deep learning model (graph neural network) and an ensemble learning model (light gradient boosting machine), which enables rapid and accurate screening of organic photovoltaic molecules. this framework establishes the relationship between molecular structure, molecular properties, and device efficiency. our framework evaluates the chemical structure of the organic photovoltaic molecules directly and accurately. since it does not involve density functional theory calculations, it makes fast predictions. the reliability of our framework is verified with data from previous reports and our newly synthesized organic molecules. our work provides an efficient method for developing new organic optoelectronic materials.","key_phrases":["organic photovoltaics","worldwide interest","their unique advantages","low-cost, lightweight, and flexible power sources","functional molecular design","synthesis","the discovery","ideal organic semiconductors","it","experimental screening","the wide organic compound space","we","a framework","a deep learning model","graph neural network","an ensemble learning model","light gradient boosting machine","which","rapid and accurate screening","organic photovoltaic molecules","this framework","the relationship","molecular structure","molecular properties","device efficiency","our framework","the chemical structure","the organic photovoltaic molecules","it","density functional theory calculations","it","fast predictions","the reliability","our framework","data","previous reports","our newly synthesized organic molecules","our work","an efficient method","new organic optoelectronic materials","the organic photovoltaic molecules"]},{"title":"Identifying strawberry appearance quality based on unsupervised deep learning","authors":["Hongfei Zhu","Xingyu Liu","Hao Zheng","Lianhe Yang","Xuchen Li","Zhongzhi Han"],"abstract":"The strawberry appearance is an essential standard for judging the quality, so it is crucial to accurately identify the strawberry appearance quality for intelligent picking. This study proposed a new strawberry appearance quality detection based on unsupervised deep learning. Firstly, using deep learning (Resnet18, Resnet50, and Resnet101) to extract the strawberry image feature information. And using the t-SNE (t-distribution stochastic neighbor embedding) to reduce the feature vectors\u2019 dimension. Finally, the unsupervised learning method (Gaussian Mixture Model) was used to cluster strawberries\u2019 feature points. The results showed that: (1) the clustering performance based on Resnet101 was effective in 2-dimensional space, the cluster accuracy was 94.89%, and the validation accuracy was 91.79%. (2) The clustering method based on Resnet50 had good performance in the 3-dimensional space, the cluster accuracy was 96.10%, and the validation accuracy was 93.08%. (3) The accuracy of deep features plus RF (random forest) was 95.00% under limited data. Thus this method will promote intelligent picking strawberry equipment and it will overcome the supervised learning drawback that divides image datasets according to prior knowledge.","doi":"10.1007\/s11119-023-10085-x","cleaned_title":"identifying strawberry appearance quality based on unsupervised deep learning","cleaned_abstract":"the strawberry appearance is an essential standard for judging the quality, so it is crucial to accurately identify the strawberry appearance quality for intelligent picking. this study proposed a new strawberry appearance quality detection based on unsupervised deep learning. firstly, using deep learning (resnet18, resnet50, and resnet101) to extract the strawberry image feature information. and using the t-sne (t-distribution stochastic neighbor embedding) to reduce the feature vectors\u2019 dimension. finally, the unsupervised learning method (gaussian mixture model) was used to cluster strawberries\u2019 feature points. the results showed that: (1) the clustering performance based on resnet101 was effective in 2-dimensional space, the cluster accuracy was 94.89%, and the validation accuracy was 91.79%. (2) the clustering method based on resnet50 had good performance in the 3-dimensional space, the cluster accuracy was 96.10%, and the validation accuracy was 93.08%. (3) the accuracy of deep features plus rf (random forest) was 95.00% under limited data. thus this method will promote intelligent picking strawberry equipment and it will overcome the supervised learning drawback that divides image datasets according to prior knowledge.","key_phrases":["the strawberry appearance","an essential standard","the quality","it","the strawberry appearance quality","intelligent picking","this study","a new strawberry appearance quality detection","unsupervised deep learning","deep learning","resnet18","resnet50","resnet101","the strawberry image feature information","(t-distribution","the feature vectors\u2019 dimension","the unsupervised learning method","gaussian mixture model","strawberries","feature points","the results","the clustering performance","resnet101","2-dimensional space","the cluster accuracy","94.89%","the validation accuracy","91.79%","(2) the clustering method","resnet50","good performance","the 3-dimensional space","the cluster accuracy","96.10%","the validation accuracy","93.08%","(3) the accuracy","deep features","rf (random forest","95.00%","limited data","this method","intelligent picking strawberry equipment","it","the supervised learning drawback","that","image datasets","prior knowledge","strawberry","firstly","resnet18","resnet50","strawberry","gaussian","1","2","94.89%","91.79%","2","resnet50","3","96.10%","93.08%","3","95.00%"]},{"title":"A survey on advanced machine learning and deep learning techniques assisting in renewable energy generation","authors":["Sri Revathi B."],"abstract":"The sustainability of the earth depends on renewable energy. Forecasting the output of renewable energy has a big impact on how we operate and manage our power networks. Accurate forecasting of renewable energy generation is crucial to ensuring grid dependability and permanence and reducing the risk and cost of the energy market and infrastructure. Although there are several approaches to forecasting solar radiation on a global scale, the two most common ones are machine learning algorithms and cloud pictures combined with physical models. The objective is to present a summary of machine learning-based techniques for solar irradiation forecasting in this context. Renewable energy is being used more and more in the world\u2019s energy grid. Numerous strategies, including hybrids, physical models, statistical approaches, and artificial intelligence techniques, have been developed to anticipate the use of renewable energy. This paper examines methods for forecasting renewable energy based on deep learning and machine learning. Review and analysis of deep learning and machine learning forecasts for renewable energy come first. The second paragraph describes metaheuristic optimization techniques for renewable energy. The third topic was the open issue of projecting renewable energy. I will wrap up with a few potential future job objectives.","doi":"10.1007\/s11356-023-29064-w","cleaned_title":"a survey on advanced machine learning and deep learning techniques assisting in renewable energy generation","cleaned_abstract":"the sustainability of the earth depends on renewable energy. forecasting the output of renewable energy has a big impact on how we operate and manage our power networks. accurate forecasting of renewable energy generation is crucial to ensuring grid dependability and permanence and reducing the risk and cost of the energy market and infrastructure. although there are several approaches to forecasting solar radiation on a global scale, the two most common ones are machine learning algorithms and cloud pictures combined with physical models. the objective is to present a summary of machine learning-based techniques for solar irradiation forecasting in this context. renewable energy is being used more and more in the world\u2019s energy grid. numerous strategies, including hybrids, physical models, statistical approaches, and artificial intelligence techniques, have been developed to anticipate the use of renewable energy. this paper examines methods for forecasting renewable energy based on deep learning and machine learning. review and analysis of deep learning and machine learning forecasts for renewable energy come first. the second paragraph describes metaheuristic optimization techniques for renewable energy. the third topic was the open issue of projecting renewable energy. i will wrap up with a few potential future job objectives.","key_phrases":["the sustainability","the earth","renewable energy","the output","renewable energy","a big impact","we","our power networks","accurate forecasting","renewable energy generation","grid dependability","permanence","the risk","cost","the energy market","infrastructure","several approaches","solar radiation","a global scale","the two most common ones","algorithms","cloud pictures","physical models","the objective","a summary","machine learning-based techniques","solar irradiation forecasting","this context","renewable energy","the world\u2019s energy grid","numerous strategies","hybrids","physical models","statistical approaches","artificial intelligence techniques","the use","renewable energy","this paper","methods","renewable energy","deep learning","machine learning","review","analysis","deep learning","machine","forecasts","renewable energy","the second paragraph","metaheuristic optimization techniques","renewable energy","the third topic","the open issue","renewable energy","i","a few potential future job objectives","earth","two","first","second","third"]},{"title":"Varicocele detection in ultrasound images using deep learning","authors":["Omar AlZoubi","Mohammad Abu Awad","Ayman M. Abdalla","Laaly Samrraie"],"abstract":"Varicocele is a disease exhibited by an abnormal dilation of the scrotal venous pampiniform plexus. It is a common cause of infertility and in some cases may cause pain or discomfort. This paper presents a new approach to automatic varicocele classification by employing convolutional neural networks with deep learning. The available dataset consists of images converted into different color modes. Each color mode dataset is partitioned, augmented, and trained by employing a deep learning network. Experiments were run with all possible combinations of four different pre-trained models and over three color modes to determine the best combination that achieves the highest performance with and without augmentation. The implementation results of training and testing were evaluated with several metrics. The analysis of the results demonstrated the efficacy of the proposed system as it identified and classified varicocele with a relatively high accuracy that surpassed the previous works with a significant improvement.","doi":"10.1007\/s11042-023-17865-7","cleaned_title":"varicocele detection in ultrasound images using deep learning","cleaned_abstract":"varicocele is a disease exhibited by an abnormal dilation of the scrotal venous pampiniform plexus. it is a common cause of infertility and in some cases may cause pain or discomfort. this paper presents a new approach to automatic varicocele classification by employing convolutional neural networks with deep learning. the available dataset consists of images converted into different color modes. each color mode dataset is partitioned, augmented, and trained by employing a deep learning network. experiments were run with all possible combinations of four different pre-trained models and over three color modes to determine the best combination that achieves the highest performance with and without augmentation. the implementation results of training and testing were evaluated with several metrics. the analysis of the results demonstrated the efficacy of the proposed system as it identified and classified varicocele with a relatively high accuracy that surpassed the previous works with a significant improvement.","key_phrases":["varicocele","a disease","an abnormal dilation","the scrotal venous pampiniform plexus","it","a common cause","infertility","some cases","pain","discomfort","this paper","a new approach","automatic varicocele classification","convolutional neural networks","deep learning","the available dataset","images","different color modes","each color mode dataset","a deep learning network","experiments","all possible combinations","four different pre-trained models","over three color modes","the best combination","that","the highest performance","augmentation","the implementation results","training","testing","several metrics","the analysis","the results","the efficacy","the proposed system","it","a relatively high accuracy","that","the previous works","a significant improvement","four","three"]},{"title":"Enhancing image retrieval through entropy-based deep metric learning","authors":["Kambiz Rahbar","Fatemeh Taheri"],"abstract":"The increasing demand for effective retrieval from image datasets has been driven by the rapid growth of digital images. Image retrieval is a method for creating a structured database based on searching for similar images to a user's query image. Effective image representation and extraction of distinctive features are among the challenges faced by image retrieval systems. The proposed approach utilizes a triplet loss function based on binary cross-entropy to train a Siamese network, allowing for deep metric learning and creating a discriminative feature space with maximum discrimination between classes and minimum intra-class distance. In this approach, a pre-trained neural network is serialized with Siamese network. Initially, image features are extracted using a pre-trained convolutional neural network. Then, a Siamese network is trained to create a discriminative feature space using deep metric learning. Learning in this method is controlled by a triplet loss function based on binary cross-entropy with anchor, positive, and negative samples. The transformation from the initial feature space to the discriminative feature space is achieved through deep metric learning in the Siamese network. The proposed approach proves to effective in discriminating features of diverse classes in the dataset. Visualization of the feature space is also demonstrated using the t-SNE statistical technique. Furthermore, the explainability of the proposed approach is presented by examining the Shapley value. The proposed approach was examined on the Corel10K, Caltech101, and Caltech256 datasets. The best reported results with the precision metric were 0.988 for the Corel10K dataset and 0.951 and 0.902 for the Caltech101 and Caltech256 datasets, respectively. Experimental results show that the proposed approach effectively improves the retrieval results of similar samples to the query image creating discrimination in the feature vectors, even with dimensionality reduction.","doi":"10.1007\/s11042-024-19296-4","cleaned_title":"enhancing image retrieval through entropy-based deep metric learning","cleaned_abstract":"the increasing demand for effective retrieval from image datasets has been driven by the rapid growth of digital images. image retrieval is a method for creating a structured database based on searching for similar images to a user's query image. effective image representation and extraction of distinctive features are among the challenges faced by image retrieval systems. the proposed approach utilizes a triplet loss function based on binary cross-entropy to train a siamese network, allowing for deep metric learning and creating a discriminative feature space with maximum discrimination between classes and minimum intra-class distance. in this approach, a pre-trained neural network is serialized with siamese network. initially, image features are extracted using a pre-trained convolutional neural network. then, a siamese network is trained to create a discriminative feature space using deep metric learning. learning in this method is controlled by a triplet loss function based on binary cross-entropy with anchor, positive, and negative samples. the transformation from the initial feature space to the discriminative feature space is achieved through deep metric learning in the siamese network. the proposed approach proves to effective in discriminating features of diverse classes in the dataset. visualization of the feature space is also demonstrated using the t-sne statistical technique. furthermore, the explainability of the proposed approach is presented by examining the shapley value. the proposed approach was examined on the corel10k, caltech101, and caltech256 datasets. the best reported results with the precision metric were 0.988 for the corel10k dataset and 0.951 and 0.902 for the caltech101 and caltech256 datasets, respectively. experimental results show that the proposed approach effectively improves the retrieval results of similar samples to the query image creating discrimination in the feature vectors, even with dimensionality reduction.","key_phrases":["the increasing demand","effective retrieval","image datasets","the rapid growth","digital images","image retrieval","a method","a structured database","similar images","a user's query image","effective image representation","extraction","distinctive features","the challenges","image retrieval systems","the proposed approach","a triplet loss function","binary cross","-","a siamese network","deep metric learning","a discriminative feature space","maximum discrimination","classes","minimum intra-class distance","this approach","a pre-trained neural network","siamese network","image features","a pre-trained convolutional neural network","a siamese network","a discriminative feature space","deep metric learning","this method","a triplet loss function","binary cross","-","anchor","negative samples","the transformation","the initial feature space","the discriminative feature space","deep metric learning","the siamese network","the proposed approach","features","diverse classes","the dataset","visualization","the feature space","the t-sne statistical technique","the explainability","the proposed approach","the shapley value","the proposed approach","the corel10k","caltech101","caltech256 datasets","the best reported results","the precision metric","the corel10k dataset","the caltech101","caltech256","datasets","experimental results","the proposed approach","the retrieval results","similar samples","the query image","discrimination","the feature vectors","dimensionality reduction","siamese","corel10k","0.988","corel10k","0.951","0.902"]},{"title":"Legal sentence boundary detection using hybrid deep learning and statistical models","authors":["Reshma Sheik","Sneha Rao Ganta","S. Jaya Nirmala"],"abstract":"Sentence boundary detection (SBD) represents an important first step in natural language processing since accurately identifying sentence boundaries significantly impacts downstream applications. Nevertheless, detecting sentence boundaries within legal texts poses a unique and challenging problem due to their distinct structural and linguistic features. Our approach utilizes deep learning models to leverage delimiter and surrounding context information as input, enabling precise detection of sentence boundaries in English legal texts. We evaluate various deep learning models, including domain-specific transformer models like LegalBERT and CaseLawBERT. To assess the efficacy of our deep learning models, we compare them with a state-of-the-art domain-specific statistical conditional random field (CRF) model. After considering model size, F1-score, and inference time, we identify the Convolutional Neural Network Model (CNN) as the top-performing deep learning model. To further enhance performance, we integrate the features of the CNN model into the subsequent CRF model, creating a hybrid architecture that combines the strengths of both models. Our experiments demonstrate that the hybrid model outperforms the baseline model, achieving a 4% improvement in the F1-score. Additional experiments showcase the superiority of the hybrid model over SBD open-source libraries when confronted with an out-of-domain test set. These findings underscore the importance of efficient SBD in legal texts and emphasize the advantages of employing deep learning models and hybrid architectures to achieve optimal performance.","doi":"10.1007\/s10506-024-09394-x","cleaned_title":"legal sentence boundary detection using hybrid deep learning and statistical models","cleaned_abstract":"sentence boundary detection (sbd) represents an important first step in natural language processing since accurately identifying sentence boundaries significantly impacts downstream applications. nevertheless, detecting sentence boundaries within legal texts poses a unique and challenging problem due to their distinct structural and linguistic features. our approach utilizes deep learning models to leverage delimiter and surrounding context information as input, enabling precise detection of sentence boundaries in english legal texts. we evaluate various deep learning models, including domain-specific transformer models like legalbert and caselawbert. to assess the efficacy of our deep learning models, we compare them with a state-of-the-art domain-specific statistical conditional random field (crf) model. after considering model size, f1-score, and inference time, we identify the convolutional neural network model (cnn) as the top-performing deep learning model. to further enhance performance, we integrate the features of the cnn model into the subsequent crf model, creating a hybrid architecture that combines the strengths of both models. our experiments demonstrate that the hybrid model outperforms the baseline model, achieving a 4% improvement in the f1-score. additional experiments showcase the superiority of the hybrid model over sbd open-source libraries when confronted with an out-of-domain test set. these findings underscore the importance of efficient sbd in legal texts and emphasize the advantages of employing deep learning models and hybrid architectures to achieve optimal performance.","key_phrases":["boundary detection","sbd","an important first step","natural language processing","sentence boundaries","downstream applications","sentence boundaries","legal texts","a unique and challenging problem","their distinct structural and linguistic features","our approach","deep learning models","delimiter","context information","input","precise detection","sentence boundaries","english legal texts","we","various deep learning models","domain-specific transformer models","legalbert","caselawbert","the efficacy","our deep learning models","we","them","the-art","crf","model size, f1-score, and inference time","we","the convolutional neural network model","cnn","the top-performing deep learning model","performance","we","the features","the cnn model","the subsequent crf model","a hybrid architecture","that","the strengths","both models","our experiments","the hybrid model","the baseline model","a 4% improvement","the f1-score","additional experiments","the superiority","the hybrid model","sbd open-source libraries","domain","these findings","the importance","efficient sbd","legal texts","the advantages","deep learning models","hybrid architectures","optimal performance","first","english","cnn","cnn","4%"]},{"title":"Deep Learning Methods for Vibration-Based Structural Health Monitoring: A Review","authors":["Hao Wang","Baoli Wang","Caixia Cui"],"abstract":"The vibration signal is an effective diagnostic tool in structural health monitoring (SHM) fields that is closely related to abnormal states. Deep learning methods have got remarkable success in utilizing vibration signals for damage detection. This paper presents a systematic review of deep learning methods for SHM, focusing on the utilization of vibration signal data from different model perspectives. In recent years, there has been a significant increase in research on deep learning for vibration-based SHM. The accuracy of such works is equivalent to that of traditional machine learning approaches, and better results could be achieved by integrating multiple approaches. Furthermore, we found that transfer learning methods yield promising results when limited data are available to train the model. This paper aims to comprehensively review deep learning research on health monitoring using vibration signal data from multiple perspectives, with a particular emphasis on transfer learning methods for SHM. It fills the gap that existing reviews lack in the discussion of transfer learning for SHM. Finally, we analyze the challenges faced by current research and provide recommendations for future work.","doi":"10.1007\/s40996-023-01287-4","cleaned_title":"deep learning methods for vibration-based structural health monitoring: a review","cleaned_abstract":"the vibration signal is an effective diagnostic tool in structural health monitoring (shm) fields that is closely related to abnormal states. deep learning methods have got remarkable success in utilizing vibration signals for damage detection. this paper presents a systematic review of deep learning methods for shm, focusing on the utilization of vibration signal data from different model perspectives. in recent years, there has been a significant increase in research on deep learning for vibration-based shm. the accuracy of such works is equivalent to that of traditional machine learning approaches, and better results could be achieved by integrating multiple approaches. furthermore, we found that transfer learning methods yield promising results when limited data are available to train the model. this paper aims to comprehensively review deep learning research on health monitoring using vibration signal data from multiple perspectives, with a particular emphasis on transfer learning methods for shm. it fills the gap that existing reviews lack in the discussion of transfer learning for shm. finally, we analyze the challenges faced by current research and provide recommendations for future work.","key_phrases":["the vibration signal","an effective diagnostic tool","structural health monitoring","shm) fields","that","abnormal states","deep learning methods","remarkable success","vibration signals","damage detection","this paper","a systematic review","deep learning methods","shm","the utilization","vibration signal data","different model perspectives","recent years","a significant increase","research","deep learning","vibration-based shm","the accuracy","such works","that","traditional machine learning approaches","better results","multiple approaches","we","transfer learning methods","promising results","limited data","the model","this paper","deep learning research","health monitoring","vibration signal data","multiple perspectives","a particular emphasis","transfer learning methods","shm","it","the gap","existing reviews","the discussion","shm","we","the challenges","current research","recommendations","future work","recent years"]},{"title":"Hyperbolic Deep Learning in Computer Vision: A Survey","authors":["Pascal Mettes","Mina Ghadimi Atigh","Martin Keller-Ressel","Jeffrey Gu","Serena Yeung"],"abstract":"Deep representation learning is a ubiquitous part of modern computer vision. While Euclidean space has been the de facto standard manifold for learning visual representations, hyperbolic space has recently gained rapid traction for learning in computer vision. Specifically, hyperbolic learning has shown a strong potential to embed hierarchical structures, learn from limited samples, quantify uncertainty, add robustness, limit error severity, and more. In this paper, we provide a categorization and in-depth overview of current literature on hyperbolic learning for computer vision. We research both supervised and unsupervised literature and identify three main research themes in each direction. We outline how hyperbolic learning is performed in all themes and discuss the main research problems that benefit from current advances in hyperbolic learning for computer vision. Moreover, we provide a high-level intuition behind hyperbolic geometry and outline open research questions to further advance research in this direction.","doi":"10.1007\/s11263-024-02043-5","cleaned_title":"hyperbolic deep learning in computer vision: a survey","cleaned_abstract":"deep representation learning is a ubiquitous part of modern computer vision. while euclidean space has been the de facto standard manifold for learning visual representations, hyperbolic space has recently gained rapid traction for learning in computer vision. specifically, hyperbolic learning has shown a strong potential to embed hierarchical structures, learn from limited samples, quantify uncertainty, add robustness, limit error severity, and more. in this paper, we provide a categorization and in-depth overview of current literature on hyperbolic learning for computer vision. we research both supervised and unsupervised literature and identify three main research themes in each direction. we outline how hyperbolic learning is performed in all themes and discuss the main research problems that benefit from current advances in hyperbolic learning for computer vision. moreover, we provide a high-level intuition behind hyperbolic geometry and outline open research questions to further advance research in this direction.","key_phrases":["deep representation learning","a ubiquitous part","modern computer vision","euclidean space","the de facto standard manifold","visual representations","hyperbolic space","rapid traction","computer vision","hyperbolic learning","a strong potential","hierarchical structures","limited samples","quantify uncertainty","robustness","error severity","this paper","we","a categorization","-depth","current literature","hyperbolic learning","computer vision","we","literature","three main research themes","each direction","we","hyperbolic learning","all themes","the main research problems","that","current advances","hyperbolic learning","computer vision","we","a high-level intuition","hyperbolic geometry","open research questions","research","this direction","three"]},{"title":"From distributed machine to distributed deep learning: a comprehensive survey","authors":["Mohammad Dehghani","Zahra Yazdanparast"],"abstract":"Artificial intelligence has made remarkable progress in handling complex tasks, thanks to advances in hardware acceleration and machine learning algorithms. However, to acquire more accurate outcomes and solve more complex issues, algorithms should be trained with more data. Processing this huge amount of data could be time-consuming and require a great deal of computation. To address these issues, distributed machine learning has been proposed, which involves distributing the data and algorithm across several machines. There has been considerable effort put into developing distributed machine learning algorithms, and different methods have been proposed so far. We divide these algorithms in classification and clustering (traditional machine learning), deep learning and deep reinforcement learning groups. Distributed deep learning has gained more attention in recent years and most of the studies have focused on this approach. Therefore, we mostly concentrate on this category. Based on the investigation of the mentioned algorithms, we highlighted the limitations that should be addressed in future research.","doi":"10.1186\/s40537-023-00829-x","cleaned_title":"from distributed machine to distributed deep learning: a comprehensive survey","cleaned_abstract":"artificial intelligence has made remarkable progress in handling complex tasks, thanks to advances in hardware acceleration and machine learning algorithms. however, to acquire more accurate outcomes and solve more complex issues, algorithms should be trained with more data. processing this huge amount of data could be time-consuming and require a great deal of computation. to address these issues, distributed machine learning has been proposed, which involves distributing the data and algorithm across several machines. there has been considerable effort put into developing distributed machine learning algorithms, and different methods have been proposed so far. we divide these algorithms in classification and clustering (traditional machine learning), deep learning and deep reinforcement learning groups. distributed deep learning has gained more attention in recent years and most of the studies have focused on this approach. therefore, we mostly concentrate on this category. based on the investigation of the mentioned algorithms, we highlighted the limitations that should be addressed in future research.","key_phrases":["artificial intelligence","remarkable progress","complex tasks","thanks","advances","hardware acceleration","machine learning","more accurate outcomes","more complex issues","algorithms","more data","this huge amount","data","a great deal","computation","these issues","distributed machine learning","which","the data","algorithm","several machines","considerable effort","developing distributed machine learning algorithms","different methods","we","these algorithms","classification","clustering","(traditional machine learning","deep learning","deep reinforcement learning groups","distributed deep learning","more attention","recent years","the studies","this approach","we","this category","the investigation","the mentioned algorithms","we","the limitations","that","future research","recent years"]},{"title":"Deep learning-assisted characterization of nanoparticle growth processes: unveiling SAXS structure evolution","authors":["Yikun Li","Lunyang Liu","Xiaoning Zhao","Shuming Zhou","Xuehui Wu","Yuecheng Lai","Zhongjun Chen","Jizhong Chen","Xueqing Xing"],"abstract":"PurposeThe purpose of this study is to explore deep learning methods for processing high-throughput small-angle X-ray scattering (SAXS) experimental data.MethodsThe deep learning algorithm was trained and validated using simulated SAXS data, which were generated in batches based on the theoretical SAXS formula using Python code. Our self-developed SAXSNET, a convolutional neural network based on PyTorch, was employed to classify SAXS data for various shapes of nanoparticles. Additionally, we conducted comparative analysis of classification algorithms including ResNet-18, ResNet-34 and Vision Transformer. Random Forest and XGboost regression algorithms were used for the nanoparticle size prediction. Finally, we evaluated the aforementioned shape classification and numerical regression methods using actual experimental data. A pipeline segment is established for the processing of SAXS data, incorporating deep learning classification algorithms and numerical regression algorithms.ResultsAfter being trained with simulated data, the four deep learning algorithms achieved a prediction accuracy of over 96% on the validation set. The fine-tuned deep learning model demonstrated robust generalization capabilities for predicting the shapes of experimental data, enabling rapid and accurate identification of morphological changes in nanoparticles during experiments. The Random Forest and XGboost regression algorithms can simultaneously provide faster and more accurate predictions of nanoparticle size.ConclusionThe pipeline segment constructed in this study, integrating deep learning classification and regression algorithms, enables real-time processing of high-throughput SAXS data. It aims to effectively mitigates the impact of human factors on data processing results and enhances the standardization, automation, and intelligence of synchrotron radiation experiments.","doi":"10.1007\/s41605-024-00471-y","cleaned_title":"deep learning-assisted characterization of nanoparticle growth processes: unveiling saxs structure evolution","cleaned_abstract":"purposethe purpose of this study is to explore deep learning methods for processing high-throughput small-angle x-ray scattering (saxs) experimental data.methodsthe deep learning algorithm was trained and validated using simulated saxs data, which were generated in batches based on the theoretical saxs formula using python code. our self-developed saxsnet, a convolutional neural network based on pytorch, was employed to classify saxs data for various shapes of nanoparticles. additionally, we conducted comparative analysis of classification algorithms including resnet-18, resnet-34 and vision transformer. random forest and xgboost regression algorithms were used for the nanoparticle size prediction. finally, we evaluated the aforementioned shape classification and numerical regression methods using actual experimental data. a pipeline segment is established for the processing of saxs data, incorporating deep learning classification algorithms and numerical regression algorithms.resultsafter being trained with simulated data, the four deep learning algorithms achieved a prediction accuracy of over 96% on the validation set. the fine-tuned deep learning model demonstrated robust generalization capabilities for predicting the shapes of experimental data, enabling rapid and accurate identification of morphological changes in nanoparticles during experiments. the random forest and xgboost regression algorithms can simultaneously provide faster and more accurate predictions of nanoparticle size.conclusionthe pipeline segment constructed in this study, integrating deep learning classification and regression algorithms, enables real-time processing of high-throughput saxs data. it aims to effectively mitigates the impact of human factors on data processing results and enhances the standardization, automation, and intelligence of synchrotron radiation experiments.","key_phrases":["purposethe purpose","this study","deep learning methods","saxs","data.methodsthe deep learning algorithm","simulated saxs data","which","batches","the theoretical saxs formula","python code","our self-developed saxsnet","a convolutional neural network","pytorch","saxs data","various shapes","nanoparticles","we","comparative analysis","classification algorithms","resnet-18","resnet-34","vision transformer","random forest","xgboost regression algorithms","the nanoparticle size prediction","we","the aforementioned shape classification and numerical regression methods","actual experimental data","a pipeline segment","the processing","saxs data","deep learning classification algorithms","numerical regression","simulated data","the four deep learning algorithms","a prediction accuracy","over 96%","the validation set","the fine-tuned deep learning model","robust generalization capabilities","the shapes","experimental data","rapid and accurate identification","morphological changes","nanoparticles","experiments","the random forest","xgboost regression algorithms","faster and more accurate predictions","nanoparticle","size.conclusionthe pipeline segment","this study","deep learning classification and regression algorithms","real-time processing","high-throughput saxs data","it","the impact","human factors","data processing results","the standardization","automation","intelligence","synchrotron radiation experiments","resnet-18","resnet-34","four","over 96%"]},{"title":"Machine learning, deep learning and hernia surgery. Are we pushing the limits of abdominal core health? A qualitative systematic review","authors":["D. L. Lima","J. Kasakewitch","D. Q. Nguyen","R. Nogueira","L. T. Cavazzola","B. T. Heniford","F. Malcher"],"abstract":"IntroductionThis systematic review aims to evaluate the use of machine learning and artificial intelligence in hernia surgery.MethodsThe PRISMA guidelines were followed throughout this systematic review. The ROBINS\u2014I and Rob 2 tools were used to perform qualitative assessment of all studies included in this review. Recommendations were then summarized for the following pre-defined key items: protocol, research question, search strategy, study eligibility, data extraction, study design, risk of bias, publication bias, and statistical analysis.ResultsA total of 13 articles were ultimately included for this review, describing the use of machine learning and deep learning for hernia surgery. All studies were published from 2020 to 2023. Articles varied regarding the population studied, type of machine learning or Deep Learning Model (DLM) used, and hernia type. Of the thirteen included studies, all included either inguinal, ventral, or incisional hernias. Four studies evaluated recognition of surgical steps during inguinal hernia repair videos. Two studies predicted outcomes using image-based DMLs. Seven studies developed and validated deep learning algorithms to predict outcomes and identify factors associated with postoperative complications.ConclusionThe use of ML for abdominal wall reconstruction has been shown to be a promising tool for predicting outcomes and identifying factors that could lead to postoperative complications.","doi":"10.1007\/s10029-024-03069-x","cleaned_title":"machine learning, deep learning and hernia surgery. are we pushing the limits of abdominal core health? a qualitative systematic review","cleaned_abstract":"introductionthis systematic review aims to evaluate the use of machine learning and artificial intelligence in hernia surgery.methodsthe prisma guidelines were followed throughout this systematic review. the robins\u2014i and rob 2 tools were used to perform qualitative assessment of all studies included in this review. recommendations were then summarized for the following pre-defined key items: protocol, research question, search strategy, study eligibility, data extraction, study design, risk of bias, publication bias, and statistical analysis.resultsa total of 13 articles were ultimately included for this review, describing the use of machine learning and deep learning for hernia surgery. all studies were published from 2020 to 2023. articles varied regarding the population studied, type of machine learning or deep learning model (dlm) used, and hernia type. of the thirteen included studies, all included either inguinal, ventral, or incisional hernias. four studies evaluated recognition of surgical steps during inguinal hernia repair videos. two studies predicted outcomes using image-based dmls. seven studies developed and validated deep learning algorithms to predict outcomes and identify factors associated with postoperative complications.conclusionthe use of ml for abdominal wall reconstruction has been shown to be a promising tool for predicting outcomes and identifying factors that could lead to postoperative complications.","key_phrases":["introductionthis systematic review","the use","machine learning","artificial intelligence","hernia surgery.methodsthe prisma guidelines","this systematic review","the robins","i","rob","2 tools","qualitative assessment","all studies","this review","recommendations","the following pre-defined key items","protocol","research question","search strategy","study eligibility","data extraction","study design","risk","bias","publication bias","statistical analysis.resultsa total","13 articles","this review","the use","machine learning","deep learning","hernia surgery","all studies","articles","the population","type","machine learning","deep learning model","dlm","hernia type","the thirteen included studies","all","either inguinal, ventral, or incisional hernias","four studies","recognition","surgical steps","inguinal hernia repair videos","two studies","outcomes","image-based dmls","seven studies","deep learning algorithms","outcomes","factors","postoperative complications.conclusionthe use","ml","abdominal wall reconstruction","a promising tool","outcomes","factors","that","postoperative complications","hernia","2","13","2020","thirteen","four","two","seven"]},{"title":"A survey: evolutionary deep learning","authors":["Yifan Li","Jing Liu"],"abstract":"Deep learning (DL) has made remarkable progress on various real-world tasks, but its construction pipeline strongly relies on human scientists. Furthermore, evolutionary computing (EC), as an optimization tool based on the biological evolution mechanism, has good performance on complex optimization problems. It provides a new way to construct DL models and has generated many sparks in the DL field, especially in automatic machine learning (AutoML). Although many reviews have been conducted on AutoML, in recent years, few comprehensive works have studied on the application of EC in DL, which is called evolutionary deep learning (EDL). After a thorough investigation, we think that EDL can be divided into four parts: (1) learning rule optimization, (2) hyperparameter optimization, (3) neural architecture search, and (4) other EDL-related works. In this work, we introduce the classic optimization methods and the challenges of EDL with respect to these four parts, review the related work, and then present the future research prospects. This work clearly and comprehensively reviews the concept and research content of EDL, which can help readers quickly find the intersection between EC and DL and seek their inspiration.","doi":"10.1007\/s00500-023-08316-4","cleaned_title":"a survey: evolutionary deep learning","cleaned_abstract":"deep learning (dl) has made remarkable progress on various real-world tasks, but its construction pipeline strongly relies on human scientists. furthermore, evolutionary computing (ec), as an optimization tool based on the biological evolution mechanism, has good performance on complex optimization problems. it provides a new way to construct dl models and has generated many sparks in the dl field, especially in automatic machine learning (automl). although many reviews have been conducted on automl, in recent years, few comprehensive works have studied on the application of ec in dl, which is called evolutionary deep learning (edl). after a thorough investigation, we think that edl can be divided into four parts: (1) learning rule optimization, (2) hyperparameter optimization, (3) neural architecture search, and (4) other edl-related works. in this work, we introduce the classic optimization methods and the challenges of edl with respect to these four parts, review the related work, and then present the future research prospects. this work clearly and comprehensively reviews the concept and research content of edl, which can help readers quickly find the intersection between ec and dl and seek their inspiration.","key_phrases":["deep learning","dl","remarkable progress","various real-world tasks","its construction pipeline","human scientists","evolutionary computing","ec","an optimization tool","the biological evolution mechanism","good performance","complex optimization problems","it","a new way","dl models","many sparks","the dl field","automatic machine learning","automl","many reviews","automl","recent years","few comprehensive works","the application","ec","dl","which","evolutionary deep learning","edl","a thorough investigation","we","edl","four parts","rule optimization","(2) hyperparameter optimization","3) neural architecture search","(4) other edl-related works","this work","we","the classic optimization methods","the challenges","edl","respect","these four parts","the related work","the future research prospects","this work","the concept","research content","edl","which","readers","the intersection","ec","dl","their inspiration","ec","recent years","ec","four","1","2","3","4","four","ec","dl"]},{"title":"Impact of the Preprocessing Steps in Deep Learning-Based Image Classifications","authors":["H. James Deva Koresh"],"abstract":"Deep learning softwares are designed using artificial neural networks for various applications by training and testing them with an appropriate dataset. The raw image samples available in the dataset may contain noisy and unclear information due to radiation, heat and poor lighting conditions. Therefore, the researchers are trying to filter and enhance such noisy images through preprocessing steps for providing a valid feature information to the neural network layers included in the deep learning software. However, there are certain claims that roam around the researchers such as an image may lose some useful information when it is not preprocessed with an appropriate filter or enhancement technique. Hence, the work reviews the efficacy of the methodologies that are designed with and without a preprocessing step. Also, the work summarizes the common reasons and statements highlighted by the researchers for using and avoiding the preprocessing steps on designing a deep learning approach. The study is conducted to provide a clarity toward the requirement and non-requirement of preprocessing step in a deep learning software.","doi":"10.1007\/s40009-023-01372-2","cleaned_title":"impact of the preprocessing steps in deep learning-based image classifications","cleaned_abstract":"deep learning softwares are designed using artificial neural networks for various applications by training and testing them with an appropriate dataset. the raw image samples available in the dataset may contain noisy and unclear information due to radiation, heat and poor lighting conditions. therefore, the researchers are trying to filter and enhance such noisy images through preprocessing steps for providing a valid feature information to the neural network layers included in the deep learning software. however, there are certain claims that roam around the researchers such as an image may lose some useful information when it is not preprocessed with an appropriate filter or enhancement technique. hence, the work reviews the efficacy of the methodologies that are designed with and without a preprocessing step. also, the work summarizes the common reasons and statements highlighted by the researchers for using and avoiding the preprocessing steps on designing a deep learning approach. the study is conducted to provide a clarity toward the requirement and non-requirement of preprocessing step in a deep learning software.","key_phrases":["deep learning softwares","artificial neural networks","various applications","training","them","an appropriate dataset","the raw image samples","the dataset","noisy and unclear information","radiation","heat","poor lighting conditions","the researchers","such noisy images","steps","a valid feature information","the neural network layers","the deep learning software","certain claims","roam","the researchers","an image","some useful information","it","an appropriate filter","enhancement technique","the work","the efficacy","the methodologies","that","a preprocessing step","the work","the common reasons","statements","the researchers","the preprocessing steps","a deep learning approach","the study","a clarity","the requirement","-","requirement","step","a deep learning software"]},{"title":"Deep learning applications in games: a survey from a data perspective","authors":["Zhipeng Hu","Yu Ding","Runze Wu","Lincheng Li","Rongsheng Zhang","Yujing Hu","Feng Qiu","Zhimeng Zhang","Kai Wang","Shiwei Zhao","Yongqiang Zhang","Ji Jiang","Yadong Xi","Jiashu Pu","Wei Zhang","Suzhen Wang","Ke Chen","Tianze Zhou","Jiarui Chen","Yan Song","Tangjie Lv","Changjie Fan"],"abstract":"This paper presents a comprehensive review of deep learning applications in the video game industry, focusing on how these techniques can be utilized in game development, experience, and operation. As relying on computation techniques, the game world can be viewed as an integration of various complex data. This examines the use of deep learning in processing various types of game data. The paper classifies the game data into asset data, interaction data, and player data, according to their utilization in game development, experience, and operation, respectively. Specifically, this paper discusses deep learning applications in generating asset data such as object images, 3D scenes, avatar models, and facial animations; enhancing interaction data through improved text-based conversations and decision-making behaviors; and analyzing player data for cheat detection and match-making purposes. Although this review may not cover all existing applications of deep learning, it aims to provide a thorough presentation of the current state of deep learning in the gaming industry and its potential to revolutionize game production by reducing costs and improving the overall player experience.","doi":"10.1007\/s10489-023-05094-2","cleaned_title":"deep learning applications in games: a survey from a data perspective","cleaned_abstract":"this paper presents a comprehensive review of deep learning applications in the video game industry, focusing on how these techniques can be utilized in game development, experience, and operation. as relying on computation techniques, the game world can be viewed as an integration of various complex data. this examines the use of deep learning in processing various types of game data. the paper classifies the game data into asset data, interaction data, and player data, according to their utilization in game development, experience, and operation, respectively. specifically, this paper discusses deep learning applications in generating asset data such as object images, 3d scenes, avatar models, and facial animations; enhancing interaction data through improved text-based conversations and decision-making behaviors; and analyzing player data for cheat detection and match-making purposes. although this review may not cover all existing applications of deep learning, it aims to provide a thorough presentation of the current state of deep learning in the gaming industry and its potential to revolutionize game production by reducing costs and improving the overall player experience.","key_phrases":["this paper","a comprehensive review","deep learning applications","the video game industry","these techniques","game development","experience","operation","computation techniques","the game world","an integration","various complex data","this","the use","deep learning","various types","game data","the paper","the game data","asset data","interaction data","player data","their utilization","game development","experience","operation","this paper","deep learning applications","asset data","object images","3d scenes","avatar models","facial animations","interaction data","improved text-based conversations","decision-making behaviors","player data","cheat detection","match-making purposes","this review","all existing applications","deep learning","it","a thorough presentation","the current state","deep learning","the gaming industry","its potential","game production","costs","the overall player experience","3d"]},{"title":"Classification of hazelnut varieties based on bigtransfer deep learning model","authors":["Emrah D\u00f6nmez","Serhat K\u0131l\u0131\u00e7arslan","Aykut Diker"],"abstract":"Hazelnut is an agricultural product that contributes greatly to the economy of the countries where it is grown. The human factor plays a major role in hazelnut classification. The typical approach involves manual inspection of each sample by experts, a process that is both labor-intensive and time-consuming, and often suffers from limited sensitivity. The deep learning techniques are extremely important in the classification and detection of agricultural products. Deep learning has great potential in the agricultural sector. This technology can improve product quality, increase productivity, and offer farmers the ability to classify and detect their produce more effectively. This is important for sustainability and efficiency in the agricultural industry. In this paper aims to the application of deep learning algorithms to streamline hazelnut classification, reducing the need for manual labor, time, and cost in the sorting process. The study utilized hazelnut images from three different varieties: Giresun, Ordu, and Van, comprising a dataset of 1165 images for Giresun, 1324 for Ordu, and 1138 for Van hazelnuts. This dataset is an open-access dataset. In the study, experiments were carried out on the determination of hazelnut varieties with BigTransfer (BiT)-M R50\u2009\u00d7\u20091, BiT-M R101\u2009\u00d7\u20093 and BiT-M R152\u2009\u00d7\u20094 models. Deep learning models, including big transfer was employed for classification. The classification task involved 3627 nut images and resulted in a remarkable accuracy of 99.49% with the BiT-M R152\u2009\u00d7\u20094 model. These innovative methods can also lead to patentable products and devices in various industries, thereby boosting the economic value of the country.","doi":"10.1007\/s00217-024-04468-1","cleaned_title":"classification of hazelnut varieties based on bigtransfer deep learning model","cleaned_abstract":"hazelnut is an agricultural product that contributes greatly to the economy of the countries where it is grown. the human factor plays a major role in hazelnut classification. the typical approach involves manual inspection of each sample by experts, a process that is both labor-intensive and time-consuming, and often suffers from limited sensitivity. the deep learning techniques are extremely important in the classification and detection of agricultural products. deep learning has great potential in the agricultural sector. this technology can improve product quality, increase productivity, and offer farmers the ability to classify and detect their produce more effectively. this is important for sustainability and efficiency in the agricultural industry. in this paper aims to the application of deep learning algorithms to streamline hazelnut classification, reducing the need for manual labor, time, and cost in the sorting process. the study utilized hazelnut images from three different varieties: giresun, ordu, and van, comprising a dataset of 1165 images for giresun, 1324 for ordu, and 1138 for van hazelnuts. this dataset is an open-access dataset. in the study, experiments were carried out on the determination of hazelnut varieties with bigtransfer (bit)-m r50 \u00d7 1, bit-m r101 \u00d7 3 and bit-m r152 \u00d7 4 models. deep learning models, including big transfer was employed for classification. the classification task involved 3627 nut images and resulted in a remarkable accuracy of 99.49% with the bit-m r152 \u00d7 4 model. these innovative methods can also lead to patentable products and devices in various industries, thereby boosting the economic value of the country.","key_phrases":["hazelnut","an agricultural product","that","the economy","the countries","it","the human factor","a major role","hazelnut classification","the typical approach","manual inspection","each sample","experts","a process","that","limited sensitivity","the deep learning techniques","the classification","detection","agricultural products","deep learning","great potential","the agricultural sector","this technology","product quality","productivity","farmers","the ability","their produce","this","sustainability","efficiency","the agricultural industry","this paper","the application","deep learning algorithms","hazelnut classification","the need","manual labor","time","cost","the sorting process","the study","hazelnut images","three different varieties","giresun","ordu","van","a dataset","1165 images","giresun","ordu","van hazelnuts","this dataset","an open-access dataset","the study","experiments","the determination","hazelnut varieties","bigtransfer","r101","r152 \u00d7 4 models","deep learning models","big transfer","classification","the classification task","3627 nut images","a remarkable accuracy","99.49%","the bit-m","r152","4 model","these innovative methods","patentable products","devices","various industries","the economic value","the country","three","van","1165","1324","1138","bit)-m","\u00d7 4","3627","99.49%","\u00d7 4"]},{"title":"A Novel Cyber Security Model Using Deep Transfer Learning","authors":["\u00dcnal \u00c7avu\u015fo\u011flu","Devrim Akgun","Selman Hizal"],"abstract":"Preventing attackers from interrupting or totally stopping critical services in cloud systems is a vital and challenging task. Today, machine learning-based algorithms and models are widely used, especially for the intelligent detection of zero-day attacks. Recently, deep learning methods that provide automatic feature extraction are designed to detect attacks automatically. In this study, we constructed a new deep learning model based on transfer learning for detecting and protecting cloud systems from malicious attacks. The developed deep transfer learning-based IDS converts network traffic into 2D preprocessed feature maps.Then the feature maps are processed with the transferred and fine-tuned convolutional layers of the deep learning model before the dense layer for detection and classification of traffic data. The results computed using the NSL-KDD test dataset reveal that the developed models achieve 89.74% multiclass and 92.58% binary classification accuracy. We performed another evaluation using only 20% of the training dataset as test data, and 80% for training. In this case, the model achieved 99.83% and 99.85% multiclass and binary classification accuracy, respectively.","doi":"10.1007\/s13369-023-08092-1","cleaned_title":"a novel cyber security model using deep transfer learning","cleaned_abstract":"preventing attackers from interrupting or totally stopping critical services in cloud systems is a vital and challenging task. today, machine learning-based algorithms and models are widely used, especially for the intelligent detection of zero-day attacks. recently, deep learning methods that provide automatic feature extraction are designed to detect attacks automatically. in this study, we constructed a new deep learning model based on transfer learning for detecting and protecting cloud systems from malicious attacks. the developed deep transfer learning-based ids converts network traffic into 2d preprocessed feature maps.then the feature maps are processed with the transferred and fine-tuned convolutional layers of the deep learning model before the dense layer for detection and classification of traffic data. the results computed using the nsl-kdd test dataset reveal that the developed models achieve 89.74% multiclass and 92.58% binary classification accuracy. we performed another evaluation using only 20% of the training dataset as test data, and 80% for training. in this case, the model achieved 99.83% and 99.85% multiclass and binary classification accuracy, respectively.","key_phrases":["attackers","critical services","cloud systems","a vital and challenging task","machine learning-based algorithms","models","the intelligent detection","zero-day attacks","deep learning methods","that","automatic feature extraction","attacks","this study","we","a new deep learning model","cloud systems","malicious attacks","the developed deep transfer learning-based ids converts network traffic","the feature maps","the transferred and fine-tuned convolutional layers","the deep learning model","the dense layer","detection","classification","traffic data","the results","the nsl-kdd test","the developed models","89.74% multiclass","92.58%","binary classification accuracy","we","another evaluation","only 20%","the training dataset","test data","80%","training","this case","the model","99.83%","99.85% multiclass and binary classification accuracy","today","zero-day","2d","89.74%","92.58%","only 20%","80%","99.83%","99.85%"]},{"title":"Software fault prediction using deep learning techniques","authors":["Iqra Batool","Tamim Ahmed Khan"],"abstract":"Software fault prediction (SFP) techniques identify faults at the early stages of the software development life cycle (SDLC). We find machine learning techniques commonly used for SFP compared to deep learning methods, which can produce more accurate results. Deep learning offers exceptional results in various domains, such as computer vision, natural language processing, and speech recognition. In this study, we use three deep learning methods, namely, long short-term memory (LSTM), bidirectional LSTM (BILSTM), and radial basis function network (RBFN) to predict software faults and compare our results with existing models to show how our results are more accurate. Our study uses Chidamber and Kemerer (CK) metrics-based datasets to conduct experiments and test our proposed algorithm. We conclude that LSTM and BILSTM perform better, whereas RBFN is faster in producing the required results. We use k-fold cross-validation to do the model evaluation. Our proposed models provide software developers with a more accurate and efficient SFP mechanism.","doi":"10.1007\/s11219-023-09642-4","cleaned_title":"software fault prediction using deep learning techniques","cleaned_abstract":"software fault prediction (sfp) techniques identify faults at the early stages of the software development life cycle (sdlc). we find machine learning techniques commonly used for sfp compared to deep learning methods, which can produce more accurate results. deep learning offers exceptional results in various domains, such as computer vision, natural language processing, and speech recognition. in this study, we use three deep learning methods, namely, long short-term memory (lstm), bidirectional lstm (bilstm), and radial basis function network (rbfn) to predict software faults and compare our results with existing models to show how our results are more accurate. our study uses chidamber and kemerer (ck) metrics-based datasets to conduct experiments and test our proposed algorithm. we conclude that lstm and bilstm perform better, whereas rbfn is faster in producing the required results. we use k-fold cross-validation to do the model evaluation. our proposed models provide software developers with a more accurate and efficient sfp mechanism.","key_phrases":["software fault prediction","(sfp) techniques","faults","the early stages","the software development life cycle","sdlc","we","techniques","sfp","deep learning methods","which","more accurate results","deep learning","exceptional results","various domains","computer vision","natural language processing","speech recognition","this study","we","three deep learning methods","namely, long short-term memory","lstm","bidirectional lstm","bilstm","radial basis function network","rbfn","software faults","our results","existing models","our results","our study","chidamber","kemerer (ck) metrics-based datasets","experiments","our proposed algorithm","we","lstm","bilstm","rbfn","the required results","we","fold cross","-","validation","the model evaluation","our proposed models","software developers","a more accurate and efficient sfp mechanism","three"]},{"title":"Software fault prediction using deep learning techniques","authors":["Iqra Batool","Tamim Ahmed Khan"],"abstract":"Software fault prediction (SFP) techniques identify faults at the early stages of the software development life cycle (SDLC). We find machine learning techniques commonly used for SFP compared to deep learning methods, which can produce more accurate results. Deep learning offers exceptional results in various domains, such as computer vision, natural language processing, and speech recognition. In this study, we use three deep learning methods, namely, long short-term memory (LSTM), bidirectional LSTM (BILSTM), and radial basis function network (RBFN) to predict software faults and compare our results with existing models to show how our results are more accurate. Our study uses Chidamber and Kemerer (CK) metrics-based datasets to conduct experiments and test our proposed algorithm. We conclude that LSTM and BILSTM perform better, whereas RBFN is faster in producing the required results. We use k-fold cross-validation to do the model evaluation. Our proposed models provide software developers with a more accurate and efficient SFP mechanism.","doi":"10.1007\/s11219-023-09642-4","cleaned_title":"software fault prediction using deep learning techniques","cleaned_abstract":"software fault prediction (sfp) techniques identify faults at the early stages of the software development life cycle (sdlc). we find machine learning techniques commonly used for sfp compared to deep learning methods, which can produce more accurate results. deep learning offers exceptional results in various domains, such as computer vision, natural language processing, and speech recognition. in this study, we use three deep learning methods, namely, long short-term memory (lstm), bidirectional lstm (bilstm), and radial basis function network (rbfn) to predict software faults and compare our results with existing models to show how our results are more accurate. our study uses chidamber and kemerer (ck) metrics-based datasets to conduct experiments and test our proposed algorithm. we conclude that lstm and bilstm perform better, whereas rbfn is faster in producing the required results. we use k-fold cross-validation to do the model evaluation. our proposed models provide software developers with a more accurate and efficient sfp mechanism.","key_phrases":["software fault prediction","(sfp) techniques","faults","the early stages","the software development life cycle","sdlc","we","techniques","sfp","deep learning methods","which","more accurate results","deep learning","exceptional results","various domains","computer vision","natural language processing","speech recognition","this study","we","three deep learning methods","namely, long short-term memory","lstm","bidirectional lstm","bilstm","radial basis function network","rbfn","software faults","our results","existing models","our results","our study","chidamber","kemerer (ck) metrics-based datasets","experiments","our proposed algorithm","we","lstm","bilstm","rbfn","the required results","we","fold cross","-","validation","the model evaluation","our proposed models","software developers","a more accurate and efficient sfp mechanism","three"]},{"title":"Imbalcbl: addressing deep learning challenges with small and imbalanced datasets","authors":["Saqib ul Sabha","Assif Assad","Sadaf Shafi","Nusrat Mohi Ud Din","Rayees Ahmad Dar","Muzafar Rasool Bhat"],"abstract":"Deep learning, while transformative for computer vision, frequently falters when confronted with small and imbalanced datasets. Despite substantial progress in this domain, prevailing models often underachieve under these constraints. Addressing this, we introduce an innovative contrast-based learning strategy for small and imbalanced data that significantly bolsters the proficiency of deep learning architectures on these challenging datasets. By ingeniously concatenating training images, the effective training dataset expands from n to \\(n^2\\), affording richer data for model training, even when n is very small. Remarkably, our solution remains indifferent to specific loss functions or network architectures, endorsing its adaptability for diverse classification scenarios. Rigorously benchmarked against four benchmark datasets, our approach was juxtaposed with state-of-the-art oversampling paradigms. The empirical evidence underscores our method\u2019s superior efficacy, outshining contemporaries across metrics like Balanced accuracy, F1 score, and Geometric mean. Noteworthy increments include 7\u201316% on the Covid-19 dataset, 4\u201320% for Honey bees, 1\u20136% on CIFAR-10, and 1\u20139% on FashionMNIST. In essence, our proposed method offers a potent remedy for the perennial issues stemming from scanty and skewed data in deep learning.","doi":"10.1007\/s13198-024-02346-3","cleaned_title":"imbalcbl: addressing deep learning challenges with small and imbalanced datasets","cleaned_abstract":"deep learning, while transformative for computer vision, frequently falters when confronted with small and imbalanced datasets. despite substantial progress in this domain, prevailing models often underachieve under these constraints. addressing this, we introduce an innovative contrast-based learning strategy for small and imbalanced data that significantly bolsters the proficiency of deep learning architectures on these challenging datasets. by ingeniously concatenating training images, the effective training dataset expands from n to \\(n^2\\), affording richer data for model training, even when n is very small. remarkably, our solution remains indifferent to specific loss functions or network architectures, endorsing its adaptability for diverse classification scenarios. rigorously benchmarked against four benchmark datasets, our approach was juxtaposed with state-of-the-art oversampling paradigms. the empirical evidence underscores our method\u2019s superior efficacy, outshining contemporaries across metrics like balanced accuracy, f1 score, and geometric mean. noteworthy increments include 7\u201316% on the covid-19 dataset, 4\u201320% for honey bees, 1\u20136% on cifar-10, and 1\u20139% on fashionmnist. in essence, our proposed method offers a potent remedy for the perennial issues stemming from scanty and skewed data in deep learning.","key_phrases":["deep learning","computer vision","small and imbalanced datasets","substantial progress","this domain","prevailing models","these constraints","this","we","an innovative contrast-based learning strategy","small and imbalanced data","that","the proficiency","deep learning architectures","these challenging datasets","training images","the effective training dataset","n","\\(n^2\\","richer data","model training","n","our solution","specific loss functions","network architectures","its adaptability","diverse classification scenarios","four benchmark datasets","our approach","the-art","the empirical evidence","our method\u2019s superior efficacy","contemporaries","metrics","balanced accuracy","f1 score","geometric mean","noteworthy increments","7\u201316%","the covid-19 dataset","4\u201320%","honey bees","1\u20136%","cifar-10","1\u20139%","fashionmnist","essence","our proposed method","a potent remedy","the perennial issues","scanty","skewed data","deep learning","four","noteworthy","7\u201316%","covid-19","4\u201320%","1\u20136%","cifar-10","1\u20139%"]},{"title":"Deep Learning-Based Speed Breaker Detection","authors":["Mohamed Anas VT","Mohd Omar","Jameel Ahamad","Khaleel Ahmad","Mohd Anas Khan"],"abstract":"Traffic operation factors include vehicle and cargo damage, environmental effects, access for emergency vehicles, transit routes, and traffic speeds and volumes. In India, traffic rule violations and high-speed vehicles contribute to a large number of accidents and fatalities on the road. To mitigate this issue, the government has implemented speed breakers as a safety measure. In most cases, speed humps are only advised for use on streets with a speed restriction of 30\u00a0mph (50\u00a0km\/h) or less. However, many speed breakers lack proper signboards, and drivers often do not see them, leading to accidents. In light of this, a speed breaker detection system using deep learning is relevant and necessary in the current scenario. This research proposes a system that uses deep learning image detection algorithms to detect speed breakers in real-time, thereby increasing their visibility and reducing the risk of accidents. The system is trained using an Indian dataset and will be tested in various driving scenarios to evaluate its performance. The results are expected to show that the proposed system will outperform existing methods and have the potential to significantly improve road safety in India.","doi":"10.1007\/s42979-024-02891-5","cleaned_title":"deep learning-based speed breaker detection","cleaned_abstract":"traffic operation factors include vehicle and cargo damage, environmental effects, access for emergency vehicles, transit routes, and traffic speeds and volumes. in india, traffic rule violations and high-speed vehicles contribute to a large number of accidents and fatalities on the road. to mitigate this issue, the government has implemented speed breakers as a safety measure. in most cases, speed humps are only advised for use on streets with a speed restriction of 30 mph (50 km\/h) or less. however, many speed breakers lack proper signboards, and drivers often do not see them, leading to accidents. in light of this, a speed breaker detection system using deep learning is relevant and necessary in the current scenario. this research proposes a system that uses deep learning image detection algorithms to detect speed breakers in real-time, thereby increasing their visibility and reducing the risk of accidents. the system is trained using an indian dataset and will be tested in various driving scenarios to evaluate its performance. the results are expected to show that the proposed system will outperform existing methods and have the potential to significantly improve road safety in india.","key_phrases":["traffic operation factors","vehicle and cargo damage","environmental effects","access","emergency vehicles","transit routes","traffic speeds","volumes","india","traffic rule violations","high-speed vehicles","a large number","accidents","fatalities","the road","this issue","the government","speed breakers","a safety measure","most cases","speed humps","use","streets","a speed restriction","30 mph","50 km\/h","many speed breakers","proper signboards","drivers","them","accidents","light","this","a speed breaker detection system","deep learning","the current scenario","this research","a system","that","deep learning image detection algorithms","speed breakers","real-time","their visibility","the risk","accidents","the system","an indian dataset","various driving scenarios","its performance","the results","the proposed system","existing methods","the potential","road safety","india","india","30 mph","50 km","indian","india"]},{"title":"Manifold learning by a deep Gaussian process autoencoder","authors":["Francesco Camastra","Angelo Casolaro","Gennaro Iannuzzo"],"abstract":"The paper presents a novel manifold learning algorithm, the deep Gaussian process autoencoder (DPGA), based on deep Gaussian processes. Deep Gaussian process autoencoder algorithm has the following two main characteristics. The former is a bottleneck structure, borrowed by variational autoencoders and the latter is based on the so-called doubly stochastic variational inference for deep Gaussian processes architecture (DSVI). The main novelties of the paper consist in DGPA algorithm and the experimental protocol for evaluating it. In fact, to the best of our knowledge, deep Gaussian processes algorithms have not been applied to manifold learning, yet. Besides, an experimental protocol is introduced, the so-called manifold learning performance protocol (MLPP), to compare quantitatively the geometric preserved properties of manifold learning projections of the proposed deep Gaussian process autoencoder with the ones of state-of-the-art manifold learning algorithms. Extensive experimental tests on eleven synthetic and five real datasets show that deep Gaussian process autoencoder compares favorably with the other manifold learning competitors.","doi":"10.1007\/s00521-023-08536-7","cleaned_title":"manifold learning by a deep gaussian process autoencoder","cleaned_abstract":"the paper presents a novel manifold learning algorithm, the deep gaussian process autoencoder (dpga), based on deep gaussian processes. deep gaussian process autoencoder algorithm has the following two main characteristics. the former is a bottleneck structure, borrowed by variational autoencoders and the latter is based on the so-called doubly stochastic variational inference for deep gaussian processes architecture (dsvi). the main novelties of the paper consist in dgpa algorithm and the experimental protocol for evaluating it. in fact, to the best of our knowledge, deep gaussian processes algorithms have not been applied to manifold learning, yet. besides, an experimental protocol is introduced, the so-called manifold learning performance protocol (mlpp), to compare quantitatively the geometric preserved properties of manifold learning projections of the proposed deep gaussian process autoencoder with the ones of state-of-the-art manifold learning algorithms. extensive experimental tests on eleven synthetic and five real datasets show that deep gaussian process autoencoder compares favorably with the other manifold learning competitors.","key_phrases":["the paper","a novel manifold learning algorithm","the deep gaussian process autoencoder","dpga","deep gaussian processes","deep gaussian process","autoencoder algorithm","the following two main characteristics","a bottleneck structure","variational autoencoders","the so-called doubly stochastic variational inference","deep gaussian processes architecture","the main novelties","the paper consist","dgpa algorithm","the experimental protocol","it","fact","our knowledge","deep gaussian processes algorithms","manifold learning","an experimental protocol","mlpp","the geometric preserved properties","manifold learning projections","the proposed deep gaussian process","the ones","the-art","extensive experimental tests","eleven synthetic","five real datasets","deep gaussian process autoencoder","the other manifold learning competitors","deep gaussian","two","deep gaussian","eleven","five","deep gaussian"]},{"title":"A survey of deep learning-based 3D shape generation","authors":["Qun-Ce Xu","Tai-Jiang Mu","Yong-Liang Yang"],"abstract":"Deep learning has been successfully used for tasks in the 2D image domain. Research on 3D computer vision and deep geometry learning has also attracted attention. Considerable achievements have been made regarding feature extraction and discrimination of 3D shapes. Following recent advances in deep generative models such as generative adversarial networks, effective generation of 3D shapes has become an active research topic. Unlike 2D images with a regular grid structure, 3D shapes have various representations, such as voxels, point clouds, meshes, and implicit functions. For deep learning of 3D shapes, shape representation has to be taken into account as there is no unified representation that can cover all tasks well. Factors such as the representativeness of geometry and topology often largely affect the quality of the generated 3D shapes. In this survey, we comprehensively review works on deep-learning-based 3D shape generation by classifying and discussing them in terms of the underlying shape representation and the architecture of the shape generator. The advantages and disadvantages of each class are further analyzed. We also consider the 3D shape datasets commonly used for shape generation. Finally, we present several potential research directions that hopefully can inspire future works on this topic.","doi":"10.1007\/s41095-022-0321-5","cleaned_title":"a survey of deep learning-based 3d shape generation","cleaned_abstract":"deep learning has been successfully used for tasks in the 2d image domain. research on 3d computer vision and deep geometry learning has also attracted attention. considerable achievements have been made regarding feature extraction and discrimination of 3d shapes. following recent advances in deep generative models such as generative adversarial networks, effective generation of 3d shapes has become an active research topic. unlike 2d images with a regular grid structure, 3d shapes have various representations, such as voxels, point clouds, meshes, and implicit functions. for deep learning of 3d shapes, shape representation has to be taken into account as there is no unified representation that can cover all tasks well. factors such as the representativeness of geometry and topology often largely affect the quality of the generated 3d shapes. in this survey, we comprehensively review works on deep-learning-based 3d shape generation by classifying and discussing them in terms of the underlying shape representation and the architecture of the shape generator. the advantages and disadvantages of each class are further analyzed. we also consider the 3d shape datasets commonly used for shape generation. finally, we present several potential research directions that hopefully can inspire future works on this topic.","key_phrases":["deep learning","tasks","the 2d image domain","research","3d computer vision","deep geometry learning","attention","considerable achievements","feature extraction","discrimination","3d shapes","recent advances","deep generative models","generative adversarial networks","effective generation","3d shapes","an active research topic","2d images","a regular grid structure","3d shapes","various representations","voxels","point clouds","meshes","implicit functions","deep learning","3d shapes","shape representation","account","no unified representation","that","all tasks","factors","the representativeness","geometry","topology","the quality","the generated 3d shapes","this survey","we","deep-learning-based 3d shape generation","them","terms","the underlying shape representation","the architecture","the shape generator","the advantages","disadvantages","each class","we","the 3d shape datasets","shape generation","we","several potential research directions","that","future works","this topic","2d","3d","3d","3d","2d","3d","3d","3d","3d","3d"]},{"title":"Accelerated cardiac magnetic resonance imaging using deep learning for volumetric assessment in children","authors":["Melina Koechli","Fraser M. Callaghan","Barbara E. U. Burkhardt","Ma\u00e9l\u00e8ne Loh\u00e9zic","Xucheng Zhu","Beate R\u00fccker","Emanuela R. Valsangiacomo Buechel","Christian J. Kellenberger","Julia Geiger"],"abstract":"BackgroundVentricular volumetry using a short-axis stack of two-dimensional (D) cine balanced steady-state free precession (bSSFP) sequences is crucial in any cardiac magnetic resonance imaging (MRI) examination. This task becomes particularly challenging in children due to multiple breath-holds.ObjectiveTo assess the diagnostic performance of accelerated 3-RR cine MRI sequences using deep learning reconstruction compared with standard 2-D cine bSSFP sequences.Material and methodsTwenty-nine consecutive patients (mean age 11\u2009\u00b1\u20095, median 12, range 1\u201317\u00a0years) undergoing cardiac MRI were scanned with a conventional segmented 2-D cine and a deep learning accelerated cine (three heartbeats) acquisition on a 1.5-tesla scanner. Short-axis volumetrics were performed (semi-)automatically in both datasets retrospectively by two experienced\u00a0readers\u00a0who visually assessed image quality employing a 4-point grading scale. Scan times and image quality were compared using the Wilcoxon rank-sum test. Volumetrics were assessed with linear regression and Bland\u2013Altman analyses, and measurement agreement with intraclass correlation coefficient (ICC).ResultsMean acquisition time was significantly reduced with the 3-RR deep learning cine compared to the standard cine sequence (45.5\u2009\u00b1\u200913.8\u00a0s vs. 218.3\u2009\u00b1\u200944.8\u00a0s; P\u2009<\u20090.001). No significant differences in biventricular volumetrics were found. Left ventricular (LV) mass was increased in the deep learning cine compared with the standard cine sequence (71.4\u2009\u00b1\u200933.1\u00a0g vs. 69.9\u2009\u00b1\u200932.5\u00a0g; P\u2009<\u20090.05). All volumetric measurements had an excellent agreement with ICC\u2009>\u20090.9 except for ejection fraction (EF) (LVEF 0.81, RVEF 0.73). The image quality of deep learning cine images was decreased for end-diastolic and end-systolic contours, papillary muscles, and valve depiction (2.9\u2009\u00b1\u20090.5 vs. 3.5\u2009\u00b1\u20090.4; P\u2009<\u20090.05).ConclusionDeep learning cine volumetrics did not differ significantly from standard cine results except for LV mass, which was slightly overestimated with deep learning cine. Deep learning cine sequences result in a significant reduction in scan time with only slightly lower image quality.Graphical Abstract","doi":"10.1007\/s00247-024-05978-6","cleaned_title":"accelerated cardiac magnetic resonance imaging using deep learning for volumetric assessment in children","cleaned_abstract":"backgroundventricular volumetry using a short-axis stack of two-dimensional (d) cine balanced steady-state free precession (bssfp) sequences is crucial in any cardiac magnetic resonance imaging (mri) examination. this task becomes particularly challenging in children due to multiple breath-holds.objectiveto assess the diagnostic performance of accelerated 3-rr cine mri sequences using deep learning reconstruction compared with standard 2-d cine bssfp sequences.material and methodstwenty-nine consecutive patients (mean age 11 \u00b1 5, median 12, range 1\u201317 years) undergoing cardiac mri were scanned with a conventional segmented 2-d cine and a deep learning accelerated cine (three heartbeats) acquisition on a 1.5-tesla scanner. short-axis volumetrics were performed (semi-)automatically in both datasets retrospectively by two experienced readers who visually assessed image quality employing a 4-point grading scale. scan times and image quality were compared using the wilcoxon rank-sum test. volumetrics were assessed with linear regression and bland\u2013altman analyses, and measurement agreement with intraclass correlation coefficient (icc).resultsmean acquisition time was significantly reduced with the 3-rr deep learning cine compared to the standard cine sequence (45.5 \u00b1 13.8 s vs. 218.3 \u00b1 44.8 s; p < 0.001). no significant differences in biventricular volumetrics were found. left ventricular (lv) mass was increased in the deep learning cine compared with the standard cine sequence (71.4 \u00b1 33.1 g vs. 69.9 \u00b1 32.5 g; p < 0.05). all volumetric measurements had an excellent agreement with icc > 0.9 except for ejection fraction (ef) (lvef 0.81, rvef 0.73). the image quality of deep learning cine images was decreased for end-diastolic and end-systolic contours, papillary muscles, and valve depiction (2.9 \u00b1 0.5 vs. 3.5 \u00b1 0.4; p < 0.05).conclusiondeep learning cine volumetrics did not differ significantly from standard cine results except for lv mass, which was slightly overestimated with deep learning cine. deep learning cine sequences result in a significant reduction in scan time with only slightly lower image quality.graphical abstract","key_phrases":["backgroundventricular volumetry","a short-axis stack","two-dimensional (d) cine","steady-state free precession (bssfp) sequences","imaging","this task","children","multiple breath-holds.objectiveto","the diagnostic performance","accelerated 3-rr cine mri sequences","deep learning reconstruction","standard 2-d cine bssfp sequences.material and methodstwenty-nine consecutive patients","(mean age","median","range 1\u201317 years","cardiac mri","a conventional segmented 2-d cine","a deep learning","cine","(three heartbeats) acquisition","a 1.5-tesla scanner","short-axis volumetrics","both datasets","two experienced readers","who","image quality","a 4-point grading scale","scan times","image quality","the wilcoxon rank-sum test","volumetrics","linear regression","bland\u2013altman analyses","measurement agreement","intraclass correlation","coefficient","icc).resultsmean acquisition time","the 3-rr deep learning cine","the standard cine sequence","45.5 \u00b1 13.8 s","218.3 \u00b1","44.8 s","no significant differences","biventricular volumetrics","ventricular (lv) mass","the deep learning cine","the standard cine sequence","33.1 g","69.9 \u00b1","p","all volumetric measurements","an excellent agreement","icc","ejection fraction","ef","lvef","rvef","the image quality","deep learning cine images","end-diastolic and end-systolic contours","papillary muscles","valve depiction","2.9 \u00b1","3.5 \u00b1","cine volumetrics","standard cine results","lv mass","which","deep learning cine","cine sequences","a significant reduction","scan time","only slightly lower image","backgroundventricular volumetry","two","breath-holds.objectiveto","3","2","age 11 \u00b1 5","12","1\u201317 years","2","three","1.5","two","4","scan times","3","218.3","44.8","p < 0.001","ventricular","33.1","69.9","32.5","0.9","0.81","0.73","2.9 \u00b1 0.5","3.5","0.4"]},{"title":"Comparative Analysis of Diabetic Retinopathy Classification Approaches Using Machine Learning and Deep Learning Techniques","authors":["Ruchika Bala","Arun Sharma","Nidhi Goel"],"abstract":"Diabetic retinopathy (DR) is an eye disease caused due to excess of sugar in retinal blood vessels and obstructs vision. Regular and timely diagnosis can prevent the severity of diabetic retinopathy at an initial stage. Manual diagnosis of diabetic retinopathy is time-consuming and thus a plethora of work has been done by researchers to automate the classification of diabetic retinopathy using machine learning and deep learning techniques. The present review pivots around the research papers covering recent and effective automated DR classification techniques from 2011 to 2022. A comparative analysis of these papers highlights the summary of DR classification datasets, pre-processing techniques, various advanced classification algorithms, and their performance. Along with the summary and analysis of the classification techniques, the present paper demonstrates the experimentation of eight pre-trained convolution neural network models on DR benchmark classification datasets. This is to help the researchers to choose the best pre-trained classification model for the corresponding DR dataset. The use of deep learning and machine learning algorithms demonstrated excellent performance; however, researchers still need to explore the design constraints of the classification models to have effective results. Attention mechanisms and vision transformers are recent breakthroughs that can be used to solve classification challenges. The objective of this article is to provide a single platform for researchers to access state-of-the-art work in the classification of diabetic retinopathy, the results of various pre-trained models on DR benchmark classification datasets, and future research prospects for the researchers working in this challenging area.","doi":"10.1007\/s11831-023-10002-5","cleaned_title":"comparative analysis of diabetic retinopathy classification approaches using machine learning and deep learning techniques","cleaned_abstract":"diabetic retinopathy (dr) is an eye disease caused due to excess of sugar in retinal blood vessels and obstructs vision. regular and timely diagnosis can prevent the severity of diabetic retinopathy at an initial stage. manual diagnosis of diabetic retinopathy is time-consuming and thus a plethora of work has been done by researchers to automate the classification of diabetic retinopathy using machine learning and deep learning techniques. the present review pivots around the research papers covering recent and effective automated dr classification techniques from 2011 to 2022. a comparative analysis of these papers highlights the summary of dr classification datasets, pre-processing techniques, various advanced classification algorithms, and their performance. along with the summary and analysis of the classification techniques, the present paper demonstrates the experimentation of eight pre-trained convolution neural network models on dr benchmark classification datasets. this is to help the researchers to choose the best pre-trained classification model for the corresponding dr dataset. the use of deep learning and machine learning algorithms demonstrated excellent performance; however, researchers still need to explore the design constraints of the classification models to have effective results. attention mechanisms and vision transformers are recent breakthroughs that can be used to solve classification challenges. the objective of this article is to provide a single platform for researchers to access state-of-the-art work in the classification of diabetic retinopathy, the results of various pre-trained models on dr benchmark classification datasets, and future research prospects for the researchers working in this challenging area.","key_phrases":["diabetic retinopathy","dr","an eye disease","excess","sugar","retinal blood vessels","vision","regular and timely diagnosis","the severity","diabetic retinopathy","an initial stage","manual diagnosis","diabetic retinopathy","a plethora","work","researchers","the classification","machine learning","deep learning techniques","the present review pivots","the research papers","recent and effective automated dr classification techniques","a comparative analysis","these papers","the summary","dr classification datasets","pre-processing techniques","various advanced classification algorithms","their performance","the summary","analysis","the classification techniques","the present paper","the experimentation","eight pre-trained convolution neural network models","dr benchmark classification datasets","this","the researchers","the best pre-trained classification model","the corresponding dr dataset","the use","deep learning and machine learning algorithms","excellent performance","researchers","the design constraints","the classification models","effective results","attention mechanisms","vision transformers","recent breakthroughs","that","classification challenges","the objective","this article","a single platform","researchers","the-art","the classification","diabetic retinopathy","the results","various pre-trained models","dr benchmark classification datasets","future research prospects","the researchers","this challenging area","2011 to 2022","eight"]},{"title":"Heart disease prediction using machine learning, deep Learning and optimization techniques-A semantic review","authors":["Girish Shrikrushnarao Bhavekar","Agam Das Goswami","Chafle Pratiksha Vasantrao","Amit K. Gaikwad","Amol V. Zade","Harsha Vyawahare"],"abstract":"Cardiovascular disease holds the position of being the foremost cause of death worldwide. Heart Disease Prediction (HDP) is a difficult task as it needs advanced knowledge with better experience. Moreover, it encounters numerous significant challenges in clinical data analysis. While many researchers have focused on predicting heart disease, the performance metric, namely prediction accuracy, remains suboptimal. The accurate HDP can help the person to prevent himself from life threats and at the same time, inaccurate prediction can prove to be fatal. To solve these issues, in this review work several Deep Learning (DL), Machine Learning (ML) and optimization based HDP techniques are discussed. In recent times, many researchers have been utilizing different DL and ML algorithms to help the professionals and health care industry for the prediction of heart disease. Further, it discussed about various optimization-based algorithms and its performance analysis. Therefore, this review paper suggests that the optimization-based HDP algorithm could assist doctors in predicting the occurrence of heart disease in advance and offering suitable treatment.","doi":"10.1007\/s11042-024-19680-0","cleaned_title":"heart disease prediction using machine learning, deep learning and optimization techniques-a semantic review","cleaned_abstract":"cardiovascular disease holds the position of being the foremost cause of death worldwide. heart disease prediction (hdp) is a difficult task as it needs advanced knowledge with better experience. moreover, it encounters numerous significant challenges in clinical data analysis. while many researchers have focused on predicting heart disease, the performance metric, namely prediction accuracy, remains suboptimal. the accurate hdp can help the person to prevent himself from life threats and at the same time, inaccurate prediction can prove to be fatal. to solve these issues, in this review work several deep learning (dl), machine learning (ml) and optimization based hdp techniques are discussed. in recent times, many researchers have been utilizing different dl and ml algorithms to help the professionals and health care industry for the prediction of heart disease. further, it discussed about various optimization-based algorithms and its performance analysis. therefore, this review paper suggests that the optimization-based hdp algorithm could assist doctors in predicting the occurrence of heart disease in advance and offering suitable treatment.","key_phrases":["cardiovascular disease","the position","the foremost cause","death","heart disease prediction","hdp","a difficult task","it","advanced knowledge","better experience","it","numerous significant challenges","clinical data analysis","many researchers","heart disease","the performance","namely prediction accuracy","the accurate hdp","the person","himself","life threats","the same time","inaccurate prediction","these issues","this review","several deep learning","dl","machine learning","optimization","hdp techniques","recent times","many researchers","different dl","ml","the professionals","health care industry","the prediction","heart disease","it","various optimization-based algorithms","its performance analysis","this review paper","the optimization-based hdp algorithm","doctors","the occurrence","heart disease","advance","suitable treatment"]},{"title":"A domain knowledge-based interpretable deep learning system for improving clinical breast ultrasound diagnosis","authors":["Lin Yan","Zhiying Liang","Hao Zhang","Gaosong Zhang","Weiwei Zheng","Chunguang Han","Dongsheng Yu","Hanqi Zhang","Xinxin Xie","Chang Liu","Wenxin Zhang","Hui Zheng","Jing Pei","Dinggang Shen","Xuejun Qian"],"abstract":"BackgroundThough deep learning has consistently demonstrated advantages in the automatic interpretation of breast ultrasound images, its black-box nature hinders potential interactions with radiologists, posing obstacles for clinical deployment.MethodsWe proposed a domain knowledge-based interpretable deep learning system for improving breast cancer risk prediction via paired multimodal ultrasound images. The deep learning system was developed on 4320 multimodal breast ultrasound images of 1440 biopsy-confirmed lesions from 1348 prospectively enrolled patients across two hospitals between August 2019 and December 2022. The lesions were allocated to 70% training cohort, 10% validation cohort, and 20% test cohort based on case recruitment date.ResultsHere, we show that the interpretable deep learning system can predict breast cancer risk as accurately as experienced radiologists, with an area under the receiver operating characteristic curve of 0.902 (95% confidence interval = 0.882 \u2013 0.921), sensitivity of 75.2%, and specificity of 91.8% on the test cohort. With the aid of the deep learning system, particularly its inherent explainable features, junior radiologists tend to achieve better clinical outcomes, while senior radiologists experience increased confidence levels. Multimodal ultrasound images augmented with domain knowledge-based reasoning cues enable an effective human-machine collaboration at a high level of prediction performance.ConclusionsSuch a clinically applicable deep learning system may be incorporated into future breast cancer screening and support assisted or second-read workflows.","doi":"10.1038\/s43856-024-00518-7","cleaned_title":"a domain knowledge-based interpretable deep learning system for improving clinical breast ultrasound diagnosis","cleaned_abstract":"backgroundthough deep learning has consistently demonstrated advantages in the automatic interpretation of breast ultrasound images, its black-box nature hinders potential interactions with radiologists, posing obstacles for clinical deployment.methodswe proposed a domain knowledge-based interpretable deep learning system for improving breast cancer risk prediction via paired multimodal ultrasound images. the deep learning system was developed on 4320 multimodal breast ultrasound images of 1440 biopsy-confirmed lesions from 1348 prospectively enrolled patients across two hospitals between august 2019 and december 2022. the lesions were allocated to 70% training cohort, 10% validation cohort, and 20% test cohort based on case recruitment date.resultshere, we show that the interpretable deep learning system can predict breast cancer risk as accurately as experienced radiologists, with an area under the receiver operating characteristic curve of 0.902 (95% confidence interval = 0.882 \u2013 0.921), sensitivity of 75.2%, and specificity of 91.8% on the test cohort. with the aid of the deep learning system, particularly its inherent explainable features, junior radiologists tend to achieve better clinical outcomes, while senior radiologists experience increased confidence levels. multimodal ultrasound images augmented with domain knowledge-based reasoning cues enable an effective human-machine collaboration at a high level of prediction performance.conclusionssuch a clinically applicable deep learning system may be incorporated into future breast cancer screening and support assisted or second-read workflows.","key_phrases":["deep learning","advantages","the automatic interpretation","breast ultrasound images","its black-box nature","potential interactions","radiologists","obstacles","clinical deployment.methodswe","a domain knowledge-based interpretable deep learning system","breast cancer risk prediction","paired multimodal ultrasound images","the deep learning system","4320 multimodal breast ultrasound images","1440 biopsy-confirmed lesions","prospectively enrolled patients","two hospitals","august","december","the lesions","70% training cohort","10% validation cohort","20% test cohort","case recruitment","we","the interpretable deep learning system","breast cancer risk","experienced radiologists","an area","the receiver operating characteristic curve","(95% confidence interval","75.2%","91.8%","the test cohort","the aid","the deep learning system","particularly its inherent explainable features","junior radiologists","better clinical outcomes","senior radiologists","increased confidence levels","multimodal ultrasound images","domain knowledge-based reasoning cues","an effective human-machine collaboration","a high level","prediction","a clinically applicable deep learning system","future breast cancer screening","support","second-read workflows","4320","1440","1348","two","august 2019","december 2022","70%","10%","20%","0.902","95%","0.882","0.921","75.2%","91.8%","second"]},{"title":"Deep learning for osteoporosis screening using an anteroposterior hip radiograph image","authors":["Artit Boonrod","Prarinthorn Piyaprapaphan","Nut Kittipongphat","Daris Theerakulpisut","Arunnit Boonrod"],"abstract":"PurposeOsteoporosis is a common bone disorder characterized by decreased bone mineral density (BMD) and increased bone fragility, which can lead to fractures and eventually cause morbidity and mortality. It is of great concern that the one-year mortality rate for osteoporotic hip fractures could be as high as 22%, regardless of the treatment. Currently, BMD measurement is the standard method for osteoporosis diagnosis, but it is costly and requires special equipment. While a plain radiograph can be obtained more simply and inexpensively, it is not used for diagnosis. Deep learning technologies had been applied to various medical contexts, yet few to osteoporosis unless they were trained on the advanced investigative images, such as computed tomography. The purpose of this study was to develop a deep learning model using the anteroposterior hip radiograph images and measure its diagnostic accuracy for osteoporosis.MethodsWe retrospectively collected all anteroposterior hip radiograph images of patients from 2013 to 2021 at a tertiary care hospital. The BMD measurements of the included patients were reviewed, and the radiograph images that had a time interval of more than two years from the measurements were excluded. All images were randomized using a computer-generated unequal allocation into two datasets, i.e., 80% of images were used for the training dataset and the remaining 20% for the test dataset. The T score of BMD obtained from the ipsilateral femoral neck of the same patient closest to the date of the performed radiograph was chosen. The T score cutoff value of\u2009\u2212\u20092.5 was used to diagnose osteoporosis. Five deep learning models were trained on the training dataset, and their diagnostic performances were evaluated using the test dataset. Finally, the best model was determined by the area under the curves (AUC).ResultsA total of 363 anteroposterior hip radiograph images were identified. The average time interval between the performed radiograph and the BMD measurement was 6.6\u00a0months. Two-hundred-thirteen images were labeled as non-osteoporosis (T score\u2009>\u2009\u2009\u2212\u20092.5), and the other 150 images as osteoporosis (T score \u2264\u2009 \u2212\u20092.5). The best-selected deep learning model achieved an AUC of 0.91 and accuracy of 0.82.ConclusionsThis study demonstrates the potential of deep learning for osteoporosis screening using anteroposterior hip radiographs. The results suggest that the deep learning model might potentially be used as a screening tool to find patients at risk for osteoporosis to perform further BMD measurement.","doi":"10.1007\/s00590-024-04032-3","cleaned_title":"deep learning for osteoporosis screening using an anteroposterior hip radiograph image","cleaned_abstract":"purposeosteoporosis is a common bone disorder characterized by decreased bone mineral density (bmd) and increased bone fragility, which can lead to fractures and eventually cause morbidity and mortality. it is of great concern that the one-year mortality rate for osteoporotic hip fractures could be as high as 22%, regardless of the treatment. currently, bmd measurement is the standard method for osteoporosis diagnosis, but it is costly and requires special equipment. while a plain radiograph can be obtained more simply and inexpensively, it is not used for diagnosis. deep learning technologies had been applied to various medical contexts, yet few to osteoporosis unless they were trained on the advanced investigative images, such as computed tomography. the purpose of this study was to develop a deep learning model using the anteroposterior hip radiograph images and measure its diagnostic accuracy for osteoporosis.methodswe retrospectively collected all anteroposterior hip radiograph images of patients from 2013 to 2021 at a tertiary care hospital. the bmd measurements of the included patients were reviewed, and the radiograph images that had a time interval of more than two years from the measurements were excluded. all images were randomized using a computer-generated unequal allocation into two datasets, i.e., 80% of images were used for the training dataset and the remaining 20% for the test dataset. the t score of bmd obtained from the ipsilateral femoral neck of the same patient closest to the date of the performed radiograph was chosen. the t score cutoff value of \u2212 2.5 was used to diagnose osteoporosis. five deep learning models were trained on the training dataset, and their diagnostic performances were evaluated using the test dataset. finally, the best model was determined by the area under the curves (auc).resultsa total of 363 anteroposterior hip radiograph images were identified. the average time interval between the performed radiograph and the bmd measurement was 6.6 months. two-hundred-thirteen images were labeled as non-osteoporosis (t score > \u2212 2.5), and the other 150 images as osteoporosis (t score \u2264 \u2212 2.5). the best-selected deep learning model achieved an auc of 0.91 and accuracy of 0.82.conclusionsthis study demonstrates the potential of deep learning for osteoporosis screening using anteroposterior hip radiographs. the results suggest that the deep learning model might potentially be used as a screening tool to find patients at risk for osteoporosis to perform further bmd measurement.","key_phrases":["purposeosteoporosis","a common bone disorder","decreased bone mineral density","bmd","increased bone fragility","which","fractures","morbidity","mortality","it","great concern","the one-year mortality rate","osteoporotic hip fractures","22%","the treatment","bmd measurement","the standard method","osteoporosis diagnosis","it","special equipment","a plain radiograph","it","diagnosis","deep learning technologies","various medical contexts","they","the advanced investigative images","computed tomography","the purpose","this study","a deep learning model","the anteroposterior hip radiograph images","its diagnostic accuracy","osteoporosis.methodswe","all anteroposterior hip radiograph images","patients","a tertiary care hospital","the bmd measurements","the included patients","the radiograph images","that","a time interval","more than two years","the measurements","all images","a computer-generated unequal allocation","two datasets","80%","images","the training dataset","the remaining 20%","the test dataset","the t score","bmd","the ipsilateral femoral neck","the same patient","the date","the performed radiograph","the t","cutoff value","\u2212","osteoporosis","five deep learning models","the training dataset","their diagnostic performances","the test dataset","the best model","the area","the curves","auc).resultsa total","363 anteroposterior hip radiograph images","the average time interval","the performed radiograph","the bmd measurement","6.6 months","two-hundred-thirteen images","osteoporosis","t","\u2212","the other 150 images","osteoporosis","t","\u2212","the best-selected deep learning model","an auc","accuracy","0.82.conclusionsthis study","the potential","deep learning","anteroposterior hip radiographs","the results","the deep learning model","a screening tool","patients","risk","osteoporosis","further bmd measurement","one-year","22%","2013","2021","tertiary","more than two years","two","80%","the remaining 20%","\u2212 2.5","five","363","6.6 months","two-hundred-thirteen","2.5","150","2.5","0.91"]},{"title":"Deep learning-based solution for smart contract vulnerabilities detection","authors":["Xueyan Tang","Yuying Du","Alan Lai","Ze Zhang","Lingzhi Shi"],"abstract":"This paper aims to explore the application of deep learning in smart contract vulnerabilities detection. Smart contracts are an essential part of blockchain technology and are crucial for developing decentralized applications. However, smart contract vulnerabilities can cause financial losses and system crashes. Static analysis tools are frequently used to detect vulnerabilities in smart contracts, but they often result in false positives and false negatives because of their high reliance on predefined rules and lack of semantic analysis capabilities. Furthermore, these predefined rules quickly become obsolete and fail to adapt or generalize to new data. In contrast, deep learning methods do not require predefined detection rules and can learn the features of vulnerabilities during the training process. In this paper, we introduce a solution called Lightning Cat which is based on deep learning techniques. We train three deep learning models for detecting vulnerabilities in smart contract: Optimized-CodeBERT, Optimized-LSTM, and Optimized-CNN. Experimental results show that, in the Lightning Cat we propose, Optimized-CodeBERT model surpasses other methods, achieving an f1-score of 93.53%. To precisely extract vulnerability features, we acquire segments of vulnerable code functions to retain critical vulnerability features. Using the CodeBERT pre-training model for data preprocessing, we could capture the syntax and semantics of the code more accurately. To demonstrate the feasibility of our proposed solution, we evaluate its performance using the SolidiFI-benchmark dataset, which consists of 9369 vulnerable contracts injected with vulnerabilities from seven different types.","doi":"10.1038\/s41598-023-47219-0","cleaned_title":"deep learning-based solution for smart contract vulnerabilities detection","cleaned_abstract":"this paper aims to explore the application of deep learning in smart contract vulnerabilities detection. smart contracts are an essential part of blockchain technology and are crucial for developing decentralized applications. however, smart contract vulnerabilities can cause financial losses and system crashes. static analysis tools are frequently used to detect vulnerabilities in smart contracts, but they often result in false positives and false negatives because of their high reliance on predefined rules and lack of semantic analysis capabilities. furthermore, these predefined rules quickly become obsolete and fail to adapt or generalize to new data. in contrast, deep learning methods do not require predefined detection rules and can learn the features of vulnerabilities during the training process. in this paper, we introduce a solution called lightning cat which is based on deep learning techniques. we train three deep learning models for detecting vulnerabilities in smart contract: optimized-codebert, optimized-lstm, and optimized-cnn. experimental results show that, in the lightning cat we propose, optimized-codebert model surpasses other methods, achieving an f1-score of 93.53%. to precisely extract vulnerability features, we acquire segments of vulnerable code functions to retain critical vulnerability features. using the codebert pre-training model for data preprocessing, we could capture the syntax and semantics of the code more accurately. to demonstrate the feasibility of our proposed solution, we evaluate its performance using the solidifi-benchmark dataset, which consists of 9369 vulnerable contracts injected with vulnerabilities from seven different types.","key_phrases":["this paper","the application","deep learning","smart contract vulnerabilities detection","smart contracts","an essential part","blockchain technology","decentralized applications","smart contract vulnerabilities","financial losses","system crashes","static analysis tools","vulnerabilities","smart contracts","they","false positives","false negatives","their high reliance","predefined rules","lack","semantic analysis capabilities","these predefined rules","new data","contrast","deep learning methods","predefined detection rules","the features","vulnerabilities","the training process","this paper","we","a solution","lightning cat","which","deep learning techniques","we","three deep learning models","vulnerabilities","smart contract","optimized-codebert","optimized-lstm","cnn","experimental results","the lightning cat","we","optimized-codebert model","other methods","an f1-score","93.53%","vulnerability features","we","segments","vulnerable code functions","critical vulnerability features","the codebert pre-training model","data","we","the syntax","semantics","the code","the feasibility","our proposed solution","we","its performance","the solidifi-benchmark dataset","which","9369 vulnerable contracts","vulnerabilities","seven different types","three","93.53%","9369","seven"]},{"title":"Image deep learning in fault diagnosis of mechanical equipment","authors":["Chuanhao Wang","Yongjian Sun","Xiaohong Wang"],"abstract":"With the development of industry, more and more crucial mechanical machinery generate wildness demand of effective fault diagnosis to ensure the safe operation. Over the past few decades, researchers have explored and developed a variety of approaches. In recent years, fault diagnosis based on deep learning has developed rapidly, which achieved satisfied results in the filed of mechanical equipment fault diagnosis. However, there is few review to systematically summarize and sort out these special image deep learning methods. In order to fill this gap, this paper concentrates on reviewing comprehensively the development of special image deep learning for mechanical equipment fault diagnosis in past 5 years. In general, a typical image fault diagnosis based on fault image deep learning generally consists of data acquisition, signal processing, model construction, feature learning and decision-making. Firstly, the method of signal preprocessing is introduced, and several common methods of converting signals into images are briefly compared and analyzed. Then, the principles and variants of deep learning models are expounded. Further more, the difficulties and challenges encountered at this stage are summarized. Last but not least, the future development and potential trends of the work are concluded, and it hoped that this work will facilitate and inspire further exploration for researchers in this area.","doi":"10.1007\/s10845-023-02176-3","cleaned_title":"image deep learning in fault diagnosis of mechanical equipment","cleaned_abstract":"with the development of industry, more and more crucial mechanical machinery generate wildness demand of effective fault diagnosis to ensure the safe operation. over the past few decades, researchers have explored and developed a variety of approaches. in recent years, fault diagnosis based on deep learning has developed rapidly, which achieved satisfied results in the filed of mechanical equipment fault diagnosis. however, there is few review to systematically summarize and sort out these special image deep learning methods. in order to fill this gap, this paper concentrates on reviewing comprehensively the development of special image deep learning for mechanical equipment fault diagnosis in past 5 years. in general, a typical image fault diagnosis based on fault image deep learning generally consists of data acquisition, signal processing, model construction, feature learning and decision-making. firstly, the method of signal preprocessing is introduced, and several common methods of converting signals into images are briefly compared and analyzed. then, the principles and variants of deep learning models are expounded. further more, the difficulties and challenges encountered at this stage are summarized. last but not least, the future development and potential trends of the work are concluded, and it hoped that this work will facilitate and inspire further exploration for researchers in this area.","key_phrases":["the development","industry","more and more crucial mechanical machinery","wildness demand","effective fault diagnosis","the safe operation","the past few decades","researchers","a variety","approaches","recent years","diagnosis","deep learning","which","satisfied results","mechanical equipment fault diagnosis","few review","these special image deep learning methods","order","this gap","this paper","the development","special image","deep learning","mechanical equipment fault diagnosis","past 5 years","in general, a typical image fault diagnosis","fault image","deep learning","data acquisition","signal processing","model construction","feature learning","decision-making","the method","signal preprocessing","several common methods","signals","images","the principles","variants","deep learning models","the difficulties","challenges","this stage","the future development","potential trends","the work","it","this work","further exploration","researchers","this area","the past few decades","recent years","past 5 years","firstly"]},{"title":"A deep learning framework for non-functional requirement classification","authors":["Kiramat Rahman","Anwar Ghani","Sanjay Misra","Arif Ur Rahman"],"abstract":"Analyzing, identifying, and classifying nonfunctional requirements from requirement documents is time-consuming and challenging. Machine learning-based approaches have been proposed to minimize analysts\u2019 efforts, labor, and stress. However, the traditional approach of supervised machine learning necessitates manual feature extraction, which is time-consuming. This study presents a novel deep-learning framework for NFR classification to overcome these limitations. The framework leverages a more profound architecture that naturally captures feature structures, possesses enhanced representational power, and efficiently captures a broader context than shallower structures. To evaluate the effectiveness of the proposed method, an experiment was conducted on two widely-used datasets, encompassing 914 NFR instances. Performance analysis was performed on the applied models, and the results were evaluated using various metrics. Notably, the DReqANN model outperforms the other models in classifying NFR, achieving precision between 81 and 99.8%, recall between 74 and 89%, and F1-score between 83 and 89%. These significant results highlight the exceptional efficacy of the proposed deep learning framework in addressing NFR classification tasks, showcasing its potential for advancing the field of NFR analysis and classification.","doi":"10.1038\/s41598-024-52802-0","cleaned_title":"a deep learning framework for non-functional requirement classification","cleaned_abstract":"analyzing, identifying, and classifying nonfunctional requirements from requirement documents is time-consuming and challenging. machine learning-based approaches have been proposed to minimize analysts\u2019 efforts, labor, and stress. however, the traditional approach of supervised machine learning necessitates manual feature extraction, which is time-consuming. this study presents a novel deep-learning framework for nfr classification to overcome these limitations. the framework leverages a more profound architecture that naturally captures feature structures, possesses enhanced representational power, and efficiently captures a broader context than shallower structures. to evaluate the effectiveness of the proposed method, an experiment was conducted on two widely-used datasets, encompassing 914 nfr instances. performance analysis was performed on the applied models, and the results were evaluated using various metrics. notably, the dreqann model outperforms the other models in classifying nfr, achieving precision between 81 and 99.8%, recall between 74 and 89%, and f1-score between 83 and 89%. these significant results highlight the exceptional efficacy of the proposed deep learning framework in addressing nfr classification tasks, showcasing its potential for advancing the field of nfr analysis and classification.","key_phrases":["nonfunctional requirements","requirement documents","machine learning-based approaches","analysts\u2019 efforts","labor","stress","however, the traditional approach","supervised machine learning necessitates manual feature extraction","which","this study","a novel deep-learning framework","nfr classification","these limitations","the framework","a more profound architecture","that","feature structures","enhanced representational power","a broader context","shallower structures","the effectiveness","the proposed method","an experiment","two widely-used datasets","914 nfr instances","performance analysis","the applied models","the results","various metrics","the dreqann model","the other models","precision","99.8%","74 and 89%","83 and 89%","these significant results","the exceptional efficacy","the proposed deep learning framework","nfr classification tasks","its potential","the field","nfr analysis","classification","two","914","between 81","99.8%","between 74 and","89%","89%"]},{"title":"Deep Learning-Enabled Image Classification for the Determination of Aluminum Ions","authors":["Ce Wang","Zhaoliang Wang","Yifei Lu","Tingting Hao","Yufang Hu","Sui Wang","Zhiyong Guo"],"abstract":"AbstractIn this work, an image classification based on deep learning for quantitative field determination of aluminum ions (Al3+) was developed. Carbon quantum dots with yellow fluorescence were synthesized by a one-pot hydrothermal method which could specifically recognize Al3+ and produce enhanced green fluorescence. Using the convolutional neural network model in deep learning, an image classification was constructed to classify Al3+ samples at different concentrations. Then, a fitting method for classification information was proposed for the first time which could convert discontinuous, semi-quantitative concentration classification information into continuous, quantitative, and accurate concentration information. Recoveries of 92.0\u2013110.3% in the concentration range of 0.3\u2013320 \u03bcM were obtained with a lower limit of detection of 0.3 \u03bcM, exhibiting excellent accuracy and sensitivity. It could be completed in 2 min simply without requiring large equipment. Thus, the deep learning-enabled image classification paves a new way for the determination of metal ions.","doi":"10.1134\/S1061934823110114","cleaned_title":"deep learning-enabled image classification for the determination of aluminum ions","cleaned_abstract":"abstractin this work, an image classification based on deep learning for quantitative field determination of aluminum ions (al3+) was developed. carbon quantum dots with yellow fluorescence were synthesized by a one-pot hydrothermal method which could specifically recognize al3+ and produce enhanced green fluorescence. using the convolutional neural network model in deep learning, an image classification was constructed to classify al3+ samples at different concentrations. then, a fitting method for classification information was proposed for the first time which could convert discontinuous, semi-quantitative concentration classification information into continuous, quantitative, and accurate concentration information. recoveries of 92.0\u2013110.3% in the concentration range of 0.3\u2013320 \u03bcm were obtained with a lower limit of detection of 0.3 \u03bcm, exhibiting excellent accuracy and sensitivity. it could be completed in 2 min simply without requiring large equipment. thus, the deep learning-enabled image classification paves a new way for the determination of metal ions.","key_phrases":["abstractin","this work","an image classification","deep learning","quantitative field determination","aluminum ions","al3","+","carbon quantum dots","yellow fluorescence","a one-pot hydrothermal method","which","al3","enhanced green fluorescence","the convolutional neural network model","deep learning","an image classification","al3","samples","different concentrations","a fitting method","classification information","the first time","which","discontinuous, semi-quantitative concentration classification information","continuous, quantitative, and accurate concentration information","recoveries","92.0\u2013110.3%","the concentration range","0.3\u2013320 \u03bcm","a lower limit","detection","0.3 \u03bcm","excellent accuracy","sensitivity","it","2 min","large equipment","the deep learning-enabled image classification","a new way","the determination","metal ions","abstractin","quantum","one","first","92.0\u2013110.3%","0.3","2"]},{"title":"A comparative analysis of classical machine learning and deep learning techniques for predicting lung cancer survivability","authors":["Shigao Huang","Ibrahim Arpaci","Mostafa Al-Emran","Serhat K\u0131l\u0131\u00e7arslan","Mohammed A. Al-Sharafi"],"abstract":"Lung cancer, one of the deadliest forms of cancer, can significantly improve patient survival rates by 60\u201370% if detected in its early stages. The prediction of lung cancer patient survival has grown to be a popular area of research among medical and computer science experts. This study aims to predict the survival period of lung cancer patients using 12 demographic and clinical features. This is achieved through a comparative analysis between traditional machine learning and deep learning techniques, deviating from previous studies that primarily used CT or X-ray images. The dataset included 10,001 lung cancer patients, and the data attributes involved gender, age, race, T (tumor size), M (tumor dissemination to other organs), N (lymph node involvement), Chemo, DX-Bone, DX-Brain, DX-Liver, DX-Lung, and survival months. Six supervised machine learning and deep learning techniques were applied, including logistic-regression (Logistic), Bayes classifier (BayesNet), lazy-classifier (LWL), meta-classifier (AttributeSelectedClassifier (ASC)), rule-learner (OneR), decision-tree (J48), and deep neural network (DNN). The findings suggest that DNN surpassed the performance of the six traditional machine learning models in accurately predicting the survival duration of lung cancer patients, achieving an accuracy rate of 88.58%. This evidence is thought to assist healthcare experts in cost management and timely treatment provision.","doi":"10.1007\/s11042-023-16349-y","cleaned_title":"a comparative analysis of classical machine learning and deep learning techniques for predicting lung cancer survivability","cleaned_abstract":"lung cancer, one of the deadliest forms of cancer, can significantly improve patient survival rates by 60\u201370% if detected in its early stages. the prediction of lung cancer patient survival has grown to be a popular area of research among medical and computer science experts. this study aims to predict the survival period of lung cancer patients using 12 demographic and clinical features. this is achieved through a comparative analysis between traditional machine learning and deep learning techniques, deviating from previous studies that primarily used ct or x-ray images. the dataset included 10,001 lung cancer patients, and the data attributes involved gender, age, race, t (tumor size), m (tumor dissemination to other organs), n (lymph node involvement), chemo, dx-bone, dx-brain, dx-liver, dx-lung, and survival months. six supervised machine learning and deep learning techniques were applied, including logistic-regression (logistic), bayes classifier (bayesnet), lazy-classifier (lwl), meta-classifier (attributeselectedclassifier (asc)), rule-learner (oner), decision-tree (j48), and deep neural network (dnn). the findings suggest that dnn surpassed the performance of the six traditional machine learning models in accurately predicting the survival duration of lung cancer patients, achieving an accuracy rate of 88.58%. this evidence is thought to assist healthcare experts in cost management and timely treatment provision.","key_phrases":["lung cancer","the deadliest forms","cancer","patient survival rates","60\u201370%","its early stages","the prediction","lung cancer patient survival","a popular area","research","medical and computer science experts","this study","the survival period","lung cancer patients","12 demographic and clinical features","this","a comparative analysis","traditional machine learning","deep learning techniques","previous studies","that","ct or x-ray images","the dataset","10,001 lung cancer patients","the data attributes","gender","age","race","t","tumor size","m","tumor dissemination","other organs","(lymph node involvement","chemo","dx","brain","dx-liver","dx-lung","survival months","six supervised machine learning","deep learning techniques","bayes classifier","lazy-classifier (lwl","meta-classifier","(attributeselectedclassifier","asc","rule-learner","(oner","decision-tree","(j48","deep neural network","dnn","the findings","dnn","the performance","the six traditional machine learning models","the survival duration","lung cancer patients","an accuracy rate","88.58%","this evidence","healthcare experts","cost management","timely treatment provision","one","60\u201370%","12","10,001","months","six","bayes classifier","meta-classifier","six","88.58%"]},{"title":"A comprehensive review of deep learning power in steady-state visual evoked potentials","authors":["Z. T. Al-Qaysi","A. S. Albahri","M. A. Ahmed","Rula A. Hamid","M. A. Alsalem","O. S. Albahri","A. H. Alamoodi","Raad Z. Homod","Ghadeer Ghazi Shayea","Ali M. Duhaim"],"abstract":"Brain\u2013computer interfacing (BCI) research, fueled by deep learning, integrates insights from diverse domains. A notable focus is on steady-state visual evoked potential (SSVEP) in BCI applications, requiring in-depth assessment through deep learning. EEG research frequently employs SSVEPs, which are regarded as normal brain responses to visual stimuli, particularly in investigations of visual perception and attention. This paper tries to give an in-depth analysis of the implications of deep learning for SSVEP-adapted BCI. A systematic search across four stable databases (Web of Science, PubMed, ScienceDirect, and IEEE) was developed to assemble a vast reservoir of relevant theoretical and scientific knowledge. A comprehensive search yielded 177 papers that appeared between 2010 and 2023. Thence a strict screening method from predetermined inclusion criteria finally generated 39 records. These selected works were the basis of the study, presenting alternate views, obstacles, limitations and interesting ideas. By providing a systematic presentation of the material, it has made a key scholarly contribution. It focuses on the technical aspects of SSVEP-based BCI, EEG technologies and complex applications of deep learning technology in these areas. The study delivers more penetrating reporting on the latest deep learning pattern recognition techniques than its predecessors, together with progress in data acquisition and recording means suitable for SSVEP-based BCI devices. Especially in the realms of deep learning technology orchestration, pattern recognition techniques, and EEG data collection, it has effectively closed four important research gaps. To increase the accessibility of this critical material, the results of the study take the form of easy-to-read tables just generated. Applying deep learning techniques in SSVEP-based BCI applications, as the research shows, also has its downsides. The study concludes that a radical framework will be presented which, includes intelligent decision-making tools for evaluation and benchmarking. Rather than just finding a comparable or similar analogy, this framework is intended to help guide future research and pragmatic applications, and to determine which SSVEP-based BCI applications have succeeded at responsibility for what they set out with.","doi":"10.1007\/s00521-024-10143-z","cleaned_title":"a comprehensive review of deep learning power in steady-state visual evoked potentials","cleaned_abstract":"brain\u2013computer interfacing (bci) research, fueled by deep learning, integrates insights from diverse domains. a notable focus is on steady-state visual evoked potential (ssvep) in bci applications, requiring in-depth assessment through deep learning. eeg research frequently employs ssveps, which are regarded as normal brain responses to visual stimuli, particularly in investigations of visual perception and attention. this paper tries to give an in-depth analysis of the implications of deep learning for ssvep-adapted bci. a systematic search across four stable databases (web of science, pubmed, sciencedirect, and ieee) was developed to assemble a vast reservoir of relevant theoretical and scientific knowledge. a comprehensive search yielded 177 papers that appeared between 2010 and 2023. thence a strict screening method from predetermined inclusion criteria finally generated 39 records. these selected works were the basis of the study, presenting alternate views, obstacles, limitations and interesting ideas. by providing a systematic presentation of the material, it has made a key scholarly contribution. it focuses on the technical aspects of ssvep-based bci, eeg technologies and complex applications of deep learning technology in these areas. the study delivers more penetrating reporting on the latest deep learning pattern recognition techniques than its predecessors, together with progress in data acquisition and recording means suitable for ssvep-based bci devices. especially in the realms of deep learning technology orchestration, pattern recognition techniques, and eeg data collection, it has effectively closed four important research gaps. to increase the accessibility of this critical material, the results of the study take the form of easy-to-read tables just generated. applying deep learning techniques in ssvep-based bci applications, as the research shows, also has its downsides. the study concludes that a radical framework will be presented which, includes intelligent decision-making tools for evaluation and benchmarking. rather than just finding a comparable or similar analogy, this framework is intended to help guide future research and pragmatic applications, and to determine which ssvep-based bci applications have succeeded at responsibility for what they set out with.","key_phrases":["brain","computer","(bci) research","deep learning","insights","diverse domains","a notable focus","steady-state visual evoked potential","ssvep","bci applications","depth","deep learning","eeg research","ssveps","which","normal brain responses","visual stimuli","investigations","visual perception","attention","this paper","an in-depth analysis","the implications","deep learning","ssvep-adapted bci","a systematic search","four stable databases","web","science","sciencedirect","ieee","a vast reservoir","relevant theoretical and scientific knowledge","a comprehensive search","177 papers","that","a strict screening method","predetermined inclusion criteria","39 records","these selected works","the basis","the study","alternate views","obstacles","limitations","interesting ideas","a systematic presentation","the material","it","a key scholarly contribution","it","the technical aspects","ssvep-based bci","eeg technologies","complex applications","deep learning technology","these areas","the study","the latest deep learning pattern recognition techniques","its predecessors","progress","data acquisition","recording","ssvep-based bci devices","the realms","deep learning technology orchestration","pattern recognition techniques","eeg data collection","it","four important research gaps","the accessibility","this critical material","the results","the study","the form","read","deep learning techniques","ssvep-based bci applications","its downsides","the study","a radical framework","which","intelligent decision-making tools","evaluation","benchmarking","a comparable or similar analogy","this framework","future research and pragmatic applications","which","ssvep-based bci applications","responsibility","what","they","four","177","between 2010 and 2023","39","four"]},{"title":"Quantum deep learning-based anomaly detection for enhanced network security","authors":["Moe Hdaib","Sutharshan Rajasegarar","Lei Pan"],"abstract":"Identifying and mitigating aberrant activities within the network traffic is important to prevent adverse consequences caused by cyber security incidents, which have been increasing significantly in recent times. Existing research mainly focuses on classical machine learning and deep learning-based approaches for detecting such attacks. However, exploiting the power of quantum deep learning to process complex correlation of features for anomaly detection is not well explored. Hence, in this paper, we investigate quantum machine learning and quantum deep learning-based anomaly detection methodologies to accurately detect network attacks. In particular, we propose three novel quantum auto-encoder-based anomaly detection frameworks. Our primary aim is to create hybrid models that leverage the strengths of both quantum and deep learning methodologies for efficient anomaly recognition. The three frameworks are formed by integrating the quantum autoencoder with a quantum one-class support vector machine, a quantum random forest, and a quantum k-nearest neighbor approach. The anomaly detection capability of the frameworks is evaluated using benchmark datasets comprising computer and Internet of Things network flows. Our evaluation demonstrates that all three frameworks have a high potential to detect the network traffic anomalies accurately, while the framework that integrates the quantum autoencoder with the quantum k-nearest neighbor yields the highest accuracy. This demonstrates the promising potential for the development of quantum frameworks for anomaly detection, underscoring their relevance for future advancements in network security.","doi":"10.1007\/s42484-024-00163-2","cleaned_title":"quantum deep learning-based anomaly detection for enhanced network security","cleaned_abstract":"identifying and mitigating aberrant activities within the network traffic is important to prevent adverse consequences caused by cyber security incidents, which have been increasing significantly in recent times. existing research mainly focuses on classical machine learning and deep learning-based approaches for detecting such attacks. however, exploiting the power of quantum deep learning to process complex correlation of features for anomaly detection is not well explored. hence, in this paper, we investigate quantum machine learning and quantum deep learning-based anomaly detection methodologies to accurately detect network attacks. in particular, we propose three novel quantum auto-encoder-based anomaly detection frameworks. our primary aim is to create hybrid models that leverage the strengths of both quantum and deep learning methodologies for efficient anomaly recognition. the three frameworks are formed by integrating the quantum autoencoder with a quantum one-class support vector machine, a quantum random forest, and a quantum k-nearest neighbor approach. the anomaly detection capability of the frameworks is evaluated using benchmark datasets comprising computer and internet of things network flows. our evaluation demonstrates that all three frameworks have a high potential to detect the network traffic anomalies accurately, while the framework that integrates the quantum autoencoder with the quantum k-nearest neighbor yields the highest accuracy. this demonstrates the promising potential for the development of quantum frameworks for anomaly detection, underscoring their relevance for future advancements in network security.","key_phrases":["aberrant activities","the network traffic","adverse consequences","cyber security incidents","which","recent times","existing research","classical machine learning","deep learning-based approaches","such attacks","the power","quantum","complex correlation","features","anomaly detection","this paper","we","quantum machine learning","quantum","deep learning-based anomaly detection methodologies","network attacks","we","three novel quantum auto-encoder-based anomaly detection frameworks","our primary aim","hybrid models","that","the strengths","both quantum","deep learning methodologies","efficient anomaly recognition","the three frameworks","the quantum","autoencoder","a quantum one-class support vector machine","a quantum random forest","a quantum k-nearest neighbor approach","the anomaly detection capability","the frameworks","benchmark datasets","computer","internet","things network","our evaluation","all three frameworks","a high potential","the network traffic anomalies","the framework","that","the quantum","the quantum k-nearest neighbor","the highest accuracy","this","the promising potential","the development","quantum frameworks","anomaly detection","their relevance","future advancements","network security","anomaly detection methodologies","three","anomaly detection frameworks","anomaly recognition","three","one","quantum","the anomaly detection capability","three","quantum","quantum"]},{"title":"An enhanced deep learning approach for vascular wall fracture analysis","authors":["Alexandros Tragoudas","Marta Alloisio","Elsayed S. Elsayed","T. Christian Gasser","Fadi Aldakheel"],"abstract":"This work outlines an efficient deep learning approach for analyzing vascular wall fractures using experimental data with openly accessible source codes (https:\/\/doi.org\/10.25835\/weuhha72) for reproduction. Vascular disease remains the primary cause of death globally to this day. Tissue damage in these vascular disorders is closely tied to how the diseases develop, which requires careful study. Therefore, the scientific community has dedicated significant efforts to capture the properties of vessel wall fractures. The symmetry-constrained compact tension (symconCT) test combined with digital image correlation (DIC) enabled the study of tissue fracture in various aorta specimens under different conditions. Main purpose of the experiments was to investigate the displacement and strain field ahead of the crack tip. These experimental data were to support the development and verification of computational models. The FEM model used the DIC information for the material parameters identification. Traditionally, the analysis of fracture processes in biological tissues involves extensive computational and experimental efforts due to the complex nature of tissue behavior under stress. These high costs have posed significant challenges, demanding efficient solutions to accelerate research progress and reduce embedded costs. Deep learning techniques have shown promise in overcoming these challenges by learning to indicate patterns and relationships between the input and label data. In this study, we integrate deep learning methodologies with the attention residual U-Net architecture to predict fracture responses in porcine aorta specimens, enhanced with a Monte Carlo dropout technique. By training the network on a sufficient amount of data, the model learns to capture the features influencing fracture progression. These parameterized datasets consist of pictures describing the evolution of tissue fracture path along with the DIC measurements. The integration of deep learning should not only enhance the predictive accuracy, but also significantly reduce the computational and experimental burden, thereby enabling a more efficient analysis of fracture response.","doi":"10.1007\/s00419-024-02589-3","cleaned_title":"an enhanced deep learning approach for vascular wall fracture analysis","cleaned_abstract":"this work outlines an efficient deep learning approach for analyzing vascular wall fractures using experimental data with openly accessible source codes (https:\/\/doi.org\/10.25835\/weuhha72) for reproduction. vascular disease remains the primary cause of death globally to this day. tissue damage in these vascular disorders is closely tied to how the diseases develop, which requires careful study. therefore, the scientific community has dedicated significant efforts to capture the properties of vessel wall fractures. the symmetry-constrained compact tension (symconct) test combined with digital image correlation (dic) enabled the study of tissue fracture in various aorta specimens under different conditions. main purpose of the experiments was to investigate the displacement and strain field ahead of the crack tip. these experimental data were to support the development and verification of computational models. the fem model used the dic information for the material parameters identification. traditionally, the analysis of fracture processes in biological tissues involves extensive computational and experimental efforts due to the complex nature of tissue behavior under stress. these high costs have posed significant challenges, demanding efficient solutions to accelerate research progress and reduce embedded costs. deep learning techniques have shown promise in overcoming these challenges by learning to indicate patterns and relationships between the input and label data. in this study, we integrate deep learning methodologies with the attention residual u-net architecture to predict fracture responses in porcine aorta specimens, enhanced with a monte carlo dropout technique. by training the network on a sufficient amount of data, the model learns to capture the features influencing fracture progression. these parameterized datasets consist of pictures describing the evolution of tissue fracture path along with the dic measurements. the integration of deep learning should not only enhance the predictive accuracy, but also significantly reduce the computational and experimental burden, thereby enabling a more efficient analysis of fracture response.","key_phrases":["this work","an efficient deep learning approach","vascular wall fractures","experimental data","openly accessible source codes","https:\/\/doi.org\/10.25835\/weuhha72","reproduction","vascular disease","the primary cause","death","this day","tissue damage","these vascular disorders","the diseases","which","careful study","the scientific community","significant efforts","the properties","vessel wall fractures","the symmetry-constrained compact tension","(symconct) test","digital image correlation","the study","tissue fracture","various aorta","specimens","different conditions","main purpose","the experiments","the displacement and strain field","the crack tip","these experimental data","the development","verification","computational models","the fem model","the dic information","the material parameters identification","the analysis","fracture processes","biological tissues","extensive computational and experimental efforts","the complex nature","tissue behavior","stress","these high costs","significant challenges","efficient solutions","research progress","embedded costs","deep learning techniques","promise","these challenges","patterns","relationships","the input and label data","this study","we","deep learning methodologies","the attention residual u-net architecture","fracture responses","porcine aorta specimens","a monte carlo dropout technique","the network","a sufficient amount","data","the model","the features","fracture progression","these parameterized datasets","pictures","the evolution","tissue fracture","path","the dic measurements","the integration","deep learning","the predictive accuracy","the computational and experimental burden","a more efficient analysis","fracture response","this day"]},{"title":"Deep Learning-Based Image and Video Inpainting: A Survey","authors":["Weize Quan","Jiaxi Chen","Yanli Liu","Dong-Ming Yan","Peter Wonka"],"abstract":"Image and video inpainting is a classic problem in computer vision and computer graphics, aiming to fill in the plausible and realistic content in the missing areas of images and videos. With the advance of deep learning, this problem has achieved significant progress recently. The goal of this paper is to comprehensively review the deep learning-based methods for image and video inpainting. Specifically, we sort existing methods into different categories from the perspective of their high-level inpainting pipeline, present different deep learning architectures, including CNN, VAE, GAN, diffusion models, etc., and summarize techniques for module design. We review the training objectives and the common benchmark datasets. We present evaluation metrics for low-level pixel and high-level perceptional similarity, conduct a performance evaluation, and discuss the strengths and weaknesses of representative inpainting methods. We also discuss related real-world applications. Finally, we discuss open challenges and suggest potential future research directions.","doi":"10.1007\/s11263-023-01977-6","cleaned_title":"deep learning-based image and video inpainting: a survey","cleaned_abstract":"image and video inpainting is a classic problem in computer vision and computer graphics, aiming to fill in the plausible and realistic content in the missing areas of images and videos. with the advance of deep learning, this problem has achieved significant progress recently. the goal of this paper is to comprehensively review the deep learning-based methods for image and video inpainting. specifically, we sort existing methods into different categories from the perspective of their high-level inpainting pipeline, present different deep learning architectures, including cnn, vae, gan, diffusion models, etc., and summarize techniques for module design. we review the training objectives and the common benchmark datasets. we present evaluation metrics for low-level pixel and high-level perceptional similarity, conduct a performance evaluation, and discuss the strengths and weaknesses of representative inpainting methods. we also discuss related real-world applications. finally, we discuss open challenges and suggest potential future research directions.","key_phrases":["image","a classic problem","computer vision","computer graphics","the plausible and realistic content","the missing areas","images","videos","the advance","deep learning","this problem","significant progress","the goal","this paper","the deep learning-based methods","image","we","existing methods","different categories","the perspective","their high-level inpainting pipeline","different deep learning architectures","cnn","vae","gan","diffusion models","techniques","module design","we","the training objectives","the common benchmark datasets","we","evaluation metrics","low-level pixel","high-level perceptional similarity","a performance evaluation","the strengths","weaknesses","representative inpainting methods","we","related real-world applications","we","open challenges","potential future research directions","cnn","vae, gan, diffusion models"]},{"title":"Deep learning for osteoporosis screening using an anteroposterior hip radiograph image","authors":["Artit Boonrod","Prarinthorn Piyaprapaphan","Nut Kittipongphat","Daris Theerakulpisut","Arunnit Boonrod"],"abstract":"PurposeOsteoporosis is a common bone disorder characterized by decreased bone mineral density (BMD) and increased bone fragility, which can lead to fractures and eventually cause morbidity and mortality. It is of great concern that the one-year mortality rate for osteoporotic hip fractures could be as high as 22%, regardless of the treatment. Currently, BMD measurement is the standard method for osteoporosis diagnosis, but it is costly and requires special equipment. While a plain radiograph can be obtained more simply and inexpensively, it is not used for diagnosis. Deep learning technologies had been applied to various medical contexts, yet few to osteoporosis unless they were trained on the advanced investigative images, such as computed tomography. The purpose of this study was to develop a deep learning model using the anteroposterior hip radiograph images and measure its diagnostic accuracy for osteoporosis.MethodsWe retrospectively collected all anteroposterior hip radiograph images of patients from 2013 to 2021 at a tertiary care hospital. The BMD measurements of the included patients were reviewed, and the radiograph images that had a time interval of more than two years from the measurements were excluded. All images were randomized using a computer-generated unequal allocation into two datasets, i.e., 80% of images were used for the training dataset and the remaining 20% for the test dataset. The T score of BMD obtained from the ipsilateral femoral neck of the same patient closest to the date of the performed radiograph was chosen. The T score cutoff value of\u2009\u2212\u20092.5 was used to diagnose osteoporosis. Five deep learning models were trained on the training dataset, and their diagnostic performances were evaluated using the test dataset. Finally, the best model was determined by the area under the curves (AUC).ResultsA total of 363 anteroposterior hip radiograph images were identified. The average time interval between the performed radiograph and the BMD measurement was 6.6\u00a0months. Two-hundred-thirteen images were labeled as non-osteoporosis (T score\u2009>\u2009\u2009\u2212\u20092.5), and the other 150 images as osteoporosis (T score \u2264\u2009 \u2212\u20092.5). The best-selected deep learning model achieved an AUC of 0.91 and accuracy of 0.82.ConclusionsThis study demonstrates the potential of deep learning for osteoporosis screening using anteroposterior hip radiographs. The results suggest that the deep learning model might potentially be used as a screening tool to find patients at risk for osteoporosis to perform further BMD measurement.","doi":"10.1007\/s00590-024-04032-3","cleaned_title":"deep learning for osteoporosis screening using an anteroposterior hip radiograph image","cleaned_abstract":"purposeosteoporosis is a common bone disorder characterized by decreased bone mineral density (bmd) and increased bone fragility, which can lead to fractures and eventually cause morbidity and mortality. it is of great concern that the one-year mortality rate for osteoporotic hip fractures could be as high as 22%, regardless of the treatment. currently, bmd measurement is the standard method for osteoporosis diagnosis, but it is costly and requires special equipment. while a plain radiograph can be obtained more simply and inexpensively, it is not used for diagnosis. deep learning technologies had been applied to various medical contexts, yet few to osteoporosis unless they were trained on the advanced investigative images, such as computed tomography. the purpose of this study was to develop a deep learning model using the anteroposterior hip radiograph images and measure its diagnostic accuracy for osteoporosis.methodswe retrospectively collected all anteroposterior hip radiograph images of patients from 2013 to 2021 at a tertiary care hospital. the bmd measurements of the included patients were reviewed, and the radiograph images that had a time interval of more than two years from the measurements were excluded. all images were randomized using a computer-generated unequal allocation into two datasets, i.e., 80% of images were used for the training dataset and the remaining 20% for the test dataset. the t score of bmd obtained from the ipsilateral femoral neck of the same patient closest to the date of the performed radiograph was chosen. the t score cutoff value of \u2212 2.5 was used to diagnose osteoporosis. five deep learning models were trained on the training dataset, and their diagnostic performances were evaluated using the test dataset. finally, the best model was determined by the area under the curves (auc).resultsa total of 363 anteroposterior hip radiograph images were identified. the average time interval between the performed radiograph and the bmd measurement was 6.6 months. two-hundred-thirteen images were labeled as non-osteoporosis (t score > \u2212 2.5), and the other 150 images as osteoporosis (t score \u2264 \u2212 2.5). the best-selected deep learning model achieved an auc of 0.91 and accuracy of 0.82.conclusionsthis study demonstrates the potential of deep learning for osteoporosis screening using anteroposterior hip radiographs. the results suggest that the deep learning model might potentially be used as a screening tool to find patients at risk for osteoporosis to perform further bmd measurement.","key_phrases":["purposeosteoporosis","a common bone disorder","decreased bone mineral density","bmd","increased bone fragility","which","fractures","morbidity","mortality","it","great concern","the one-year mortality rate","osteoporotic hip fractures","22%","the treatment","bmd measurement","the standard method","osteoporosis diagnosis","it","special equipment","a plain radiograph","it","diagnosis","deep learning technologies","various medical contexts","they","the advanced investigative images","computed tomography","the purpose","this study","a deep learning model","the anteroposterior hip radiograph images","its diagnostic accuracy","osteoporosis.methodswe","all anteroposterior hip radiograph images","patients","a tertiary care hospital","the bmd measurements","the included patients","the radiograph images","that","a time interval","more than two years","the measurements","all images","a computer-generated unequal allocation","two datasets","80%","images","the training dataset","the remaining 20%","the test dataset","the t score","bmd","the ipsilateral femoral neck","the same patient","the date","the performed radiograph","the t","cutoff value","\u2212","osteoporosis","five deep learning models","the training dataset","their diagnostic performances","the test dataset","the best model","the area","the curves","auc).resultsa total","363 anteroposterior hip radiograph images","the average time interval","the performed radiograph","the bmd measurement","6.6 months","two-hundred-thirteen images","osteoporosis","t","\u2212","the other 150 images","osteoporosis","t","\u2212","the best-selected deep learning model","an auc","accuracy","0.82.conclusionsthis study","the potential","deep learning","anteroposterior hip radiographs","the results","the deep learning model","a screening tool","patients","risk","osteoporosis","further bmd measurement","one-year","22%","2013","2021","tertiary","more than two years","two","80%","the remaining 20%","\u2212 2.5","five","363","6.6 months","two-hundred-thirteen","2.5","150","2.5","0.91"]},{"title":"Deep learning-based solution for smart contract vulnerabilities detection","authors":["Xueyan Tang","Yuying Du","Alan Lai","Ze Zhang","Lingzhi Shi"],"abstract":"This paper aims to explore the application of deep learning in smart contract vulnerabilities detection. Smart contracts are an essential part of blockchain technology and are crucial for developing decentralized applications. However, smart contract vulnerabilities can cause financial losses and system crashes. Static analysis tools are frequently used to detect vulnerabilities in smart contracts, but they often result in false positives and false negatives because of their high reliance on predefined rules and lack of semantic analysis capabilities. Furthermore, these predefined rules quickly become obsolete and fail to adapt or generalize to new data. In contrast, deep learning methods do not require predefined detection rules and can learn the features of vulnerabilities during the training process. In this paper, we introduce a solution called Lightning Cat which is based on deep learning techniques. We train three deep learning models for detecting vulnerabilities in smart contract: Optimized-CodeBERT, Optimized-LSTM, and Optimized-CNN. Experimental results show that, in the Lightning Cat we propose, Optimized-CodeBERT model surpasses other methods, achieving an f1-score of 93.53%. To precisely extract vulnerability features, we acquire segments of vulnerable code functions to retain critical vulnerability features. Using the CodeBERT pre-training model for data preprocessing, we could capture the syntax and semantics of the code more accurately. To demonstrate the feasibility of our proposed solution, we evaluate its performance using the SolidiFI-benchmark dataset, which consists of 9369 vulnerable contracts injected with vulnerabilities from seven different types.","doi":"10.1038\/s41598-023-47219-0","cleaned_title":"deep learning-based solution for smart contract vulnerabilities detection","cleaned_abstract":"this paper aims to explore the application of deep learning in smart contract vulnerabilities detection. smart contracts are an essential part of blockchain technology and are crucial for developing decentralized applications. however, smart contract vulnerabilities can cause financial losses and system crashes. static analysis tools are frequently used to detect vulnerabilities in smart contracts, but they often result in false positives and false negatives because of their high reliance on predefined rules and lack of semantic analysis capabilities. furthermore, these predefined rules quickly become obsolete and fail to adapt or generalize to new data. in contrast, deep learning methods do not require predefined detection rules and can learn the features of vulnerabilities during the training process. in this paper, we introduce a solution called lightning cat which is based on deep learning techniques. we train three deep learning models for detecting vulnerabilities in smart contract: optimized-codebert, optimized-lstm, and optimized-cnn. experimental results show that, in the lightning cat we propose, optimized-codebert model surpasses other methods, achieving an f1-score of 93.53%. to precisely extract vulnerability features, we acquire segments of vulnerable code functions to retain critical vulnerability features. using the codebert pre-training model for data preprocessing, we could capture the syntax and semantics of the code more accurately. to demonstrate the feasibility of our proposed solution, we evaluate its performance using the solidifi-benchmark dataset, which consists of 9369 vulnerable contracts injected with vulnerabilities from seven different types.","key_phrases":["this paper","the application","deep learning","smart contract vulnerabilities detection","smart contracts","an essential part","blockchain technology","decentralized applications","smart contract vulnerabilities","financial losses","system crashes","static analysis tools","vulnerabilities","smart contracts","they","false positives","false negatives","their high reliance","predefined rules","lack","semantic analysis capabilities","these predefined rules","new data","contrast","deep learning methods","predefined detection rules","the features","vulnerabilities","the training process","this paper","we","a solution","lightning cat","which","deep learning techniques","we","three deep learning models","vulnerabilities","smart contract","optimized-codebert","optimized-lstm","cnn","experimental results","the lightning cat","we","optimized-codebert model","other methods","an f1-score","93.53%","vulnerability features","we","segments","vulnerable code functions","critical vulnerability features","the codebert pre-training model","data","we","the syntax","semantics","the code","the feasibility","our proposed solution","we","its performance","the solidifi-benchmark dataset","which","9369 vulnerable contracts","vulnerabilities","seven different types","three","93.53%","9369","seven"]},{"title":"Image deep learning in fault diagnosis of mechanical equipment","authors":["Chuanhao Wang","Yongjian Sun","Xiaohong Wang"],"abstract":"With the development of industry, more and more crucial mechanical machinery generate wildness demand of effective fault diagnosis to ensure the safe operation. Over the past few decades, researchers have explored and developed a variety of approaches. In recent years, fault diagnosis based on deep learning has developed rapidly, which achieved satisfied results in the filed of mechanical equipment fault diagnosis. However, there is few review to systematically summarize and sort out these special image deep learning methods. In order to fill this gap, this paper concentrates on reviewing comprehensively the development of special image deep learning for mechanical equipment fault diagnosis in past 5 years. In general, a typical image fault diagnosis based on fault image deep learning generally consists of data acquisition, signal processing, model construction, feature learning and decision-making. Firstly, the method of signal preprocessing is introduced, and several common methods of converting signals into images are briefly compared and analyzed. Then, the principles and variants of deep learning models are expounded. Further more, the difficulties and challenges encountered at this stage are summarized. Last but not least, the future development and potential trends of the work are concluded, and it hoped that this work will facilitate and inspire further exploration for researchers in this area.","doi":"10.1007\/s10845-023-02176-3","cleaned_title":"image deep learning in fault diagnosis of mechanical equipment","cleaned_abstract":"with the development of industry, more and more crucial mechanical machinery generate wildness demand of effective fault diagnosis to ensure the safe operation. over the past few decades, researchers have explored and developed a variety of approaches. in recent years, fault diagnosis based on deep learning has developed rapidly, which achieved satisfied results in the filed of mechanical equipment fault diagnosis. however, there is few review to systematically summarize and sort out these special image deep learning methods. in order to fill this gap, this paper concentrates on reviewing comprehensively the development of special image deep learning for mechanical equipment fault diagnosis in past 5 years. in general, a typical image fault diagnosis based on fault image deep learning generally consists of data acquisition, signal processing, model construction, feature learning and decision-making. firstly, the method of signal preprocessing is introduced, and several common methods of converting signals into images are briefly compared and analyzed. then, the principles and variants of deep learning models are expounded. further more, the difficulties and challenges encountered at this stage are summarized. last but not least, the future development and potential trends of the work are concluded, and it hoped that this work will facilitate and inspire further exploration for researchers in this area.","key_phrases":["the development","industry","more and more crucial mechanical machinery","wildness demand","effective fault diagnosis","the safe operation","the past few decades","researchers","a variety","approaches","recent years","diagnosis","deep learning","which","satisfied results","mechanical equipment fault diagnosis","few review","these special image deep learning methods","order","this gap","this paper","the development","special image","deep learning","mechanical equipment fault diagnosis","past 5 years","in general, a typical image fault diagnosis","fault image","deep learning","data acquisition","signal processing","model construction","feature learning","decision-making","the method","signal preprocessing","several common methods","signals","images","the principles","variants","deep learning models","the difficulties","challenges","this stage","the future development","potential trends","the work","it","this work","further exploration","researchers","this area","the past few decades","recent years","past 5 years","firstly"]},{"title":"A deep learning framework for non-functional requirement classification","authors":["Kiramat Rahman","Anwar Ghani","Sanjay Misra","Arif Ur Rahman"],"abstract":"Analyzing, identifying, and classifying nonfunctional requirements from requirement documents is time-consuming and challenging. Machine learning-based approaches have been proposed to minimize analysts\u2019 efforts, labor, and stress. However, the traditional approach of supervised machine learning necessitates manual feature extraction, which is time-consuming. This study presents a novel deep-learning framework for NFR classification to overcome these limitations. The framework leverages a more profound architecture that naturally captures feature structures, possesses enhanced representational power, and efficiently captures a broader context than shallower structures. To evaluate the effectiveness of the proposed method, an experiment was conducted on two widely-used datasets, encompassing 914 NFR instances. Performance analysis was performed on the applied models, and the results were evaluated using various metrics. Notably, the DReqANN model outperforms the other models in classifying NFR, achieving precision between 81 and 99.8%, recall between 74 and 89%, and F1-score between 83 and 89%. These significant results highlight the exceptional efficacy of the proposed deep learning framework in addressing NFR classification tasks, showcasing its potential for advancing the field of NFR analysis and classification.","doi":"10.1038\/s41598-024-52802-0","cleaned_title":"a deep learning framework for non-functional requirement classification","cleaned_abstract":"analyzing, identifying, and classifying nonfunctional requirements from requirement documents is time-consuming and challenging. machine learning-based approaches have been proposed to minimize analysts\u2019 efforts, labor, and stress. however, the traditional approach of supervised machine learning necessitates manual feature extraction, which is time-consuming. this study presents a novel deep-learning framework for nfr classification to overcome these limitations. the framework leverages a more profound architecture that naturally captures feature structures, possesses enhanced representational power, and efficiently captures a broader context than shallower structures. to evaluate the effectiveness of the proposed method, an experiment was conducted on two widely-used datasets, encompassing 914 nfr instances. performance analysis was performed on the applied models, and the results were evaluated using various metrics. notably, the dreqann model outperforms the other models in classifying nfr, achieving precision between 81 and 99.8%, recall between 74 and 89%, and f1-score between 83 and 89%. these significant results highlight the exceptional efficacy of the proposed deep learning framework in addressing nfr classification tasks, showcasing its potential for advancing the field of nfr analysis and classification.","key_phrases":["nonfunctional requirements","requirement documents","machine learning-based approaches","analysts\u2019 efforts","labor","stress","however, the traditional approach","supervised machine learning necessitates manual feature extraction","which","this study","a novel deep-learning framework","nfr classification","these limitations","the framework","a more profound architecture","that","feature structures","enhanced representational power","a broader context","shallower structures","the effectiveness","the proposed method","an experiment","two widely-used datasets","914 nfr instances","performance analysis","the applied models","the results","various metrics","the dreqann model","the other models","precision","99.8%","74 and 89%","83 and 89%","these significant results","the exceptional efficacy","the proposed deep learning framework","nfr classification tasks","its potential","the field","nfr analysis","classification","two","914","between 81","99.8%","between 74 and","89%","89%"]},{"title":"Deep Learning-Enabled Image Classification for the Determination of Aluminum Ions","authors":["Ce Wang","Zhaoliang Wang","Yifei Lu","Tingting Hao","Yufang Hu","Sui Wang","Zhiyong Guo"],"abstract":"AbstractIn this work, an image classification based on deep learning for quantitative field determination of aluminum ions (Al3+) was developed. Carbon quantum dots with yellow fluorescence were synthesized by a one-pot hydrothermal method which could specifically recognize Al3+ and produce enhanced green fluorescence. Using the convolutional neural network model in deep learning, an image classification was constructed to classify Al3+ samples at different concentrations. Then, a fitting method for classification information was proposed for the first time which could convert discontinuous, semi-quantitative concentration classification information into continuous, quantitative, and accurate concentration information. Recoveries of 92.0\u2013110.3% in the concentration range of 0.3\u2013320 \u03bcM were obtained with a lower limit of detection of 0.3 \u03bcM, exhibiting excellent accuracy and sensitivity. It could be completed in 2 min simply without requiring large equipment. Thus, the deep learning-enabled image classification paves a new way for the determination of metal ions.","doi":"10.1134\/S1061934823110114","cleaned_title":"deep learning-enabled image classification for the determination of aluminum ions","cleaned_abstract":"abstractin this work, an image classification based on deep learning for quantitative field determination of aluminum ions (al3+) was developed. carbon quantum dots with yellow fluorescence were synthesized by a one-pot hydrothermal method which could specifically recognize al3+ and produce enhanced green fluorescence. using the convolutional neural network model in deep learning, an image classification was constructed to classify al3+ samples at different concentrations. then, a fitting method for classification information was proposed for the first time which could convert discontinuous, semi-quantitative concentration classification information into continuous, quantitative, and accurate concentration information. recoveries of 92.0\u2013110.3% in the concentration range of 0.3\u2013320 \u03bcm were obtained with a lower limit of detection of 0.3 \u03bcm, exhibiting excellent accuracy and sensitivity. it could be completed in 2 min simply without requiring large equipment. thus, the deep learning-enabled image classification paves a new way for the determination of metal ions.","key_phrases":["abstractin","this work","an image classification","deep learning","quantitative field determination","aluminum ions","al3","+","carbon quantum dots","yellow fluorescence","a one-pot hydrothermal method","which","al3","enhanced green fluorescence","the convolutional neural network model","deep learning","an image classification","al3","samples","different concentrations","a fitting method","classification information","the first time","which","discontinuous, semi-quantitative concentration classification information","continuous, quantitative, and accurate concentration information","recoveries","92.0\u2013110.3%","the concentration range","0.3\u2013320 \u03bcm","a lower limit","detection","0.3 \u03bcm","excellent accuracy","sensitivity","it","2 min","large equipment","the deep learning-enabled image classification","a new way","the determination","metal ions","abstractin","quantum","one","first","92.0\u2013110.3%","0.3","2"]},{"title":"A comparative analysis of classical machine learning and deep learning techniques for predicting lung cancer survivability","authors":["Shigao Huang","Ibrahim Arpaci","Mostafa Al-Emran","Serhat K\u0131l\u0131\u00e7arslan","Mohammed A. Al-Sharafi"],"abstract":"Lung cancer, one of the deadliest forms of cancer, can significantly improve patient survival rates by 60\u201370% if detected in its early stages. The prediction of lung cancer patient survival has grown to be a popular area of research among medical and computer science experts. This study aims to predict the survival period of lung cancer patients using 12 demographic and clinical features. This is achieved through a comparative analysis between traditional machine learning and deep learning techniques, deviating from previous studies that primarily used CT or X-ray images. The dataset included 10,001 lung cancer patients, and the data attributes involved gender, age, race, T (tumor size), M (tumor dissemination to other organs), N (lymph node involvement), Chemo, DX-Bone, DX-Brain, DX-Liver, DX-Lung, and survival months. Six supervised machine learning and deep learning techniques were applied, including logistic-regression (Logistic), Bayes classifier (BayesNet), lazy-classifier (LWL), meta-classifier (AttributeSelectedClassifier (ASC)), rule-learner (OneR), decision-tree (J48), and deep neural network (DNN). The findings suggest that DNN surpassed the performance of the six traditional machine learning models in accurately predicting the survival duration of lung cancer patients, achieving an accuracy rate of 88.58%. This evidence is thought to assist healthcare experts in cost management and timely treatment provision.","doi":"10.1007\/s11042-023-16349-y","cleaned_title":"a comparative analysis of classical machine learning and deep learning techniques for predicting lung cancer survivability","cleaned_abstract":"lung cancer, one of the deadliest forms of cancer, can significantly improve patient survival rates by 60\u201370% if detected in its early stages. the prediction of lung cancer patient survival has grown to be a popular area of research among medical and computer science experts. this study aims to predict the survival period of lung cancer patients using 12 demographic and clinical features. this is achieved through a comparative analysis between traditional machine learning and deep learning techniques, deviating from previous studies that primarily used ct or x-ray images. the dataset included 10,001 lung cancer patients, and the data attributes involved gender, age, race, t (tumor size), m (tumor dissemination to other organs), n (lymph node involvement), chemo, dx-bone, dx-brain, dx-liver, dx-lung, and survival months. six supervised machine learning and deep learning techniques were applied, including logistic-regression (logistic), bayes classifier (bayesnet), lazy-classifier (lwl), meta-classifier (attributeselectedclassifier (asc)), rule-learner (oner), decision-tree (j48), and deep neural network (dnn). the findings suggest that dnn surpassed the performance of the six traditional machine learning models in accurately predicting the survival duration of lung cancer patients, achieving an accuracy rate of 88.58%. this evidence is thought to assist healthcare experts in cost management and timely treatment provision.","key_phrases":["lung cancer","the deadliest forms","cancer","patient survival rates","60\u201370%","its early stages","the prediction","lung cancer patient survival","a popular area","research","medical and computer science experts","this study","the survival period","lung cancer patients","12 demographic and clinical features","this","a comparative analysis","traditional machine learning","deep learning techniques","previous studies","that","ct or x-ray images","the dataset","10,001 lung cancer patients","the data attributes","gender","age","race","t","tumor size","m","tumor dissemination","other organs","(lymph node involvement","chemo","dx","brain","dx-liver","dx-lung","survival months","six supervised machine learning","deep learning techniques","bayes classifier","lazy-classifier (lwl","meta-classifier","(attributeselectedclassifier","asc","rule-learner","(oner","decision-tree","(j48","deep neural network","dnn","the findings","dnn","the performance","the six traditional machine learning models","the survival duration","lung cancer patients","an accuracy rate","88.58%","this evidence","healthcare experts","cost management","timely treatment provision","one","60\u201370%","12","10,001","months","six","bayes classifier","meta-classifier","six","88.58%"]},{"title":"A comprehensive review of deep learning power in steady-state visual evoked potentials","authors":["Z. T. Al-Qaysi","A. S. Albahri","M. A. Ahmed","Rula A. Hamid","M. A. Alsalem","O. S. Albahri","A. H. Alamoodi","Raad Z. Homod","Ghadeer Ghazi Shayea","Ali M. Duhaim"],"abstract":"Brain\u2013computer interfacing (BCI) research, fueled by deep learning, integrates insights from diverse domains. A notable focus is on steady-state visual evoked potential (SSVEP) in BCI applications, requiring in-depth assessment through deep learning. EEG research frequently employs SSVEPs, which are regarded as normal brain responses to visual stimuli, particularly in investigations of visual perception and attention. This paper tries to give an in-depth analysis of the implications of deep learning for SSVEP-adapted BCI. A systematic search across four stable databases (Web of Science, PubMed, ScienceDirect, and IEEE) was developed to assemble a vast reservoir of relevant theoretical and scientific knowledge. A comprehensive search yielded 177 papers that appeared between 2010 and 2023. Thence a strict screening method from predetermined inclusion criteria finally generated 39 records. These selected works were the basis of the study, presenting alternate views, obstacles, limitations and interesting ideas. By providing a systematic presentation of the material, it has made a key scholarly contribution. It focuses on the technical aspects of SSVEP-based BCI, EEG technologies and complex applications of deep learning technology in these areas. The study delivers more penetrating reporting on the latest deep learning pattern recognition techniques than its predecessors, together with progress in data acquisition and recording means suitable for SSVEP-based BCI devices. Especially in the realms of deep learning technology orchestration, pattern recognition techniques, and EEG data collection, it has effectively closed four important research gaps. To increase the accessibility of this critical material, the results of the study take the form of easy-to-read tables just generated. Applying deep learning techniques in SSVEP-based BCI applications, as the research shows, also has its downsides. The study concludes that a radical framework will be presented which, includes intelligent decision-making tools for evaluation and benchmarking. Rather than just finding a comparable or similar analogy, this framework is intended to help guide future research and pragmatic applications, and to determine which SSVEP-based BCI applications have succeeded at responsibility for what they set out with.","doi":"10.1007\/s00521-024-10143-z","cleaned_title":"a comprehensive review of deep learning power in steady-state visual evoked potentials","cleaned_abstract":"brain\u2013computer interfacing (bci) research, fueled by deep learning, integrates insights from diverse domains. a notable focus is on steady-state visual evoked potential (ssvep) in bci applications, requiring in-depth assessment through deep learning. eeg research frequently employs ssveps, which are regarded as normal brain responses to visual stimuli, particularly in investigations of visual perception and attention. this paper tries to give an in-depth analysis of the implications of deep learning for ssvep-adapted bci. a systematic search across four stable databases (web of science, pubmed, sciencedirect, and ieee) was developed to assemble a vast reservoir of relevant theoretical and scientific knowledge. a comprehensive search yielded 177 papers that appeared between 2010 and 2023. thence a strict screening method from predetermined inclusion criteria finally generated 39 records. these selected works were the basis of the study, presenting alternate views, obstacles, limitations and interesting ideas. by providing a systematic presentation of the material, it has made a key scholarly contribution. it focuses on the technical aspects of ssvep-based bci, eeg technologies and complex applications of deep learning technology in these areas. the study delivers more penetrating reporting on the latest deep learning pattern recognition techniques than its predecessors, together with progress in data acquisition and recording means suitable for ssvep-based bci devices. especially in the realms of deep learning technology orchestration, pattern recognition techniques, and eeg data collection, it has effectively closed four important research gaps. to increase the accessibility of this critical material, the results of the study take the form of easy-to-read tables just generated. applying deep learning techniques in ssvep-based bci applications, as the research shows, also has its downsides. the study concludes that a radical framework will be presented which, includes intelligent decision-making tools for evaluation and benchmarking. rather than just finding a comparable or similar analogy, this framework is intended to help guide future research and pragmatic applications, and to determine which ssvep-based bci applications have succeeded at responsibility for what they set out with.","key_phrases":["brain","computer","(bci) research","deep learning","insights","diverse domains","a notable focus","steady-state visual evoked potential","ssvep","bci applications","depth","deep learning","eeg research","ssveps","which","normal brain responses","visual stimuli","investigations","visual perception","attention","this paper","an in-depth analysis","the implications","deep learning","ssvep-adapted bci","a systematic search","four stable databases","web","science","sciencedirect","ieee","a vast reservoir","relevant theoretical and scientific knowledge","a comprehensive search","177 papers","that","a strict screening method","predetermined inclusion criteria","39 records","these selected works","the basis","the study","alternate views","obstacles","limitations","interesting ideas","a systematic presentation","the material","it","a key scholarly contribution","it","the technical aspects","ssvep-based bci","eeg technologies","complex applications","deep learning technology","these areas","the study","the latest deep learning pattern recognition techniques","its predecessors","progress","data acquisition","recording","ssvep-based bci devices","the realms","deep learning technology orchestration","pattern recognition techniques","eeg data collection","it","four important research gaps","the accessibility","this critical material","the results","the study","the form","read","deep learning techniques","ssvep-based bci applications","its downsides","the study","a radical framework","which","intelligent decision-making tools","evaluation","benchmarking","a comparable or similar analogy","this framework","future research and pragmatic applications","which","ssvep-based bci applications","responsibility","what","they","four","177","between 2010 and 2023","39","four"]},{"title":"Quantum deep learning-based anomaly detection for enhanced network security","authors":["Moe Hdaib","Sutharshan Rajasegarar","Lei Pan"],"abstract":"Identifying and mitigating aberrant activities within the network traffic is important to prevent adverse consequences caused by cyber security incidents, which have been increasing significantly in recent times. Existing research mainly focuses on classical machine learning and deep learning-based approaches for detecting such attacks. However, exploiting the power of quantum deep learning to process complex correlation of features for anomaly detection is not well explored. Hence, in this paper, we investigate quantum machine learning and quantum deep learning-based anomaly detection methodologies to accurately detect network attacks. In particular, we propose three novel quantum auto-encoder-based anomaly detection frameworks. Our primary aim is to create hybrid models that leverage the strengths of both quantum and deep learning methodologies for efficient anomaly recognition. The three frameworks are formed by integrating the quantum autoencoder with a quantum one-class support vector machine, a quantum random forest, and a quantum k-nearest neighbor approach. The anomaly detection capability of the frameworks is evaluated using benchmark datasets comprising computer and Internet of Things network flows. Our evaluation demonstrates that all three frameworks have a high potential to detect the network traffic anomalies accurately, while the framework that integrates the quantum autoencoder with the quantum k-nearest neighbor yields the highest accuracy. This demonstrates the promising potential for the development of quantum frameworks for anomaly detection, underscoring their relevance for future advancements in network security.","doi":"10.1007\/s42484-024-00163-2","cleaned_title":"quantum deep learning-based anomaly detection for enhanced network security","cleaned_abstract":"identifying and mitigating aberrant activities within the network traffic is important to prevent adverse consequences caused by cyber security incidents, which have been increasing significantly in recent times. existing research mainly focuses on classical machine learning and deep learning-based approaches for detecting such attacks. however, exploiting the power of quantum deep learning to process complex correlation of features for anomaly detection is not well explored. hence, in this paper, we investigate quantum machine learning and quantum deep learning-based anomaly detection methodologies to accurately detect network attacks. in particular, we propose three novel quantum auto-encoder-based anomaly detection frameworks. our primary aim is to create hybrid models that leverage the strengths of both quantum and deep learning methodologies for efficient anomaly recognition. the three frameworks are formed by integrating the quantum autoencoder with a quantum one-class support vector machine, a quantum random forest, and a quantum k-nearest neighbor approach. the anomaly detection capability of the frameworks is evaluated using benchmark datasets comprising computer and internet of things network flows. our evaluation demonstrates that all three frameworks have a high potential to detect the network traffic anomalies accurately, while the framework that integrates the quantum autoencoder with the quantum k-nearest neighbor yields the highest accuracy. this demonstrates the promising potential for the development of quantum frameworks for anomaly detection, underscoring their relevance for future advancements in network security.","key_phrases":["aberrant activities","the network traffic","adverse consequences","cyber security incidents","which","recent times","existing research","classical machine learning","deep learning-based approaches","such attacks","the power","quantum","complex correlation","features","anomaly detection","this paper","we","quantum machine learning","quantum","deep learning-based anomaly detection methodologies","network attacks","we","three novel quantum auto-encoder-based anomaly detection frameworks","our primary aim","hybrid models","that","the strengths","both quantum","deep learning methodologies","efficient anomaly recognition","the three frameworks","the quantum","autoencoder","a quantum one-class support vector machine","a quantum random forest","a quantum k-nearest neighbor approach","the anomaly detection capability","the frameworks","benchmark datasets","computer","internet","things network","our evaluation","all three frameworks","a high potential","the network traffic anomalies","the framework","that","the quantum","the quantum k-nearest neighbor","the highest accuracy","this","the promising potential","the development","quantum frameworks","anomaly detection","their relevance","future advancements","network security","anomaly detection methodologies","three","anomaly detection frameworks","anomaly recognition","three","one","quantum","the anomaly detection capability","three","quantum","quantum"]},{"title":"A pedagogical study on promoting students' deep learning through design-based learning","authors":["Chunmeng Weng","Congying Chen","Xianfeng Ai"],"abstract":"This paper illustrates the design-based learning (DBL) approach to promoting the deep learning of students and improving the quality of teaching in engineering design education. We performed three aspects of research with students in a typical educational activity. The first study investigated students' deep learning before and after the DBL approach, both in terms of deep learning status and deep learning ability. The second study examined the effectiveness of the DBL approach by comparative research of a control class (traditional teaching method) and an experimental class (DBL method). The third study examined students' evaluations of the DBL approach. It is approved that the DBL approach has distinctively stimulated the students' motivation to learn, making them more actively engaged in study. The students' higher-order thinking and higher-order capabilities are enhanced, such as critical thinking ability and problem-solving ability. At the same time, they are satisfied with the DBL approach. These findings suggest that the DBL approach is effective in promoting students' deep learning and improving the quality of teaching and learning.","doi":"10.1007\/s10798-022-09789-4","cleaned_title":"a pedagogical study on promoting students' deep learning through design-based learning","cleaned_abstract":"this paper illustrates the design-based learning (dbl) approach to promoting the deep learning of students and improving the quality of teaching in engineering design education. we performed three aspects of research with students in a typical educational activity. the first study investigated students' deep learning before and after the dbl approach, both in terms of deep learning status and deep learning ability. the second study examined the effectiveness of the dbl approach by comparative research of a control class (traditional teaching method) and an experimental class (dbl method). the third study examined students' evaluations of the dbl approach. it is approved that the dbl approach has distinctively stimulated the students' motivation to learn, making them more actively engaged in study. the students' higher-order thinking and higher-order capabilities are enhanced, such as critical thinking ability and problem-solving ability. at the same time, they are satisfied with the dbl approach. these findings suggest that the dbl approach is effective in promoting students' deep learning and improving the quality of teaching and learning.","key_phrases":["this paper","the design-based learning","(dbl) approach","the deep learning","students","the quality","teaching","engineering design education","we","three aspects","research","students","a typical educational activity","the first study","students' deep learning","the dbl approach","terms","deep learning status","deep learning ability","the second study","the effectiveness","the dbl approach","comparative research","a control class","traditional teaching method","an experimental class","dbl method","the third study","students' evaluations","the dbl approach","it","the dbl approach","the students' motivation","them","study","the students' higher-order thinking","higher-order capabilities","critical thinking ability","problem-solving ability","the same time","they","the dbl approach","these findings","the dbl approach","students' deep learning","the quality","teaching","three","first","second","third"]},{"title":"Deep Learning-Based Image and Video Inpainting: A Survey","authors":["Weize Quan","Jiaxi Chen","Yanli Liu","Dong-Ming Yan","Peter Wonka"],"abstract":"Image and video inpainting is a classic problem in computer vision and computer graphics, aiming to fill in the plausible and realistic content in the missing areas of images and videos. With the advance of deep learning, this problem has achieved significant progress recently. The goal of this paper is to comprehensively review the deep learning-based methods for image and video inpainting. Specifically, we sort existing methods into different categories from the perspective of their high-level inpainting pipeline, present different deep learning architectures, including CNN, VAE, GAN, diffusion models, etc., and summarize techniques for module design. We review the training objectives and the common benchmark datasets. We present evaluation metrics for low-level pixel and high-level perceptional similarity, conduct a performance evaluation, and discuss the strengths and weaknesses of representative inpainting methods. We also discuss related real-world applications. Finally, we discuss open challenges and suggest potential future research directions.","doi":"10.1007\/s11263-023-01977-6","cleaned_title":"deep learning-based image and video inpainting: a survey","cleaned_abstract":"image and video inpainting is a classic problem in computer vision and computer graphics, aiming to fill in the plausible and realistic content in the missing areas of images and videos. with the advance of deep learning, this problem has achieved significant progress recently. the goal of this paper is to comprehensively review the deep learning-based methods for image and video inpainting. specifically, we sort existing methods into different categories from the perspective of their high-level inpainting pipeline, present different deep learning architectures, including cnn, vae, gan, diffusion models, etc., and summarize techniques for module design. we review the training objectives and the common benchmark datasets. we present evaluation metrics for low-level pixel and high-level perceptional similarity, conduct a performance evaluation, and discuss the strengths and weaknesses of representative inpainting methods. we also discuss related real-world applications. finally, we discuss open challenges and suggest potential future research directions.","key_phrases":["image","a classic problem","computer vision","computer graphics","the plausible and realistic content","the missing areas","images","videos","the advance","deep learning","this problem","significant progress","the goal","this paper","the deep learning-based methods","image","we","existing methods","different categories","the perspective","their high-level inpainting pipeline","different deep learning architectures","cnn","vae","gan","diffusion models","techniques","module design","we","the training objectives","the common benchmark datasets","we","evaluation metrics","low-level pixel","high-level perceptional similarity","a performance evaluation","the strengths","weaknesses","representative inpainting methods","we","related real-world applications","we","open challenges","potential future research directions","cnn","vae, gan, diffusion models"]},{"title":"Brain tumour detection using machine and deep learning: a systematic review","authors":["Novsheena Rasool","Javaid Iqbal Bhat"],"abstract":"Brain tumors rank as the 1oth leading cause of mortality worldwide, accounting for 85% to 95% of all primary nervous system malignancies. The prevalence of this life-threatening disease is steadily increasing worldwide, highlighting the urgent need for an early and precise diagnosis. Timely identification is critical for initiating effective treatment and improving patient survival chances. Delayed diagnosis significantly elevates the risk of mortality. However, the heterogeneous nature of tumor cells poses challenges for radiologists, making manual diagnosis from magnetic resonance imaging (MRI) images time-consuming and complex. Machine learning (ML) and deep learning (DL) have become useful tools in medical image analysis. These techniques facilitate the automated extraction of intricate patterns and features from MRI images, thereby facilitating a more accurate and efficient tumor diagnosis. Furthermore, these algorithms have demonstrated the capability to handle the intricacy and variability of brain tumor characteristics, thereby improving the diagnostic process. A range of deep learning-based algorithms have been utilized to detect brain tumors, yielding impressive results. The purpose of this paper is to provide an exhaustive examination of the latest techniques used for diagnosing brain tumors from MRI imaging, utilizing machine and deep learning technologies. Moreover, it seeks to outline potential avenues for future exploration within this field. The profound insights gleaned from this comprehensive review are poised to offer invaluable guidance and support to both researchers and medical professionals in the healthcare industry.","doi":"10.1007\/s11042-024-19333-2","cleaned_title":"brain tumour detection using machine and deep learning: a systematic review","cleaned_abstract":"brain tumors rank as the 1oth leading cause of mortality worldwide, accounting for 85% to 95% of all primary nervous system malignancies. the prevalence of this life-threatening disease is steadily increasing worldwide, highlighting the urgent need for an early and precise diagnosis. timely identification is critical for initiating effective treatment and improving patient survival chances. delayed diagnosis significantly elevates the risk of mortality. however, the heterogeneous nature of tumor cells poses challenges for radiologists, making manual diagnosis from magnetic resonance imaging (mri) images time-consuming and complex. machine learning (ml) and deep learning (dl) have become useful tools in medical image analysis. these techniques facilitate the automated extraction of intricate patterns and features from mri images, thereby facilitating a more accurate and efficient tumor diagnosis. furthermore, these algorithms have demonstrated the capability to handle the intricacy and variability of brain tumor characteristics, thereby improving the diagnostic process. a range of deep learning-based algorithms have been utilized to detect brain tumors, yielding impressive results. the purpose of this paper is to provide an exhaustive examination of the latest techniques used for diagnosing brain tumors from mri imaging, utilizing machine and deep learning technologies. moreover, it seeks to outline potential avenues for future exploration within this field. the profound insights gleaned from this comprehensive review are poised to offer invaluable guidance and support to both researchers and medical professionals in the healthcare industry.","key_phrases":["brain tumors","the 1oth leading cause","mortality","85% to 95%","all primary nervous system malignancies","the prevalence","this life-threatening disease","the urgent need","an early and precise diagnosis","timely identification","effective treatment","patient survival chances","delayed diagnosis","the risk","mortality","the heterogeneous nature","tumor cells","challenges","radiologists","manual diagnosis","magnetic resonance imaging","(mri","machine learning","ml","deep learning","dl","useful tools","medical image analysis","these techniques","the automated extraction","intricate patterns","features","mri images","a more accurate and efficient tumor diagnosis","these algorithms","the capability","the intricacy","variability","brain tumor characteristics","the diagnostic process","a range","deep learning-based algorithms","brain tumors","impressive results","the purpose","this paper","an exhaustive examination","the latest techniques","brain tumors","mri imaging","machine","deep learning technologies","it","potential avenues","future exploration","this field","the profound insights","this comprehensive review","invaluable guidance","support","both researchers","medical professionals","the healthcare industry","85% to 95%"]},{"title":"Deep learning-based streamflow prediction for western Himalayan river basins","authors":["Tabasum Majeed","Riyaz Ahmad Mir","Rayees Ahmad Dar","Mohd Anul Haq","Shabana Nargis Rasool","Assif Assad"],"abstract":"Accurate streamflow (Qflow) forecasting plays a pivotal role in water resource monitoring and management, presenting a complex challenge for water managers and engineers. Effective streamflow prediction enables the optimized operation of water resource systems in alignment with technological, financial, ethical, and political objectives. Traditional data-driven models like the Autoregressive model and Autoregressive Moving Average model are widely used for water resource management. However, these models show limitations in handling intricate nonlinear hydrological phenomena. To address these limitations, Deep Learning models emerge as promising alternatives, given their inherent ability to handle nonlinearity. Nonlinearity in time series modeling is formidable due to factors like long-term trends, seasonal variations, cyclical oscillations, and external disturbances. This study proposes a deep neural network architecture based on Neural Basis Expansion Analysis for Time Series (N-BEATS) to predict the daily Qflow of western Himalayan river basins. The study employs datasets collected from the Rampur station of the Satluj basin and the Pandoh and Manali stations of the Beas basin. The experimental results unequivocally demonstrate the superiority of the proposed deep neural network model over benchmarked conventional deep learning models such as Long Short-Term Memory, Feedforward Neural Network, Gated Recurrent Unit, and Recurrent Neural Network. The proposed deep neural network model achieves remarkable accuracy, exhibiting a root mean square error below 0.05 m3\/s when comparing actual and predicted Qflow values across all datasets. Consequently, the proposed deep neural network model based on N-BEATS emerges as an efficient and invaluable solution for precise Qflow prediction, empowering efficient water resource management and control. The results suggest that\u00a0the proposed model can serve for streamflow prediction and water management in Himalayan river basins.","doi":"10.1007\/s13198-024-02403-x","cleaned_title":"deep learning-based streamflow prediction for western himalayan river basins","cleaned_abstract":"accurate streamflow (qflow) forecasting plays a pivotal role in water resource monitoring and management, presenting a complex challenge for water managers and engineers. effective streamflow prediction enables the optimized operation of water resource systems in alignment with technological, financial, ethical, and political objectives. traditional data-driven models like the autoregressive model and autoregressive moving average model are widely used for water resource management. however, these models show limitations in handling intricate nonlinear hydrological phenomena. to address these limitations, deep learning models emerge as promising alternatives, given their inherent ability to handle nonlinearity. nonlinearity in time series modeling is formidable due to factors like long-term trends, seasonal variations, cyclical oscillations, and external disturbances. this study proposes a deep neural network architecture based on neural basis expansion analysis for time series (n-beats) to predict the daily qflow of western himalayan river basins. the study employs datasets collected from the rampur station of the satluj basin and the pandoh and manali stations of the beas basin. the experimental results unequivocally demonstrate the superiority of the proposed deep neural network model over benchmarked conventional deep learning models such as long short-term memory, feedforward neural network, gated recurrent unit, and recurrent neural network. the proposed deep neural network model achieves remarkable accuracy, exhibiting a root mean square error below 0.05 m3\/s when comparing actual and predicted qflow values across all datasets. consequently, the proposed deep neural network model based on n-beats emerges as an efficient and invaluable solution for precise qflow prediction, empowering efficient water resource management and control. the results suggest that the proposed model can serve for streamflow prediction and water management in himalayan river basins.","key_phrases":["accurate streamflow (qflow) forecasting","a pivotal role","water resource monitoring","management","a complex challenge","water managers","engineers","effective streamflow prediction","the optimized operation","water resource systems","alignment","technological, financial, ethical, and political objectives","traditional data-driven models","the autoregressive model","autoregressive moving average model","water resource management","these models","limitations","intricate nonlinear hydrological phenomena","these limitations","deep learning models","promising alternatives","their inherent ability","nonlinearity","nonlinearity","time series","modeling","factors","long-term trends","seasonal variations","cyclical oscillations","external disturbances","this study","a deep neural network architecture","neural basis expansion analysis","time series","-beats","the daily qflow","western himalayan river basins","the study","datasets","the rampur station","the satluj basin","the pandoh and manali stations","the beas basin","the experimental results","the superiority","the proposed deep neural network model","benchmarked conventional deep learning models","long short-term memory","feedforward neural network","gated recurrent unit","neural network","the proposed deep neural network model","remarkable accuracy","a root mean square error","0.05 m3","qflow values","all datasets","the proposed deep neural network model","n-beats","an efficient and invaluable solution","precise qflow prediction","efficient water resource management","control","the results","the proposed model","streamflow prediction","water management","himalayan river basins","daily","rampur","0.05","m3\/s"]},{"title":"Supervised deep learning for content-aware image retargeting with Fourier Convolutions","authors":["MohammadHossein Givkashi","MohammadReza Naderi","Nader Karimi","Shahram Shirani","Shadrokh Samavi"],"abstract":"Image retargeting aims to alter the size of the image with attention to the contents. One of the main obstacles to training deep learning models for image retargeting is the need for a vast labeled dataset. Labeled datasets are unavailable for training deep learning models in the image retargeting tasks. As a result, we present a new supervised approach for training deep learning models. We use the original images as ground truth and create inputs for the model by resizing and cropping the original images. A second challenge is generating different image sizes in inference time. However, normal convolutional neural networks cannot generate images of different sizes than the input image. To address this issue, we introduced a new method for supervised learning. In our approach, a mask is generated to show the desired size and location of the object. Then the mask and the input image are fed to the network. Comparing image retargeting methods and our proposed method demonstrates the model\u2019s ability to produce high-quality retargeted images. Afterward, we compute the image quality assessment score for each output image based on different techniques and illustrate the effectiveness of our approach.","doi":"10.1007\/s11042-024-18876-8","cleaned_title":"supervised deep learning for content-aware image retargeting with fourier convolutions","cleaned_abstract":"image retargeting aims to alter the size of the image with attention to the contents. one of the main obstacles to training deep learning models for image retargeting is the need for a vast labeled dataset. labeled datasets are unavailable for training deep learning models in the image retargeting tasks. as a result, we present a new supervised approach for training deep learning models. we use the original images as ground truth and create inputs for the model by resizing and cropping the original images. a second challenge is generating different image sizes in inference time. however, normal convolutional neural networks cannot generate images of different sizes than the input image. to address this issue, we introduced a new method for supervised learning. in our approach, a mask is generated to show the desired size and location of the object. then the mask and the input image are fed to the network. comparing image retargeting methods and our proposed method demonstrates the model\u2019s ability to produce high-quality retargeted images. afterward, we compute the image quality assessment score for each output image based on different techniques and illustrate the effectiveness of our approach.","key_phrases":["the size","the image","attention","the contents","the main obstacles","deep learning models","the need","a vast labeled dataset","labeled datasets","deep learning models","the image retargeting tasks","a result","we","a new supervised approach","deep learning models","we","the original images","ground truth","inputs","the model","the original images","a second challenge","different image sizes","inference time","normal convolutional neural networks","images","different sizes","the input image","this issue","we","a new method","supervised learning","our approach","a mask","the desired size","location","the object","the mask","the input image","the network","image retargeting methods","our proposed method","the model\u2019s ability","high-quality retargeted images","we","the image quality assessment score","each output image","different techniques","the effectiveness","our approach","one","second","fed"]},{"title":"Identifying keystone species in microbial communities using deep learning","authors":["Xu-Wen Wang","Zheng Sun","Huijue Jia","Sebastian Michel-Mata","Marco Tulio Angulo","Lei Dai","Xuesong He","Scott T. Weiss","Yang-Yu Liu"],"abstract":"Previous studies suggested that microbial communities can harbour keystone species whose removal can cause a dramatic shift in microbiome structure and functioning. Yet, an efficient method to systematically identify keystone species in microbial communities is still lacking. Here we propose a data-driven keystone species identification (DKI) framework based on deep learning to resolve this challenge. Our key idea is to implicitly learn the assembly rules of microbial communities from a particular habitat by training a deep-learning model using microbiome samples collected from this habitat. The well-trained deep-learning model enables us to quantify the community-specific keystoneness of each species in any microbiome sample from this habitat by conducting a thought experiment on species removal. We systematically validated this DKI framework using synthetic data and applied DKI to analyse real data. We found that those taxa with high median keystoneness across different communities display strong community specificity. The presented DKI framework demonstrates the power of machine learning in tackling a fundamental problem in community ecology, paving the way for the data-driven management of complex microbial communities.","doi":"10.1038\/s41559-023-02250-2","cleaned_title":"identifying keystone species in microbial communities using deep learning","cleaned_abstract":"previous studies suggested that microbial communities can harbour keystone species whose removal can cause a dramatic shift in microbiome structure and functioning. yet, an efficient method to systematically identify keystone species in microbial communities is still lacking. here we propose a data-driven keystone species identification (dki) framework based on deep learning to resolve this challenge. our key idea is to implicitly learn the assembly rules of microbial communities from a particular habitat by training a deep-learning model using microbiome samples collected from this habitat. the well-trained deep-learning model enables us to quantify the community-specific keystoneness of each species in any microbiome sample from this habitat by conducting a thought experiment on species removal. we systematically validated this dki framework using synthetic data and applied dki to analyse real data. we found that those taxa with high median keystoneness across different communities display strong community specificity. the presented dki framework demonstrates the power of machine learning in tackling a fundamental problem in community ecology, paving the way for the data-driven management of complex microbial communities.","key_phrases":["previous studies","microbial communities","keystone species","whose removal","a dramatic shift","microbiome structure","functioning","an efficient method","keystone species","microbial communities","we","a data-driven keystone species identification","(dki) framework","deep learning","this challenge","our key idea","the assembly rules","microbial communities","a particular habitat","a deep-learning model","microbiome samples","this habitat","the well-trained deep-learning model","us","the community-specific keystoneness","each species","any microbiome sample","this habitat","a thought experiment","species removal","we","this dki framework","synthetic data","applied dki","analyse real data","we","those taxa","high median keystoneness","different communities","strong community specificity","the presented dki framework","the power","a fundamental problem","community ecology","the way","the data-driven management","complex microbial communities"]},{"title":"Leveraging self-paced learning and deep sparse embedding for image clustering","authors":["Yanming Liu","Jinglei Liu"],"abstract":"Deep clustering outperforms traditional methods by incorporating feature learning. However, some existing deep clustering methods overlook the suitability of the learned features for clustering, leading to insufficient feedback received by the clustering model and hampering the accuracy improvement. To tackle these issues, we propose a joint self-paced learning and deep sparse embedding for image clustering. Our method consists of two stages: pretraining and finetuning. In the pretraining stage, the autoencoder learns basic features and constructs the feature space. In the finetuning stage, method performs two tasks: feature learning and cluster assignment. Specifically, we finetune the encoder with both original and augmented data to preserve the local structure in the feature space. Self-paced learning guarantees that the most confident features are used for each iteration and mitigates the influence of boundary samples. Furthermore, sparse embedding ensures that the model encodes only key features in feature learning tasks, thereby avoiding incorrect calculations resulting from redundant features. Finally, we jointly optimize these two tasks to complete the feature learning for clustering. Extensive experiments on various datasets demonstrate that our approach outperforms existing solutions.","doi":"10.1007\/s00521-023-09335-w","cleaned_title":"leveraging self-paced learning and deep sparse embedding for image clustering","cleaned_abstract":"deep clustering outperforms traditional methods by incorporating feature learning. however, some existing deep clustering methods overlook the suitability of the learned features for clustering, leading to insufficient feedback received by the clustering model and hampering the accuracy improvement. to tackle these issues, we propose a joint self-paced learning and deep sparse embedding for image clustering. our method consists of two stages: pretraining and finetuning. in the pretraining stage, the autoencoder learns basic features and constructs the feature space. in the finetuning stage, method performs two tasks: feature learning and cluster assignment. specifically, we finetune the encoder with both original and augmented data to preserve the local structure in the feature space. self-paced learning guarantees that the most confident features are used for each iteration and mitigates the influence of boundary samples. furthermore, sparse embedding ensures that the model encodes only key features in feature learning tasks, thereby avoiding incorrect calculations resulting from redundant features. finally, we jointly optimize these two tasks to complete the feature learning for clustering. extensive experiments on various datasets demonstrate that our approach outperforms existing solutions.","key_phrases":["deep clustering","traditional methods","feature learning","some existing deep clustering methods","the suitability","the learned features","clustering","insufficient feedback","the clustering model","the accuracy improvement","these issues","we","a joint self-paced learning","deep sparse","image clustering","our method","two stages","the pretraining stage","the autoencoder","basic features","the feature space","the finetuning stage","method","two tasks","feature learning","cluster assignment","we","the encoder","both original and augmented data","the local structure","the feature space","self-paced learning guarantees","the most confident features","each iteration","the influence","boundary samples","sparse","the model","only key features","feature learning tasks","incorrect calculations","redundant features","we","these two tasks","the feature","extensive experiments","various datasets","our approach","existing solutions","two","two","two"]},{"title":"Automated abnormalities detection in mammography using deep learning","authors":["Ghada M. El-Banby","Nourhan S. Salem","Eman A. Tafweek","Essam N. Abd El-Azziz"],"abstract":"Breast cancer is the second most prevalent cause of cancer death and the most common malignancy among women, posing a life-threatening risk. Treatment for breast cancer can be highly effective, with a survival chance of 90% or higher, especially when the disease is detected early. This paper introduces a groundbreaking deep U-Net framework for mammography breast cancer images to perform automatic detection of abnormalities. The objective is to provide segmented images that show areas of tumors more accurately than other deep learning techniques. The proposed framework consists of three steps. The first step is image preprocessing using the Li algorithm to minimize the cross-entropy between the foreground and the background, contrast enhancement using contrast-limited adaptive histogram equalization (CLAHE), normalization, and median filtering. The second step involves data augmentation to mitigate overfitting and underfitting, and the final step is implementing a convolutional encoder-decoder network-based U-Net architecture, characterized by high precision in medical image analysis. The framework has been tested on two comprehensive public datasets, namely INbreast and CBIS-DDSM. Several metrics have been adopted for quantitative performance assessment, including the Dice score, sensitivity, Hausdorff distance, Jaccard coefficient, precision, and F1 score. Quantitative results on the INbreast dataset show an average Dice score of 85.61% and a sensitivity of 81.26%. On the CBIS-DDSM dataset, the average Dice score is 87.98%, and the sensitivity reaches 90.58%. The experimental results ensure earlier and more accurate abnormality detection. Furthermore, the success of the proposed deep learning framework in mammography shows promise for broader applications in medical imaging, potentially revolutionizing various radiological practices.","doi":"10.1007\/s40747-024-01532-x","cleaned_title":"automated abnormalities detection in mammography using deep learning","cleaned_abstract":"breast cancer is the second most prevalent cause of cancer death and the most common malignancy among women, posing a life-threatening risk. treatment for breast cancer can be highly effective, with a survival chance of 90% or higher, especially when the disease is detected early. this paper introduces a groundbreaking deep u-net framework for mammography breast cancer images to perform automatic detection of abnormalities. the objective is to provide segmented images that show areas of tumors more accurately than other deep learning techniques. the proposed framework consists of three steps. the first step is image preprocessing using the li algorithm to minimize the cross-entropy between the foreground and the background, contrast enhancement using contrast-limited adaptive histogram equalization (clahe), normalization, and median filtering. the second step involves data augmentation to mitigate overfitting and underfitting, and the final step is implementing a convolutional encoder-decoder network-based u-net architecture, characterized by high precision in medical image analysis. the framework has been tested on two comprehensive public datasets, namely inbreast and cbis-ddsm. several metrics have been adopted for quantitative performance assessment, including the dice score, sensitivity, hausdorff distance, jaccard coefficient, precision, and f1 score. quantitative results on the inbreast dataset show an average dice score of 85.61% and a sensitivity of 81.26%. on the cbis-ddsm dataset, the average dice score is 87.98%, and the sensitivity reaches 90.58%. the experimental results ensure earlier and more accurate abnormality detection. furthermore, the success of the proposed deep learning framework in mammography shows promise for broader applications in medical imaging, potentially revolutionizing various radiological practices.","key_phrases":["breast cancer","the second most prevalent cause","cancer death","the most common malignancy","women","a life-threatening risk","treatment","breast cancer","a survival chance","90%","the disease","this paper","a groundbreaking deep u-net framework","mammography breast cancer images","automatic detection","abnormalities","the objective","segmented images","that","areas","tumors","other deep learning techniques","the proposed framework","three steps","the first step","image","the li algorithm","the cross","-","entropy","the foreground","the background","contrast enhancement","contrast-limited adaptive histogram equalization","clahe","normalization","median filtering","the second step","data augmentation","overfitting","the final step","a convolutional encoder-decoder network-based u-net architecture","high precision","medical image analysis","the framework","two comprehensive public datasets","namely inbreast and cbis-ddsm","several metrics","quantitative performance assessment","the dice score","sensitivity","hausdorff distance","precision","f1 score","quantitative results","the inbreast dataset","an average dice score","85.61%","a sensitivity","81.26%","the cbis-ddsm dataset","the average dice score","87.98%","the sensitivity","90.58%","the experimental results","earlier and more accurate abnormality detection","the success","the proposed deep learning framework","mammography","promise","broader applications","medical imaging","various radiological practices","second","90%","three","first","li","second","two","85.61%","81.26%","87.98%","90.58%"]},{"title":"Deep Learning Based Entropy Controlled Optimization for the Detection of Covid-19","authors":["Jiong Chen","Abdullah Alshammari","Mohammed Alonazi","Aisha M. Alqahtani","Sara A. Althubiti","Romi Fadillah Rahmat"],"abstract":"Emerging technological advancements open the door for employing deep learning-based methods in practically all spheres of human endeavor. Because of their accuracy, deep learning algorithms can be used in healthcare to categorize and identify different illnesses. The recent coronavirus (COVID-19) outbreak has significantly strained the global medical system. By using medical imaging and PCR testing, COVID-19 can be diagnosed. Since COVID-19 is highly transmissible, it is generally considered secure to analyze it with a chest X-ray. To distinguish COVID-19 infections from additional infections that are not COVID-19 infections, a deep learning-based entropy-controlled whale optimization (EWOA) with Transfer Learning is suggested in this paper. The created system comprises three stages: a preliminary processing phase to remove noise effects and resize the image, then a deep learning architecture using a pre-trained model to extract features from the pre-processed image. After extracting the features, optimization is carried out. EWOA is utilized to combine and optimize the optimum features. A softmax layer is used to reach the final categorization. Various activation functions, thresholds, and optimizers are used to assess the systems. Numerous metrics for performance are utilized to measure the performance of the offered methodologies for assessment. Through an accuracy of 97.95%, the suggested technique accurately categorizes four classes, including COVID-19, viral pneumonia, chest infection, and routine. Compared to current methodologies found in the literature, the proposed technique exhibits advantages regarding accuracy.","doi":"10.1007\/s10723-024-09766-2","cleaned_title":"deep learning based entropy controlled optimization for the detection of covid-19","cleaned_abstract":"emerging technological advancements open the door for employing deep learning-based methods in practically all spheres of human endeavor. because of their accuracy, deep learning algorithms can be used in healthcare to categorize and identify different illnesses. the recent coronavirus (covid-19) outbreak has significantly strained the global medical system. by using medical imaging and pcr testing, covid-19 can be diagnosed. since covid-19 is highly transmissible, it is generally considered secure to analyze it with a chest x-ray. to distinguish covid-19 infections from additional infections that are not covid-19 infections, a deep learning-based entropy-controlled whale optimization (ewoa) with transfer learning is suggested in this paper. the created system comprises three stages: a preliminary processing phase to remove noise effects and resize the image, then a deep learning architecture using a pre-trained model to extract features from the pre-processed image. after extracting the features, optimization is carried out. ewoa is utilized to combine and optimize the optimum features. a softmax layer is used to reach the final categorization. various activation functions, thresholds, and optimizers are used to assess the systems. numerous metrics for performance are utilized to measure the performance of the offered methodologies for assessment. through an accuracy of 97.95%, the suggested technique accurately categorizes four classes, including covid-19, viral pneumonia, chest infection, and routine. compared to current methodologies found in the literature, the proposed technique exhibits advantages regarding accuracy.","key_phrases":["emerging technological advancements","the door","deep learning-based methods","practically all spheres","human endeavor","their accuracy","deep learning algorithms","healthcare","different illnesses","the recent coronavirus (covid-19) outbreak","the global medical system","medical imaging","pcr testing","covid-19","covid-19","it","it","-","covid-19 infections","additional infections","that","covid-19 infections","a deep learning-based entropy-controlled whale optimization","(ewoa","transfer learning","this paper","the created system","three stages","a preliminary processing phase","noise effects","the image","then a deep learning architecture","a pre-trained model","features","the pre-processed image","the features","optimization","ewoa","the optimum features","a softmax layer","the final categorization","various activation functions","thresholds","optimizers","the systems","numerous metrics","performance","the performance","the offered methodologies","assessment","an accuracy","97.95%","the suggested technique","four classes","covid-19","viral pneumonia","chest infection","current methodologies","the literature","accuracy","covid-19","covid-19","covid-19","covid-19","covid-19","three","97.95%","four","covid-19"]},{"title":"Data set terminology of deep learning in medicine: a historical review and recommendation","authors":["Shannon L. Walston","Hiroshi Seki","Hirotaka Takita","Yasuhito Mitsuyama","Shingo Sato","Akifumi Hagiwara","Rintaro Ito","Shouhei Hanaoka","Yukio Miki","Daiju Ueda"],"abstract":"Medicine and deep learning-based artificial intelligence (AI) engineering represent two distinct fields each with decades of published history. The current rapid convergence of deep learning and medicine has led to significant advancements, yet it has also introduced ambiguity regarding data set terms common to both fields, potentially leading to miscommunication and methodological discrepancies. This narrative review aims to give historical context for these terms, accentuate the importance of clarity when these terms are used in medical deep learning contexts, and offer solutions to mitigate misunderstandings by readers from either field. Through an examination of historical documents, including articles, writing guidelines, and textbooks, this review traces the divergent evolution of terms for data sets and their impact. Initially, the discordant interpretations of the word \u2018validation\u2019 in medical and AI contexts are explored. We then show that in the medical field as well, terms traditionally used in the deep learning domain are becoming more common, with the data for creating models referred to as the \u2018training set\u2019, the data for tuning of parameters referred to as the \u2018validation (or tuning) set\u2019, and the data for the evaluation of models as the \u2018test set\u2019. Additionally, the test sets used for model evaluation are classified into internal (random splitting, cross-validation, and leave-one-out) sets and external (temporal and geographic) sets. This review then identifies often misunderstood terms and proposes pragmatic solutions to mitigate terminological confusion in the field of deep learning in medicine. We support the accurate and standardized description of these data sets and the explicit definition of data set splitting terminologies in each publication. These are crucial methods for demonstrating the robustness and generalizability of deep learning applications in medicine. This review aspires to enhance the precision of communication, thereby fostering more effective and transparent research methodologies in this interdisciplinary field.","doi":"10.1007\/s11604-024-01608-1","cleaned_title":"data set terminology of deep learning in medicine: a historical review and recommendation","cleaned_abstract":"medicine and deep learning-based artificial intelligence (ai) engineering represent two distinct fields each with decades of published history. the current rapid convergence of deep learning and medicine has led to significant advancements, yet it has also introduced ambiguity regarding data set terms common to both fields, potentially leading to miscommunication and methodological discrepancies. this narrative review aims to give historical context for these terms, accentuate the importance of clarity when these terms are used in medical deep learning contexts, and offer solutions to mitigate misunderstandings by readers from either field. through an examination of historical documents, including articles, writing guidelines, and textbooks, this review traces the divergent evolution of terms for data sets and their impact. initially, the discordant interpretations of the word \u2018validation\u2019 in medical and ai contexts are explored. we then show that in the medical field as well, terms traditionally used in the deep learning domain are becoming more common, with the data for creating models referred to as the \u2018training set\u2019, the data for tuning of parameters referred to as the \u2018validation (or tuning) set\u2019, and the data for the evaluation of models as the \u2018test set\u2019. additionally, the test sets used for model evaluation are classified into internal (random splitting, cross-validation, and leave-one-out) sets and external (temporal and geographic) sets. this review then identifies often misunderstood terms and proposes pragmatic solutions to mitigate terminological confusion in the field of deep learning in medicine. we support the accurate and standardized description of these data sets and the explicit definition of data set splitting terminologies in each publication. these are crucial methods for demonstrating the robustness and generalizability of deep learning applications in medicine. this review aspires to enhance the precision of communication, thereby fostering more effective and transparent research methodologies in this interdisciplinary field.","key_phrases":["medicine","deep learning-based artificial intelligence","(ai) engineering","two distinct fields","each","decades","published history","the current rapid convergence","deep learning","medicine","significant advancements","it","ambiguity","data","terms","both fields","miscommunication","methodological discrepancies","this narrative review","historical context","these terms","the importance","clarity","these terms","medical deep learning contexts","solutions","misunderstandings","readers","either field","an examination","historical documents","articles","guidelines","textbooks","this review","the divergent evolution","terms","data sets","their impact","the discordant interpretations","the word","validation","ai contexts","we","the medical field","terms","the deep learning domain","the data","models","the \u2018training set","the data","tuning","parameters","the \u2018validation (or tuning) set","the evaluation","models","the test sets","model evaluation","internal (random splitting","validation","sets","external (temporal and geographic) sets","this review","terms","pragmatic solutions","terminological confusion","the field","deep learning","medicine","we","the accurate and standardized description","these data sets","the explicit definition","data","splitting terminologies","each publication","these","crucial methods","the robustness","generalizability","deep learning applications","medicine","this review","the precision","communication","more effective and transparent research methodologies","this interdisciplinary field","two","each with decades","one"]},{"title":"A Study on Machine Learning and Deep Learning Techniques for Identifying Malicious Web Content","authors":["Sarita Mohanty","Asha Ambhakar"],"abstract":"The rapid proliferation of internet usage has led to an exponential increase in cyber threats, particularly malicious websites that can compromise user data and system integrity. Traditional methods of web security are increasingly becoming obsolete, necessitating more dynamic and adaptive approaches. This research paper presents a comprehensive comparative study of Machine Learning (ML) and Deep Learning (DL) techniques for the detection of malicious websites. Utilizing a dataset of over 420,000 web URLs, categorized into various features such as domain, subdomain, and domain suffix, the study aims to evaluate the effectiveness, precision, and computational efficiency of multiple algorithms. Two Convolutional Neural Network (CNN) models were developed and compared against traditional ML algorithms including Decision Trees, Random Forests, AdaBoost, K-Nearest Neighbors (KNN), Stochastic Gradient Descent (SGD), Extra Trees, and Gaussian Naive Bayes. The models were rigorously evaluated based on metrics such as accuracy, precision, recall, and F1-score. Preliminary results indicate that CNN models outperform traditional ML algorithms, achieving an accuracy rate of up to 98%, thereby highlighting the potential of DL in cybersecurity applications. Moreover, the study addresses the challenges posed by high cardinality and class imbalance in the dataset. Various data preprocessing techniques were employed to mitigate these issues, including feature engineering and oversampling of minority classes. The research contributes to the field by providing a detailed analysis of each algorithm\u2019s strengths and weaknesses, thereby offering valuable insights into the adaptability and scalability of ML and DL techniques in malicious web detection.","doi":"10.1007\/s42979-024-03099-3","cleaned_title":"a study on machine learning and deep learning techniques for identifying malicious web content","cleaned_abstract":"the rapid proliferation of internet usage has led to an exponential increase in cyber threats, particularly malicious websites that can compromise user data and system integrity. traditional methods of web security are increasingly becoming obsolete, necessitating more dynamic and adaptive approaches. this research paper presents a comprehensive comparative study of machine learning (ml) and deep learning (dl) techniques for the detection of malicious websites. utilizing a dataset of over 420,000 web urls, categorized into various features such as domain, subdomain, and domain suffix, the study aims to evaluate the effectiveness, precision, and computational efficiency of multiple algorithms. two convolutional neural network (cnn) models were developed and compared against traditional ml algorithms including decision trees, random forests, adaboost, k-nearest neighbors (knn), stochastic gradient descent (sgd), extra trees, and gaussian naive bayes. the models were rigorously evaluated based on metrics such as accuracy, precision, recall, and f1-score. preliminary results indicate that cnn models outperform traditional ml algorithms, achieving an accuracy rate of up to 98%, thereby highlighting the potential of dl in cybersecurity applications. moreover, the study addresses the challenges posed by high cardinality and class imbalance in the dataset. various data preprocessing techniques were employed to mitigate these issues, including feature engineering and oversampling of minority classes. the research contributes to the field by providing a detailed analysis of each algorithm\u2019s strengths and weaknesses, thereby offering valuable insights into the adaptability and scalability of ml and dl techniques in malicious web detection.","key_phrases":["the rapid proliferation","internet usage","an exponential increase","cyber threats","particularly malicious websites","that","user data and system integrity","traditional methods","web security","more dynamic and adaptive approaches","this research paper","a comprehensive comparative study","machine learning","ml","deep learning","(dl) techniques","the detection","malicious websites","a dataset","over 420,000 web urls","various features","domain","subdomain","domain suffix","the study","the effectiveness","precision","computational efficiency","multiple algorithms","two convolutional neural network (cnn) models","traditional ml algorithms","decision trees","random forests","adaboost, k-nearest neighbors",", stochastic gradient descent","sgd","extra trees","gaussian naive bayes","the models","metrics","accuracy","precision","recall","f1-score","preliminary results","cnn models","an accuracy rate","up to 98%","the potential","dl","cybersecurity applications","the study","the challenges","high cardinality","class imbalance","the dataset","various data","techniques","these issues","feature engineering","oversampling","minority classes","the research","the field","a detailed analysis","each algorithm\u2019s strengths","weaknesses","valuable insights","the adaptability","scalability","ml","dl techniques","malicious web detection","over 420,000","suffix","two","cnn","bayes","cnn","up to 98%"]},{"title":"Federal learning-based a dual-branch deep learning model for colon polyp segmentation","authors":["Xuguang Cao","Kefeng Fan","Huilin Ma"],"abstract":"The incidence of colon cancer occupies the top three places in gastrointestinal tumors, and colon polyps are an important causative factor in the development of colon cancer. Early screening for colon polyps and colon polypectomy can reduce the chances of colon cancer. The current means of colon polyp examination is through colonoscopy, taking images of the gastrointestinal tract, and then manually marking them manually, which is time-consuming and labor-intensive for doctors. Therefore, relying on advanced deep learning technology to automatically identify colon polyps in the gastrointestinal tract of the patient and segmenting the polyps is an important direction of research nowadays. Due to the privacy of medical data and the non-interoperability of disease information, this paper proposes a dual-branch colon polyp segmentation network based on federated learning, which makes it possible to achieve a better training effect under the guarantee of data independence, and secondly, the dual-branch colon polyp segmentation network proposed in this paper adopts the two different structures of convolutional neural network (CNN) and Transformer to form a dual-branch structure, and through layer-by-layer fusion embedding, the advantages between different structures are realized. In this paper, we also propose the Aggregated Attention Module (AAM) to preserve the high-dimensional semantic information and to complement the missing information in the lower layers. Ultimately our approach achieves state of the art in Kvasir-SEG and CVC-ClinicDB datasets.","doi":"10.1007\/s11042-024-19197-6","cleaned_title":"federal learning-based a dual-branch deep learning model for colon polyp segmentation","cleaned_abstract":"the incidence of colon cancer occupies the top three places in gastrointestinal tumors, and colon polyps are an important causative factor in the development of colon cancer. early screening for colon polyps and colon polypectomy can reduce the chances of colon cancer. the current means of colon polyp examination is through colonoscopy, taking images of the gastrointestinal tract, and then manually marking them manually, which is time-consuming and labor-intensive for doctors. therefore, relying on advanced deep learning technology to automatically identify colon polyps in the gastrointestinal tract of the patient and segmenting the polyps is an important direction of research nowadays. due to the privacy of medical data and the non-interoperability of disease information, this paper proposes a dual-branch colon polyp segmentation network based on federated learning, which makes it possible to achieve a better training effect under the guarantee of data independence, and secondly, the dual-branch colon polyp segmentation network proposed in this paper adopts the two different structures of convolutional neural network (cnn) and transformer to form a dual-branch structure, and through layer-by-layer fusion embedding, the advantages between different structures are realized. in this paper, we also propose the aggregated attention module (aam) to preserve the high-dimensional semantic information and to complement the missing information in the lower layers. ultimately our approach achieves state of the art in kvasir-seg and cvc-clinicdb datasets.","key_phrases":["the incidence","colon cancer","the top three places","gastrointestinal tumors","colon polyps","an important causative factor","the development","colon cancer","colon polyps","colon polypectomy","the chances","colon cancer","the current means","colon polyp examination","colonoscopy","images","the gastrointestinal tract","them","which","doctors","advanced deep learning technology","colon polyps","the gastrointestinal tract","the patient","the polyps","an important direction","research","the privacy","medical data","the non","-","interoperability","disease information","this paper","a dual-branch colon polyp segmentation network","federated learning","which","it","a better training effect","the guarantee","data independence","the dual-branch colon polyp segmentation network","this paper","the two different structures","convolutional neural network","cnn","a dual-branch structure","layer","the advantages","different structures","this paper","we","the aggregated attention module","aam","the high-dimensional semantic information","the missing information","the lower layers","our approach","state","the art","kvasir-seg and cvc-clinicdb datasets","three","secondly","two","cnn"]},{"title":"Deep learning in pulmonary nodule detection and segmentation: a systematic review","authors":["Chuan Gao","Linyu Wu","Wei Wu","Yichao Huang","Xinyue Wang","Zhichao Sun","Maosheng Xu","Chen Gao"],"abstract":"ObjectivesThe accurate detection and precise segmentation of lung nodules on computed tomography are key prerequisites for early diagnosis and appropriate treatment of lung cancer. This study was designed to compare detection and segmentation methods for pulmonary nodules using deep-learning techniques to fill methodological gaps and biases in the existing literature.MethodsThis study utilized a systematic review with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines, searching PubMed, Embase, Web of Science Core Collection, and the Cochrane Library databases up to May 10, 2023. The Quality Assessment of Diagnostic Accuracy Studies 2 criteria was used to assess the risk of bias and was adjusted with the Checklist for Artificial Intelligence in Medical Imaging. The study analyzed and extracted model performance, data sources, and task-focus information.ResultsAfter screening, we included nine studies meeting our inclusion criteria. These studies were published between 2019 and 2023 and predominantly used public datasets, with the Lung Image Database Consortium Image Collection and Image Database Resource Initiative and Lung Nodule Analysis 2016 being the most common. The studies focused on detection, segmentation, and other tasks, primarily utilizing Convolutional Neural Networks for model development. Performance evaluation covered multiple metrics, including sensitivity and the Dice coefficient.ConclusionsThis study highlights the potential power of deep learning in lung nodule detection and segmentation. It underscores the importance of standardized data processing, code and data sharing, the value of external test datasets, and the need to balance model complexity and efficiency in future research.Clinical relevance statementDeep learning demonstrates significant promise in autonomously detecting and segmenting pulmonary nodules. Future research should address methodological shortcomings and variability to enhance its clinical utility.Key Points\n\nDeep learning shows potential in the detection and segmentation of pulmonary nodules.\n\n\nThere are methodological gaps and biases present in the existing literature.\n\n\nFactors such as external validation and transparency affect the clinical application.","doi":"10.1007\/s00330-024-10907-0","cleaned_title":"deep learning in pulmonary nodule detection and segmentation: a systematic review","cleaned_abstract":"objectivesthe accurate detection and precise segmentation of lung nodules on computed tomography are key prerequisites for early diagnosis and appropriate treatment of lung cancer. this study was designed to compare detection and segmentation methods for pulmonary nodules using deep-learning techniques to fill methodological gaps and biases in the existing literature.methodsthis study utilized a systematic review with the preferred reporting items for systematic reviews and meta-analyses guidelines, searching pubmed, embase, web of science core collection, and the cochrane library databases up to may 10, 2023. the quality assessment of diagnostic accuracy studies 2 criteria was used to assess the risk of bias and was adjusted with the checklist for artificial intelligence in medical imaging. the study analyzed and extracted model performance, data sources, and task-focus information.resultsafter screening, we included nine studies meeting our inclusion criteria. these studies were published between 2019 and 2023 and predominantly used public datasets, with the lung image database consortium image collection and image database resource initiative and lung nodule analysis 2016 being the most common. the studies focused on detection, segmentation, and other tasks, primarily utilizing convolutional neural networks for model development. performance evaluation covered multiple metrics, including sensitivity and the dice coefficient.conclusionsthis study highlights the potential power of deep learning in lung nodule detection and segmentation. it underscores the importance of standardized data processing, code and data sharing, the value of external test datasets, and the need to balance model complexity and efficiency in future research.clinical relevance statementdeep learning demonstrates significant promise in autonomously detecting and segmenting pulmonary nodules. future research should address methodological shortcomings and variability to enhance its clinical utility.key points deep learning shows potential in the detection and segmentation of pulmonary nodules. there are methodological gaps and biases present in the existing literature. factors such as external validation and transparency affect the clinical application.","key_phrases":["objectivesthe accurate detection","precise segmentation","lung nodules","computed tomography","key prerequisites","early diagnosis","appropriate treatment","lung cancer","this study","detection and segmentation methods","pulmonary nodules","deep-learning techniques","methodological gaps","biases","the existing literature.methodsthis study","a systematic review","the preferred reporting items","systematic reviews","meta-analyses guidelines","embase, web","science core collection","the cochrane library","may","the quality assessment","diagnostic accuracy studies","2 criteria","the risk","bias","the checklist","artificial intelligence","medical imaging","the study","model performance","data sources","task-focus information.resultsafter screening","we","nine studies","our inclusion criteria","these studies","public datasets","the lung image database consortium image collection and image database resource initiative","lung","the studies","detection","segmentation","other tasks","convolutional neural networks","model development","performance evaluation","multiple metrics","sensitivity","the dice coefficient.conclusionsthis study","the potential power","deep learning","lung nodule detection","segmentation","it","the importance","standardized data processing","code","data","sharing","the value","external test datasets","the need","model complexity","efficiency","future research.clinical relevance statementdeep learning","significant promise","segmenting pulmonary nodules","future research","methodological shortcomings","variability","its clinical utility.key points","deep learning","potential","the detection","segmentation","pulmonary nodules","methodological gaps","biases","the existing literature","factors","external validation","transparency","the clinical application","2","nine","between 2019 and 2023","2016"]},{"title":"Towards optimized tensor code generation for deep learning on sunway many-core processor","authors":["Mingzhen Li","Changxi Liu","Jianjin Liao","Xuegui Zheng","Hailong Yang","Rujun Sun","Jun Xu","Lin Gan","Guangwen Yang","Zhongzhi Luan","Depei Qian"],"abstract":"The flourish of deep learning frameworks and hardware platforms has been demanding an efficient compiler that can shield the diversity in both software and hardware in order to provide application portability. Among the existing deep learning compilers, TVM is well known for its efficiency in code generation and optimization across diverse hardware devices. In the meanwhile, the Sunway many-core processor renders itself as a competitive candidate for its attractive computational power in both scientific computing and deep learning workloads. This paper combines the trends in these two directions. Specifically, we propose swTVM} that extends the original TVM to support ahead-of-time compilation for architecture requiring cross-compilation such as Sunway. In addition, we leverage the architecture features during the compilation such as core group for massive parallelism, DMA for high bandwidth memory transfer and local device memory for data locality, in order to generate efficient codes for deep learning workloads on Sunway. The experiment results show that the codes generated by swTVM} achieve 1.79\u00d7 improvement of inference latency on average compared to the state-of-the-art deep learning framework on Sunway, across eight representative benchmarks. This work is the first attempt from the compiler perspective to bridge the gap of deep learning and Sunway processor particularly with productivity and efficiency in mind. We believe this work will encourage more people to embrace the power of deep learning and Sunway many-core processor.","doi":"10.1007\/s11704-022-2440-7","cleaned_title":"towards optimized tensor code generation for deep learning on sunway many-core processor","cleaned_abstract":"the flourish of deep learning frameworks and hardware platforms has been demanding an efficient compiler that can shield the diversity in both software and hardware in order to provide application portability. among the existing deep learning compilers, tvm is well known for its efficiency in code generation and optimization across diverse hardware devices. in the meanwhile, the sunway many-core processor renders itself as a competitive candidate for its attractive computational power in both scientific computing and deep learning workloads. this paper combines the trends in these two directions. specifically, we propose swtvm} that extends the original tvm to support ahead-of-time compilation for architecture requiring cross-compilation such as sunway. in addition, we leverage the architecture features during the compilation such as core group for massive parallelism, dma for high bandwidth memory transfer and local device memory for data locality, in order to generate efficient codes for deep learning workloads on sunway. the experiment results show that the codes generated by swtvm} achieve 1.79\u00d7 improvement of inference latency on average compared to the state-of-the-art deep learning framework on sunway, across eight representative benchmarks. this work is the first attempt from the compiler perspective to bridge the gap of deep learning and sunway processor particularly with productivity and efficiency in mind. we believe this work will encourage more people to embrace the power of deep learning and sunway many-core processor.","key_phrases":["the flourish","deep learning frameworks","hardware platforms","an efficient compiler","that","the diversity","both software","hardware","order","application portability","the existing deep learning compilers","tvm","its efficiency","code generation","optimization","diverse hardware devices","the meanwhile","the sunway many-core processor","itself","a competitive candidate","its attractive computational power","both scientific computing","deep learning workloads","this paper","the trends","these two directions","we","swtvm","that","the original tvm","time","architecture","cross","-","compilation","sunway","addition","we","the architecture","the compilation","core group","massive parallelism","dma","high bandwidth memory transfer","local device memory","data locality","order","efficient codes","deep learning workloads","sunway","the experiment results","the codes","swtvm","1.79\u00d7 improvement","inference latency","the-art","sunway","eight representative benchmarks","this work","the first attempt","the compiler perspective","the gap","deep learning and sunway processor","productivity","efficiency","mind","we","this work","more people","the power","deep learning","many-core processor","two","1.79\u00d7","eight","first"]},{"title":"Discovering cholinesterase inhibitors from Chinese herbal medicine with deep learning models","authors":["Fulu Pan","Yang Liu","Zhiqiang Luo","Guopeng Wang","Xueyan Li","Huining Liu","Shuang Yu","Dongying Qi","Xinyu Wang","Xiaoyu Chai","Qianqian Wang","Renfang Yin","Yanli Pan"],"abstract":"Traditional Chinese medicine (TCM) holds distinctive advantages in the management of Alzheimer\u2019s disease. Nonetheless, a considerable gap remains in our understanding of its pharmacologically active constituents. In this study, we harnessed the potential of deep learning models to swiftly and precisely predict drug-target interactions. We conducted a systematic screening of cholinesterase (ChE) inhibitors from an extensive array of TCM ingredients, followed by rigorous validation through in vitro experiments. We constructed both a drug-target interactions (DTI) model and a blood-brain barrier permeability (BBBP) model, with both models achieving an AUPRC score exceeding 0.9. Subsequently, we conducted a screening process that identified six compounds for in vitro ChE inhibitory assay. Notably, all six compounds exhibited a robust inhibitory effect on acetylcholinesterase (AChE), while four of the six compounds demonstrated potent inhibitory activity against butyrylcholinesterase (BChE). Our findings underscore the promise of leveraging deep learning to discover inhibitors from TCM.","doi":"10.1007\/s00044-024-03238-8","cleaned_title":"discovering cholinesterase inhibitors from chinese herbal medicine with deep learning models","cleaned_abstract":"traditional chinese medicine (tcm) holds distinctive advantages in the management of alzheimer\u2019s disease. nonetheless, a considerable gap remains in our understanding of its pharmacologically active constituents. in this study, we harnessed the potential of deep learning models to swiftly and precisely predict drug-target interactions. we conducted a systematic screening of cholinesterase (che) inhibitors from an extensive array of tcm ingredients, followed by rigorous validation through in vitro experiments. we constructed both a drug-target interactions (dti) model and a blood-brain barrier permeability (bbbp) model, with both models achieving an auprc score exceeding 0.9. subsequently, we conducted a screening process that identified six compounds for in vitro che inhibitory assay. notably, all six compounds exhibited a robust inhibitory effect on acetylcholinesterase (ache), while four of the six compounds demonstrated potent inhibitory activity against butyrylcholinesterase (bche). our findings underscore the promise of leveraging deep learning to discover inhibitors from tcm.","key_phrases":["traditional chinese medicine","tcm","distinctive advantages","the management","alzheimer\u2019s disease","a considerable gap","our understanding","its pharmacologically active constituents","this study","we","the potential","deep learning models","drug-target interactions","we","a systematic screening","cholinesterase (che) inhibitors","an extensive array","tcm ingredients","rigorous validation","vitro experiments","we","both a drug-target interactions","(dti) model","a blood-brain barrier permeability","(bbbp) model","both models","an auprc score","we","a screening process","that","six compounds","in vitro che inhibitory assay","all six compounds","a robust inhibitory effect","acetylcholinesterase","ache","the six compounds","potent inhibitory activity","butyrylcholinesterase","bche","our findings","the promise","deep learning","inhibitors","tcm","chinese","cholinesterase (che","dti","0.9","six","six","four","six"]},{"title":"Predict customer churn using combination deep learning networks model","authors":["Van-Hieu Vu"],"abstract":"Customers churn is an important issue that is always concerned by banks, and is put at the forefront of the bank\u2019s policies. The fact that banks can identify customers who are intending to leave the service can help banks promptly make policies to retain customers. In this paper, we propose a combined deep learning network models to predict customers leaving or staying at the bank. The proposed model consists of two levels, Level 0 consists of three basic models using three Deep Learning Neural Networks, and Level 1 is a logistic regression model. The proposed model has obtained evaluation results with accuracy metrics of 96.60%, precision metrics of 90.26%, recall metrics of 91.91% and F1 score of 91.07% on the dataset \u201cBank Customer Churn Prediction\u201d.","doi":"10.1007\/s00521-023-09327-w","cleaned_title":"predict customer churn using combination deep learning networks model","cleaned_abstract":"customers churn is an important issue that is always concerned by banks, and is put at the forefront of the bank\u2019s policies. the fact that banks can identify customers who are intending to leave the service can help banks promptly make policies to retain customers. in this paper, we propose a combined deep learning network models to predict customers leaving or staying at the bank. the proposed model consists of two levels, level 0 consists of three basic models using three deep learning neural networks, and level 1 is a logistic regression model. the proposed model has obtained evaluation results with accuracy metrics of 96.60%, precision metrics of 90.26%, recall metrics of 91.91% and f1 score of 91.07% on the dataset \u201cbank customer churn prediction\u201d.","key_phrases":["customers","an important issue","that","banks","the forefront","the bank\u2019s policies","the fact","banks","customers","who","the service","banks","policies","customers","this paper","we","a combined deep learning network models","customers","the bank","the proposed model","two levels","level","three basic models","three deep learning neural networks","level","a logistic regression model","the proposed model","evaluation results","accuracy metrics","96.60%","precision metrics","90.26%","recall metrics","91.91%","f1 score","91.07%","the dataset \u201cbank customer churn prediction","two","0","three","three","1","96.60%","90.26%","91.91%","91.07%"]},{"title":"Genome analysis through image processing with deep learning models","authors":["Yao-zhong Zhang","Seiya Imoto"],"abstract":"Genomic sequences are traditionally represented as strings of characters: A (adenine), C (cytosine), G (guanine), and T (thymine). However, an alternative approach involves depicting sequence-related information through image representations, such as Chaos Game Representation (CGR) and read pileup images. With rapid advancements in deep learning (DL) methods within computer vision and natural language processing, there is growing interest in applying image-based DL methods to genomic sequence analysis. These methods involve encoding genomic information as images or integrating spatial information from images into the analytical process. In this review, we summarize three typical applications that use image processing with DL models for genome analysis. We examine the utilization and advantages of these image-based approaches.","doi":"10.1038\/s10038-024-01275-0","cleaned_title":"genome analysis through image processing with deep learning models","cleaned_abstract":"genomic sequences are traditionally represented as strings of characters: a (adenine), c (cytosine), g (guanine), and t (thymine). however, an alternative approach involves depicting sequence-related information through image representations, such as chaos game representation (cgr) and read pileup images. with rapid advancements in deep learning (dl) methods within computer vision and natural language processing, there is growing interest in applying image-based dl methods to genomic sequence analysis. these methods involve encoding genomic information as images or integrating spatial information from images into the analytical process. in this review, we summarize three typical applications that use image processing with dl models for genome analysis. we examine the utilization and advantages of these image-based approaches.","key_phrases":["genomic sequences","strings","characters","a (adenine","c","cytosine","g","guanine","t","(thymine","an alternative approach","sequence-related information","image representations","chaos game representation","pileup images","rapid advancements","deep learning","computer vision","natural language processing","interest","image-based dl methods","genomic sequence analysis","these methods","genomic information","images","spatial information","images","the analytical process","this review","we","three typical applications","that","image processing","dl models","genome analysis","we","the utilization","advantages","these image-based approaches","three"]},{"title":"Prediction of mechanical properties for deep drawing steel by deep learning","authors":["Gang Xu","Jinshan He","Zhimin L\u00fc","Min Li","Jinwu Xu"],"abstract":"At present, iron and steel enterprises mainly use \u201cafter spot test ward\u201d to control final product quality. However, it is impossible to realize on-line quality predetermining for all products by this traditional approach, hence claims and returns often occur, resulting in major economic losses of enterprises. In order to realize the on-line quality predetermining for steel products during manufacturing process, the prediction models of mechanical properties based on deep learning have been proposed in this work. First, the mechanical properties of deep drawing steels were predicted by using LSTM (long short team memory), GRU (gated recurrent unit) network, and GPR (Gaussian process regression) model, and prediction accuracy and learning efficiency for different models were also discussed. Then, on-line re-learning methods for transfer learning models and model parameters were proposed. The experimental results show that not only the prediction accuracy of optimized transfer learning models has been improved, but also predetermining time was shortened to meet real time requirements of on-line property predetermining. The industrial production data of interstitial-free (IF) steel was used to demonstrate that R2 value of GRU model in training stage reaches more than 0.99, and R2 value in testing stage is more than 0.96.","doi":"10.1007\/s12613-022-2547-8","cleaned_title":"prediction of mechanical properties for deep drawing steel by deep learning","cleaned_abstract":"at present, iron and steel enterprises mainly use \u201cafter spot test ward\u201d to control final product quality. however, it is impossible to realize on-line quality predetermining for all products by this traditional approach, hence claims and returns often occur, resulting in major economic losses of enterprises. in order to realize the on-line quality predetermining for steel products during manufacturing process, the prediction models of mechanical properties based on deep learning have been proposed in this work. first, the mechanical properties of deep drawing steels were predicted by using lstm (long short team memory), gru (gated recurrent unit) network, and gpr (gaussian process regression) model, and prediction accuracy and learning efficiency for different models were also discussed. then, on-line re-learning methods for transfer learning models and model parameters were proposed. the experimental results show that not only the prediction accuracy of optimized transfer learning models has been improved, but also predetermining time was shortened to meet real time requirements of on-line property predetermining. the industrial production data of interstitial-free (if) steel was used to demonstrate that r2 value of gru model in training stage reaches more than 0.99, and r2 value in testing stage is more than 0.96.","key_phrases":["iron and steel enterprises","spot test ward","final product quality","it","line","all products","this traditional approach","returns","major economic losses","enterprises","order","line","steel products","manufacturing process","the prediction models","mechanical properties","deep learning","this work","the mechanical properties","deep drawing steels","lstm","long short team memory","gru","gated recurrent unit","network","gpr","gaussian process regression","model","prediction accuracy","efficiency","different models","line","-learning methods","transfer learning models","model parameters","the experimental results","not only the prediction accuracy","optimized transfer learning models","predetermining time","real time requirements","line","the industrial production data",") steel","r2 value","gru model","training stage","r2 value","testing stage","first","gpr","gaussian","more than 0.99"]},{"title":"Supervised and Unsupervised Deep Learning Approaches for EEG Seizure Prediction","authors":["Zakary Georgis-Yap","Milos R. Popovic","Shehroz S. Khan"],"abstract":"Epilepsy affects more than 50 million people worldwide, making it one of the world\u2019s most prevalent neurological diseases. The main symptom of epilepsy is seizures, which occur abruptly and can cause serious injury or death. The ability to predict the occurrence of an epileptic seizure could alleviate many risks and stresses people with epilepsy face. We formulate the problem of detecting preictal (or pre-seizure) with reference to normal EEG as a precursor to incoming seizure. To this end, we developed several supervised deep learning approaches model to identify preictal EEG from normal EEG. We further develop novel unsupervised deep learning approaches to train the models on only normal EEG, and detecting pre-seizure EEG as an anomalous event. These deep learning models were trained and evaluated on two large EEG seizure datasets in a person-specific manner. We found that both supervised and unsupervised approaches are feasible; however, their performance varies depending on the patient, approach and architecture. This new line of research has the potential to develop therapeutic interventions and save human lives.","doi":"10.1007\/s41666-024-00160-x","cleaned_title":"supervised and unsupervised deep learning approaches for eeg seizure prediction","cleaned_abstract":"epilepsy affects more than 50 million people worldwide, making it one of the world\u2019s most prevalent neurological diseases. the main symptom of epilepsy is seizures, which occur abruptly and can cause serious injury or death. the ability to predict the occurrence of an epileptic seizure could alleviate many risks and stresses people with epilepsy face. we formulate the problem of detecting preictal (or pre-seizure) with reference to normal eeg as a precursor to incoming seizure. to this end, we developed several supervised deep learning approaches model to identify preictal eeg from normal eeg. we further develop novel unsupervised deep learning approaches to train the models on only normal eeg, and detecting pre-seizure eeg as an anomalous event. these deep learning models were trained and evaluated on two large eeg seizure datasets in a person-specific manner. we found that both supervised and unsupervised approaches are feasible; however, their performance varies depending on the patient, approach and architecture. this new line of research has the potential to develop therapeutic interventions and save human lives.","key_phrases":["epilepsy","more than 50 million people","it","the world\u2019s most prevalent neurological diseases","the main symptom","epilepsy","seizures","which","serious injury","death","the ability","the occurrence","an epileptic seizure","many risks","people","epilepsy face","we","the problem","reference","normal eeg","a precursor","incoming seizure","this end","we","several supervised deep learning approaches model","preictal eeg","normal eeg","we","novel unsupervised deep learning approaches","the models","only normal eeg","pre-seizure eeg","an anomalous event","these deep learning models","two large eeg seizure datasets","a person-specific manner","we","both","approaches","their performance","the patient","approach","architecture","this new line","research","the potential","therapeutic interventions","human lives","more than 50 million","two"]},{"title":"Leveraging deep learning-assisted attacks against image obfuscation via federated learning","authors":["Jimmy Tekli","Bechara Al Bouna","Gilbert Tekli","Rapha\u00ebl Couturier","Antoine Charbel"],"abstract":"Obfuscation techniques (e.g., blurring) are employed to protect sensitive information (SI) in images such as individuals\u2019 faces. Recent works demonstrated that adversaries can perform deep learning-assisted (DL) attacks to re-identify obfuscated face images. Adversaries are modeled by their goals, knowledge (e.g., background knowledge), and capabilities (e.g., DL-assisted attacks). Nevertheless, enhancing the evaluation methodology of obfuscation techniques and improving the defense strategies against adversaries requires considering more \"pessimistic\u201d attacking scenario, i.e., stronger adversaries. According to a 2019 article published by the European Union Agency for Cybersecurity (ENISA), adversaries tend to perform more sophisticated and dangerous attacks when collaborating together. To address these concerns, our paper investigates a novel privacy challenge in the context of image obfuscation. Specifically, we examine whether adversaries, when collaborating together, can amplify their DL-assisted attacks and cause additional privacy breaches against a target dataset of obfuscated images. We empirically demonstrate that federated learning (FL) can be used as a collaborative attack\/adversarial strategy to (i) leverage the attacking capabilities of an adversary, (ii) increase the privacy breaches, and (iii) remedy the lack of background knowledge and data shortage without the need to share\/disclose the local training datasets in a centralized location. To the best of our knowledge, we are the first to consider collaborative and more specifically FL-based attacks in the context of face obfuscation.","doi":"10.1007\/s00521-024-09703-0","cleaned_title":"leveraging deep learning-assisted attacks against image obfuscation via federated learning","cleaned_abstract":"obfuscation techniques (e.g., blurring) are employed to protect sensitive information (si) in images such as individuals\u2019 faces. recent works demonstrated that adversaries can perform deep learning-assisted (dl) attacks to re-identify obfuscated face images. adversaries are modeled by their goals, knowledge (e.g., background knowledge), and capabilities (e.g., dl-assisted attacks). nevertheless, enhancing the evaluation methodology of obfuscation techniques and improving the defense strategies against adversaries requires considering more \"pessimistic\u201d attacking scenario, i.e., stronger adversaries. according to a 2019 article published by the european union agency for cybersecurity (enisa), adversaries tend to perform more sophisticated and dangerous attacks when collaborating together. to address these concerns, our paper investigates a novel privacy challenge in the context of image obfuscation. specifically, we examine whether adversaries, when collaborating together, can amplify their dl-assisted attacks and cause additional privacy breaches against a target dataset of obfuscated images. we empirically demonstrate that federated learning (fl) can be used as a collaborative attack\/adversarial strategy to (i) leverage the attacking capabilities of an adversary, (ii) increase the privacy breaches, and (iii) remedy the lack of background knowledge and data shortage without the need to share\/disclose the local training datasets in a centralized location. to the best of our knowledge, we are the first to consider collaborative and more specifically fl-based attacks in the context of face obfuscation.","key_phrases":["obfuscation techniques","e.g., blurring","sensitive information","images","recent works","adversaries","dl","obfuscated face images","adversaries","their goals","knowledge","capabilities","(e.g., dl-assisted attacks","the evaluation methodology","obfuscation techniques","the defense strategies","adversaries","scenario","i.e., stronger adversaries","a 2019 article","the european union agency","cybersecurity","enisa","adversaries","more sophisticated and dangerous attacks","these concerns","our paper","a novel privacy challenge","the context","image obfuscation","we","adversaries","their dl-assisted attacks","additional privacy breaches","a target dataset","obfuscated images","we","federated learning","a collaborative attack\/adversarial strategy","(i","the attacking capabilities","an adversary","ii","the privacy breaches","(iii","the lack","background knowledge and data shortage","the need","the local training datasets","a centralized location","our knowledge","we","collaborative and more specifically fl-based attacks","the context","face obfuscation","2019","the european union agency","first"]},{"title":"Are deep learning classification results obtained on CT scans fair and interpretable?","authors":["Mohamad M. A. Ashames","Ahmet Demir","Omer N. Gerek","Mehmet Fidan","M. Bilginer Gulmezoglu","Semih Ergin","Rifat Edizkan","Mehmet Koc","Atalay Barkana","Cuneyt Calisir"],"abstract":"Following the great success of various deep learning methods in image and object classification, the biomedical image processing society is also overwhelmed with their applications to various automatic diagnosis cases. Unfortunately, most of the deep learning-based classification attempts in the literature solely focus on the aim of extreme accuracy scores, without considering interpretability, or patient-wise separation of training and test data. For example, most lung nodule classification papers using deep learning randomly shuffle data and split it into training, validation, and test sets, causing certain images from the Computed Tomography (CT) scan of a person to be in the training set, while other images of the same person to be in the validation or testing image sets. This can result in reporting misleading accuracy rates and the learning of irrelevant features, ultimately reducing the real-life usability of these models. When the deep neural networks trained on the traditional, unfair data shuffling method are challenged with new patient images, it is observed that the trained models perform poorly. In contrast, deep neural networks trained with strict patient-level separation maintain their accuracy rates even when new patient images are tested. Heat map visualizations of the activations of the deep neural networks trained with strict patient-level separation indicate a higher degree of focus on the relevant nodules. We argue that the research question posed in the title has a positive answer only if the deep neural networks are trained with images of patients that are strictly isolated from the validation and testing patient sets.","doi":"10.1007\/s13246-024-01419-8","cleaned_title":"are deep learning classification results obtained on ct scans fair and interpretable?","cleaned_abstract":"following the great success of various deep learning methods in image and object classification, the biomedical image processing society is also overwhelmed with their applications to various automatic diagnosis cases. unfortunately, most of the deep learning-based classification attempts in the literature solely focus on the aim of extreme accuracy scores, without considering interpretability, or patient-wise separation of training and test data. for example, most lung nodule classification papers using deep learning randomly shuffle data and split it into training, validation, and test sets, causing certain images from the computed tomography (ct) scan of a person to be in the training set, while other images of the same person to be in the validation or testing image sets. this can result in reporting misleading accuracy rates and the learning of irrelevant features, ultimately reducing the real-life usability of these models. when the deep neural networks trained on the traditional, unfair data shuffling method are challenged with new patient images, it is observed that the trained models perform poorly. in contrast, deep neural networks trained with strict patient-level separation maintain their accuracy rates even when new patient images are tested. heat map visualizations of the activations of the deep neural networks trained with strict patient-level separation indicate a higher degree of focus on the relevant nodules. we argue that the research question posed in the title has a positive answer only if the deep neural networks are trained with images of patients that are strictly isolated from the validation and testing patient sets.","key_phrases":["the great success","various deep learning methods","image and object classification","the biomedical image processing society","their applications","various automatic diagnosis cases","the deep learning-based classification attempts","the literature","the aim","extreme accuracy scores","interpretability","patient-wise separation","training","test data","example","most lung","classification papers","deep learning randomly shuffle data","it","training","validation","test sets","certain images","the computed tomography (ct) scan","a person","the training set","the same person","the validation or testing image sets","this","misleading accuracy rates","the learning","irrelevant features","the real-life usability","these models","the deep neural networks","the traditional, unfair data","method","new patient images","it","the trained models","contrast","deep neural networks","strict patient-level separation","their accuracy rates","new patient images","heat map visualizations","the activations","the deep neural networks","strict patient-level separation","a higher degree","focus","the relevant nodules","we","the research question","the title","a positive answer","the deep neural networks","images","patients","that","the validation and testing patient sets"]},{"title":"A Novel Approach Using Transfer Learning Architectural Models Based Deep Learning Techniques for Identification and Classification of Malignant Skin Cancer","authors":["Balambigai Subramanian","Suresh Muthusamy","Kokilavani Thangaraj","Hitesh Panchal","Elavarasi Kasirajan","Abarna Marimuthu","Abinaya Ravi"],"abstract":"Melanoma, a form of skin cancer originating in melanocyte cells, poses a significant health risk, although it is less prevalent than other types of skin cancer. Its detection presents challenges, even under expert observation. To enhance the classification accuracy of skin lesions, a Deep Convolutional Neural Network, Visual Geometry Group model has been proposed. However, deep learning methods typically require substantial training time. To mitigate this, transfer learning techniques are employed, reducing training duration. Data sets sourced from the International Skin Imaging Collaboration are utilized to train the model within this proposed approach. Evaluation of classification performance involves metrics such as Accuracy, Positive Predictive Value, Negative Predictive Value, Specificity, and Sensitivity. The classifier\u2019s performance on test data is depicted through a confusion matrix. The introduction of transfer learning techniques into the Deep Convolutional Neural Network has resulted in an improved classification accuracy of 85%, compared to the 81% achieved by a standard Convolutional Neural Network.","doi":"10.1007\/s11277-024-11006-5","cleaned_title":"a novel approach using transfer learning architectural models based deep learning techniques for identification and classification of malignant skin cancer","cleaned_abstract":"melanoma, a form of skin cancer originating in melanocyte cells, poses a significant health risk, although it is less prevalent than other types of skin cancer. its detection presents challenges, even under expert observation. to enhance the classification accuracy of skin lesions, a deep convolutional neural network, visual geometry group model has been proposed. however, deep learning methods typically require substantial training time. to mitigate this, transfer learning techniques are employed, reducing training duration. data sets sourced from the international skin imaging collaboration are utilized to train the model within this proposed approach. evaluation of classification performance involves metrics such as accuracy, positive predictive value, negative predictive value, specificity, and sensitivity. the classifier\u2019s performance on test data is depicted through a confusion matrix. the introduction of transfer learning techniques into the deep convolutional neural network has resulted in an improved classification accuracy of 85%, compared to the 81% achieved by a standard convolutional neural network.","key_phrases":["melanoma","a form","melanocyte cells","a significant health risk","it","other types","skin cancer","its detection","challenges","expert observation","the classification accuracy","skin lesions","a deep convolutional neural network","visual geometry group model","deep learning methods","substantial training time","this","transfer learning techniques","training duration","data sets","the international skin imaging collaboration","the model","this proposed approach","evaluation","classification performance","metrics","accuracy","positive predictive value","negative predictive value","specificity","sensitivity","the classifier\u2019s performance","test data","a confusion matrix","the introduction","techniques","the deep convolutional neural network","an improved classification accuracy","85%","the 81%","a standard convolutional neural network","85%","the 81%"]},{"title":"Deep contrastive learning of molecular conformation for efficient property prediction","authors":["Yang Jeong Park","HyunGi Kim","Jeonghee Jo","Sungroh Yoon"],"abstract":"Data-driven deep learning algorithms provide accurate prediction of high-level quantum-chemical molecular properties. However, their inputs must be constrained to the same quantum-chemical level of geometric relaxation as the training dataset, limiting their flexibility. Adopting alternative cost-effective conformation generative methods introduces domain-shift problems, deteriorating prediction accuracy. Here we propose a deep contrastive learning-based domain-adaptation method called Local Atomic environment Contrastive Learning (LACL). LACL learns to alleviate the disparities in distribution between the two geometric conformations by comparing different conformation-generation methods. We found that LACL forms a domain-agnostic latent space that encapsulates the semantics of an atom\u2019s local atomic environment. LACL achieves quantum-chemical accuracy while circumventing the geometric relaxation bottleneck and could enable future application scenarios such as inverse molecular engineering and large-scale screening. Our approach is also generalizable from small organic molecules to long chains of biological and pharmacological molecules.","doi":"10.1038\/s43588-023-00560-w","cleaned_title":"deep contrastive learning of molecular conformation for efficient property prediction","cleaned_abstract":"data-driven deep learning algorithms provide accurate prediction of high-level quantum-chemical molecular properties. however, their inputs must be constrained to the same quantum-chemical level of geometric relaxation as the training dataset, limiting their flexibility. adopting alternative cost-effective conformation generative methods introduces domain-shift problems, deteriorating prediction accuracy. here we propose a deep contrastive learning-based domain-adaptation method called local atomic environment contrastive learning (lacl). lacl learns to alleviate the disparities in distribution between the two geometric conformations by comparing different conformation-generation methods. we found that lacl forms a domain-agnostic latent space that encapsulates the semantics of an atom\u2019s local atomic environment. lacl achieves quantum-chemical accuracy while circumventing the geometric relaxation bottleneck and could enable future application scenarios such as inverse molecular engineering and large-scale screening. our approach is also generalizable from small organic molecules to long chains of biological and pharmacological molecules.","key_phrases":["data-driven deep learning algorithms","accurate prediction","high-level quantum-chemical molecular properties","their inputs","the same quantum-chemical level","geometric relaxation","the training dataset","their flexibility","alternative cost-effective conformation generative methods","domain-shift problems","deteriorating prediction accuracy","we","a deep contrastive learning-based domain-adaptation method","local atomic environment","contrastive learning","lacl","lacl","the disparities","distribution","the two geometric conformations","different conformation-generation methods","we","lacl","a domain-agnostic latent space","that","the semantics","an atom\u2019s local atomic environment","lacl","quantum-chemical accuracy","the geometric relaxation bottleneck","future application scenarios","inverse molecular engineering and large-scale screening","our approach","small organic molecules","long chains","biological and pharmacological molecules","two"]},{"title":"A pedagogical study on promoting students' deep learning through design-based learning","authors":["Chunmeng Weng","Congying Chen","Xianfeng Ai"],"abstract":"This paper illustrates the design-based learning (DBL) approach to promoting the deep learning of students and improving the quality of teaching in engineering design education. We performed three aspects of research with students in a typical educational activity. The first study investigated students' deep learning before and after the DBL approach, both in terms of deep learning status and deep learning ability. The second study examined the effectiveness of the DBL approach by comparative research of a control class (traditional teaching method) and an experimental class (DBL method). The third study examined students' evaluations of the DBL approach. It is approved that the DBL approach has distinctively stimulated the students' motivation to learn, making them more actively engaged in study. The students' higher-order thinking and higher-order capabilities are enhanced, such as critical thinking ability and problem-solving ability. At the same time, they are satisfied with the DBL approach. These findings suggest that the DBL approach is effective in promoting students' deep learning and improving the quality of teaching and learning.","doi":"10.1007\/s10798-022-09789-4","cleaned_title":"a pedagogical study on promoting students' deep learning through design-based learning","cleaned_abstract":"this paper illustrates the design-based learning (dbl) approach to promoting the deep learning of students and improving the quality of teaching in engineering design education. we performed three aspects of research with students in a typical educational activity. the first study investigated students' deep learning before and after the dbl approach, both in terms of deep learning status and deep learning ability. the second study examined the effectiveness of the dbl approach by comparative research of a control class (traditional teaching method) and an experimental class (dbl method). the third study examined students' evaluations of the dbl approach. it is approved that the dbl approach has distinctively stimulated the students' motivation to learn, making them more actively engaged in study. the students' higher-order thinking and higher-order capabilities are enhanced, such as critical thinking ability and problem-solving ability. at the same time, they are satisfied with the dbl approach. these findings suggest that the dbl approach is effective in promoting students' deep learning and improving the quality of teaching and learning.","key_phrases":["this paper","the design-based learning","(dbl) approach","the deep learning","students","the quality","teaching","engineering design education","we","three aspects","research","students","a typical educational activity","the first study","students' deep learning","the dbl approach","terms","deep learning status","deep learning ability","the second study","the effectiveness","the dbl approach","comparative research","a control class","traditional teaching method","an experimental class","dbl method","the third study","students' evaluations","the dbl approach","it","the dbl approach","the students' motivation","them","study","the students' higher-order thinking","higher-order capabilities","critical thinking ability","problem-solving ability","the same time","they","the dbl approach","these findings","the dbl approach","students' deep learning","the quality","teaching","three","first","second","third"]},{"title":"Machine and deep learning techniques for the prediction of diabetics: a review","authors":["Sandip Kumar Singh Modak","Vijay Kumar Jha"],"abstract":"Diabetes has become one of the significant reasons for public sickness and death in worldwide. By 2019, diabetes had affected more than 463 million people worldwide. According to the International Diabetes Federation report, this figure is expected to rise to more than 700 million in 2040, so early screening and diagnosis of diabetes patients have great significance in detecting and treating diabetes on time. Diabetes is a multi factorial metabolic disease, its diagnostic criteria are difficult to cover all the ethology, damage degree, pathogenesis and other factors, so there is a situation for uncertainty and imprecision under various aspects of the medical diagnosis process. With the development of Data mining, researchers find that machine learning and deep learning, playing an important role in diabetes prediction research. This paper is an in-depth study on the application of machine learning and deep learning techniques in the prediction of diabetics. In addition, this paper also discusses the different methodology used in machine and deep learning for prediction of diabetics since last two decades and examines the methods used, to explore their successes and failure. This review would help researchers and practitioners understand the current state-of-the-art methods and identify gaps in the literature.","doi":"10.1007\/s11042-024-19766-9","cleaned_title":"machine and deep learning techniques for the prediction of diabetics: a review","cleaned_abstract":"diabetes has become one of the significant reasons for public sickness and death in worldwide. by 2019, diabetes had affected more than 463 million people worldwide. according to the international diabetes federation report, this figure is expected to rise to more than 700 million in 2040, so early screening and diagnosis of diabetes patients have great significance in detecting and treating diabetes on time. diabetes is a multi factorial metabolic disease, its diagnostic criteria are difficult to cover all the ethology, damage degree, pathogenesis and other factors, so there is a situation for uncertainty and imprecision under various aspects of the medical diagnosis process. with the development of data mining, researchers find that machine learning and deep learning, playing an important role in diabetes prediction research. this paper is an in-depth study on the application of machine learning and deep learning techniques in the prediction of diabetics. in addition, this paper also discusses the different methodology used in machine and deep learning for prediction of diabetics since last two decades and examines the methods used, to explore their successes and failure. this review would help researchers and practitioners understand the current state-of-the-art methods and identify gaps in the literature.","key_phrases":["diabetes","the significant reasons","public sickness","death","diabetes","more than 463 million people","the international diabetes federation report","this figure","early screening","diagnosis","diabetes patients","great significance","diabetes","time","diabetes","a multi factorial metabolic disease","its diagnostic criteria","all the ethology","damage degree","pathogenesis","other factors","a situation","uncertainty","imprecision","various aspects","the medical diagnosis process","the development","data mining","researchers","an important role","diabetes prediction research","this paper","-depth","the application","machine learning","deep learning techniques","the prediction","diabetics","addition","this paper","the different methodology","machine","deep learning","prediction","diabetics","last two decades","the methods","their successes","failure","this review","researchers","practitioners","the-art","gaps","the literature","2019","more than 463 million","more than 700 million","2040","last two decades"]},{"title":"Deep encoder\u2013decoder-based shared learning for multi-criteria recommendation systems","authors":["Salam Fraihat","Bushra Abu Tahon","Bushra Alhijawi","Arafat Awajan"],"abstract":"A recommendation system (RS) can help overcome information overload issues by offering personalized predictions for users. Typically, RS considers the overall ratings of users on items to generate recommendations for them. However, users may consider several aspects when evaluating items. Hence, a multi-criteria RS considers n-aspects of items to generate more accurate recommendations than a single-criteria RS. This research paper proposes two deep encoder\u2013decoder models based on shared learning for a multi-criteria RS, multi-modal deep encoder\u2013decoder-based shared learning (MMEDSL) and multi-criteria deep encoder\u2013decoder-based shared learning (MCEDSL). MMEDSL employs the shared learning technique by concentrating on the multi-modality concept in deep learning, while MCEDSL focuses on the training process to apply the shared learning technique. The shared learning captures useful shared information during the learning process since the multi-criteria may have hidden inter-relationships. A set of experiments were conducted to compare the proposed models with recent baseline approaches. The Yahoo! Movies multi-criteria dataset was utilized. The results demonstrate that the proposed models outperform other algorithms. In addition, the results show that integrating the shared learning technique with the RS produces precise recommendation predictions.","doi":"10.1007\/s00521-023-09007-9","cleaned_title":"deep encoder\u2013decoder-based shared learning for multi-criteria recommendation systems","cleaned_abstract":"a recommendation system (rs) can help overcome information overload issues by offering personalized predictions for users. typically, rs considers the overall ratings of users on items to generate recommendations for them. however, users may consider several aspects when evaluating items. hence, a multi-criteria rs considers n-aspects of items to generate more accurate recommendations than a single-criteria rs. this research paper proposes two deep encoder\u2013decoder models based on shared learning for a multi-criteria rs, multi-modal deep encoder\u2013decoder-based shared learning (mmedsl) and multi-criteria deep encoder\u2013decoder-based shared learning (mcedsl). mmedsl employs the shared learning technique by concentrating on the multi-modality concept in deep learning, while mcedsl focuses on the training process to apply the shared learning technique. the shared learning captures useful shared information during the learning process since the multi-criteria may have hidden inter-relationships. a set of experiments were conducted to compare the proposed models with recent baseline approaches. the yahoo! movies multi-criteria dataset was utilized. the results demonstrate that the proposed models outperform other algorithms. in addition, the results show that integrating the shared learning technique with the rs produces precise recommendation predictions.","key_phrases":["a recommendation system","information overload issues","personalized predictions","users","rs","the overall ratings","users","items","recommendations","them","users","several aspects","items","a multi-criteria rs","n","-aspects","items","more accurate recommendations","a single-criteria rs","this research paper","two deep encoder","decoder models","shared learning","a multi-criteria rs, multi-modal deep encoder","decoder-based shared learning","mmedsl","multi-criteria deep encoder","decoder-based shared learning","mcedsl","the shared learning technique","the multi-modality concept","deep learning","the training process","the shared learning technique","the shared learning captures","useful shared information","the learning process","-","criteria","inter","-","relationships","a set","experiments","the proposed models","recent baseline approaches","the yahoo","movies multi-criteria dataset","the results","the proposed models","other algorithms","addition","the results","the shared learning technique","the rs","precise recommendation predictions","two","yahoo"]},{"title":"A Deep Learning Based Approach for Biomedical Named Entity Recognition Using Multitasking Transfer Learning with BiLSTM, BERT and CRF","authors":["H. Pooja","M. P. Prabhudev Jagadeesh"],"abstract":"The named entity recognition (NER) is a method for locating references to rigid designators in text that fall into well-established semantic categories like person, place, organisation, etc., Many natural language e applications, like summarization of text, question\u2013answer models and machine translation, always include NER at their core. Early NER systems were quite successful in reaching high performance at the expense of using human engineers to create features and rules that were particular to a certain domain. Currently, the biomedical data is soaring expeditiously and extracting the useful information can help to facilitate the appropriate diagnosis. Therefore, these systems are widely adopted in biomedical domain. However, the traditional rule-based, dictionary based and machine learning based methods suffer from computational complexity and out-of-vocabulary (OOV)issues Deep learning has recently been used in NER systems, achieving the state-of-the-art outcome. The present work proposes a novel deep learning based approach which uses Bidirectional Long Short Term (BiLSTM), Bidirectional Encoder Representation (BERT) and Conditional Random Field mode (CRF) model along with transfer learning and multi-tasking model to solve the OOV problem in biomedical domain. The transfer learning architecture uses shared and task specific layers to achieve the multi-task transfer learning task. The shared layer consists of lexicon encoder and transformer encoder followed by embedding vectors. Finally, we define a training loss function based on the BERT model. The proposed Multi-task TLBBC approach is compared with numerous prevailing methods. The proposed Multi-task TLBBC approach realizes average accuracy as 97.30%, 97.20%, 96.80% and 97.50% for NCBI, BC5CDR, JNLPBA, and s800 dataset, respectively.","doi":"10.1007\/s42979-024-02835-z","cleaned_title":"a deep learning based approach for biomedical named entity recognition using multitasking transfer learning with bilstm, bert and crf","cleaned_abstract":"the named entity recognition (ner) is a method for locating references to rigid designators in text that fall into well-established semantic categories like person, place, organisation, etc., many natural language e applications, like summarization of text, question\u2013answer models and machine translation, always include ner at their core. early ner systems were quite successful in reaching high performance at the expense of using human engineers to create features and rules that were particular to a certain domain. currently, the biomedical data is soaring expeditiously and extracting the useful information can help to facilitate the appropriate diagnosis. therefore, these systems are widely adopted in biomedical domain. however, the traditional rule-based, dictionary based and machine learning based methods suffer from computational complexity and out-of-vocabulary (oov)issues deep learning has recently been used in ner systems, achieving the state-of-the-art outcome. the present work proposes a novel deep learning based approach which uses bidirectional long short term (bilstm), bidirectional encoder representation (bert) and conditional random field mode (crf) model along with transfer learning and multi-tasking model to solve the oov problem in biomedical domain. the transfer learning architecture uses shared and task specific layers to achieve the multi-task transfer learning task. the shared layer consists of lexicon encoder and transformer encoder followed by embedding vectors. finally, we define a training loss function based on the bert model. the proposed multi-task tlbbc approach is compared with numerous prevailing methods. the proposed multi-task tlbbc approach realizes average accuracy as 97.30%, 97.20%, 96.80% and 97.50% for ncbi, bc5cdr, jnlpba, and s800 dataset, respectively.","key_phrases":["the named entity recognition","ner","a method","references","rigid designators","text","that","well-established semantic categories","person","place","organisation","many natural language e applications","summarization","text","question\u2013answer models","machine translation","ner","their core","early ner systems","high performance","the expense","human engineers","features","rules","that","a certain domain","the biomedical data","the useful information","the appropriate diagnosis","these systems","biomedical domain","the traditional rule-based, dictionary based and machine learning based methods","computational complexity","deep learning","ner systems","the-art","the present work","a novel deep learning based approach","which","bidirectional long short term","bilstm","bert","crf","transfer learning","multi-tasking model","the oov problem","biomedical domain","the transfer learning architecture","shared and task specific layers","the multi-task transfer learning task","the shared layer","lexicon encoder","transformer encoder","vectors","we","a training loss function","the bert model","the proposed multi-task tlbbc approach","numerous prevailing methods","the proposed multi-task tlbbc approach","average accuracy","97.30%","97.20%","96.80%","97.50%","bc5cdr","jnlpba","s800 dataset","ner","as 97.30%","97.20%","96.80%","97.50%"]},{"title":"Is deep learning good enough for software defect prediction?","authors":["Sushant Kumar Pandey","Arya Haldar","Anil Kumar Tripathi"],"abstract":"Due to high impact of internet technology and rapid change in software systems, it has been a tough challenge for us to detect software defects with high accuracy. Traditional software defect prediction research mainly concentrates on manually designing features (e.g., complexity metrics) and inputting them into machine learning classifiers to distinguish defective code. To gain high prediction accuracy, researchers have developed several deep learning or high computational models for software defect prediction. However, there are several critical conditions and theoretical problems in order to achieve better results. This article explores the investigation of SDP using two deep learning techniques, i.e., SqueezeNet and Bottleneck models. We employed seven different open-source datasets from NASA Repository to perform this comparative study. We use F-Measure as a performance evaluator and found that these methods statistically outperform eight state-of-the-art methods with mean F-Measure of 0.93 \u00b1 0.014 and 0.90 \u00b1 0.013, respectively. We found that these two methods are significantly more effective in terms of F-Measure over large- and moderate-size projects. But they are computationally expensive in terms of training time. As the size of projects is getting immense and sophisticated, such deep learning methods are worth applying.","doi":"10.1007\/s11334-023-00542-1","cleaned_title":"is deep learning good enough for software defect prediction?","cleaned_abstract":"due to high impact of internet technology and rapid change in software systems, it has been a tough challenge for us to detect software defects with high accuracy. traditional software defect prediction research mainly concentrates on manually designing features (e.g., complexity metrics) and inputting them into machine learning classifiers to distinguish defective code. to gain high prediction accuracy, researchers have developed several deep learning or high computational models for software defect prediction. however, there are several critical conditions and theoretical problems in order to achieve better results. this article explores the investigation of sdp using two deep learning techniques, i.e., squeezenet and bottleneck models. we employed seven different open-source datasets from nasa repository to perform this comparative study. we use f-measure as a performance evaluator and found that these methods statistically outperform eight state-of-the-art methods with mean f-measure of 0.93 \u00b1 0.014 and 0.90 \u00b1 0.013, respectively. we found that these two methods are significantly more effective in terms of f-measure over large- and moderate-size projects. but they are computationally expensive in terms of training time. as the size of projects is getting immense and sophisticated, such deep learning methods are worth applying.","key_phrases":["high impact","internet technology","rapid change","software systems","it","a tough challenge","us","software defects","high accuracy","traditional software defect prediction research","features","e.g., complexity metrics","them","machine learning classifiers","defective code","high prediction accuracy","researchers","several deep learning","high computational models","software defect prediction","several critical conditions","theoretical problems","order","better results","this article","the investigation","sdp","two deep learning techniques","i.e., squeezenet and bottleneck models","we","seven different open-source datasets","nasa repository","this comparative study","we","f-measure","a performance evaluator","these methods","the-art","mean f-measure","0.93 \u00b1","0.90 \u00b1","we","these two methods","terms","f-measure","large- and moderate-size projects","they","terms","training time","the size","projects","such deep learning methods","two","seven","nasa","eight","0.93","0.014","0.90","0.013","two"]},{"title":"Forecasting oil price in times of crisis: a new evidence from machine learning versus deep learning models","authors":["Haithem Awijen","Hachmi Ben Ameur","Zied Ftiti","Wa\u00ebl Louhichi"],"abstract":"This study investigates oil price forecasting during a time of crisis, from December 2007 to December 2021. As the oil market has experienced various shocks (exogenous versus endogenous), modelling and forecasting its prices dynamics become more complex based on conventional (econometric and structural) models. A new strand of literature has been attracting more attention during the last decades dealing with artificial intelligence methods. However, this literature is unanimous regarding the performance accuracy between machine learning and deep learning methods. We aim in this study to contribute to this literature by investigating the oil price forecasting based on these two approaches. Based on the stylized facts of oil prices dynamics, we select the support vector machine and long short-term memory approach, as two main models of Machine Learning and deep learning methods, respectively. Our findings support the superiority of the Deep Learning method compared to the Machine Learning approach. Interestingly, our results show that the Deep LSTM-prediction has a close pattern to the observed oil prices, demonstrating robust fitting accuracy at mid-to-long forecast horizons during crisis events. However, our results show that SVM machine learning has poor memory ability to establish a clearer understanding of time-dependent volatility and the dynamic co-movements between actual and predicted data. Moreover, our results show that the power of SVM to learn for long-term predictions is reduced, which potentially lead to distortions of forecasting performance.","doi":"10.1007\/s10479-023-05400-8","cleaned_title":"forecasting oil price in times of crisis: a new evidence from machine learning versus deep learning models","cleaned_abstract":"this study investigates oil price forecasting during a time of crisis, from december 2007 to december 2021. as the oil market has experienced various shocks (exogenous versus endogenous), modelling and forecasting its prices dynamics become more complex based on conventional (econometric and structural) models. a new strand of literature has been attracting more attention during the last decades dealing with artificial intelligence methods. however, this literature is unanimous regarding the performance accuracy between machine learning and deep learning methods. we aim in this study to contribute to this literature by investigating the oil price forecasting based on these two approaches. based on the stylized facts of oil prices dynamics, we select the support vector machine and long short-term memory approach, as two main models of machine learning and deep learning methods, respectively. our findings support the superiority of the deep learning method compared to the machine learning approach. interestingly, our results show that the deep lstm-prediction has a close pattern to the observed oil prices, demonstrating robust fitting accuracy at mid-to-long forecast horizons during crisis events. however, our results show that svm machine learning has poor memory ability to establish a clearer understanding of time-dependent volatility and the dynamic co-movements between actual and predicted data. moreover, our results show that the power of svm to learn for long-term predictions is reduced, which potentially lead to distortions of forecasting performance.","key_phrases":["this study","oil price forecasting","a time","crisis","december","december","the oil market","various shocks","its prices dynamics","conventional (econometric and structural) models","a new strand","literature","more attention","the last decades","artificial intelligence methods","this literature","the performance accuracy","machine learning","deep learning methods","we","this study","this literature","the oil price forecasting","these two approaches","the stylized facts","we","the support vector machine","long short-term memory approach","two main models","machine learning","deep learning methods","our findings","the superiority","the deep learning method","the machine learning approach","our results","the deep lstm-prediction","a close pattern","the observed oil prices","robust fitting accuracy","mid","forecast horizons","crisis events","our results","svm machine learning","poor memory ability","a clearer understanding","time-dependent volatility","the dynamic co","-","movements","actual and predicted data","our results","the power","svm","long-term predictions","which","distortions","forecasting performance","december 2007 to december 2021","the last decades","two","two"]},{"title":"A machine learning based deep convective trigger for climate models","authors":["Siddharth Kumar","P Mukhopadhyay","C Balaji"],"abstract":"The present study focuses on addressing the issue of too frequent triggers of deep convection in climate models, which are primarily based on physics-based classical trigger functions such as convective available potential energy (CAPE) or cloud work function (CWF). To overcome this problem, the study proposes using machine learning (ML) based deep convective triggers as an alternative. The deep convective trigger is formulated as a binary classification problem, where the goal is to predict whether deep convection will occur or not. Two elementary classification algorithms, namely support vector machines and neural networks, are adopted in this study. Additionally, a novel method is proposed to rank the importance of input variables for the classification problem, which may aid in understanding the underlying mechanisms and factors influencing deep convection. The accuracy of the ML-based methods is compared with the widely used convective available potential energy (CAPE)-based and dynamic generation of CAPE (dCAPE) trigger function found in many convective parameterization schemes. Results demonstrate that the elementary machine learning-based algorithms can outperform the classical CAPE-based triggers, indicating the potential effectiveness of ML-based approaches in dealing with this issue. Furthermore, a method based on the Mahalanobis distance is presented for binary classification, which is easy to interpret and implement. The Mahalanobis distance-based approach shows accuracy comparable to other ML-based methods, suggesting its viability as an alternative method for deep convective triggers. By correcting for deep convective triggers using ML-based approaches, the study proposes a possible solution to improve the probability density of rain in the climate model. This improvement may help overcome the issue of excessive drizzle often observed in many climate models.","doi":"10.1007\/s00382-024-07332-w","cleaned_title":"a machine learning based deep convective trigger for climate models","cleaned_abstract":"the present study focuses on addressing the issue of too frequent triggers of deep convection in climate models, which are primarily based on physics-based classical trigger functions such as convective available potential energy (cape) or cloud work function (cwf). to overcome this problem, the study proposes using machine learning (ml) based deep convective triggers as an alternative. the deep convective trigger is formulated as a binary classification problem, where the goal is to predict whether deep convection will occur or not. two elementary classification algorithms, namely support vector machines and neural networks, are adopted in this study. additionally, a novel method is proposed to rank the importance of input variables for the classification problem, which may aid in understanding the underlying mechanisms and factors influencing deep convection. the accuracy of the ml-based methods is compared with the widely used convective available potential energy (cape)-based and dynamic generation of cape (dcape) trigger function found in many convective parameterization schemes. results demonstrate that the elementary machine learning-based algorithms can outperform the classical cape-based triggers, indicating the potential effectiveness of ml-based approaches in dealing with this issue. furthermore, a method based on the mahalanobis distance is presented for binary classification, which is easy to interpret and implement. the mahalanobis distance-based approach shows accuracy comparable to other ml-based methods, suggesting its viability as an alternative method for deep convective triggers. by correcting for deep convective triggers using ml-based approaches, the study proposes a possible solution to improve the probability density of rain in the climate model. this improvement may help overcome the issue of excessive drizzle often observed in many climate models.","key_phrases":["the present study","the issue","too frequent triggers","deep convection","climate models","which","physics-based classical trigger functions","convective available potential energy","cape","cloud work function","cwf","this problem","the study","machine learning","ml","deep convective triggers","an alternative","the deep convective trigger","a binary classification problem","the goal","deep convection","two elementary classification algorithms","vector machines","neural networks","this study","a novel method","the importance","input variables","the classification problem","which","the underlying mechanisms","factors","deep convection","the accuracy","the ml-based methods","the widely used convective available potential energy","cape)-based and dynamic generation","cape","dcape","trigger function","many convective parameterization schemes","results","the elementary machine learning-based algorithms","the classical cape-based triggers","the potential effectiveness","ml-based approaches","this issue","a method","the mahalanobis distance","binary classification","which","the mahalanobis distance-based approach","accuracy","other ml-based methods","its viability","an alternative method","deep convective triggers","deep convective triggers","ml-based approaches","the study","a possible solution","the probability density","rain","the climate model","this improvement","the issue","excessive drizzle","many climate models","cwf","two"]},{"title":"Deep learning theory of distribution regression with CNNs","authors":["Zhan Yu","Ding-Xuan Zhou"],"abstract":"We establish a deep learning theory for distribution regression with deep convolutional neural networks (DCNNs). Deep learning based on structured deep neural networks has been powerful in practical applications. Generalization analysis for regression with DCNNs has been carried out very recently. However, for the distribution regression problem in which the input variables are probability measures, there is no mathematical model or theoretical analysis of DCNN-based learning theory. One of the difficulties is that the classical neural network structure requires the input variable to be a Euclidean vector. When the input samples are probability distributions, the traditional neural network structure cannot be directly used. A well-defined DCNN framework for distribution regression is desirable. In this paper, we overcome the difficulty and establish a novel DCNN-based learning theory for a two-stage distribution regression model. Firstly, we realize an approximation theory for functionals defined on the set of Borel probability measures with the proposed DCNN framework. Then, we show that the hypothesis space is well-defined by rigorously proving its compactness. Furthermore, in the hypothesis space induced by the general DCNN framework with distribution inputs, by using a two-stage error decomposition technique, we derive a novel DCNN-based two-stage oracle inequality and optimal learning rates (up to a logarithmic factor) for the proposed algorithm for distribution regression.","doi":"10.1007\/s10444-023-10054-y","cleaned_title":"deep learning theory of distribution regression with cnns","cleaned_abstract":"we establish a deep learning theory for distribution regression with deep convolutional neural networks (dcnns). deep learning based on structured deep neural networks has been powerful in practical applications. generalization analysis for regression with dcnns has been carried out very recently. however, for the distribution regression problem in which the input variables are probability measures, there is no mathematical model or theoretical analysis of dcnn-based learning theory. one of the difficulties is that the classical neural network structure requires the input variable to be a euclidean vector. when the input samples are probability distributions, the traditional neural network structure cannot be directly used. a well-defined dcnn framework for distribution regression is desirable. in this paper, we overcome the difficulty and establish a novel dcnn-based learning theory for a two-stage distribution regression model. firstly, we realize an approximation theory for functionals defined on the set of borel probability measures with the proposed dcnn framework. then, we show that the hypothesis space is well-defined by rigorously proving its compactness. furthermore, in the hypothesis space induced by the general dcnn framework with distribution inputs, by using a two-stage error decomposition technique, we derive a novel dcnn-based two-stage oracle inequality and optimal learning rates (up to a logarithmic factor) for the proposed algorithm for distribution regression.","key_phrases":["we","a deep learning theory","distribution regression","deep convolutional neural networks","dcnns","deep learning","structured deep neural networks","practical applications","generalization analysis","regression","dcnns","the distribution regression problem","which","the input variables","probability measures","no mathematical model","theoretical analysis","dcnn-based learning theory","the difficulties","the classical neural network structure","the input variable","a euclidean vector","the input samples","probability distributions","the traditional neural network structure","a well-defined dcnn framework","distribution regression","this paper","we","the difficulty","a novel dcnn-based learning theory","a two-stage distribution regression model","we","an approximation theory","functionals","the set","borel probability measures","the proposed dcnn framework","we","the hypothesis space","its compactness","the hypothesis space","the general dcnn framework","distribution inputs","a two-stage error decomposition technique","we","a novel dcnn-based two-stage oracle inequality","optimal learning rates","a logarithmic factor","the proposed algorithm","distribution regression","one","two","firstly","two","two"]},{"title":"Deep learning for quadratic hedging in incomplete jump market","authors":["Nacira Agram","Bernt \u00d8ksendal","Jan Rems"],"abstract":"We propose a deep learning approach to study the minimal variance pricing and hedging problem in an incomplete jump diffusion market. It is based on a rigorous stochastic calculus derivation of the optimal hedging portfolio, optimal option price, and the corresponding equivalent martingale measure through the means of the Stackelberg game approach. A deep learning algorithm based on the combination of the feed-forward and LSTM neural networks is tested on three different market models, two of which are incomplete. In contrast, the complete market Black\u2013Scholes model serves as a benchmark for the algorithm\u2019s performance. The results that indicate the algorithm\u2019s good performance are presented and discussed. In particular, we apply our results to the special incomplete market model studied by Merton and give a detailed comparison between our results based on the minimal variance principle and the results obtained by Merton based on a different pricing principle. Using deep learning, we find that the minimal variance principle leads to typically higher option prices than those deduced from the Merton principle. On the other hand, the minimal variance principle leads to lower losses than the Merton principle.","doi":"10.1007\/s42521-024-00112-5","cleaned_title":"deep learning for quadratic hedging in incomplete jump market","cleaned_abstract":"we propose a deep learning approach to study the minimal variance pricing and hedging problem in an incomplete jump diffusion market. it is based on a rigorous stochastic calculus derivation of the optimal hedging portfolio, optimal option price, and the corresponding equivalent martingale measure through the means of the stackelberg game approach. a deep learning algorithm based on the combination of the feed-forward and lstm neural networks is tested on three different market models, two of which are incomplete. in contrast, the complete market black\u2013scholes model serves as a benchmark for the algorithm\u2019s performance. the results that indicate the algorithm\u2019s good performance are presented and discussed. in particular, we apply our results to the special incomplete market model studied by merton and give a detailed comparison between our results based on the minimal variance principle and the results obtained by merton based on a different pricing principle. using deep learning, we find that the minimal variance principle leads to typically higher option prices than those deduced from the merton principle. on the other hand, the minimal variance principle leads to lower losses than the merton principle.","key_phrases":["we","a deep learning approach","the minimal variance pricing","hedging problem","an incomplete jump diffusion market","it","a rigorous stochastic calculus derivation","the optimal hedging portfolio","optimal option price","the corresponding equivalent martingale measure","the means","the stackelberg game approach","a deep learning algorithm","the combination","the feed-forward","lstm neural networks","three different market models","which","contrast","the complete market black\u2013scholes model","a benchmark","the algorithm\u2019s performance","the results","that","the algorithm\u2019s good performance","we","our results","the special incomplete market model","merton","a detailed comparison","our results","the minimal variance principle","the results","merton","a different pricing principle","deep learning","we","the minimal variance principle","typically higher option prices","those","the merton principle","the other hand","the minimal variance principle","lower losses","the merton principle","three","two","merton","merton"]},{"title":"Structure-based, deep-learning models for protein-ligand binding affinity prediction","authors":["Debby D. Wang","Wenhui Wu","Ran Wang"],"abstract":"The launch of AlphaFold series has brought deep-learning techniques into the molecular structural science. As another crucial problem, structure-based prediction of protein-ligand binding affinity urgently calls for advanced computational techniques. Is deep learning ready to decode this problem? Here we review mainstream structure-based, deep-learning approaches for this problem, focusing on molecular representations, learning architectures and model interpretability. A model taxonomy has been generated. To compensate for the lack of valid comparisons among those models, we realized and evaluated representatives from a uniform basis, with the advantages and shortcomings discussed. This review will potentially benefit structure-based drug discovery and related areas.Graphical Abstract","doi":"10.1186\/s13321-023-00795-9","cleaned_title":"structure-based, deep-learning models for protein-ligand binding affinity prediction","cleaned_abstract":"the launch of alphafold series has brought deep-learning techniques into the molecular structural science. as another crucial problem, structure-based prediction of protein-ligand binding affinity urgently calls for advanced computational techniques. is deep learning ready to decode this problem? here we review mainstream structure-based, deep-learning approaches for this problem, focusing on molecular representations, learning architectures and model interpretability. a model taxonomy has been generated. to compensate for the lack of valid comparisons among those models, we realized and evaluated representatives from a uniform basis, with the advantages and shortcomings discussed. this review will potentially benefit structure-based drug discovery and related areas.graphical abstract","key_phrases":["the launch","alphafold series","deep-learning techniques","the molecular structural science","another crucial problem","structure-based prediction","protein-ligand binding affinity","advanced computational techniques","this problem","we","mainstream structure-based, deep-learning approaches","this problem","molecular representations","architectures","model interpretability","a model taxonomy","the lack","valid comparisons","those models","we","evaluated representatives","a uniform basis","the advantages","shortcomings","this review","structure-based drug discovery","related areas.graphical abstract"]},{"title":"Deep learning in magnetic resonance enterography for Crohn\u2019s disease assessment: a systematic review","authors":["Ofir Brem","David Elisha","Eli Konen","Michal Amitai","Eyal Klang"],"abstract":"Crohn\u2019s disease (CD) poses significant morbidity, underscoring the need for effective, non-invasive inflammatory assessment using magnetic resonance enterography (MRE). This literature review evaluates recent publications on the role of deep learning in improving MRE for CD assessment. We searched MEDLINE\/PUBMED for studies that reported the use of deep learning algorithms for assessment of CD activity. The study was conducted according to the PRISMA guidelines. The risk of bias was evaluated using the QUADAS\u20102 tool. Five eligible studies, encompassing 468 subjects, were identified. Our study suggests that diverse deep learning applications, including image quality enhancement, bowel segmentation for disease burden quantification, and 3D reconstruction for surgical planning are useful and promising for CD assessment. However, most of the studies are preliminary, retrospective studies, and have a high risk of bias in at least one category. Future research is needed to assess how deep learning can impact CD patient diagnostics, particularly when considering the increasing integration of such models into hospital systems.","doi":"10.1007\/s00261-024-04326-4","cleaned_title":"deep learning in magnetic resonance enterography for crohn\u2019s disease assessment: a systematic review","cleaned_abstract":"crohn\u2019s disease (cd) poses significant morbidity, underscoring the need for effective, non-invasive inflammatory assessment using magnetic resonance enterography (mre). this literature review evaluates recent publications on the role of deep learning in improving mre for cd assessment. we searched medline\/pubmed for studies that reported the use of deep learning algorithms for assessment of cd activity. the study was conducted according to the prisma guidelines. the risk of bias was evaluated using the quadas\u20102 tool. five eligible studies, encompassing 468 subjects, were identified. our study suggests that diverse deep learning applications, including image quality enhancement, bowel segmentation for disease burden quantification, and 3d reconstruction for surgical planning are useful and promising for cd assessment. however, most of the studies are preliminary, retrospective studies, and have a high risk of bias in at least one category. future research is needed to assess how deep learning can impact cd patient diagnostics, particularly when considering the increasing integration of such models into hospital systems.","key_phrases":["crohn\u2019s disease","cd","significant morbidity","the need","effective, non-invasive inflammatory assessment","magnetic resonance enterography","mre","this literature review","recent publications","the role","deep learning","mre","cd assessment","we","medline","studies","that","the use","deep learning algorithms","assessment","cd activity","the study","the prisma guidelines","the risk","bias","the quadas\u20102 tool","five eligible studies","468 subjects","our study","diverse deep learning applications","image quality enhancement","bowel segmentation","disease burden quantification","3d reconstruction","surgical planning","cd assessment","the studies","preliminary, retrospective studies","a high risk","bias","at least one category","future research","how deep learning","cd patient diagnostics","the increasing integration","such models","hospital systems","crohn","five","468","3d","at least one"]},{"title":"Is deep learning good enough for software defect prediction?","authors":["Sushant Kumar Pandey","Arya Haldar","Anil Kumar Tripathi"],"abstract":"Due to high impact of internet technology and rapid change in software systems, it has been a tough challenge for us to detect software defects with high accuracy. Traditional software defect prediction research mainly concentrates on manually designing features (e.g., complexity metrics) and inputting them into machine learning classifiers to distinguish defective code. To gain high prediction accuracy, researchers have developed several deep learning or high computational models for software defect prediction. However, there are several critical conditions and theoretical problems in order to achieve better results. This article explores the investigation of SDP using two deep learning techniques, i.e., SqueezeNet and Bottleneck models. We employed seven different open-source datasets from NASA Repository to perform this comparative study. We use F-Measure as a performance evaluator and found that these methods statistically outperform eight state-of-the-art methods with mean F-Measure of 0.93 \u00b1 0.014 and 0.90 \u00b1 0.013, respectively. We found that these two methods are significantly more effective in terms of F-Measure over large- and moderate-size projects. But they are computationally expensive in terms of training time. As the size of projects is getting immense and sophisticated, such deep learning methods are worth applying.","doi":"10.1007\/s11334-023-00542-1","cleaned_title":"is deep learning good enough for software defect prediction?","cleaned_abstract":"due to high impact of internet technology and rapid change in software systems, it has been a tough challenge for us to detect software defects with high accuracy. traditional software defect prediction research mainly concentrates on manually designing features (e.g., complexity metrics) and inputting them into machine learning classifiers to distinguish defective code. to gain high prediction accuracy, researchers have developed several deep learning or high computational models for software defect prediction. however, there are several critical conditions and theoretical problems in order to achieve better results. this article explores the investigation of sdp using two deep learning techniques, i.e., squeezenet and bottleneck models. we employed seven different open-source datasets from nasa repository to perform this comparative study. we use f-measure as a performance evaluator and found that these methods statistically outperform eight state-of-the-art methods with mean f-measure of 0.93 \u00b1 0.014 and 0.90 \u00b1 0.013, respectively. we found that these two methods are significantly more effective in terms of f-measure over large- and moderate-size projects. but they are computationally expensive in terms of training time. as the size of projects is getting immense and sophisticated, such deep learning methods are worth applying.","key_phrases":["high impact","internet technology","rapid change","software systems","it","a tough challenge","us","software defects","high accuracy","traditional software defect prediction research","features","e.g., complexity metrics","them","machine learning classifiers","defective code","high prediction accuracy","researchers","several deep learning","high computational models","software defect prediction","several critical conditions","theoretical problems","order","better results","this article","the investigation","sdp","two deep learning techniques","i.e., squeezenet and bottleneck models","we","seven different open-source datasets","nasa repository","this comparative study","we","f-measure","a performance evaluator","these methods","the-art","mean f-measure","0.93 \u00b1","0.90 \u00b1","we","these two methods","terms","f-measure","large- and moderate-size projects","they","terms","training time","the size","projects","such deep learning methods","two","seven","nasa","eight","0.93","0.014","0.90","0.013","two"]},{"title":"Forecasting oil price in times of crisis: a new evidence from machine learning versus deep learning models","authors":["Haithem Awijen","Hachmi Ben Ameur","Zied Ftiti","Wa\u00ebl Louhichi"],"abstract":"This study investigates oil price forecasting during a time of crisis, from December 2007 to December 2021. As the oil market has experienced various shocks (exogenous versus endogenous), modelling and forecasting its prices dynamics become more complex based on conventional (econometric and structural) models. A new strand of literature has been attracting more attention during the last decades dealing with artificial intelligence methods. However, this literature is unanimous regarding the performance accuracy between machine learning and deep learning methods. We aim in this study to contribute to this literature by investigating the oil price forecasting based on these two approaches. Based on the stylized facts of oil prices dynamics, we select the support vector machine and long short-term memory approach, as two main models of Machine Learning and deep learning methods, respectively. Our findings support the superiority of the Deep Learning method compared to the Machine Learning approach. Interestingly, our results show that the Deep LSTM-prediction has a close pattern to the observed oil prices, demonstrating robust fitting accuracy at mid-to-long forecast horizons during crisis events. However, our results show that SVM machine learning has poor memory ability to establish a clearer understanding of time-dependent volatility and the dynamic co-movements between actual and predicted data. Moreover, our results show that the power of SVM to learn for long-term predictions is reduced, which potentially lead to distortions of forecasting performance.","doi":"10.1007\/s10479-023-05400-8","cleaned_title":"forecasting oil price in times of crisis: a new evidence from machine learning versus deep learning models","cleaned_abstract":"this study investigates oil price forecasting during a time of crisis, from december 2007 to december 2021. as the oil market has experienced various shocks (exogenous versus endogenous), modelling and forecasting its prices dynamics become more complex based on conventional (econometric and structural) models. a new strand of literature has been attracting more attention during the last decades dealing with artificial intelligence methods. however, this literature is unanimous regarding the performance accuracy between machine learning and deep learning methods. we aim in this study to contribute to this literature by investigating the oil price forecasting based on these two approaches. based on the stylized facts of oil prices dynamics, we select the support vector machine and long short-term memory approach, as two main models of machine learning and deep learning methods, respectively. our findings support the superiority of the deep learning method compared to the machine learning approach. interestingly, our results show that the deep lstm-prediction has a close pattern to the observed oil prices, demonstrating robust fitting accuracy at mid-to-long forecast horizons during crisis events. however, our results show that svm machine learning has poor memory ability to establish a clearer understanding of time-dependent volatility and the dynamic co-movements between actual and predicted data. moreover, our results show that the power of svm to learn for long-term predictions is reduced, which potentially lead to distortions of forecasting performance.","key_phrases":["this study","oil price forecasting","a time","crisis","december","december","the oil market","various shocks","its prices dynamics","conventional (econometric and structural) models","a new strand","literature","more attention","the last decades","artificial intelligence methods","this literature","the performance accuracy","machine learning","deep learning methods","we","this study","this literature","the oil price forecasting","these two approaches","the stylized facts","we","the support vector machine","long short-term memory approach","two main models","machine learning","deep learning methods","our findings","the superiority","the deep learning method","the machine learning approach","our results","the deep lstm-prediction","a close pattern","the observed oil prices","robust fitting accuracy","mid","forecast horizons","crisis events","our results","svm machine learning","poor memory ability","a clearer understanding","time-dependent volatility","the dynamic co","-","movements","actual and predicted data","our results","the power","svm","long-term predictions","which","distortions","forecasting performance","december 2007 to december 2021","the last decades","two","two"]},{"title":"A machine learning based deep convective trigger for climate models","authors":["Siddharth Kumar","P Mukhopadhyay","C Balaji"],"abstract":"The present study focuses on addressing the issue of too frequent triggers of deep convection in climate models, which are primarily based on physics-based classical trigger functions such as convective available potential energy (CAPE) or cloud work function (CWF). To overcome this problem, the study proposes using machine learning (ML) based deep convective triggers as an alternative. The deep convective trigger is formulated as a binary classification problem, where the goal is to predict whether deep convection will occur or not. Two elementary classification algorithms, namely support vector machines and neural networks, are adopted in this study. Additionally, a novel method is proposed to rank the importance of input variables for the classification problem, which may aid in understanding the underlying mechanisms and factors influencing deep convection. The accuracy of the ML-based methods is compared with the widely used convective available potential energy (CAPE)-based and dynamic generation of CAPE (dCAPE) trigger function found in many convective parameterization schemes. Results demonstrate that the elementary machine learning-based algorithms can outperform the classical CAPE-based triggers, indicating the potential effectiveness of ML-based approaches in dealing with this issue. Furthermore, a method based on the Mahalanobis distance is presented for binary classification, which is easy to interpret and implement. The Mahalanobis distance-based approach shows accuracy comparable to other ML-based methods, suggesting its viability as an alternative method for deep convective triggers. By correcting for deep convective triggers using ML-based approaches, the study proposes a possible solution to improve the probability density of rain in the climate model. This improvement may help overcome the issue of excessive drizzle often observed in many climate models.","doi":"10.1007\/s00382-024-07332-w","cleaned_title":"a machine learning based deep convective trigger for climate models","cleaned_abstract":"the present study focuses on addressing the issue of too frequent triggers of deep convection in climate models, which are primarily based on physics-based classical trigger functions such as convective available potential energy (cape) or cloud work function (cwf). to overcome this problem, the study proposes using machine learning (ml) based deep convective triggers as an alternative. the deep convective trigger is formulated as a binary classification problem, where the goal is to predict whether deep convection will occur or not. two elementary classification algorithms, namely support vector machines and neural networks, are adopted in this study. additionally, a novel method is proposed to rank the importance of input variables for the classification problem, which may aid in understanding the underlying mechanisms and factors influencing deep convection. the accuracy of the ml-based methods is compared with the widely used convective available potential energy (cape)-based and dynamic generation of cape (dcape) trigger function found in many convective parameterization schemes. results demonstrate that the elementary machine learning-based algorithms can outperform the classical cape-based triggers, indicating the potential effectiveness of ml-based approaches in dealing with this issue. furthermore, a method based on the mahalanobis distance is presented for binary classification, which is easy to interpret and implement. the mahalanobis distance-based approach shows accuracy comparable to other ml-based methods, suggesting its viability as an alternative method for deep convective triggers. by correcting for deep convective triggers using ml-based approaches, the study proposes a possible solution to improve the probability density of rain in the climate model. this improvement may help overcome the issue of excessive drizzle often observed in many climate models.","key_phrases":["the present study","the issue","too frequent triggers","deep convection","climate models","which","physics-based classical trigger functions","convective available potential energy","cape","cloud work function","cwf","this problem","the study","machine learning","ml","deep convective triggers","an alternative","the deep convective trigger","a binary classification problem","the goal","deep convection","two elementary classification algorithms","vector machines","neural networks","this study","a novel method","the importance","input variables","the classification problem","which","the underlying mechanisms","factors","deep convection","the accuracy","the ml-based methods","the widely used convective available potential energy","cape)-based and dynamic generation","cape","dcape","trigger function","many convective parameterization schemes","results","the elementary machine learning-based algorithms","the classical cape-based triggers","the potential effectiveness","ml-based approaches","this issue","a method","the mahalanobis distance","binary classification","which","the mahalanobis distance-based approach","accuracy","other ml-based methods","its viability","an alternative method","deep convective triggers","deep convective triggers","ml-based approaches","the study","a possible solution","the probability density","rain","the climate model","this improvement","the issue","excessive drizzle","many climate models","cwf","two"]},{"title":"English\u2013Vietnamese Machine Translation Using Deep Learning for Chatbot Applications","authors":["Nguyen Minh Tuan","Phayung Meesad","Ha Huy Cuong Nguyen"],"abstract":"Recently, artificial intelligence-based machine translation has been much improved over traditional methods. A machine translator is very useful for translating text or speech from one language to another. Machine translators have replaced the word mechanism in one language for words in another with verbatim translations. However, a good translation should be employed as both a sentence and a word that has a completed meaning in accordance with the context of the relevant sentence. In this paper, we studied English\u2013Vietnamese translation using deep learning methods including recurrent neural network, long short-term memory, gated recurrent units, attention, and transformer. The deep learning-based machine translators were compared based on the test accuracy of the result translation. It was found that the best deep learning-based machine translator model was the Attention mechanism, and the Transformer yielded the second rank.","doi":"10.1007\/s42979-023-02339-2","cleaned_title":"english\u2013vietnamese machine translation using deep learning for chatbot applications","cleaned_abstract":"recently, artificial intelligence-based machine translation has been much improved over traditional methods. a machine translator is very useful for translating text or speech from one language to another. machine translators have replaced the word mechanism in one language for words in another with verbatim translations. however, a good translation should be employed as both a sentence and a word that has a completed meaning in accordance with the context of the relevant sentence. in this paper, we studied english\u2013vietnamese translation using deep learning methods including recurrent neural network, long short-term memory, gated recurrent units, attention, and transformer. the deep learning-based machine translators were compared based on the test accuracy of the result translation. it was found that the best deep learning-based machine translator model was the attention mechanism, and the transformer yielded the second rank.","key_phrases":["artificial intelligence-based machine translation","traditional methods","a machine translator","text","speech","one language","another","machine translators","the word mechanism","one language","words","another","verbatim translations","a good translation","both a sentence","a word","that","a completed meaning","accordance","the context","the relevant sentence","this paper","we","english\u2013vietnamese translation","deep learning methods","recurrent neural network","long short-term memory","gated recurrent units","attention","transformer","the deep learning-based machine translators","the test accuracy","the result translation","it","the best deep learning-based machine translator model","the attention mechanism","the transformer","the second rank","one language","english","vietnamese","second"]},{"title":"Role of machine learning and deep learning techniques in EEG-based BCI emotion recognition system: a review","authors":["Priyadarsini Samal","Mohammad Farukh Hashmi"],"abstract":"Emotion is a subjective psychophysiological reaction coming from external stimuli which impacts every aspect of our daily lives. Due to the continuing development of non-invasive and portable sensor technologies, such as brain-computer interfaces (BCI), intellectuals from several fields have been interested in emotion recognition techniques. Human emotions can be recognised using a variety of behavioural cues, including gestures and body language, voice, and physiological markers. The first three, however, might be ineffective because people sometimes conceal their genuine emotions either intentionally or unknowingly. More precise and objective emotion recognition can be accomplished using physiological signals. Among other physiological signals, Electroencephalogram (EEG) is more responsive and sensitive to variation in affective states. Various EEG-based emotion recognition methods have recently been introduced. This study reviews EEG-based BCIs for emotion identification and gives an outline of the progress made in this field. A summary of the datasets and techniques utilised to evoke human emotions and various emotion models\u00a0is also given. We discuss several EEG feature extractions, feature selection\/reduction, machine learning, and deep learning algorithms in accordance with standard emotional identification process. We provide an overview of the human brain's EEG rhythms, which are closely related to emotional states. We also go over a number of EEG-based emotion identification research and compare numerous machine learning and deep learning techniques. In conclusion, this study highlights the\u00a0applications,\u00a0challenges and potential areas for future research in identification and classification of human emotional states.","doi":"10.1007\/s10462-023-10690-2","cleaned_title":"role of machine learning and deep learning techniques in eeg-based bci emotion recognition system: a review","cleaned_abstract":"emotion is a subjective psychophysiological reaction coming from external stimuli which impacts every aspect of our daily lives. due to the continuing development of non-invasive and portable sensor technologies, such as brain-computer interfaces (bci), intellectuals from several fields have been interested in emotion recognition techniques. human emotions can be recognised using a variety of behavioural cues, including gestures and body language, voice, and physiological markers. the first three, however, might be ineffective because people sometimes conceal their genuine emotions either intentionally or unknowingly. more precise and objective emotion recognition can be accomplished using physiological signals. among other physiological signals, electroencephalogram (eeg) is more responsive and sensitive to variation in affective states. various eeg-based emotion recognition methods have recently been introduced. this study reviews eeg-based bcis for emotion identification and gives an outline of the progress made in this field. a summary of the datasets and techniques utilised to evoke human emotions and various emotion models is also given. we discuss several eeg feature extractions, feature selection\/reduction, machine learning, and deep learning algorithms in accordance with standard emotional identification process. we provide an overview of the human brain's eeg rhythms, which are closely related to emotional states. we also go over a number of eeg-based emotion identification research and compare numerous machine learning and deep learning techniques. in conclusion, this study highlights the applications, challenges and potential areas for future research in identification and classification of human emotional states.","key_phrases":["emotion","a subjective psychophysiological reaction","external stimuli","which","every aspect","our daily lives","the continuing development","non-invasive and portable sensor technologies","brain-computer interfaces","bci","intellectuals","several fields","emotion recognition techniques","human emotions","a variety","behavioural cues","gestures","body language","voice","physiological markers","people","their genuine emotions","more precise and objective emotion recognition","physiological signals","other physiological signals","electroencephalogram","eeg","variation","affective states","various eeg-based emotion recognition methods","this study","eeg-based bcis","emotion identification","an outline","the progress","this field","a summary","the datasets","techniques","human emotions","various emotion models","we","several eeg feature extractions","feature selection\/reduction","machine learning","deep learning algorithms","accordance","standard emotional identification process","we","an overview","the human brain's eeg rhythms","which","emotional states","we","a number","eeg-based emotion identification research","numerous machine learning","deep learning techniques","conclusion","this study","the applications","challenges","potential areas","future research","identification","classification","human emotional states","first","three"]},{"title":"Enhanced Hybrid Intrusion Detection System with Attention Mechanism using Deep Learning","authors":["Pundalik Chavan","H. Hanumanthappa","E. G. Satish","Sunil Manoli","S. Supreeth","S. Rohith","H. C. Ramaprasad"],"abstract":"The introduction of the Attention mechanism by the Internet of Things\u2014or WSN-IoT\u2014in the sector has greatly enhanced the intrusion detection mechanism capabilities, whereas the deep learning techniques, together with the attention mechanism, enhance the efficacy and efficiency of IDS within WSNs. The proposed \u201cEnhanced Hybrid Intrusion Detection System with Attention Mechanism\u201d (EHID-SCA) underlying insight of this is that the characteristics of WSN data validate this including Convolutional Neural Networks (CNNs) and other deep learning models into an effective and coherent architecture. The design of the deep hybrid network with attention incorporates Channel Attention and Spatial Attention as some of the main components. The model biased capacity is enlarged to choose and stress important spatial and channel information from the sensor input. The method is based on the consideration of these things and it cuts noise. Therefore, the proposed technique can be made to pave the way for enabling the Internet of Things intrusion detection systems to do the automatic extraction of useful information from sensor data through the utilization of deep learning along with the Attention mechanism. The attack might therefore be better situated for mitigation in the network design that allows analysis of the geographical and temporal context of events, which are then more properly termed as intrusion events.","doi":"10.1007\/s42979-024-02852-y","cleaned_title":"enhanced hybrid intrusion detection system with attention mechanism using deep learning","cleaned_abstract":"the introduction of the attention mechanism by the internet of things\u2014or wsn-iot\u2014in the sector has greatly enhanced the intrusion detection mechanism capabilities, whereas the deep learning techniques, together with the attention mechanism, enhance the efficacy and efficiency of ids within wsns. the proposed \u201cenhanced hybrid intrusion detection system with attention mechanism\u201d (ehid-sca) underlying insight of this is that the characteristics of wsn data validate this including convolutional neural networks (cnns) and other deep learning models into an effective and coherent architecture. the design of the deep hybrid network with attention incorporates channel attention and spatial attention as some of the main components. the model biased capacity is enlarged to choose and stress important spatial and channel information from the sensor input. the method is based on the consideration of these things and it cuts noise. therefore, the proposed technique can be made to pave the way for enabling the internet of things intrusion detection systems to do the automatic extraction of useful information from sensor data through the utilization of deep learning along with the attention mechanism. the attack might therefore be better situated for mitigation in the network design that allows analysis of the geographical and temporal context of events, which are then more properly termed as intrusion events.","key_phrases":["the introduction","the attention mechanism","the internet","things","wsn-iot","the sector","the intrusion detection mechanism capabilities","the deep learning techniques","the attention mechanism","the efficacy","efficiency","ids","wsns","the proposed \u201cenhanced hybrid intrusion detection system","attention mechanism","ehid-sca) underlying insight","this","the characteristics","wsn data","this","convolutional neural networks","cnns","other deep learning models","an effective and coherent architecture","the design","the deep hybrid network","attention","channel attention","spatial attention","some","the main components","the model biased capacity","important spatial and channel information","the sensor input","the method","the consideration","these things","it","noise","the proposed technique","the way","the internet","things intrusion detection systems","the automatic extraction","useful information","sensor data","the utilization","deep learning","the attention mechanism","the attack","mitigation","the network design","that","analysis","the geographical and temporal context","events","which","intrusion events"]},{"title":"Federal learning-based a dual-branch deep learning model for colon polyp segmentation","authors":["Xuguang Cao","Kefeng Fan","Huilin Ma"],"abstract":"The incidence of colon cancer occupies the top three places in gastrointestinal tumors, and colon polyps are an important causative factor in the development of colon cancer. Early screening for colon polyps and colon polypectomy can reduce the chances of colon cancer. The current means of colon polyp examination is through colonoscopy, taking images of the gastrointestinal tract, and then manually marking them manually, which is time-consuming and labor-intensive for doctors. Therefore, relying on advanced deep learning technology to automatically identify colon polyps in the gastrointestinal tract of the patient and segmenting the polyps is an important direction of research nowadays. Due to the privacy of medical data and the non-interoperability of disease information, this paper proposes a dual-branch colon polyp segmentation network based on federated learning, which makes it possible to achieve a better training effect under the guarantee of data independence, and secondly, the dual-branch colon polyp segmentation network proposed in this paper adopts the two different structures of convolutional neural network (CNN) and Transformer to form a dual-branch structure, and through layer-by-layer fusion embedding, the advantages between different structures are realized. In this paper, we also propose the Aggregated Attention Module (AAM) to preserve the high-dimensional semantic information and to complement the missing information in the lower layers. Ultimately our approach achieves state of the art in Kvasir-SEG and CVC-ClinicDB datasets.","doi":"10.1007\/s11042-024-19197-6","cleaned_title":"federal learning-based a dual-branch deep learning model for colon polyp segmentation","cleaned_abstract":"the incidence of colon cancer occupies the top three places in gastrointestinal tumors, and colon polyps are an important causative factor in the development of colon cancer. early screening for colon polyps and colon polypectomy can reduce the chances of colon cancer. the current means of colon polyp examination is through colonoscopy, taking images of the gastrointestinal tract, and then manually marking them manually, which is time-consuming and labor-intensive for doctors. therefore, relying on advanced deep learning technology to automatically identify colon polyps in the gastrointestinal tract of the patient and segmenting the polyps is an important direction of research nowadays. due to the privacy of medical data and the non-interoperability of disease information, this paper proposes a dual-branch colon polyp segmentation network based on federated learning, which makes it possible to achieve a better training effect under the guarantee of data independence, and secondly, the dual-branch colon polyp segmentation network proposed in this paper adopts the two different structures of convolutional neural network (cnn) and transformer to form a dual-branch structure, and through layer-by-layer fusion embedding, the advantages between different structures are realized. in this paper, we also propose the aggregated attention module (aam) to preserve the high-dimensional semantic information and to complement the missing information in the lower layers. ultimately our approach achieves state of the art in kvasir-seg and cvc-clinicdb datasets.","key_phrases":["the incidence","colon cancer","the top three places","gastrointestinal tumors","colon polyps","an important causative factor","the development","colon cancer","colon polyps","colon polypectomy","the chances","colon cancer","the current means","colon polyp examination","colonoscopy","images","the gastrointestinal tract","them","which","doctors","advanced deep learning technology","colon polyps","the gastrointestinal tract","the patient","the polyps","an important direction","research","the privacy","medical data","the non","-","interoperability","disease information","this paper","a dual-branch colon polyp segmentation network","federated learning","which","it","a better training effect","the guarantee","data independence","the dual-branch colon polyp segmentation network","this paper","the two different structures","convolutional neural network","cnn","a dual-branch structure","layer","the advantages","different structures","this paper","we","the aggregated attention module","aam","the high-dimensional semantic information","the missing information","the lower layers","our approach","state","the art","kvasir-seg and cvc-clinicdb datasets","three","secondly","two","cnn"]},{"title":"Improving Access Trust in Healthcare Through Multimodal Deep Learning for Affective Computing","authors":["I. Sakthidevi","G. Fathima"],"abstract":"In healthcare domain, access trust is of prime importance paramount to ensure effective delivery of medical services. It also fosters positive patient-provider relationships. With the advancement of technology, affective computing has emerged as a promising approach to enhance access trust. It enables systems to understand and respond to human emotions. The research work investigates the application of multimodal deep learning techniques in affective computing to improve access trust in healthcare environment. A novel algorithm, \"Belief-Emo-Fusion,\" is proposed, aiming to enhance the understanding and interpretation of emotions in healthcare. The research conducts a comprehensive simulation analysis, comparing the performance of Belief-Emo-Fusion with existing algorithms using simulation metrics: modal accuracy, \u0131nference time, and F1-score. The study emphasizes the importance of emotion recognition and understanding in healthcare settings. The work highlights the role of deep learning models in facilitating empathetic and emotionally intelligent technologies. By addressing the challenges associated with affective computing, the proposed approach contributes to the development of more effective and reliable healthcare systems. The findings offer valuable insights for researchers and practitioners seeking to leverage deep learning techniques for enhancing trust and communication in healthcare environments.","doi":"10.1007\/s44230-024-00080-4","cleaned_title":"improving access trust in healthcare through multimodal deep learning for affective computing","cleaned_abstract":"in healthcare domain, access trust is of prime importance paramount to ensure effective delivery of medical services. it also fosters positive patient-provider relationships. with the advancement of technology, affective computing has emerged as a promising approach to enhance access trust. it enables systems to understand and respond to human emotions. the research work investigates the application of multimodal deep learning techniques in affective computing to improve access trust in healthcare environment. a novel algorithm, \"belief-emo-fusion,\" is proposed, aiming to enhance the understanding and interpretation of emotions in healthcare. the research conducts a comprehensive simulation analysis, comparing the performance of belief-emo-fusion with existing algorithms using simulation metrics: modal accuracy, \u0131nference time, and f1-score. the study emphasizes the importance of emotion recognition and understanding in healthcare settings. the work highlights the role of deep learning models in facilitating empathetic and emotionally intelligent technologies. by addressing the challenges associated with affective computing, the proposed approach contributes to the development of more effective and reliable healthcare systems. the findings offer valuable insights for researchers and practitioners seeking to leverage deep learning techniques for enhancing trust and communication in healthcare environments.","key_phrases":["healthcare domain","access trust","prime importance","effective delivery","medical services","it","positive patient-provider relationships","the advancement","technology","affective computing","a promising approach","access trust","it","systems","human emotions","the research work","the application","techniques","affective computing","access trust","healthcare environment","a novel algorithm","\"belief-emo-fusion","the understanding","interpretation","emotions","healthcare","the research","a comprehensive simulation analysis","the performance","belief-emo-fusion","existing algorithms","simulation metrics","modal accuracy","\u0131nference time","f1-score","the study","the importance","emotion recognition","understanding","healthcare settings","the work","the role","deep learning models","empathetic and emotionally intelligent technologies","the challenges","affective computing","the proposed approach","the development","more effective and reliable healthcare systems","the findings","valuable insights","researchers","practitioners","deep learning techniques","trust","communication","healthcare environments"]},{"title":"Grapevine fruits disease detection using different deep learning models","authors":["Om G","Saketh Ram Billa","Vishal Malik","Eslavath Bharath","Sanjeev Sharma"],"abstract":"In India, grapes are one of the most important crops for business. Grapes and their byproducts are one of India\u2019s leading exports. The leaves of grapes are susceptible to a variety of diseases. Large-scale production of grapes can be affected by these diseases. The purpose of this research paper is to classify and detect disease on leaves using deep learning before taking a big picture. It describes the advances in detecting leaf disease and shows how to improve results. A total of seven different types of pre-trained deep Convolutional Neural Networks (CNNs) were used for transfer learning: MobileNet, InceptionResNetV2, DenseNet121, InceptionV3, Xception, VGG16 and ResNet101V2. There are 4062 images of grapevine leaves in total, divided into four classes: Leaf Blight, Black Rot, Black Measles, and Healthy. The major challenge was to detect if the leaf was healthy or had been infected, followed by classifying the type of disease of the leaf. In order to learn and detect the disease, these images were trained to seven different transfer learning models. The image classification accuracy obtained is 99.672% by DenseNet121 and it is the highest accuracy obtained compared to any accuracy reported in the literature. It is followed by 99.500% of Xception and 99.345% of VGG16, 99.345% of InceptionV3. Therefore, the proposed research paper can be useful for detecting Grapevine disease at an early stage and preventing its spread. This process can increase the production rate and profit for farmers.","doi":"10.1007\/s11042-024-19036-8","cleaned_title":"grapevine fruits disease detection using different deep learning models","cleaned_abstract":"in india, grapes are one of the most important crops for business. grapes and their byproducts are one of india\u2019s leading exports. the leaves of grapes are susceptible to a variety of diseases. large-scale production of grapes can be affected by these diseases. the purpose of this research paper is to classify and detect disease on leaves using deep learning before taking a big picture. it describes the advances in detecting leaf disease and shows how to improve results. a total of seven different types of pre-trained deep convolutional neural networks (cnns) were used for transfer learning: mobilenet, inceptionresnetv2, densenet121, inceptionv3, xception, vgg16 and resnet101v2. there are 4062 images of grapevine leaves in total, divided into four classes: leaf blight, black rot, black measles, and healthy. the major challenge was to detect if the leaf was healthy or had been infected, followed by classifying the type of disease of the leaf. in order to learn and detect the disease, these images were trained to seven different transfer learning models. the image classification accuracy obtained is 99.672% by densenet121 and it is the highest accuracy obtained compared to any accuracy reported in the literature. it is followed by 99.500% of xception and 99.345% of vgg16, 99.345% of inceptionv3. therefore, the proposed research paper can be useful for detecting grapevine disease at an early stage and preventing its spread. this process can increase the production rate and profit for farmers.","key_phrases":["india","grapes","the most important crops","business","grapes","their byproducts","india\u2019s leading exports","the leaves","grapes","a variety","diseases","large-scale production","grapes","these diseases","the purpose","this research paper","disease","leaves","deep learning","a big picture","it","the advances","leaf disease","results","a total","seven different types","pre-trained deep convolutional neural networks","cnns","transfer learning","mobilenet","inceptionresnetv2","densenet121","inceptionv3","xception","vgg16","resnet101v2","4062 images","grapevine leaves","four classes","leaf blight","black rot","black measles","the major challenge","the leaf","the type","disease","the leaf","order","the disease","these images","seven different transfer learning models","the image classification accuracy","99.672%","densenet121","it","the highest accuracy","any accuracy","the literature","it","99.500%","xception","99.345%","vgg16","99.345%","inceptionv3","the proposed research paper","grapevine disease","an early stage","its spread","this process","the production rate","profit","farmers","india","one","one","india","seven","inceptionresnetv2","inceptionv3","4062","four","seven","99.672%","99.500%","99.345%","99.345%","inceptionv3"]},{"title":"Deep learning in pulmonary nodule detection and segmentation: a systematic review","authors":["Chuan Gao","Linyu Wu","Wei Wu","Yichao Huang","Xinyue Wang","Zhichao Sun","Maosheng Xu","Chen Gao"],"abstract":"ObjectivesThe accurate detection and precise segmentation of lung nodules on computed tomography are key prerequisites for early diagnosis and appropriate treatment of lung cancer. This study was designed to compare detection and segmentation methods for pulmonary nodules using deep-learning techniques to fill methodological gaps and biases in the existing literature.MethodsThis study utilized a systematic review with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines, searching PubMed, Embase, Web of Science Core Collection, and the Cochrane Library databases up to May 10, 2023. The Quality Assessment of Diagnostic Accuracy Studies 2 criteria was used to assess the risk of bias and was adjusted with the Checklist for Artificial Intelligence in Medical Imaging. The study analyzed and extracted model performance, data sources, and task-focus information.ResultsAfter screening, we included nine studies meeting our inclusion criteria. These studies were published between 2019 and 2023 and predominantly used public datasets, with the Lung Image Database Consortium Image Collection and Image Database Resource Initiative and Lung Nodule Analysis 2016 being the most common. The studies focused on detection, segmentation, and other tasks, primarily utilizing Convolutional Neural Networks for model development. Performance evaluation covered multiple metrics, including sensitivity and the Dice coefficient.ConclusionsThis study highlights the potential power of deep learning in lung nodule detection and segmentation. It underscores the importance of standardized data processing, code and data sharing, the value of external test datasets, and the need to balance model complexity and efficiency in future research.Clinical relevance statementDeep learning demonstrates significant promise in autonomously detecting and segmenting pulmonary nodules. Future research should address methodological shortcomings and variability to enhance its clinical utility.Key Points\n\nDeep learning shows potential in the detection and segmentation of pulmonary nodules.\n\n\nThere are methodological gaps and biases present in the existing literature.\n\n\nFactors such as external validation and transparency affect the clinical application.","doi":"10.1007\/s00330-024-10907-0","cleaned_title":"deep learning in pulmonary nodule detection and segmentation: a systematic review","cleaned_abstract":"objectivesthe accurate detection and precise segmentation of lung nodules on computed tomography are key prerequisites for early diagnosis and appropriate treatment of lung cancer. this study was designed to compare detection and segmentation methods for pulmonary nodules using deep-learning techniques to fill methodological gaps and biases in the existing literature.methodsthis study utilized a systematic review with the preferred reporting items for systematic reviews and meta-analyses guidelines, searching pubmed, embase, web of science core collection, and the cochrane library databases up to may 10, 2023. the quality assessment of diagnostic accuracy studies 2 criteria was used to assess the risk of bias and was adjusted with the checklist for artificial intelligence in medical imaging. the study analyzed and extracted model performance, data sources, and task-focus information.resultsafter screening, we included nine studies meeting our inclusion criteria. these studies were published between 2019 and 2023 and predominantly used public datasets, with the lung image database consortium image collection and image database resource initiative and lung nodule analysis 2016 being the most common. the studies focused on detection, segmentation, and other tasks, primarily utilizing convolutional neural networks for model development. performance evaluation covered multiple metrics, including sensitivity and the dice coefficient.conclusionsthis study highlights the potential power of deep learning in lung nodule detection and segmentation. it underscores the importance of standardized data processing, code and data sharing, the value of external test datasets, and the need to balance model complexity and efficiency in future research.clinical relevance statementdeep learning demonstrates significant promise in autonomously detecting and segmenting pulmonary nodules. future research should address methodological shortcomings and variability to enhance its clinical utility.key points deep learning shows potential in the detection and segmentation of pulmonary nodules. there are methodological gaps and biases present in the existing literature. factors such as external validation and transparency affect the clinical application.","key_phrases":["objectivesthe accurate detection","precise segmentation","lung nodules","computed tomography","key prerequisites","early diagnosis","appropriate treatment","lung cancer","this study","detection and segmentation methods","pulmonary nodules","deep-learning techniques","methodological gaps","biases","the existing literature.methodsthis study","a systematic review","the preferred reporting items","systematic reviews","meta-analyses guidelines","embase, web","science core collection","the cochrane library","may","the quality assessment","diagnostic accuracy studies","2 criteria","the risk","bias","the checklist","artificial intelligence","medical imaging","the study","model performance","data sources","task-focus information.resultsafter screening","we","nine studies","our inclusion criteria","these studies","public datasets","the lung image database consortium image collection and image database resource initiative","lung","the studies","detection","segmentation","other tasks","convolutional neural networks","model development","performance evaluation","multiple metrics","sensitivity","the dice coefficient.conclusionsthis study","the potential power","deep learning","lung nodule detection","segmentation","it","the importance","standardized data processing","code","data","sharing","the value","external test datasets","the need","model complexity","efficiency","future research.clinical relevance statementdeep learning","significant promise","segmenting pulmonary nodules","future research","methodological shortcomings","variability","its clinical utility.key points","deep learning","potential","the detection","segmentation","pulmonary nodules","methodological gaps","biases","the existing literature","factors","external validation","transparency","the clinical application","2","nine","between 2019 and 2023","2016"]},{"title":"Emotion Detection-Based Video Recommendation System Using Machine Learning and Deep Learning Framework","authors":["Anuja Bokhare","Tripti Kothari"],"abstract":"Emotions play an important role in identification of the interest of user with respect to product, technology and other parameter. A human being has so many emotions which is used to express emotions such as anger, surprise, happiness, sadness and fear. In today's fast growing world of technology, there is necessity of expert system which helps to recognize the interest of user based on emotions. These systems would be helpful for various applications. Techniques such as machine learning or deep learning are the emerging ones\u2019 for the same. These are the technologies that have been used in applications like in automatic cars, for music recommendation and movie recommendation. Human emotions cannot be expressed but it can be detected by the help of techniques that is machine learning and deep learning. Current study emphases on human face detection and emotion detection to recommend video which motivates the person for changing the mood. Current system attention is on detection of face and emotion, using Haar cascade and deep face technology. Based on the outcome, an appropriate video is proposed.","doi":"10.1007\/s42979-022-01619-7","cleaned_title":"emotion detection-based video recommendation system using machine learning and deep learning framework","cleaned_abstract":"emotions play an important role in identification of the interest of user with respect to product, technology and other parameter. a human being has so many emotions which is used to express emotions such as anger, surprise, happiness, sadness and fear. in today's fast growing world of technology, there is necessity of expert system which helps to recognize the interest of user based on emotions. these systems would be helpful for various applications. techniques such as machine learning or deep learning are the emerging ones\u2019 for the same. these are the technologies that have been used in applications like in automatic cars, for music recommendation and movie recommendation. human emotions cannot be expressed but it can be detected by the help of techniques that is machine learning and deep learning. current study emphases on human face detection and emotion detection to recommend video which motivates the person for changing the mood. current system attention is on detection of face and emotion, using haar cascade and deep face technology. based on the outcome, an appropriate video is proposed.","key_phrases":["emotions","an important role","identification","the interest","user","respect","product","technology","other parameter","a human being","so many emotions","which","emotions","anger","surprise","happiness","sadness","fear","today's fast growing world","technology","necessity","expert system","which","the interest","user","emotions","these systems","various applications","techniques","machine learning","deep learning","the emerging ones","these","the technologies","that","applications","automatic cars","music recommendation and movie recommendation","human emotions","it","the help","techniques","that","machine learning","deep learning","current study","human face detection","emotion detection","video","which","the person","the mood","current system attention","detection","face","emotion","haar cascade","deep face technology","the outcome","an appropriate video","today"]},{"title":"Genome analysis through image processing with deep learning models","authors":["Yao-zhong Zhang","Seiya Imoto"],"abstract":"Genomic sequences are traditionally represented as strings of characters: A (adenine), C (cytosine), G (guanine), and T (thymine). However, an alternative approach involves depicting sequence-related information through image representations, such as Chaos Game Representation (CGR) and read pileup images. With rapid advancements in deep learning (DL) methods within computer vision and natural language processing, there is growing interest in applying image-based DL methods to genomic sequence analysis. These methods involve encoding genomic information as images or integrating spatial information from images into the analytical process. In this review, we summarize three typical applications that use image processing with DL models for genome analysis. We examine the utilization and advantages of these image-based approaches.","doi":"10.1038\/s10038-024-01275-0","cleaned_title":"genome analysis through image processing with deep learning models","cleaned_abstract":"genomic sequences are traditionally represented as strings of characters: a (adenine), c (cytosine), g (guanine), and t (thymine). however, an alternative approach involves depicting sequence-related information through image representations, such as chaos game representation (cgr) and read pileup images. with rapid advancements in deep learning (dl) methods within computer vision and natural language processing, there is growing interest in applying image-based dl methods to genomic sequence analysis. these methods involve encoding genomic information as images or integrating spatial information from images into the analytical process. in this review, we summarize three typical applications that use image processing with dl models for genome analysis. we examine the utilization and advantages of these image-based approaches.","key_phrases":["genomic sequences","strings","characters","a (adenine","c","cytosine","g","guanine","t","(thymine","an alternative approach","sequence-related information","image representations","chaos game representation","pileup images","rapid advancements","deep learning","computer vision","natural language processing","interest","image-based dl methods","genomic sequence analysis","these methods","genomic information","images","spatial information","images","the analytical process","this review","we","three typical applications","that","image processing","dl models","genome analysis","we","the utilization","advantages","these image-based approaches","three"]},{"title":"Deep learning based System for automatic motorcycle license plates detection and recognition","authors":["Abdolhossein Fathi","Babak Moradi","Iman Zarei","Afshin Shirbandi"],"abstract":"Nowadays, by increased the utilization of motorcycle the detection and recognition of its license plate play a very important role in intelligent transportation systems (ITS). ITS can be used for traffic control, violation monitoring, e-payment systems in the toll pay and parking. Several algorithms have been developed for this task and each of them has advantages and disadvantages under different circumstances and situations. By emerging deep learning based methods, they were employed to tackle the issue of automatic license plate detection and recognition. Since the deep learning models need a high volume of data for efficient training, and also each country has its license plate template, at first, it is crucial to collect proper dataset and then trains an efficient model on it. To this end, this research collected and introduced a new dataset, and then, designed a deep learning-based system for automatically detecting and identifying Iranian motorcycle license plates. At first, images that have different dimensions, angles, levels of lighting (daytime and nighttime images), were collected from various cities. Then two datasets for detection and identification are annotated and constructed from these images. Finally for implementing an efficient deep learning-based system, three networks YOLOv8, SSD, and Faster RCNN were investigated for detection and identification of license plates. The obtained results showed that the YOLOv8 network has the best result with 98.5% accuracy in the detection stage and 99% accuracy in the identification stage. The proposed YOLOv8 model was compared with deep learning-based methods and showed better performance on the collected dataset. The collected dataset and the source code of the investigated models are publicly available.","doi":"10.1007\/s11760-024-03514-5","cleaned_title":"deep learning based system for automatic motorcycle license plates detection and recognition","cleaned_abstract":"nowadays, by increased the utilization of motorcycle the detection and recognition of its license plate play a very important role in intelligent transportation systems (its). its can be used for traffic control, violation monitoring, e-payment systems in the toll pay and parking. several algorithms have been developed for this task and each of them has advantages and disadvantages under different circumstances and situations. by emerging deep learning based methods, they were employed to tackle the issue of automatic license plate detection and recognition. since the deep learning models need a high volume of data for efficient training, and also each country has its license plate template, at first, it is crucial to collect proper dataset and then trains an efficient model on it. to this end, this research collected and introduced a new dataset, and then, designed a deep learning-based system for automatically detecting and identifying iranian motorcycle license plates. at first, images that have different dimensions, angles, levels of lighting (daytime and nighttime images), were collected from various cities. then two datasets for detection and identification are annotated and constructed from these images. finally for implementing an efficient deep learning-based system, three networks yolov8, ssd, and faster rcnn were investigated for detection and identification of license plates. the obtained results showed that the yolov8 network has the best result with 98.5% accuracy in the detection stage and 99% accuracy in the identification stage. the proposed yolov8 model was compared with deep learning-based methods and showed better performance on the collected dataset. the collected dataset and the source code of the investigated models are publicly available.","key_phrases":["the utilization","motorcycle","the detection","recognition","its license plate","a very important role","intelligent transportation systems","its","its","traffic control","violation monitoring","e-payment systems","the toll pay","parking","several algorithms","this task","each","them","advantages","disadvantages","different circumstances","situations","deep learning based methods","they","the issue","automatic license plate detection","recognition","the deep learning models","a high volume","data","efficient training","each country","its license plate template","it","proper dataset","an efficient model","it","this end","this research","a new dataset","a deep learning-based system","iranian motorcycle license plates","images","that","different dimensions","angles","levels","lighting","daytime and nighttime images","various cities","two datasets","detection","identification","these images","an efficient deep learning-based system","three networks","ssd","faster rcnn","detection","identification","license plates","the obtained results","the yolov8 network","the best result","98.5% accuracy","the detection stage","99% accuracy","the identification stage","the proposed yolov8 model","deep learning-based methods","better performance","the collected dataset","the collected dataset","the source code","the investigated models","first","iranian","first","daytime","nighttime","two","three","yolov8","rcnn","yolov8","98.5%","99%","yolov8 model"]},{"title":"A Novel Approach Using Transfer Learning Architectural Models Based Deep Learning Techniques for Identification and Classification of Malignant Skin Cancer","authors":["Balambigai Subramanian","Suresh Muthusamy","Kokilavani Thangaraj","Hitesh Panchal","Elavarasi Kasirajan","Abarna Marimuthu","Abinaya Ravi"],"abstract":"Melanoma, a form of skin cancer originating in melanocyte cells, poses a significant health risk, although it is less prevalent than other types of skin cancer. Its detection presents challenges, even under expert observation. To enhance the classification accuracy of skin lesions, a Deep Convolutional Neural Network, Visual Geometry Group model has been proposed. However, deep learning methods typically require substantial training time. To mitigate this, transfer learning techniques are employed, reducing training duration. Data sets sourced from the International Skin Imaging Collaboration are utilized to train the model within this proposed approach. Evaluation of classification performance involves metrics such as Accuracy, Positive Predictive Value, Negative Predictive Value, Specificity, and Sensitivity. The classifier\u2019s performance on test data is depicted through a confusion matrix. The introduction of transfer learning techniques into the Deep Convolutional Neural Network has resulted in an improved classification accuracy of 85%, compared to the 81% achieved by a standard Convolutional Neural Network.","doi":"10.1007\/s11277-024-11006-5","cleaned_title":"a novel approach using transfer learning architectural models based deep learning techniques for identification and classification of malignant skin cancer","cleaned_abstract":"melanoma, a form of skin cancer originating in melanocyte cells, poses a significant health risk, although it is less prevalent than other types of skin cancer. its detection presents challenges, even under expert observation. to enhance the classification accuracy of skin lesions, a deep convolutional neural network, visual geometry group model has been proposed. however, deep learning methods typically require substantial training time. to mitigate this, transfer learning techniques are employed, reducing training duration. data sets sourced from the international skin imaging collaboration are utilized to train the model within this proposed approach. evaluation of classification performance involves metrics such as accuracy, positive predictive value, negative predictive value, specificity, and sensitivity. the classifier\u2019s performance on test data is depicted through a confusion matrix. the introduction of transfer learning techniques into the deep convolutional neural network has resulted in an improved classification accuracy of 85%, compared to the 81% achieved by a standard convolutional neural network.","key_phrases":["melanoma","a form","melanocyte cells","a significant health risk","it","other types","skin cancer","its detection","challenges","expert observation","the classification accuracy","skin lesions","a deep convolutional neural network","visual geometry group model","deep learning methods","substantial training time","this","transfer learning techniques","training duration","data sets","the international skin imaging collaboration","the model","this proposed approach","evaluation","classification performance","metrics","accuracy","positive predictive value","negative predictive value","specificity","sensitivity","the classifier\u2019s performance","test data","a confusion matrix","the introduction","techniques","the deep convolutional neural network","an improved classification accuracy","85%","the 81%","a standard convolutional neural network","85%","the 81%"]},{"title":"An Adaptive Learning Rate Deep Learning Optimizer Using Long and Short-Term Gradients Based on G\u2013L Fractional-Order Derivative","authors":["Shuang Chen","Changlun Zhang","Haibing Mu"],"abstract":"Deep learning model is a multi-layered network structure, and the network parameters that evaluate the final performance of the model must be trained by a deep learning optimizer. In comparison to the mainstream optimizers that utilize integer-order derivatives reflecting only local information, fractional-order derivatives optimizers, which can capture global information, are gradually gaining attention. However, relying solely on the long-term estimated gradients computed from fractional-order derivatives while disregarding the influence of recent gradients on the optimization process can sometimes lead to issues such as local optima and slower optimization speeds. In this paper, we design an adaptive learning rate optimizer called AdaGL based on the Gr\u00fcnwald\u2013Letnikov (G\u2013L) fractional-order derivative. It changes the direction and step size of parameter updating dynamically according to the long-term and short-term gradients information, addressing the problem of falling into local minima or saddle points. To be specific, by utilizing the global memory of fractional-order calculus, we replace the gradient of parameter update with G\u2013L fractional-order approximated gradient, making better use of the long-term curvature information in the past. Furthermore, considering that the recent gradient information often impacts the optimization phase significantly, we propose a step size control coefficient to adjust the learning rate in real-time. To compare the performance of the proposed AdaGL with the current advanced optimizers, we conduct several different deep learning tasks, including image classification on CNNs, node classification and graph classification on GNNs, image generation on GANs, and language modeling on LSTM. Extensive experimental results demonstrate that AdaGL achieves stable and fast convergence, excellent accuracy, and good generalization performance.","doi":"10.1007\/s11063-024-11571-7","cleaned_title":"an adaptive learning rate deep learning optimizer using long and short-term gradients based on g\u2013l fractional-order derivative","cleaned_abstract":"deep learning model is a multi-layered network structure, and the network parameters that evaluate the final performance of the model must be trained by a deep learning optimizer. in comparison to the mainstream optimizers that utilize integer-order derivatives reflecting only local information, fractional-order derivatives optimizers, which can capture global information, are gradually gaining attention. however, relying solely on the long-term estimated gradients computed from fractional-order derivatives while disregarding the influence of recent gradients on the optimization process can sometimes lead to issues such as local optima and slower optimization speeds. in this paper, we design an adaptive learning rate optimizer called adagl based on the gr\u00fcnwald\u2013letnikov (g\u2013l) fractional-order derivative. it changes the direction and step size of parameter updating dynamically according to the long-term and short-term gradients information, addressing the problem of falling into local minima or saddle points. to be specific, by utilizing the global memory of fractional-order calculus, we replace the gradient of parameter update with g\u2013l fractional-order approximated gradient, making better use of the long-term curvature information in the past. furthermore, considering that the recent gradient information often impacts the optimization phase significantly, we propose a step size control coefficient to adjust the learning rate in real-time. to compare the performance of the proposed adagl with the current advanced optimizers, we conduct several different deep learning tasks, including image classification on cnns, node classification and graph classification on gnns, image generation on gans, and language modeling on lstm. extensive experimental results demonstrate that adagl achieves stable and fast convergence, excellent accuracy, and good generalization performance.","key_phrases":["deep learning model","a multi-layered network structure","the network parameters","that","the final performance","the model","a deep learning optimizer","comparison","the mainstream optimizers","that","integer-order derivatives","only local information, fractional-order derivatives optimizers","which","global information","attention","the long-term estimated gradients","fractional-order derivatives","the influence","recent gradients","the optimization process","issues","local optima","slower optimization speeds","this paper","we","an adaptive learning rate optimizer","adagl","the gr\u00fcnwald","\u2013letnikov (g\u2013l) fractional-order derivative","it","the direction","step size","parameter","the long-term and short-term gradients information","the problem","local minima or saddle points","the global memory","fractional-order calculus","we","the gradient","parameter update","g","fractional-order","gradient","better use","the long-term curvature information","the past","the recent gradient information","the optimization phase","we","a step size control","coefficient","the learning rate","real-time","the performance","the proposed adagl","the current advanced optimizers","we","several different deep learning tasks","image classification","cnns","node classification","graph classification","gnns","image generation","gans","language modeling","lstm","extensive experimental results","adagl","stable and fast convergence","excellent accuracy","good generalization performance","adagl","adagl"]},{"title":"Deep learning-based point cloud upsampling: a review of recent trends","authors":["Soonjo Kwon","Ji-Hyeon Hur","Hyungki Kim"],"abstract":"Point clouds are acquired primarily using 3D scanners and are used for product inspection and reverse engineering. The quality of the point cloud varies depending on the scanning environment and scanner specifications. The quality of the point cloud has a significant impact on the accuracy of automatic or manual modeling. In response, various point cloud post-processing technologies are being developed. Point cloud upsampling is a technique to improve the resolution of point clouds, and the purpose of upsampling is to generate additional points to express the target object more accurately and in higher detail. This technology is important in areas where high-resolution 3D representation is required, and approaches based on deep learning have been recently gaining attention. Deep learning-based point cloud upsampling research can be classified as surface consolidation or edge consolidation research depending on the target regions to be consolidated, and as supervised or self-supervised learning depending on the type of learning approaches. This study examines the latest research trends in deep learning-based point cloud sampling, analyzes the issues and limitations of each research category, and proposes future research directions.Graphical abstract","doi":"10.1007\/s42791-023-00058-6","cleaned_title":"deep learning-based point cloud upsampling: a review of recent trends","cleaned_abstract":"point clouds are acquired primarily using 3d scanners and are used for product inspection and reverse engineering. the quality of the point cloud varies depending on the scanning environment and scanner specifications. the quality of the point cloud has a significant impact on the accuracy of automatic or manual modeling. in response, various point cloud post-processing technologies are being developed. point cloud upsampling is a technique to improve the resolution of point clouds, and the purpose of upsampling is to generate additional points to express the target object more accurately and in higher detail. this technology is important in areas where high-resolution 3d representation is required, and approaches based on deep learning have been recently gaining attention. deep learning-based point cloud upsampling research can be classified as surface consolidation or edge consolidation research depending on the target regions to be consolidated, and as supervised or self-supervised learning depending on the type of learning approaches. this study examines the latest research trends in deep learning-based point cloud sampling, analyzes the issues and limitations of each research category, and proposes future research directions.graphical abstract","key_phrases":["point clouds","3d scanners","product inspection","engineering","the quality","the point cloud","the scanning environment","scanner specifications","the quality","the point cloud","a significant impact","the accuracy","automatic or manual modeling","response","various point cloud post-processing technologies","point cloud upsampling","a technique","the resolution","point clouds","the purpose","upsampling","additional points","the target object","higher detail","this technology","areas","high-resolution 3d representation","approaches","deep learning","attention","deep learning-based point cloud upsampling research","surface consolidation","edge consolidation research","the target regions","supervised or self-supervised learning","the type","approaches","this study","the latest research trends","deep learning-based point cloud sampling","the issues","limitations","each research category","future research","directions.graphical abstract","3d","3d"]},{"title":"Determination of scintillation pixel location through deep learning using a two-layer DOI detector","authors":["Byungdu Jo","Seung-Jae Lee"],"abstract":"Small gantries and long, thin scintillation pixels are used in preclinical positron emission tomography, resulting in parallax errors outside the system\u2019s field of view. To solve this problem, a detector for measuring the depth of interaction (DOI) was developed. In addition, conduct of research on methods for DOI measurement through deep learning is underway. In this study, we designed a detector for measurement of DOI, consisting of two layers of scintillation pixel arrays and developed a method for specifying 3-dimensional (3D) position through deep learning. DETECT2000 simulation was performed to assess the 3D-positioning accuracy of the designed detector. Data acquired through DETECT2000 simulation wereused for learning a deep learning model, and assessment of location specification accuracy was performed using data generated at a new location and the deep learning model. According to the result, the 3D-position measurement accuracy was calculated as 94.48% on average.","doi":"10.1007\/s40042-024-01134-3","cleaned_title":"determination of scintillation pixel location through deep learning using a two-layer doi detector","cleaned_abstract":"small gantries and long, thin scintillation pixels are used in preclinical positron emission tomography, resulting in parallax errors outside the system\u2019s field of view. to solve this problem, a detector for measuring the depth of interaction (doi) was developed. in addition, conduct of research on methods for doi measurement through deep learning is underway. in this study, we designed a detector for measurement of doi, consisting of two layers of scintillation pixel arrays and developed a method for specifying 3-dimensional (3d) position through deep learning. detect2000 simulation was performed to assess the 3d-positioning accuracy of the designed detector. data acquired through detect2000 simulation wereused for learning a deep learning model, and assessment of location specification accuracy was performed using data generated at a new location and the deep learning model. according to the result, the 3d-position measurement accuracy was calculated as 94.48% on average.","key_phrases":["small gantries","long, thin scintillation pixels","preclinical positron emission tomography","parallax errors","the system\u2019s field","view","this problem","a detector","the depth","interaction","doi","addition","conduct","research","methods","doi measurement","deep learning","this study","we","a detector","measurement","doi","two layers","scintillation pixel arrays","a method","3-dimensional (3d) position","deep learning","detect2000 simulation","the 3d-positioning accuracy","the designed detector","data","detect2000 simulation","a deep learning model","assessment","location specification accuracy","data","a new location","the deep learning model","the result","the 3d-position measurement accuracy","94.48%","two","3","3d","3d","detect2000","3d","94.48%"]},{"title":"Sentiment analysis using a deep ensemble learning model","authors":["Muhammet Sinan Ba\u015farslan","Fatih Kayaalp"],"abstract":"The coronavirus pandemic has kept people away from social life and this has led to an increase in the use of social media over the past two years. Thanks to social media, people can now instantly share their thoughts on various topics such as their favourite movies, restaurants, hotels, etc. This has created a huge amount of data and many researchers from different sciences have focused on analysing this data. Natural Language Processing (NLP) is one of these areas of computer science that uses artificial technologies. Sentiment analysis is also one of the tasks of NLP, which is based on extracting emotions from huge post data. In this study, sentiment analysis was performed on two datasets of tweets about coronavirus and TripAdvisor hotel reviews. A frequency-based word representation method (Term Frequency-Inverse Document Frequency (TF-IDF)) and a prediction-based Word2Vec word embedding method were used to vectorise the datasets. Sentiment analysis models were then built using single machine learning methods (Decision Trees-DT, K-Nearest Neighbour-KNN, Naive Bayes-NB and Support Vector Machine-SVM), single deep learning methods (Long Short Term Memory-LSTM, Recurrent Neural Network-RNN) and heterogeneous ensemble learning methods (Stacking and Majority Voting) based on these single machine learning and deep learning methods. Accuracy was used as a performance measure. The heterogeneous model with stacking (LSTM-RNN) has outperformed the other models with accuracy values of 0.864 on the coronavirus dataset and 0.898 on the Trip Advisor dataset and they have been evaluated as promising results when compared to the literature. It has been observed that the use of single methods as an ensemble gives better results, which is consistent with the literature, which is a step forward in the detection of sentiments through posts. Investigating the performance of heterogeneous ensemble learning models based on different algorithms in sentiment analysis tasks is planned as future work.","doi":"10.1007\/s11042-023-17278-6","cleaned_title":"sentiment analysis using a deep ensemble learning model","cleaned_abstract":"the coronavirus pandemic has kept people away from social life and this has led to an increase in the use of social media over the past two years. thanks to social media, people can now instantly share their thoughts on various topics such as their favourite movies, restaurants, hotels, etc. this has created a huge amount of data and many researchers from different sciences have focused on analysing this data. natural language processing (nlp) is one of these areas of computer science that uses artificial technologies. sentiment analysis is also one of the tasks of nlp, which is based on extracting emotions from huge post data. in this study, sentiment analysis was performed on two datasets of tweets about coronavirus and tripadvisor hotel reviews. a frequency-based word representation method (term frequency-inverse document frequency (tf-idf)) and a prediction-based word2vec word embedding method were used to vectorise the datasets. sentiment analysis models were then built using single machine learning methods (decision trees-dt, k-nearest neighbour-knn, naive bayes-nb and support vector machine-svm), single deep learning methods (long short term memory-lstm, recurrent neural network-rnn) and heterogeneous ensemble learning methods (stacking and majority voting) based on these single machine learning and deep learning methods. accuracy was used as a performance measure. the heterogeneous model with stacking (lstm-rnn) has outperformed the other models with accuracy values of 0.864 on the coronavirus dataset and 0.898 on the trip advisor dataset and they have been evaluated as promising results when compared to the literature. it has been observed that the use of single methods as an ensemble gives better results, which is consistent with the literature, which is a step forward in the detection of sentiments through posts. investigating the performance of heterogeneous ensemble learning models based on different algorithms in sentiment analysis tasks is planned as future work.","key_phrases":["people","social life","this","an increase","the use","social media","the past two years","social media","people","their thoughts","various topics","their favourite movies","restaurants","hotels","this","a huge amount","data","many researchers","different sciences","this data","natural language processing","nlp","these areas","computer science","that","artificial technologies","sentiment analysis","the tasks","nlp","which","emotions","huge post data","this study","sentiment analysis","two datasets","tweets","coronavirus and tripadvisor hotel reviews","a frequency-based word representation method","term frequency-inverse document frequency","tf-idf","a prediction-based word2vec word","embedding method","the datasets","sentiment analysis models","single machine learning methods","decision trees","dt","k-nearest neighbour-knn, naive bayes","vector machine-svm","single deep learning methods","long short term memory-lstm","recurrent neural network-rnn","heterogeneous ensemble learning methods","stacking and majority voting","these single machine learning","deep learning methods","accuracy","a performance measure","the heterogeneous model","lstm-rnn","the other models","accuracy values","the coronavirus dataset","the trip advisor dataset","they","results","the literature","it","the use","single methods","an ensemble","better results","which","the literature","which","a step","the detection","sentiments","posts","the performance","heterogeneous ensemble learning models","different algorithms","sentiment analysis tasks","future work","the past two years","two","accuracy","0.864","0.898"]},{"title":"Differential testing for machine learning: an analysis for classification algorithms beyond deep learning","authors":["Steffen Herbold","Steffen Tunkel"],"abstract":"Differential testing is a useful approach that uses different implementations of the same algorithms and compares the results for software testing. In recent years, this approach was successfully used for test campaigns of deep learning frameworks. There is little knowledge about the application of differential testing beyond deep learning. Within this article, we want to close this gap for classification algorithms. We conduct a case study using Scikit-learn, Weka, Spark MLlib, and Caret in which we identify the potential of differential testing by considering which algorithms are available in multiple frameworks, the feasibility by identifying pairs of algorithms that should exhibit the same behavior, and the effectiveness by executing tests for the identified pairs and analyzing the deviations. While we found a large potential for popular algorithms, the feasibility seems limited because, often, it is not possible to determine configurations that are the same in other frameworks. The execution of the feasible tests revealed that there is a large number of deviations for the scores and classes. Only a lenient approach based on statistical significance of classes does not lead to a huge amount of test failures. The potential of differential testing beyond deep learning seems limited for research into the quality of machine learning libraries. Practitioners may still use the approach if they have deep knowledge about implementations, especially if a coarse oracle that only considers significant differences of classes is sufficient.","doi":"10.1007\/s10664-022-10273-9","cleaned_title":"differential testing for machine learning: an analysis for classification algorithms beyond deep learning","cleaned_abstract":"differential testing is a useful approach that uses different implementations of the same algorithms and compares the results for software testing. in recent years, this approach was successfully used for test campaigns of deep learning frameworks. there is little knowledge about the application of differential testing beyond deep learning. within this article, we want to close this gap for classification algorithms. we conduct a case study using scikit-learn, weka, spark mllib, and caret in which we identify the potential of differential testing by considering which algorithms are available in multiple frameworks, the feasibility by identifying pairs of algorithms that should exhibit the same behavior, and the effectiveness by executing tests for the identified pairs and analyzing the deviations. while we found a large potential for popular algorithms, the feasibility seems limited because, often, it is not possible to determine configurations that are the same in other frameworks. the execution of the feasible tests revealed that there is a large number of deviations for the scores and classes. only a lenient approach based on statistical significance of classes does not lead to a huge amount of test failures. the potential of differential testing beyond deep learning seems limited for research into the quality of machine learning libraries. practitioners may still use the approach if they have deep knowledge about implementations, especially if a coarse oracle that only considers significant differences of classes is sufficient.","key_phrases":["differential testing","a useful approach","that","different implementations","the same algorithms","the results","software testing","recent years","this approach","test campaigns","deep learning frameworks","little knowledge","the application","differential testing","deep learning","this article","we","this gap","classification algorithms","we","a case study","scikit-learn","weka","spark mllib","which","we","the potential","differential testing","which algorithms","multiple frameworks","the feasibility","pairs","algorithms","that","the same behavior","the effectiveness","tests","the identified pairs","the deviations","we","a large potential","popular algorithms","the feasibility","it","configurations","that","other frameworks","the execution","the feasible tests","a large number","deviations","the scores","classes","only a lenient approach","statistical significance","classes","a huge amount","test failures","the potential","differential testing","deep learning","research","the quality","machine learning libraries","practitioners","the approach","they","deep knowledge","implementations","that","significant differences","classes","recent years","weka","spark mllib"]},{"title":"Deep encoder\u2013decoder-based shared learning for multi-criteria recommendation systems","authors":["Salam Fraihat","Bushra Abu Tahon","Bushra Alhijawi","Arafat Awajan"],"abstract":"A recommendation system (RS) can help overcome information overload issues by offering personalized predictions for users. Typically, RS considers the overall ratings of users on items to generate recommendations for them. However, users may consider several aspects when evaluating items. Hence, a multi-criteria RS considers n-aspects of items to generate more accurate recommendations than a single-criteria RS. This research paper proposes two deep encoder\u2013decoder models based on shared learning for a multi-criteria RS, multi-modal deep encoder\u2013decoder-based shared learning (MMEDSL) and multi-criteria deep encoder\u2013decoder-based shared learning (MCEDSL). MMEDSL employs the shared learning technique by concentrating on the multi-modality concept in deep learning, while MCEDSL focuses on the training process to apply the shared learning technique. The shared learning captures useful shared information during the learning process since the multi-criteria may have hidden inter-relationships. A set of experiments were conducted to compare the proposed models with recent baseline approaches. The Yahoo! Movies multi-criteria dataset was utilized. The results demonstrate that the proposed models outperform other algorithms. In addition, the results show that integrating the shared learning technique with the RS produces precise recommendation predictions.","doi":"10.1007\/s00521-023-09007-9","cleaned_title":"deep encoder\u2013decoder-based shared learning for multi-criteria recommendation systems","cleaned_abstract":"a recommendation system (rs) can help overcome information overload issues by offering personalized predictions for users. typically, rs considers the overall ratings of users on items to generate recommendations for them. however, users may consider several aspects when evaluating items. hence, a multi-criteria rs considers n-aspects of items to generate more accurate recommendations than a single-criteria rs. this research paper proposes two deep encoder\u2013decoder models based on shared learning for a multi-criteria rs, multi-modal deep encoder\u2013decoder-based shared learning (mmedsl) and multi-criteria deep encoder\u2013decoder-based shared learning (mcedsl). mmedsl employs the shared learning technique by concentrating on the multi-modality concept in deep learning, while mcedsl focuses on the training process to apply the shared learning technique. the shared learning captures useful shared information during the learning process since the multi-criteria may have hidden inter-relationships. a set of experiments were conducted to compare the proposed models with recent baseline approaches. the yahoo! movies multi-criteria dataset was utilized. the results demonstrate that the proposed models outperform other algorithms. in addition, the results show that integrating the shared learning technique with the rs produces precise recommendation predictions.","key_phrases":["a recommendation system","information overload issues","personalized predictions","users","rs","the overall ratings","users","items","recommendations","them","users","several aspects","items","a multi-criteria rs","n","-aspects","items","more accurate recommendations","a single-criteria rs","this research paper","two deep encoder","decoder models","shared learning","a multi-criteria rs, multi-modal deep encoder","decoder-based shared learning","mmedsl","multi-criteria deep encoder","decoder-based shared learning","mcedsl","the shared learning technique","the multi-modality concept","deep learning","the training process","the shared learning technique","the shared learning captures","useful shared information","the learning process","-","criteria","inter","-","relationships","a set","experiments","the proposed models","recent baseline approaches","the yahoo","movies multi-criteria dataset","the results","the proposed models","other algorithms","addition","the results","the shared learning technique","the rs","precise recommendation predictions","two","yahoo"]},{"title":"Federated Deep Learning for Solving an Image Classification Problem on a Desktop Grid System","authors":["I. I. Kurochkin","A. I. Prun","A. A. Balaev"],"abstract":"AbstractThe paper considers the adaptation of federated deep learning on a desktop grid system using the example of an image classification problem. Restrictions are imposed on data transfer between the nodes of the desktop grid only for a part of the dataset. The implementation of federated deep learning on a desktop grid system based on the BOINC platform is considered. Methods for generating local datasets for desktop grid nodes are discussed. The results of numerical experiments are presented.","doi":"10.1134\/S1063779624030560","cleaned_title":"federated deep learning for solving an image classification problem on a desktop grid system","cleaned_abstract":"abstractthe paper considers the adaptation of federated deep learning on a desktop grid system using the example of an image classification problem. restrictions are imposed on data transfer between the nodes of the desktop grid only for a part of the dataset. the implementation of federated deep learning on a desktop grid system based on the boinc platform is considered. methods for generating local datasets for desktop grid nodes are discussed. the results of numerical experiments are presented.","key_phrases":["abstractthe paper","the adaptation","federated deep learning","a desktop grid system","the example","an image classification problem","restrictions","data transfer","the nodes","the desktop grid","a part","the dataset","the implementation","federated deep learning","a desktop grid system","the boinc platform","methods","local datasets","desktop grid nodes","the results","numerical experiments"]},{"title":"Machine and deep learning techniques for the prediction of diabetics: a review","authors":["Sandip Kumar Singh Modak","Vijay Kumar Jha"],"abstract":"Diabetes has become one of the significant reasons for public sickness and death in worldwide. By 2019, diabetes had affected more than 463 million people worldwide. According to the International Diabetes Federation report, this figure is expected to rise to more than 700 million in 2040, so early screening and diagnosis of diabetes patients have great significance in detecting and treating diabetes on time. Diabetes is a multi factorial metabolic disease, its diagnostic criteria are difficult to cover all the ethology, damage degree, pathogenesis and other factors, so there is a situation for uncertainty and imprecision under various aspects of the medical diagnosis process. With the development of Data mining, researchers find that machine learning and deep learning, playing an important role in diabetes prediction research. This paper is an in-depth study on the application of machine learning and deep learning techniques in the prediction of diabetics. In addition, this paper also discusses the different methodology used in machine and deep learning for prediction of diabetics since last two decades and examines the methods used, to explore their successes and failure. This review would help researchers and practitioners understand the current state-of-the-art methods and identify gaps in the literature.","doi":"10.1007\/s11042-024-19766-9","cleaned_title":"machine and deep learning techniques for the prediction of diabetics: a review","cleaned_abstract":"diabetes has become one of the significant reasons for public sickness and death in worldwide. by 2019, diabetes had affected more than 463 million people worldwide. according to the international diabetes federation report, this figure is expected to rise to more than 700 million in 2040, so early screening and diagnosis of diabetes patients have great significance in detecting and treating diabetes on time. diabetes is a multi factorial metabolic disease, its diagnostic criteria are difficult to cover all the ethology, damage degree, pathogenesis and other factors, so there is a situation for uncertainty and imprecision under various aspects of the medical diagnosis process. with the development of data mining, researchers find that machine learning and deep learning, playing an important role in diabetes prediction research. this paper is an in-depth study on the application of machine learning and deep learning techniques in the prediction of diabetics. in addition, this paper also discusses the different methodology used in machine and deep learning for prediction of diabetics since last two decades and examines the methods used, to explore their successes and failure. this review would help researchers and practitioners understand the current state-of-the-art methods and identify gaps in the literature.","key_phrases":["diabetes","the significant reasons","public sickness","death","diabetes","more than 463 million people","the international diabetes federation report","this figure","early screening","diagnosis","diabetes patients","great significance","diabetes","time","diabetes","a multi factorial metabolic disease","its diagnostic criteria","all the ethology","damage degree","pathogenesis","other factors","a situation","uncertainty","imprecision","various aspects","the medical diagnosis process","the development","data mining","researchers","an important role","diabetes prediction research","this paper","-depth","the application","machine learning","deep learning techniques","the prediction","diabetics","addition","this paper","the different methodology","machine","deep learning","prediction","diabetics","last two decades","the methods","their successes","failure","this review","researchers","practitioners","the-art","gaps","the literature","2019","more than 463 million","more than 700 million","2040","last two decades"]},{"title":"Deep learning-based PET image denoising and reconstruction: a review","authors":["Fumio Hashimoto","Yuya Onishi","Kibo Ote","Hideaki Tashima","Andrew J. Reader","Taiga Yamaya"],"abstract":"This review focuses on positron emission tomography (PET) imaging algorithms and traces the evolution of PET image reconstruction methods. First, we provide an overview of conventional PET image reconstruction methods from filtered backprojection through to recent iterative PET image reconstruction algorithms, and then review deep learning methods for PET data up to the latest innovations within three main categories. The first category involves post-processing methods for PET image denoising. The second category comprises direct image reconstruction methods that learn mappings from sinograms to the reconstructed images in an end-to-end manner. The third category comprises iterative reconstruction methods that combine conventional iterative image reconstruction with neural-network enhancement. We discuss future perspectives on PET imaging and deep learning technology.","doi":"10.1007\/s12194-024-00780-3","cleaned_title":"deep learning-based pet image denoising and reconstruction: a review","cleaned_abstract":"this review focuses on positron emission tomography (pet) imaging algorithms and traces the evolution of pet image reconstruction methods. first, we provide an overview of conventional pet image reconstruction methods from filtered backprojection through to recent iterative pet image reconstruction algorithms, and then review deep learning methods for pet data up to the latest innovations within three main categories. the first category involves post-processing methods for pet image denoising. the second category comprises direct image reconstruction methods that learn mappings from sinograms to the reconstructed images in an end-to-end manner. the third category comprises iterative reconstruction methods that combine conventional iterative image reconstruction with neural-network enhancement. we discuss future perspectives on pet imaging and deep learning technology.","key_phrases":["this review","positron emission tomography","pet","the evolution","pet image reconstruction methods","we","an overview","conventional pet image reconstruction methods","filtered backprojection","recent iterative pet image reconstruction algorithms","deep learning methods","pet data","the latest innovations","three main categories","the first category","post-processing methods","pet image denoising","the second category","direct image reconstruction methods","that","mappings","sinograms","the reconstructed images","end","the third category","iterative reconstruction methods","that","conventional iterative image reconstruction","neural-network enhancement","we","future perspectives","pet imaging","deep learning technology","first","three","first","second","third"]},{"title":"Deep learning algorithms for hedging with frictions","authors":["Xiaofei Shi","Daran Xu","Zhanhao Zhang"],"abstract":"This work studies the deep learning-based numerical algorithms for optimal hedging problems in markets with general convex transaction costs. Our main focus is on how these algorithms scale with the length of the trading time horizon. Based on the comparison results of the FBSDE solver by Han, Jentzen, and E (2018) and the Deep Hedging algorithm by Buehler, Gonon, Teichmann, and Wood (2019), we propose a Stable-Transfer Hedging (ST-Hedging) algorithm, to aggregate the convenience of the leading-order approximation formulas and the accuracy of the deep learning-based algorithms. Our ST-Hedging algorithm achieves the same state-of-the-art performance in short and moderately long time horizon as FBSDE solver and Deep Hedging, and generalize well to long time horizon when previous algorithms become suboptimal. With the transfer learning technique, ST-Hedging drastically reduce the training time, and shows great scalability to high-dimensional settings. This opens up new possibilities in model-based deep learning algorithms in economics, finance, and operational research, which takes advantage of the domain expert knowledge and the accuracy of the learning-based methods.","doi":"10.1007\/s42521-023-00075-z","cleaned_title":"deep learning algorithms for hedging with frictions","cleaned_abstract":"this work studies the deep learning-based numerical algorithms for optimal hedging problems in markets with general convex transaction costs. our main focus is on how these algorithms scale with the length of the trading time horizon. based on the comparison results of the fbsde solver by han, jentzen, and e (2018) and the deep hedging algorithm by buehler, gonon, teichmann, and wood (2019), we propose a stable-transfer hedging (st-hedging) algorithm, to aggregate the convenience of the leading-order approximation formulas and the accuracy of the deep learning-based algorithms. our st-hedging algorithm achieves the same state-of-the-art performance in short and moderately long time horizon as fbsde solver and deep hedging, and generalize well to long time horizon when previous algorithms become suboptimal. with the transfer learning technique, st-hedging drastically reduce the training time, and shows great scalability to high-dimensional settings. this opens up new possibilities in model-based deep learning algorithms in economics, finance, and operational research, which takes advantage of the domain expert knowledge and the accuracy of the learning-based methods.","key_phrases":["this work","the deep learning-based numerical algorithms","optimal hedging problems","markets","general convex transaction costs","our main focus","these algorithms","the length","the trading time horizon","the comparison results","the fbsde","han","jentzen","e","buehler","gonon","teichmann","wood","we","a stable-transfer hedging","st-hedging","the convenience","the leading-order approximation formulas","the accuracy","the deep learning-based algorithms","our st-hedging algorithm","the-art","short and moderately long time horizon","fbsde","long time horizon","previous algorithms","the transfer learning technique","st-hedging","the training time","great scalability","high-dimensional settings","this","new possibilities","model-based deep learning algorithms","economics","finance","operational research","which","advantage","the domain expert knowledge","the accuracy","the learning-based methods","han","2018","teichmann","2019"]},{"title":"Enhancing trash classification in smart cities using federated deep learning","authors":["Haroon Ahmed Khan","Syed Saud Naqvi","Abeer A. K. Alharbi","Salihah Alotaibi","Mohammed Alkhathami"],"abstract":"Efficient Waste management plays a crucial role to ensure clean and green environment in the smart cities. This study investigates the critical role of efficient trash classification in achieving sustainable solid waste management within smart city environments. We conduct a comparative analysis of various trash classification methods utilizing deep learning models built on convolutional neural networks (CNNs). Leveraging the PyTorch open-source framework and the TrashBox dataset, we perform experiments involving ten unique deep neural network models. Our approach aims to maximize training accuracy. Through extensive experimentation, we observe the consistent superiority of the ResNext-101 model compared to others, achieving exceptional training, validation, and test accuracies. These findings illuminate the potential of CNN-based techniques in significantly advancing trash classification for optimized solid waste management within smart city initiatives. Lastly, this study presents a distributed framework based on federated learning that can be used to optimize the performance of a combination of CNN models for trash detection.","doi":"10.1038\/s41598-024-62003-4","cleaned_title":"enhancing trash classification in smart cities using federated deep learning","cleaned_abstract":"efficient waste management plays a crucial role to ensure clean and green environment in the smart cities. this study investigates the critical role of efficient trash classification in achieving sustainable solid waste management within smart city environments. we conduct a comparative analysis of various trash classification methods utilizing deep learning models built on convolutional neural networks (cnns). leveraging the pytorch open-source framework and the trashbox dataset, we perform experiments involving ten unique deep neural network models. our approach aims to maximize training accuracy. through extensive experimentation, we observe the consistent superiority of the resnext-101 model compared to others, achieving exceptional training, validation, and test accuracies. these findings illuminate the potential of cnn-based techniques in significantly advancing trash classification for optimized solid waste management within smart city initiatives. lastly, this study presents a distributed framework based on federated learning that can be used to optimize the performance of a combination of cnn models for trash detection.","key_phrases":["efficient waste management","a crucial role","clean and green environment","the smart cities","this study","the critical role","efficient trash classification","sustainable solid waste management","smart city environments","we","a comparative analysis","various trash classification methods","deep learning models","convolutional neural networks","cnns","the pytorch open-source framework","the trashbox dataset","we","experiments","ten unique deep neural network models","our approach","training accuracy","extensive experimentation","we","the consistent superiority","the resnext-101 model","others","exceptional training","validation","test accuracies","these findings","the potential","cnn-based techniques","significantly advancing trash classification","optimized solid waste management","smart city initiatives","this study","a distributed framework","federated learning","that","the performance","a combination","cnn models","trash detection","smart city","cnn","smart city","cnn"]},{"title":"Interpreting and generalizing deep learning in physics-based problems with functional linear models","authors":["Amirhossein Arzani","Lingxiao Yuan","Pania Newell","Bei Wang"],"abstract":"Although deep learning has achieved remarkable success in various scientific machine learning applications, its opaque nature poses concerns regarding interpretability and generalization capabilities beyond the training data. Interpretability is crucial and often desired in modeling physical systems. Moreover, acquiring extensive datasets that encompass the entire range of input features is challenging in many physics-based learning tasks, leading to increased errors when encountering out-of-distribution (OOD) data. In this work, motivated by the field of functional data analysis (FDA), we propose generalized functional linear models as an interpretable surrogate for a trained deep learning model. We demonstrate that our model could be trained either based on a trained neural network (post-hoc interpretation) or directly from training data (interpretable operator learning). A library of generalized functional linear models with different kernel functions is considered and sparse regression is used to discover an interpretable surrogate model that could be analytically presented. We present test cases in solid mechanics, fluid mechanics, and transport. Our results demonstrate that our model can achieve comparable accuracy to deep learning and can improve OOD generalization while providing more transparency and interpretability. Our study underscores the significance of interpretable representation in scientific machine learning and showcases the potential of functional linear models as a tool for interpreting and generalizing deep learning.","doi":"10.1007\/s00366-024-01987-z","cleaned_title":"interpreting and generalizing deep learning in physics-based problems with functional linear models","cleaned_abstract":"although deep learning has achieved remarkable success in various scientific machine learning applications, its opaque nature poses concerns regarding interpretability and generalization capabilities beyond the training data. interpretability is crucial and often desired in modeling physical systems. moreover, acquiring extensive datasets that encompass the entire range of input features is challenging in many physics-based learning tasks, leading to increased errors when encountering out-of-distribution (ood) data. in this work, motivated by the field of functional data analysis (fda), we propose generalized functional linear models as an interpretable surrogate for a trained deep learning model. we demonstrate that our model could be trained either based on a trained neural network (post-hoc interpretation) or directly from training data (interpretable operator learning). a library of generalized functional linear models with different kernel functions is considered and sparse regression is used to discover an interpretable surrogate model that could be analytically presented. we present test cases in solid mechanics, fluid mechanics, and transport. our results demonstrate that our model can achieve comparable accuracy to deep learning and can improve ood generalization while providing more transparency and interpretability. our study underscores the significance of interpretable representation in scientific machine learning and showcases the potential of functional linear models as a tool for interpreting and generalizing deep learning.","key_phrases":["deep learning","remarkable success","various scientific machine learning applications","its opaque nature","concerns","interpretability and generalization capabilities","the training data","interpretability","physical systems","extensive datasets","that","the entire range","input features","many physics-based learning tasks","increased errors","distribution","this work","the field","functional data analysis","fda","we","generalized functional linear models","an interpretable surrogate","a trained deep learning model","we","our model","a trained neural network","post-hoc interpretation","training data","interpretable operator learning","a library","generalized functional linear models","different kernel functions","sparse regression","an interpretable surrogate model","that","we","test cases","solid mechanics","fluid mechanics","transport","our results","our model","comparable accuracy","deep learning","ood generalization","more transparency","interpretability","our study","the significance","interpretable representation","scientific machine learning","the potential","functional linear models","a tool","interpreting","deep learning","fda","linear","linear"]},{"title":"Can supervised deep learning architecture outperform autoencoders in building propensity score models for matching?","authors":["Mohammad Ehsanul Karim"],"abstract":"PurposePropensity score matching is vital in epidemiological studies using observational data, yet its estimates relies on correct model-specification. This study assesses supervised deep learning models and unsupervised autoencoders for propensity score estimation, comparing them with traditional methods for bias and variance accuracy in treatment effect estimations.MethodsUtilizing a plasmode simulation based on the Right Heart Catheterization dataset, under a variety of settings, we evaluated (1) a supervised deep learning architecture and (2) an unsupervised autoencoder, alongside two traditional methods: logistic regression and a spline-based method in estimating propensity scores for matching. Performance metrics included bias, standard errors, and coverage probability. The analysis was also extended to real-world data, with estimates compared to those obtained via a double robust approach.ResultsThe analysis revealed that supervised deep learning models outperformed unsupervised autoencoders in variance estimation while maintaining comparable levels of bias. These results were supported by analyses of real-world data, where the supervised model\u2019s estimates closely matched those derived from conventional methods. Additionally, deep learning models performed well compared to traditional methods in settings where exposure was rare.ConclusionSupervised deep learning models hold promise in refining propensity score estimations in epidemiological research, offering nuanced confounder adjustment, especially in complex datasets. We endorse integrating supervised deep learning into epidemiological research and share reproducible codes for widespread use and methodological transparency.","doi":"10.1186\/s12874-024-02284-5","cleaned_title":"can supervised deep learning architecture outperform autoencoders in building propensity score models for matching?","cleaned_abstract":"purposepropensity score matching is vital in epidemiological studies using observational data, yet its estimates relies on correct model-specification. this study assesses supervised deep learning models and unsupervised autoencoders for propensity score estimation, comparing them with traditional methods for bias and variance accuracy in treatment effect estimations.methodsutilizing a plasmode simulation based on the right heart catheterization dataset, under a variety of settings, we evaluated (1) a supervised deep learning architecture and (2) an unsupervised autoencoder, alongside two traditional methods: logistic regression and a spline-based method in estimating propensity scores for matching. performance metrics included bias, standard errors, and coverage probability. the analysis was also extended to real-world data, with estimates compared to those obtained via a double robust approach.resultsthe analysis revealed that supervised deep learning models outperformed unsupervised autoencoders in variance estimation while maintaining comparable levels of bias. these results were supported by analyses of real-world data, where the supervised model\u2019s estimates closely matched those derived from conventional methods. additionally, deep learning models performed well compared to traditional methods in settings where exposure was rare.conclusionsupervised deep learning models hold promise in refining propensity score estimations in epidemiological research, offering nuanced confounder adjustment, especially in complex datasets. we endorse integrating supervised deep learning into epidemiological research and share reproducible codes for widespread use and methodological transparency.","key_phrases":["purposepropensity score matching","epidemiological studies","observational data","its estimates","correct model-specification","this study","deep learning models","autoencoders","propensity score estimation","them","traditional methods","bias","variance accuracy","treatment effect","a plasmode simulation","the right heart catheterization dataset","a variety","settings","we","1) a supervised deep learning architecture","(2) an unsupervised autoencoder","two traditional methods","logistic regression","a spline-based method","propensity scores","performance metrics","bias","standard errors","coverage probability","the analysis","real-world data","estimates","those","a double robust","approach.resultsthe analysis","supervised deep learning models","unsupervised autoencoders","variance estimation","comparable levels","bias","these results","analyses","real-world data","the supervised model\u2019s estimates","those","conventional methods","deep learning models","traditional methods","settings","exposure","deep learning models","promise","propensity score estimations","epidemiological research","nuanced confounder adjustment","complex datasets","we","supervised deep learning","epidemiological research","reproducible codes","widespread use","methodological transparency","1","2","two"]},{"title":"Estimating infant age from skull X-ray images using deep learning","authors":["Heui Seung Lee","Jaewoong Kang","So Eui Kim","Ji Hee Kim","Bum-Joo Cho"],"abstract":"This study constructed deep learning models using plain skull radiograph images to predict the accurate postnatal age of infants under 12\u00a0months. Utilizing the results of the trained deep learning models, it aimed to evaluate the feasibility of employing major changes visible in skull X-ray images for assessing postnatal cranial development through gradient-weighted class activation mapping. We developed DenseNet-121 and EfficientNet-v2-M convolutional neural network models to analyze 4933 skull X-ray images collected from 1343 infants. Notably, allowing for a \u00b1\u20091\u00a0month error margin, DenseNet-121 reached a maximum corrected accuracy of 79.4% for anteroposterior (AP) views (average: 78.0\u2009\u00b1\u20091.5%) and 84.2% for lateral views (average: 81.1\u2009\u00b1\u20092.9%). EfficientNet-v2-M reached a maximum corrected accuracy 79.1% for AP views (average: 77.0\u2009\u00b1\u20092.3%) and 87.3% for lateral views (average: 85.1\u2009\u00b1\u20092.5%). Saliency maps identified critical discriminative areas in skull radiographs, including the coronal, sagittal, and metopic sutures in AP skull X-ray images, and the lambdoid suture and cortical bone density in lateral images, marking them as indicators for evaluating cranial development. These findings highlight the precision of deep learning in estimating infant age through non-invasive methods, offering the progress for clinical diagnostics and developmental assessment tools.","doi":"10.1038\/s41598-024-64489-4","cleaned_title":"estimating infant age from skull x-ray images using deep learning","cleaned_abstract":"this study constructed deep learning models using plain skull radiograph images to predict the accurate postnatal age of infants under 12 months. utilizing the results of the trained deep learning models, it aimed to evaluate the feasibility of employing major changes visible in skull x-ray images for assessing postnatal cranial development through gradient-weighted class activation mapping. we developed densenet-121 and efficientnet-v2-m convolutional neural network models to analyze 4933 skull x-ray images collected from 1343 infants. notably, allowing for a \u00b1 1 month error margin, densenet-121 reached a maximum corrected accuracy of 79.4% for anteroposterior (ap) views (average: 78.0 \u00b1 1.5%) and 84.2% for lateral views (average: 81.1 \u00b1 2.9%). efficientnet-v2-m reached a maximum corrected accuracy 79.1% for ap views (average: 77.0 \u00b1 2.3%) and 87.3% for lateral views (average: 85.1 \u00b1 2.5%). saliency maps identified critical discriminative areas in skull radiographs, including the coronal, sagittal, and metopic sutures in ap skull x-ray images, and the lambdoid suture and cortical bone density in lateral images, marking them as indicators for evaluating cranial development. these findings highlight the precision of deep learning in estimating infant age through non-invasive methods, offering the progress for clinical diagnostics and developmental assessment tools.","key_phrases":["this study","deep learning models","plain skull radiograph images","the accurate postnatal age","infants","12 months","the results","the trained deep learning models","it","the feasibility","major changes","skull x-ray images","postnatal cranial development","gradient-weighted class activation mapping","we","densenet-121 and efficientnet-v2-m","convolutional neural network models","4933 skull x-ray images","1343 infants","a \u00b1 1 month error margin","densenet-121","a maximum corrected accuracy","79.4%","anteroposterior (ap) views","1.5%","lateral views","81.1 \u00b1","2.9%","efficientnet-v2","m","a maximum corrected accuracy","ap views","2.3%","lateral views","2.5%","saliency maps","critical discriminative areas","the coronal, sagittal, and metopic sutures","ap skull x-ray images","the lambdoid suture","cortical bone density","lateral images","them","indicators","cranial development","these findings","the precision","deep learning","infant age","non-invasive methods","the progress","clinical diagnostics","developmental assessment tools","under 12 months","4933","1343","1 month","79.4%","78.0 \u00b1","1.5%","84.2%","81.1","2.9%","79.1%","77.0","2.3%","87.3%","85.1","2.5%"]},{"title":"Deep learning based damage detection of concrete structures","authors":["Maheswara Rao Bandi","Laxmi Narayana Pasupuleti","Tanmay Das","Shyamal Guchhait"],"abstract":"Damage detection of any civil engineering structure is a part of the interest in the field of engineering from which one can estimate the stability and lifetime of the structure. With the advancement of technology in the field of infrastructure, damage assessment of any structure with the help of convolutional neural networks (CNNs) and deep learning is gaining importance because of the ease with which it can detect the damage. By using these computer-aided techniques, we can reduce manpower and detect damage in inaccessible places that we cannot see directly. In the current study, we used ResNet-50, which is a part of a deep convolutional neural network with 50 layers deep in it. ResNet-50 is a subclass of convolutional neural networks most popularly used for image classification. Furthermore, we used the image data set collected from the Utah State University, Logan, Utah, USA. This data set contains nearly 56,000 annotated images of both cracked and non-cracked bridge deck images, wall images, and pavement images. These images contain a variety of obstructions, such as shadows and surface roughness in some cases. The present study aimed to train and test the images and compare the accuracy of the results among themselves with an increase in training data. We devised an algorithm to measure cracks as soon as we detected them. The results show that the Resnet-50 architecture was in good agreement with the developed algorithm.","doi":"10.1007\/s42107-024-01106-9","cleaned_title":"deep learning based damage detection of concrete structures","cleaned_abstract":"damage detection of any civil engineering structure is a part of the interest in the field of engineering from which one can estimate the stability and lifetime of the structure. with the advancement of technology in the field of infrastructure, damage assessment of any structure with the help of convolutional neural networks (cnns) and deep learning is gaining importance because of the ease with which it can detect the damage. by using these computer-aided techniques, we can reduce manpower and detect damage in inaccessible places that we cannot see directly. in the current study, we used resnet-50, which is a part of a deep convolutional neural network with 50 layers deep in it. resnet-50 is a subclass of convolutional neural networks most popularly used for image classification. furthermore, we used the image data set collected from the utah state university, logan, utah, usa. this data set contains nearly 56,000 annotated images of both cracked and non-cracked bridge deck images, wall images, and pavement images. these images contain a variety of obstructions, such as shadows and surface roughness in some cases. the present study aimed to train and test the images and compare the accuracy of the results among themselves with an increase in training data. we devised an algorithm to measure cracks as soon as we detected them. the results show that the resnet-50 architecture was in good agreement with the developed algorithm.","key_phrases":["damage detection","any civil engineering structure","a part","the interest","the field","engineering","which","one","the stability","lifetime","the structure","the advancement","technology","the field","infrastructure","damage assessment","any structure","the help","convolutional neural networks","cnns","deep learning","importance","the ease","which","it","the damage","these computer-aided techniques","we","manpower","damage","inaccessible places","that","we","the current study","we","resnet-50","which","a part","a deep convolutional neural network","50 layers","it","resnet-50","a subclass","convolutional neural networks","image classification","we","the utah state university","usa","this data","nearly 56,000 annotated images","both cracked and non-cracked bridge deck images","wall images","pavement images","these images","a variety","obstructions","shadows","surface roughness","some cases","the present study","the images","the accuracy","the results","themselves","an increase","training data","we","an algorithm","cracks","we","them","the results","the resnet-50 architecture","good agreement","the developed algorithm","resnet-50","50","resnet-50","the utah state university","utah","usa","nearly 56,000","resnet-50"]},{"title":"A Comparative Analysis of Deep Learning-Based Approaches for Classifying Dental Implants Decision Support System","authors":["Mohammed A. H. Lubbad","Ikbal Leblebicioglu Kurtulus","Dervis Karaboga","Kerem Kilic","Alper Basturk","Bahriye Akay","Ozkan Ufuk Nalbantoglu","Ozden Melis Durmaz Yilmaz","Mustafa Ayata","Serkan Yilmaz","Ishak Pacal"],"abstract":"This study aims to provide an effective solution for the autonomous identification of dental implant brands through a deep learning-based computer diagnostic system. It also seeks to ascertain the system\u2019s potential in clinical practices and to offer a strategic framework for improving diagnosis and treatment processes in implantology. This study employed a total of 28 different deep learning models, including 18 convolutional neural network (CNN) models (VGG, ResNet, DenseNet, EfficientNet, RegNet, ConvNeXt) and 10 vision transformer models (Swin and Vision Transformer). The dataset comprises 1258 panoramic radiographs from patients who received implant treatments at Erciyes University Faculty of Dentistry between 2012 and 2023. It is utilized for the training and evaluation process of deep learning models and consists of prototypes from six different implant systems provided by six manufacturers. The deep learning-based dental implant system provided high classification accuracy for different dental implant brands using deep learning models. Furthermore, among all the architectures evaluated, the small model of the ConvNeXt architecture achieved an impressive accuracy rate of 94.2%, demonstrating a high level of classification success.This study emphasizes the effectiveness of deep learning-based systems in achieving high classification accuracy in dental implant types. These findings pave the way for integrating advanced deep learning tools into clinical practice, promising significant improvements in patient care and treatment outcomes.","doi":"10.1007\/s10278-024-01086-x","cleaned_title":"a comparative analysis of deep learning-based approaches for classifying dental implants decision support system","cleaned_abstract":"this study aims to provide an effective solution for the autonomous identification of dental implant brands through a deep learning-based computer diagnostic system. it also seeks to ascertain the system\u2019s potential in clinical practices and to offer a strategic framework for improving diagnosis and treatment processes in implantology. this study employed a total of 28 different deep learning models, including 18 convolutional neural network (cnn) models (vgg, resnet, densenet, efficientnet, regnet, convnext) and 10 vision transformer models (swin and vision transformer). the dataset comprises 1258 panoramic radiographs from patients who received implant treatments at erciyes university faculty of dentistry between 2012 and 2023. it is utilized for the training and evaluation process of deep learning models and consists of prototypes from six different implant systems provided by six manufacturers. the deep learning-based dental implant system provided high classification accuracy for different dental implant brands using deep learning models. furthermore, among all the architectures evaluated, the small model of the convnext architecture achieved an impressive accuracy rate of 94.2%, demonstrating a high level of classification success.this study emphasizes the effectiveness of deep learning-based systems in achieving high classification accuracy in dental implant types. these findings pave the way for integrating advanced deep learning tools into clinical practice, promising significant improvements in patient care and treatment outcomes.","key_phrases":["this study","an effective solution","the autonomous identification","dental implant brands","a deep learning-based computer diagnostic system","it","the system\u2019s potential","clinical practices","a strategic framework","diagnosis and treatment processes","implantology","this study","a total","28 different deep learning models","18 convolutional neural network (cnn) models","vgg","resnet","densenet","efficientnet","regnet","convnext","10 vision transformer models","swin and vision transformer","the dataset","patients","who","implant treatments","erciyes university faculty","dentistry","it","the training and evaluation process","deep learning models","prototypes","six different implant systems","six manufacturers","the deep learning-based dental implant system","high classification accuracy","different dental implant brands","deep learning models","all the architectures","the small model","the convnext architecture","an impressive accuracy rate","94.2%","a high level","classification success.this study","the effectiveness","deep learning-based systems","high classification accuracy","dental implant types","these findings","the way","advanced deep learning tools","clinical practice","significant improvements","patient care and treatment outcomes","28","18","cnn","10","1258","erciyes university faculty of dentistry","between 2012 and 2023","six","six","94.2%"]},{"title":"A radiomics-boosted deep-learning for risk assessment of synchronous peritoneal metastasis in colorectal cancer","authors":["Ding Zhang","BingShu Zheng","LiuWei Xu","YiCong Wu","Chen Shen","ShanLei Bao","ZhongHua Tan","ChunFeng Sun"],"abstract":"ObjectivesSynchronous colorectal cancer peritoneal metastasis (CRPM) has a poor prognosis. This study aimed to create a radiomics-boosted deep learning model by PET\/CT image for risk assessment of synchronous CRPM.MethodsA total of 220 colorectal cancer (CRC) cases were enrolled in this study. We mapped the feature maps (Radiomic feature maps (RFMs)) of radiomic features across CT and PET image patches by a 2D sliding kernel. Based on ResNet50, a radiomics-boosted deep learning model was trained using PET\/CT image patches and RFMs. Besides that, we explored whether the peritumoral region contributes to the assessment of CRPM. In this study, the performance of each model was evaluated by the area under the curves (AUC).ResultsThe AUCs of the radiomics-boosted deep learning model in the training, internal, external, and all validation datasets were 0.926 (95% confidence interval (CI): 0.874\u20130.978), 0.897 (95% CI: 0.801\u20130.994), 0.885 (95% CI: 0.795\u20130.975), and 0.889 (95% CI: 0.823\u20130.954), respectively. This model exhibited consistency in the calibration curve, the Delong test and IDI identified it as the most predictive model.ConclusionsThe radiomics-boosted deep learning model showed superior estimated performance in preoperative prediction of synchronous CRPM from pre-treatment PET\/CT, offering potential assistance in the development of more personalized treatment methods and follow-up plans.Critical relevance statementThe onset of synchronous colorectal CRPM is insidious, and using a radiomics-boosted deep learning model to assess the risk of CRPM before treatment can help make personalized clinical treatment decisions or choose more sensitive follow-up plans.Key Points\n\nPrognosis for patients with CRPM is bleak, and early detection poses challenges.\n\n\nThe synergy between radiomics and deep learning proves advantageous in evaluating CRPM.\n\n\nThe radiomics-boosted deep-learning model proves valuable in tailoring treatment approaches for CRC patients.\n\nGraphical Abstract","doi":"10.1186\/s13244-024-01733-5","cleaned_title":"a radiomics-boosted deep-learning for risk assessment of synchronous peritoneal metastasis in colorectal cancer","cleaned_abstract":"objectivessynchronous colorectal cancer peritoneal metastasis (crpm) has a poor prognosis. this study aimed to create a radiomics-boosted deep learning model by pet\/ct image for risk assessment of synchronous crpm.methodsa total of 220 colorectal cancer (crc) cases were enrolled in this study. we mapped the feature maps (radiomic feature maps (rfms)) of radiomic features across ct and pet image patches by a 2d sliding kernel. based on resnet50, a radiomics-boosted deep learning model was trained using pet\/ct image patches and rfms. besides that, we explored whether the peritumoral region contributes to the assessment of crpm. in this study, the performance of each model was evaluated by the area under the curves (auc).resultsthe aucs of the radiomics-boosted deep learning model in the training, internal, external, and all validation datasets were 0.926 (95% confidence interval (ci): 0.874\u20130.978), 0.897 (95% ci: 0.801\u20130.994), 0.885 (95% ci: 0.795\u20130.975), and 0.889 (95% ci: 0.823\u20130.954), respectively. this model exhibited consistency in the calibration curve, the delong test and idi identified it as the most predictive model.conclusionsthe radiomics-boosted deep learning model showed superior estimated performance in preoperative prediction of synchronous crpm from pre-treatment pet\/ct, offering potential assistance in the development of more personalized treatment methods and follow-up plans.critical relevance statementthe onset of synchronous colorectal crpm is insidious, and using a radiomics-boosted deep learning model to assess the risk of crpm before treatment can help make personalized clinical treatment decisions or choose more sensitive follow-up plans.key points prognosis for patients with crpm is bleak, and early detection poses challenges. the synergy between radiomics and deep learning proves advantageous in evaluating crpm. the radiomics-boosted deep-learning model proves valuable in tailoring treatment approaches for crc patients. graphical abstract","key_phrases":["objectivessynchronous colorectal cancer peritoneal metastasis","crpm","a poor prognosis","this study","a radiomics-boosted deep learning model","pet\/ct image","risk assessment","synchronous crpm.methodsa total","220 colorectal cancer (crc) cases","this study","we","the feature maps","radiomic feature maps","rfms","radiomic features","ct","pet image","a 2d sliding kernel","resnet50","a radiomics-boosted deep learning model","pet\/ct image patches","rfms","that","we","the peritumoral region","the assessment","crpm","this study","the performance","each model","the area","the curves","auc).resultsthe aucs","the radiomics","the training","all validation datasets","95% confidence interval","ci","0.874\u20130.978","95% ci","95% ci","0.889 (95% ci","this model","consistency","the calibration curve","the delong test","idi","it","the most predictive model.conclusionsthe radiomics-boosted deep learning model","superior estimated performance","preoperative prediction","synchronous crpm","treatment pet","ct","potential assistance","the development","more personalized treatment methods","follow-up plans.critical relevance","statementthe onset","synchronous colorectal crpm","a radiomics-boosted deep learning model","the risk","crpm","treatment","personalized clinical treatment decisions","prognosis","patients","crpm","early detection poses challenges","the synergy","radiomics","deep learning","crpm","the radiomics","treatment approaches","crc patients","graphical abstract","crpm.methodsa","220","2d","resnet50","0.926","95%","0.897","95%","0.885","95%","0.795\u20130.975","0.889","95%"]},{"title":"Transfer learning-based quantized deep learning models for nail melanoma classification","authors":["Mujahid Hussain","Makhmoor Fiza","Aiman Khalil","Asad Ali Siyal","Fayaz Ali Dharejo","Waheeduddin Hyder","Antonella Guzzo","Moez Krichen","Giancarlo Fortino"],"abstract":"Skin cancer, particularly melanoma, has remained a severe issue for many years due to its increasing incidences. The rising mortality rate associated with melanoma demands immediate attention at early stages to facilitate timely diagnosis and effective treatment. Due to the similar visual appearance of malignant tumors and normal cells, the detection and classification of melanoma are considered to be one of the most challenging tasks. Detecting melanoma accurately and promptly is essential to diagnosis and treatment, which can contribute significantly to patient survival. A new dataset, Nailmelonma, is presented in this study in order to train and evaluate various deep learning models applying transfer learning for an indigenous nail melanoma localization dataset. Using the dermoscopic image datasets, seven CNN-based DL architectures (viz., VGG19, ResNet101, ResNet152V2, Xception, InceptionV3, MobileNet, and MobileNetv2) have been trained and tested for the classification of skin lesions for melanoma detection. The trained models have been validated, and key performance parameters (i.e., accuracy, recall, specificity, precision, and F1-score) are systematically evaluated to test the performance of each transfer learning model. The results indicated that the proposed workflow could realize and achieve more than 95% accuracy. In addition, we show how the quantization of such models can enable them for memory-constrained mobile\/edge devices. To facilitate an accurate, timely, and faster diagnosis of nail melanoma and to evaluate the early detection of other types of skin cancer, the proposed workflow can be readily applied and robust to the early detection of nail melanoma.","doi":"10.1007\/s00521-023-08925-y","cleaned_title":"transfer learning-based quantized deep learning models for nail melanoma classification","cleaned_abstract":"skin cancer, particularly melanoma, has remained a severe issue for many years due to its increasing incidences. the rising mortality rate associated with melanoma demands immediate attention at early stages to facilitate timely diagnosis and effective treatment. due to the similar visual appearance of malignant tumors and normal cells, the detection and classification of melanoma are considered to be one of the most challenging tasks. detecting melanoma accurately and promptly is essential to diagnosis and treatment, which can contribute significantly to patient survival. a new dataset, nailmelonma, is presented in this study in order to train and evaluate various deep learning models applying transfer learning for an indigenous nail melanoma localization dataset. using the dermoscopic image datasets, seven cnn-based dl architectures (viz., vgg19, resnet101, resnet152v2, xception, inceptionv3, mobilenet, and mobilenetv2) have been trained and tested for the classification of skin lesions for melanoma detection. the trained models have been validated, and key performance parameters (i.e., accuracy, recall, specificity, precision, and f1-score) are systematically evaluated to test the performance of each transfer learning model. the results indicated that the proposed workflow could realize and achieve more than 95% accuracy. in addition, we show how the quantization of such models can enable them for memory-constrained mobile\/edge devices. to facilitate an accurate, timely, and faster diagnosis of nail melanoma and to evaluate the early detection of other types of skin cancer, the proposed workflow can be readily applied and robust to the early detection of nail melanoma.","key_phrases":["skin cancer","particularly melanoma","a severe issue","many years","its increasing incidences","the rising mortality rate","melanoma","immediate attention","early stages","timely diagnosis and effective treatment","the similar visual appearance","malignant tumors","normal cells","the detection","classification","melanoma","the most challenging tasks","melanoma","diagnosis","treatment","which","patient survival","a new dataset","nailmelonma","this study","order","various deep learning models","an indigenous nail melanoma localization dataset","the dermoscopic image datasets","seven cnn-based dl architectures","viz",".","vgg19","xception","inceptionv3","mobilenet","mobilenetv2","the classification","skin lesions","melanoma detection","the trained models","key performance parameters","i.e., accuracy","recall","specificity","precision","f1-score","the performance","each transfer learning model","the results","the proposed workflow","more than 95% accuracy","addition","we","the quantization","such models","them","memory-constrained mobile\/edge devices","an accurate, timely, and faster diagnosis","nail melanoma","the early detection","other types","skin cancer","the proposed workflow","the early detection","nail melanoma","seven","cnn","inceptionv3","mobilenetv2","more than 95%"]},{"title":"RETRACTED ARTICLE: Multimedia Lu Xun literature online learning based on deep learning","authors":["Wang Hongsheng"],"abstract":"As a great Chinese thinker and writer in the twentieth century, Lu Xun and his literary works are widely known. However, as a successful cultural communication activist, editor and publisher, we still need to conduct in-depth research on Lu Xun in many aspects. Therefore, based on deep learning, this paper constructs an online multimedia learning system of Lu Xun literature. This system takes the relationship between classical Lu Xun literature and modern multimedia technology as the research object, and compare the calculation effect of other different types of algorithms and this dhraa algorithm. Through the availability of data, the dhraa algorithm is significantly better than other algorithms in the recommendation accuracy, thus proving its effectiveness. This system is managed by two servers and one system. The two servers are database server and web server, respectively. After testing, the system has good bearing capacity, can make up for the limited processing capacity of the server, and ensure the system has high performance. Its performance characteristics also show that the system achieved the expected performance. This paper systematically combines Lu Xun\u2019s literature with modern multimedia, which can provide online learning services for Lu Xun\u2019s literature lovers, thus helping scholars to expand Lu Xun\u2019s research field and academic vision. This paper designs an effective online learning system of Lu Xun\u2019s literature by combining deep learning, multimedia technology and Lu Xun\u2019s literature.","doi":"10.1007\/s00500-023-08118-8","cleaned_title":"retracted article: multimedia lu xun literature online learning based on deep learning","cleaned_abstract":"as a great chinese thinker and writer in the twentieth century, lu xun and his literary works are widely known. however, as a successful cultural communication activist, editor and publisher, we still need to conduct in-depth research on lu xun in many aspects. therefore, based on deep learning, this paper constructs an online multimedia learning system of lu xun literature. this system takes the relationship between classical lu xun literature and modern multimedia technology as the research object, and compare the calculation effect of other different types of algorithms and this dhraa algorithm. through the availability of data, the dhraa algorithm is significantly better than other algorithms in the recommendation accuracy, thus proving its effectiveness. this system is managed by two servers and one system. the two servers are database server and web server, respectively. after testing, the system has good bearing capacity, can make up for the limited processing capacity of the server, and ensure the system has high performance. its performance characteristics also show that the system achieved the expected performance. this paper systematically combines lu xun\u2019s literature with modern multimedia, which can provide online learning services for lu xun\u2019s literature lovers, thus helping scholars to expand lu xun\u2019s research field and academic vision. this paper designs an effective online learning system of lu xun\u2019s literature by combining deep learning, multimedia technology and lu xun\u2019s literature.","key_phrases":["a great chinese thinker","writer","the twentieth century","lu xun","his literary works","a successful cultural communication activist","editor","publisher","we","depth","lu xun","many aspects","deep learning","this paper","an online multimedia learning system","lu xun literature","this system","the relationship","classical lu xun literature","modern multimedia technology","the research object","the calculation effect","other different types","algorithms","this dhraa algorithm","the availability","data","the dhraa algorithm","other algorithms","the recommendation accuracy","its effectiveness","this system","two servers","one system","the two servers","database server and web server","testing","the system","good bearing capacity","the limited processing capacity","the server","the system","high performance","its performance characteristics","the system","the expected performance","this paper","lu xun\u2019s literature","modern multimedia","which","online learning services","lu xun\u2019s literature lovers","scholars","lu xun\u2019s research field","academic vision","this paper","an effective online learning system","lu xun\u2019s literature","deep learning","multimedia technology","lu xun\u2019s literature","chinese","the twentieth century","lu xun","lu xun","lu xun literature","two","one","two","xun","xun\u2019s","xun","xun"]},{"title":"Deep learning approaches for lyme disease detection: leveraging progressive resizing and self-supervised learning models","authors":["Daryl Jacob Jerrish","Om Nankar","Shilpa Gite","Shruti Patil","Ketan Kotecha","Ganeshsree Selvachandran","Ajith Abraham"],"abstract":"Lyme disease diagnosis poses a significant challenge, with blood tests exhibiting an alarming inaccuracy rate of nearly 60% in detecting early-stage infections. As a result, there is an urgent need for improved diagnostic methods that can offer more accurate detection outcomes. To address this pressing issue, our study focuses on harnessing the potential of deep learning approaches, specifically by employing model pipelining through progressive resizing and multiple self-supervised learning models. In this paper, we present a comprehensive exploration of self-supervised learning models, including SimCLR, SwAV, MoCo, and BYOL, tailored to the context of Lyme disease detection using medical imaging. The effectiveness and performance of these models are evaluated using standard metrics such as F1 score, precision, recall, and accuracy. Furthermore, we emphasize the significance of progressive resizing and its implications when dealing with convolutional neural networks (CNNs) for medical image analysis. By leveraging deep learning approaches, progressive resizing, and self-supervised learning models, the challenges associated with Lyme disease detection are effectively addressed in this study. The application of our novel methodology and the execution of a comprehensive evaluation framework contribute invaluable insights, fostering the development of more efficient and accurate diagnostic methods for Lyme disease. It is firmly believed that our research will serve as a catalyst, inspiring interdisciplinary collaborations that accelerate progress at the convergence of medicine, computing, and technology, ultimately benefiting public health.","doi":"10.1007\/s11042-023-16306-9","cleaned_title":"deep learning approaches for lyme disease detection: leveraging progressive resizing and self-supervised learning models","cleaned_abstract":"lyme disease diagnosis poses a significant challenge, with blood tests exhibiting an alarming inaccuracy rate of nearly 60% in detecting early-stage infections. as a result, there is an urgent need for improved diagnostic methods that can offer more accurate detection outcomes. to address this pressing issue, our study focuses on harnessing the potential of deep learning approaches, specifically by employing model pipelining through progressive resizing and multiple self-supervised learning models. in this paper, we present a comprehensive exploration of self-supervised learning models, including simclr, swav, moco, and byol, tailored to the context of lyme disease detection using medical imaging. the effectiveness and performance of these models are evaluated using standard metrics such as f1 score, precision, recall, and accuracy. furthermore, we emphasize the significance of progressive resizing and its implications when dealing with convolutional neural networks (cnns) for medical image analysis. by leveraging deep learning approaches, progressive resizing, and self-supervised learning models, the challenges associated with lyme disease detection are effectively addressed in this study. the application of our novel methodology and the execution of a comprehensive evaluation framework contribute invaluable insights, fostering the development of more efficient and accurate diagnostic methods for lyme disease. it is firmly believed that our research will serve as a catalyst, inspiring interdisciplinary collaborations that accelerate progress at the convergence of medicine, computing, and technology, ultimately benefiting public health.","key_phrases":["lyme disease diagnosis","a significant challenge","blood tests","an alarming inaccuracy rate","nearly 60%","early-stage infections","a result","an urgent need","improved diagnostic methods","that","more accurate detection outcomes","this pressing issue","our study","the potential","deep learning approaches","model","progressive resizing","multiple self-supervised learning models","this paper","we","a comprehensive exploration","self-supervised learning models","simclr","swav","moco","byol","the context","lyme disease detection","medical imaging","the effectiveness","performance","these models","standard metrics","f1 score","precision","recall","accuracy","we","the significance","progressive resizing","its implications","convolutional neural networks","cnns","medical image analysis","deep learning approaches","progressive resizing","self-supervised learning models","the challenges","lyme disease detection","this study","the application","our novel methodology","the execution","a comprehensive evaluation framework","invaluable insights","the development","more efficient and accurate diagnostic methods","lyme disease","it","our research","a catalyst","interdisciplinary collaborations","that","progress","the convergence","medicine","computing","technology","public health","nearly 60%"]},{"title":"Enhancing trash classification in smart cities using federated deep learning","authors":["Haroon Ahmed Khan","Syed Saud Naqvi","Abeer A. K. Alharbi","Salihah Alotaibi","Mohammed Alkhathami"],"abstract":"Efficient Waste management plays a crucial role to ensure clean and green environment in the smart cities. This study investigates the critical role of efficient trash classification in achieving sustainable solid waste management within smart city environments. We conduct a comparative analysis of various trash classification methods utilizing deep learning models built on convolutional neural networks (CNNs). Leveraging the PyTorch open-source framework and the TrashBox dataset, we perform experiments involving ten unique deep neural network models. Our approach aims to maximize training accuracy. Through extensive experimentation, we observe the consistent superiority of the ResNext-101 model compared to others, achieving exceptional training, validation, and test accuracies. These findings illuminate the potential of CNN-based techniques in significantly advancing trash classification for optimized solid waste management within smart city initiatives. Lastly, this study presents a distributed framework based on federated learning that can be used to optimize the performance of a combination of CNN models for trash detection.","doi":"10.1038\/s41598-024-62003-4","cleaned_title":"enhancing trash classification in smart cities using federated deep learning","cleaned_abstract":"efficient waste management plays a crucial role to ensure clean and green environment in the smart cities. this study investigates the critical role of efficient trash classification in achieving sustainable solid waste management within smart city environments. we conduct a comparative analysis of various trash classification methods utilizing deep learning models built on convolutional neural networks (cnns). leveraging the pytorch open-source framework and the trashbox dataset, we perform experiments involving ten unique deep neural network models. our approach aims to maximize training accuracy. through extensive experimentation, we observe the consistent superiority of the resnext-101 model compared to others, achieving exceptional training, validation, and test accuracies. these findings illuminate the potential of cnn-based techniques in significantly advancing trash classification for optimized solid waste management within smart city initiatives. lastly, this study presents a distributed framework based on federated learning that can be used to optimize the performance of a combination of cnn models for trash detection.","key_phrases":["efficient waste management","a crucial role","clean and green environment","the smart cities","this study","the critical role","efficient trash classification","sustainable solid waste management","smart city environments","we","a comparative analysis","various trash classification methods","deep learning models","convolutional neural networks","cnns","the pytorch open-source framework","the trashbox dataset","we","experiments","ten unique deep neural network models","our approach","training accuracy","extensive experimentation","we","the consistent superiority","the resnext-101 model","others","exceptional training","validation","test accuracies","these findings","the potential","cnn-based techniques","significantly advancing trash classification","optimized solid waste management","smart city initiatives","this study","a distributed framework","federated learning","that","the performance","a combination","cnn models","trash detection","smart city","cnn","smart city","cnn"]},{"title":"Can supervised deep learning architecture outperform autoencoders in building propensity score models for matching?","authors":["Mohammad Ehsanul Karim"],"abstract":"PurposePropensity score matching is vital in epidemiological studies using observational data, yet its estimates relies on correct model-specification. This study assesses supervised deep learning models and unsupervised autoencoders for propensity score estimation, comparing them with traditional methods for bias and variance accuracy in treatment effect estimations.MethodsUtilizing a plasmode simulation based on the Right Heart Catheterization dataset, under a variety of settings, we evaluated (1) a supervised deep learning architecture and (2) an unsupervised autoencoder, alongside two traditional methods: logistic regression and a spline-based method in estimating propensity scores for matching. Performance metrics included bias, standard errors, and coverage probability. The analysis was also extended to real-world data, with estimates compared to those obtained via a double robust approach.ResultsThe analysis revealed that supervised deep learning models outperformed unsupervised autoencoders in variance estimation while maintaining comparable levels of bias. These results were supported by analyses of real-world data, where the supervised model\u2019s estimates closely matched those derived from conventional methods. Additionally, deep learning models performed well compared to traditional methods in settings where exposure was rare.ConclusionSupervised deep learning models hold promise in refining propensity score estimations in epidemiological research, offering nuanced confounder adjustment, especially in complex datasets. We endorse integrating supervised deep learning into epidemiological research and share reproducible codes for widespread use and methodological transparency.","doi":"10.1186\/s12874-024-02284-5","cleaned_title":"can supervised deep learning architecture outperform autoencoders in building propensity score models for matching?","cleaned_abstract":"purposepropensity score matching is vital in epidemiological studies using observational data, yet its estimates relies on correct model-specification. this study assesses supervised deep learning models and unsupervised autoencoders for propensity score estimation, comparing them with traditional methods for bias and variance accuracy in treatment effect estimations.methodsutilizing a plasmode simulation based on the right heart catheterization dataset, under a variety of settings, we evaluated (1) a supervised deep learning architecture and (2) an unsupervised autoencoder, alongside two traditional methods: logistic regression and a spline-based method in estimating propensity scores for matching. performance metrics included bias, standard errors, and coverage probability. the analysis was also extended to real-world data, with estimates compared to those obtained via a double robust approach.resultsthe analysis revealed that supervised deep learning models outperformed unsupervised autoencoders in variance estimation while maintaining comparable levels of bias. these results were supported by analyses of real-world data, where the supervised model\u2019s estimates closely matched those derived from conventional methods. additionally, deep learning models performed well compared to traditional methods in settings where exposure was rare.conclusionsupervised deep learning models hold promise in refining propensity score estimations in epidemiological research, offering nuanced confounder adjustment, especially in complex datasets. we endorse integrating supervised deep learning into epidemiological research and share reproducible codes for widespread use and methodological transparency.","key_phrases":["purposepropensity score matching","epidemiological studies","observational data","its estimates","correct model-specification","this study","deep learning models","autoencoders","propensity score estimation","them","traditional methods","bias","variance accuracy","treatment effect","a plasmode simulation","the right heart catheterization dataset","a variety","settings","we","1) a supervised deep learning architecture","(2) an unsupervised autoencoder","two traditional methods","logistic regression","a spline-based method","propensity scores","performance metrics","bias","standard errors","coverage probability","the analysis","real-world data","estimates","those","a double robust","approach.resultsthe analysis","supervised deep learning models","unsupervised autoencoders","variance estimation","comparable levels","bias","these results","analyses","real-world data","the supervised model\u2019s estimates","those","conventional methods","deep learning models","traditional methods","settings","exposure","deep learning models","promise","propensity score estimations","epidemiological research","nuanced confounder adjustment","complex datasets","we","supervised deep learning","epidemiological research","reproducible codes","widespread use","methodological transparency","1","2","two"]},{"title":"A systematic review of deep learning based image segmentation to detect polyp","authors":["Mayuri Gupta","Ashish Mishra"],"abstract":"Among the world\u2019s most common cancers, colorectal cancer is the third most severe form of cancer. Early polyp detection reduces the risk of colorectal cancer, vital for effective treatment. Artificial intelligence methods such as deep learning have emerged as leading techniques for polyp image segmentation that have gained success in advancing medical image diagnosis. This study aims to provide a review of the most recent research studies that have used deep learning methods and models for polyp segmentation. A comprehensive review of deep learning-based image segmentation models is provided based on existing research studies that are essential for polyp segmentation. Convolutional neural networks, encoder\u2013decoder models, recurrent neural networks, attention-based models, and generative models were the most popular deep learning models which play an essential role in detecting and diagnosing polyp at an early stage. Additionally, this study also aims to provide a detailed classification of prominently used polyp image and video datasets. The evaluation metrics for assessing the effectiveness of different methods, models, and techniques are identified and discussed. A statistical analysis of deep learning models based on polyp datasets and performance metrics is presented, with a discussion of future research trends and limitations.","doi":"10.1007\/s10462-023-10621-1","cleaned_title":"a systematic review of deep learning based image segmentation to detect polyp","cleaned_abstract":"among the world\u2019s most common cancers, colorectal cancer is the third most severe form of cancer. early polyp detection reduces the risk of colorectal cancer, vital for effective treatment. artificial intelligence methods such as deep learning have emerged as leading techniques for polyp image segmentation that have gained success in advancing medical image diagnosis. this study aims to provide a review of the most recent research studies that have used deep learning methods and models for polyp segmentation. a comprehensive review of deep learning-based image segmentation models is provided based on existing research studies that are essential for polyp segmentation. convolutional neural networks, encoder\u2013decoder models, recurrent neural networks, attention-based models, and generative models were the most popular deep learning models which play an essential role in detecting and diagnosing polyp at an early stage. additionally, this study also aims to provide a detailed classification of prominently used polyp image and video datasets. the evaluation metrics for assessing the effectiveness of different methods, models, and techniques are identified and discussed. a statistical analysis of deep learning models based on polyp datasets and performance metrics is presented, with a discussion of future research trends and limitations.","key_phrases":["the world\u2019s most common cancers","colorectal cancer","the third most severe form","cancer","early polyp detection","the risk","colorectal cancer","effective treatment","artificial intelligence methods","deep learning","leading techniques","polyp image segmentation","that","success","medical image diagnosis","this study","a review","the most recent research studies","that","deep learning methods","models","polyp segmentation","a comprehensive review","deep learning-based image segmentation models","existing research studies","that","polyp segmentation","convolutional neural networks","encoder","decoder models","recurrent neural networks","attention-based models","generative models","the most popular deep learning models","which","an essential role","polyp","an early stage","this study","a detailed classification","prominently used polyp image","video datasets","the evaluation metrics","the effectiveness","different methods","models","techniques","a statistical analysis","deep learning models","polyp datasets","performance metrics","a discussion","future research trends","limitations","third"]},{"title":"Deep learning based damage detection of concrete structures","authors":["Maheswara Rao Bandi","Laxmi Narayana Pasupuleti","Tanmay Das","Shyamal Guchhait"],"abstract":"Damage detection of any civil engineering structure is a part of the interest in the field of engineering from which one can estimate the stability and lifetime of the structure. With the advancement of technology in the field of infrastructure, damage assessment of any structure with the help of convolutional neural networks (CNNs) and deep learning is gaining importance because of the ease with which it can detect the damage. By using these computer-aided techniques, we can reduce manpower and detect damage in inaccessible places that we cannot see directly. In the current study, we used ResNet-50, which is a part of a deep convolutional neural network with 50 layers deep in it. ResNet-50 is a subclass of convolutional neural networks most popularly used for image classification. Furthermore, we used the image data set collected from the Utah State University, Logan, Utah, USA. This data set contains nearly 56,000 annotated images of both cracked and non-cracked bridge deck images, wall images, and pavement images. These images contain a variety of obstructions, such as shadows and surface roughness in some cases. The present study aimed to train and test the images and compare the accuracy of the results among themselves with an increase in training data. We devised an algorithm to measure cracks as soon as we detected them. The results show that the Resnet-50 architecture was in good agreement with the developed algorithm.","doi":"10.1007\/s42107-024-01106-9","cleaned_title":"deep learning based damage detection of concrete structures","cleaned_abstract":"damage detection of any civil engineering structure is a part of the interest in the field of engineering from which one can estimate the stability and lifetime of the structure. with the advancement of technology in the field of infrastructure, damage assessment of any structure with the help of convolutional neural networks (cnns) and deep learning is gaining importance because of the ease with which it can detect the damage. by using these computer-aided techniques, we can reduce manpower and detect damage in inaccessible places that we cannot see directly. in the current study, we used resnet-50, which is a part of a deep convolutional neural network with 50 layers deep in it. resnet-50 is a subclass of convolutional neural networks most popularly used for image classification. furthermore, we used the image data set collected from the utah state university, logan, utah, usa. this data set contains nearly 56,000 annotated images of both cracked and non-cracked bridge deck images, wall images, and pavement images. these images contain a variety of obstructions, such as shadows and surface roughness in some cases. the present study aimed to train and test the images and compare the accuracy of the results among themselves with an increase in training data. we devised an algorithm to measure cracks as soon as we detected them. the results show that the resnet-50 architecture was in good agreement with the developed algorithm.","key_phrases":["damage detection","any civil engineering structure","a part","the interest","the field","engineering","which","one","the stability","lifetime","the structure","the advancement","technology","the field","infrastructure","damage assessment","any structure","the help","convolutional neural networks","cnns","deep learning","importance","the ease","which","it","the damage","these computer-aided techniques","we","manpower","damage","inaccessible places","that","we","the current study","we","resnet-50","which","a part","a deep convolutional neural network","50 layers","it","resnet-50","a subclass","convolutional neural networks","image classification","we","the utah state university","usa","this data","nearly 56,000 annotated images","both cracked and non-cracked bridge deck images","wall images","pavement images","these images","a variety","obstructions","shadows","surface roughness","some cases","the present study","the images","the accuracy","the results","themselves","an increase","training data","we","an algorithm","cracks","we","them","the results","the resnet-50 architecture","good agreement","the developed algorithm","resnet-50","50","resnet-50","the utah state university","utah","usa","nearly 56,000","resnet-50"]},{"title":"A Comparative Analysis of Deep Learning-Based Approaches for Classifying Dental Implants Decision Support System","authors":["Mohammed A. H. Lubbad","Ikbal Leblebicioglu Kurtulus","Dervis Karaboga","Kerem Kilic","Alper Basturk","Bahriye Akay","Ozkan Ufuk Nalbantoglu","Ozden Melis Durmaz Yilmaz","Mustafa Ayata","Serkan Yilmaz","Ishak Pacal"],"abstract":"This study aims to provide an effective solution for the autonomous identification of dental implant brands through a deep learning-based computer diagnostic system. It also seeks to ascertain the system\u2019s potential in clinical practices and to offer a strategic framework for improving diagnosis and treatment processes in implantology. This study employed a total of 28 different deep learning models, including 18 convolutional neural network (CNN) models (VGG, ResNet, DenseNet, EfficientNet, RegNet, ConvNeXt) and 10 vision transformer models (Swin and Vision Transformer). The dataset comprises 1258 panoramic radiographs from patients who received implant treatments at Erciyes University Faculty of Dentistry between 2012 and 2023. It is utilized for the training and evaluation process of deep learning models and consists of prototypes from six different implant systems provided by six manufacturers. The deep learning-based dental implant system provided high classification accuracy for different dental implant brands using deep learning models. Furthermore, among all the architectures evaluated, the small model of the ConvNeXt architecture achieved an impressive accuracy rate of 94.2%, demonstrating a high level of classification success.This study emphasizes the effectiveness of deep learning-based systems in achieving high classification accuracy in dental implant types. These findings pave the way for integrating advanced deep learning tools into clinical practice, promising significant improvements in patient care and treatment outcomes.","doi":"10.1007\/s10278-024-01086-x","cleaned_title":"a comparative analysis of deep learning-based approaches for classifying dental implants decision support system","cleaned_abstract":"this study aims to provide an effective solution for the autonomous identification of dental implant brands through a deep learning-based computer diagnostic system. it also seeks to ascertain the system\u2019s potential in clinical practices and to offer a strategic framework for improving diagnosis and treatment processes in implantology. this study employed a total of 28 different deep learning models, including 18 convolutional neural network (cnn) models (vgg, resnet, densenet, efficientnet, regnet, convnext) and 10 vision transformer models (swin and vision transformer). the dataset comprises 1258 panoramic radiographs from patients who received implant treatments at erciyes university faculty of dentistry between 2012 and 2023. it is utilized for the training and evaluation process of deep learning models and consists of prototypes from six different implant systems provided by six manufacturers. the deep learning-based dental implant system provided high classification accuracy for different dental implant brands using deep learning models. furthermore, among all the architectures evaluated, the small model of the convnext architecture achieved an impressive accuracy rate of 94.2%, demonstrating a high level of classification success.this study emphasizes the effectiveness of deep learning-based systems in achieving high classification accuracy in dental implant types. these findings pave the way for integrating advanced deep learning tools into clinical practice, promising significant improvements in patient care and treatment outcomes.","key_phrases":["this study","an effective solution","the autonomous identification","dental implant brands","a deep learning-based computer diagnostic system","it","the system\u2019s potential","clinical practices","a strategic framework","diagnosis and treatment processes","implantology","this study","a total","28 different deep learning models","18 convolutional neural network (cnn) models","vgg","resnet","densenet","efficientnet","regnet","convnext","10 vision transformer models","swin and vision transformer","the dataset","patients","who","implant treatments","erciyes university faculty","dentistry","it","the training and evaluation process","deep learning models","prototypes","six different implant systems","six manufacturers","the deep learning-based dental implant system","high classification accuracy","different dental implant brands","deep learning models","all the architectures","the small model","the convnext architecture","an impressive accuracy rate","94.2%","a high level","classification success.this study","the effectiveness","deep learning-based systems","high classification accuracy","dental implant types","these findings","the way","advanced deep learning tools","clinical practice","significant improvements","patient care and treatment outcomes","28","18","cnn","10","1258","erciyes university faculty of dentistry","between 2012 and 2023","six","six","94.2%"]},{"title":"A radiomics-boosted deep-learning for risk assessment of synchronous peritoneal metastasis in colorectal cancer","authors":["Ding Zhang","BingShu Zheng","LiuWei Xu","YiCong Wu","Chen Shen","ShanLei Bao","ZhongHua Tan","ChunFeng Sun"],"abstract":"ObjectivesSynchronous colorectal cancer peritoneal metastasis (CRPM) has a poor prognosis. This study aimed to create a radiomics-boosted deep learning model by PET\/CT image for risk assessment of synchronous CRPM.MethodsA total of 220 colorectal cancer (CRC) cases were enrolled in this study. We mapped the feature maps (Radiomic feature maps (RFMs)) of radiomic features across CT and PET image patches by a 2D sliding kernel. Based on ResNet50, a radiomics-boosted deep learning model was trained using PET\/CT image patches and RFMs. Besides that, we explored whether the peritumoral region contributes to the assessment of CRPM. In this study, the performance of each model was evaluated by the area under the curves (AUC).ResultsThe AUCs of the radiomics-boosted deep learning model in the training, internal, external, and all validation datasets were 0.926 (95% confidence interval (CI): 0.874\u20130.978), 0.897 (95% CI: 0.801\u20130.994), 0.885 (95% CI: 0.795\u20130.975), and 0.889 (95% CI: 0.823\u20130.954), respectively. This model exhibited consistency in the calibration curve, the Delong test and IDI identified it as the most predictive model.ConclusionsThe radiomics-boosted deep learning model showed superior estimated performance in preoperative prediction of synchronous CRPM from pre-treatment PET\/CT, offering potential assistance in the development of more personalized treatment methods and follow-up plans.Critical relevance statementThe onset of synchronous colorectal CRPM is insidious, and using a radiomics-boosted deep learning model to assess the risk of CRPM before treatment can help make personalized clinical treatment decisions or choose more sensitive follow-up plans.Key Points\n\nPrognosis for patients with CRPM is bleak, and early detection poses challenges.\n\n\nThe synergy between radiomics and deep learning proves advantageous in evaluating CRPM.\n\n\nThe radiomics-boosted deep-learning model proves valuable in tailoring treatment approaches for CRC patients.\n\nGraphical Abstract","doi":"10.1186\/s13244-024-01733-5","cleaned_title":"a radiomics-boosted deep-learning for risk assessment of synchronous peritoneal metastasis in colorectal cancer","cleaned_abstract":"objectivessynchronous colorectal cancer peritoneal metastasis (crpm) has a poor prognosis. this study aimed to create a radiomics-boosted deep learning model by pet\/ct image for risk assessment of synchronous crpm.methodsa total of 220 colorectal cancer (crc) cases were enrolled in this study. we mapped the feature maps (radiomic feature maps (rfms)) of radiomic features across ct and pet image patches by a 2d sliding kernel. based on resnet50, a radiomics-boosted deep learning model was trained using pet\/ct image patches and rfms. besides that, we explored whether the peritumoral region contributes to the assessment of crpm. in this study, the performance of each model was evaluated by the area under the curves (auc).resultsthe aucs of the radiomics-boosted deep learning model in the training, internal, external, and all validation datasets were 0.926 (95% confidence interval (ci): 0.874\u20130.978), 0.897 (95% ci: 0.801\u20130.994), 0.885 (95% ci: 0.795\u20130.975), and 0.889 (95% ci: 0.823\u20130.954), respectively. this model exhibited consistency in the calibration curve, the delong test and idi identified it as the most predictive model.conclusionsthe radiomics-boosted deep learning model showed superior estimated performance in preoperative prediction of synchronous crpm from pre-treatment pet\/ct, offering potential assistance in the development of more personalized treatment methods and follow-up plans.critical relevance statementthe onset of synchronous colorectal crpm is insidious, and using a radiomics-boosted deep learning model to assess the risk of crpm before treatment can help make personalized clinical treatment decisions or choose more sensitive follow-up plans.key points prognosis for patients with crpm is bleak, and early detection poses challenges. the synergy between radiomics and deep learning proves advantageous in evaluating crpm. the radiomics-boosted deep-learning model proves valuable in tailoring treatment approaches for crc patients. graphical abstract","key_phrases":["objectivessynchronous colorectal cancer peritoneal metastasis","crpm","a poor prognosis","this study","a radiomics-boosted deep learning model","pet\/ct image","risk assessment","synchronous crpm.methodsa total","220 colorectal cancer (crc) cases","this study","we","the feature maps","radiomic feature maps","rfms","radiomic features","ct","pet image","a 2d sliding kernel","resnet50","a radiomics-boosted deep learning model","pet\/ct image patches","rfms","that","we","the peritumoral region","the assessment","crpm","this study","the performance","each model","the area","the curves","auc).resultsthe aucs","the radiomics","the training","all validation datasets","95% confidence interval","ci","0.874\u20130.978","95% ci","95% ci","0.889 (95% ci","this model","consistency","the calibration curve","the delong test","idi","it","the most predictive model.conclusionsthe radiomics-boosted deep learning model","superior estimated performance","preoperative prediction","synchronous crpm","treatment pet","ct","potential assistance","the development","more personalized treatment methods","follow-up plans.critical relevance","statementthe onset","synchronous colorectal crpm","a radiomics-boosted deep learning model","the risk","crpm","treatment","personalized clinical treatment decisions","prognosis","patients","crpm","early detection poses challenges","the synergy","radiomics","deep learning","crpm","the radiomics","treatment approaches","crc patients","graphical abstract","crpm.methodsa","220","2d","resnet50","0.926","95%","0.897","95%","0.885","95%","0.795\u20130.975","0.889","95%"]},{"title":"Transfer learning-based quantized deep learning models for nail melanoma classification","authors":["Mujahid Hussain","Makhmoor Fiza","Aiman Khalil","Asad Ali Siyal","Fayaz Ali Dharejo","Waheeduddin Hyder","Antonella Guzzo","Moez Krichen","Giancarlo Fortino"],"abstract":"Skin cancer, particularly melanoma, has remained a severe issue for many years due to its increasing incidences. The rising mortality rate associated with melanoma demands immediate attention at early stages to facilitate timely diagnosis and effective treatment. Due to the similar visual appearance of malignant tumors and normal cells, the detection and classification of melanoma are considered to be one of the most challenging tasks. Detecting melanoma accurately and promptly is essential to diagnosis and treatment, which can contribute significantly to patient survival. A new dataset, Nailmelonma, is presented in this study in order to train and evaluate various deep learning models applying transfer learning for an indigenous nail melanoma localization dataset. Using the dermoscopic image datasets, seven CNN-based DL architectures (viz., VGG19, ResNet101, ResNet152V2, Xception, InceptionV3, MobileNet, and MobileNetv2) have been trained and tested for the classification of skin lesions for melanoma detection. The trained models have been validated, and key performance parameters (i.e., accuracy, recall, specificity, precision, and F1-score) are systematically evaluated to test the performance of each transfer learning model. The results indicated that the proposed workflow could realize and achieve more than 95% accuracy. In addition, we show how the quantization of such models can enable them for memory-constrained mobile\/edge devices. To facilitate an accurate, timely, and faster diagnosis of nail melanoma and to evaluate the early detection of other types of skin cancer, the proposed workflow can be readily applied and robust to the early detection of nail melanoma.","doi":"10.1007\/s00521-023-08925-y","cleaned_title":"transfer learning-based quantized deep learning models for nail melanoma classification","cleaned_abstract":"skin cancer, particularly melanoma, has remained a severe issue for many years due to its increasing incidences. the rising mortality rate associated with melanoma demands immediate attention at early stages to facilitate timely diagnosis and effective treatment. due to the similar visual appearance of malignant tumors and normal cells, the detection and classification of melanoma are considered to be one of the most challenging tasks. detecting melanoma accurately and promptly is essential to diagnosis and treatment, which can contribute significantly to patient survival. a new dataset, nailmelonma, is presented in this study in order to train and evaluate various deep learning models applying transfer learning for an indigenous nail melanoma localization dataset. using the dermoscopic image datasets, seven cnn-based dl architectures (viz., vgg19, resnet101, resnet152v2, xception, inceptionv3, mobilenet, and mobilenetv2) have been trained and tested for the classification of skin lesions for melanoma detection. the trained models have been validated, and key performance parameters (i.e., accuracy, recall, specificity, precision, and f1-score) are systematically evaluated to test the performance of each transfer learning model. the results indicated that the proposed workflow could realize and achieve more than 95% accuracy. in addition, we show how the quantization of such models can enable them for memory-constrained mobile\/edge devices. to facilitate an accurate, timely, and faster diagnosis of nail melanoma and to evaluate the early detection of other types of skin cancer, the proposed workflow can be readily applied and robust to the early detection of nail melanoma.","key_phrases":["skin cancer","particularly melanoma","a severe issue","many years","its increasing incidences","the rising mortality rate","melanoma","immediate attention","early stages","timely diagnosis and effective treatment","the similar visual appearance","malignant tumors","normal cells","the detection","classification","melanoma","the most challenging tasks","melanoma","diagnosis","treatment","which","patient survival","a new dataset","nailmelonma","this study","order","various deep learning models","an indigenous nail melanoma localization dataset","the dermoscopic image datasets","seven cnn-based dl architectures","viz",".","vgg19","xception","inceptionv3","mobilenet","mobilenetv2","the classification","skin lesions","melanoma detection","the trained models","key performance parameters","i.e., accuracy","recall","specificity","precision","f1-score","the performance","each transfer learning model","the results","the proposed workflow","more than 95% accuracy","addition","we","the quantization","such models","them","memory-constrained mobile\/edge devices","an accurate, timely, and faster diagnosis","nail melanoma","the early detection","other types","skin cancer","the proposed workflow","the early detection","nail melanoma","seven","cnn","inceptionv3","mobilenetv2","more than 95%"]},{"title":"RETRACTED ARTICLE: Multimedia Lu Xun literature online learning based on deep learning","authors":["Wang Hongsheng"],"abstract":"As a great Chinese thinker and writer in the twentieth century, Lu Xun and his literary works are widely known. However, as a successful cultural communication activist, editor and publisher, we still need to conduct in-depth research on Lu Xun in many aspects. Therefore, based on deep learning, this paper constructs an online multimedia learning system of Lu Xun literature. This system takes the relationship between classical Lu Xun literature and modern multimedia technology as the research object, and compare the calculation effect of other different types of algorithms and this dhraa algorithm. Through the availability of data, the dhraa algorithm is significantly better than other algorithms in the recommendation accuracy, thus proving its effectiveness. This system is managed by two servers and one system. The two servers are database server and web server, respectively. After testing, the system has good bearing capacity, can make up for the limited processing capacity of the server, and ensure the system has high performance. Its performance characteristics also show that the system achieved the expected performance. This paper systematically combines Lu Xun\u2019s literature with modern multimedia, which can provide online learning services for Lu Xun\u2019s literature lovers, thus helping scholars to expand Lu Xun\u2019s research field and academic vision. This paper designs an effective online learning system of Lu Xun\u2019s literature by combining deep learning, multimedia technology and Lu Xun\u2019s literature.","doi":"10.1007\/s00500-023-08118-8","cleaned_title":"retracted article: multimedia lu xun literature online learning based on deep learning","cleaned_abstract":"as a great chinese thinker and writer in the twentieth century, lu xun and his literary works are widely known. however, as a successful cultural communication activist, editor and publisher, we still need to conduct in-depth research on lu xun in many aspects. therefore, based on deep learning, this paper constructs an online multimedia learning system of lu xun literature. this system takes the relationship between classical lu xun literature and modern multimedia technology as the research object, and compare the calculation effect of other different types of algorithms and this dhraa algorithm. through the availability of data, the dhraa algorithm is significantly better than other algorithms in the recommendation accuracy, thus proving its effectiveness. this system is managed by two servers and one system. the two servers are database server and web server, respectively. after testing, the system has good bearing capacity, can make up for the limited processing capacity of the server, and ensure the system has high performance. its performance characteristics also show that the system achieved the expected performance. this paper systematically combines lu xun\u2019s literature with modern multimedia, which can provide online learning services for lu xun\u2019s literature lovers, thus helping scholars to expand lu xun\u2019s research field and academic vision. this paper designs an effective online learning system of lu xun\u2019s literature by combining deep learning, multimedia technology and lu xun\u2019s literature.","key_phrases":["a great chinese thinker","writer","the twentieth century","lu xun","his literary works","a successful cultural communication activist","editor","publisher","we","depth","lu xun","many aspects","deep learning","this paper","an online multimedia learning system","lu xun literature","this system","the relationship","classical lu xun literature","modern multimedia technology","the research object","the calculation effect","other different types","algorithms","this dhraa algorithm","the availability","data","the dhraa algorithm","other algorithms","the recommendation accuracy","its effectiveness","this system","two servers","one system","the two servers","database server and web server","testing","the system","good bearing capacity","the limited processing capacity","the server","the system","high performance","its performance characteristics","the system","the expected performance","this paper","lu xun\u2019s literature","modern multimedia","which","online learning services","lu xun\u2019s literature lovers","scholars","lu xun\u2019s research field","academic vision","this paper","an effective online learning system","lu xun\u2019s literature","deep learning","multimedia technology","lu xun\u2019s literature","chinese","the twentieth century","lu xun","lu xun","lu xun literature","two","one","two","xun","xun\u2019s","xun","xun"]},{"title":"Deep learning approaches for lyme disease detection: leveraging progressive resizing and self-supervised learning models","authors":["Daryl Jacob Jerrish","Om Nankar","Shilpa Gite","Shruti Patil","Ketan Kotecha","Ganeshsree Selvachandran","Ajith Abraham"],"abstract":"Lyme disease diagnosis poses a significant challenge, with blood tests exhibiting an alarming inaccuracy rate of nearly 60% in detecting early-stage infections. As a result, there is an urgent need for improved diagnostic methods that can offer more accurate detection outcomes. To address this pressing issue, our study focuses on harnessing the potential of deep learning approaches, specifically by employing model pipelining through progressive resizing and multiple self-supervised learning models. In this paper, we present a comprehensive exploration of self-supervised learning models, including SimCLR, SwAV, MoCo, and BYOL, tailored to the context of Lyme disease detection using medical imaging. The effectiveness and performance of these models are evaluated using standard metrics such as F1 score, precision, recall, and accuracy. Furthermore, we emphasize the significance of progressive resizing and its implications when dealing with convolutional neural networks (CNNs) for medical image analysis. By leveraging deep learning approaches, progressive resizing, and self-supervised learning models, the challenges associated with Lyme disease detection are effectively addressed in this study. The application of our novel methodology and the execution of a comprehensive evaluation framework contribute invaluable insights, fostering the development of more efficient and accurate diagnostic methods for Lyme disease. It is firmly believed that our research will serve as a catalyst, inspiring interdisciplinary collaborations that accelerate progress at the convergence of medicine, computing, and technology, ultimately benefiting public health.","doi":"10.1007\/s11042-023-16306-9","cleaned_title":"deep learning approaches for lyme disease detection: leveraging progressive resizing and self-supervised learning models","cleaned_abstract":"lyme disease diagnosis poses a significant challenge, with blood tests exhibiting an alarming inaccuracy rate of nearly 60% in detecting early-stage infections. as a result, there is an urgent need for improved diagnostic methods that can offer more accurate detection outcomes. to address this pressing issue, our study focuses on harnessing the potential of deep learning approaches, specifically by employing model pipelining through progressive resizing and multiple self-supervised learning models. in this paper, we present a comprehensive exploration of self-supervised learning models, including simclr, swav, moco, and byol, tailored to the context of lyme disease detection using medical imaging. the effectiveness and performance of these models are evaluated using standard metrics such as f1 score, precision, recall, and accuracy. furthermore, we emphasize the significance of progressive resizing and its implications when dealing with convolutional neural networks (cnns) for medical image analysis. by leveraging deep learning approaches, progressive resizing, and self-supervised learning models, the challenges associated with lyme disease detection are effectively addressed in this study. the application of our novel methodology and the execution of a comprehensive evaluation framework contribute invaluable insights, fostering the development of more efficient and accurate diagnostic methods for lyme disease. it is firmly believed that our research will serve as a catalyst, inspiring interdisciplinary collaborations that accelerate progress at the convergence of medicine, computing, and technology, ultimately benefiting public health.","key_phrases":["lyme disease diagnosis","a significant challenge","blood tests","an alarming inaccuracy rate","nearly 60%","early-stage infections","a result","an urgent need","improved diagnostic methods","that","more accurate detection outcomes","this pressing issue","our study","the potential","deep learning approaches","model","progressive resizing","multiple self-supervised learning models","this paper","we","a comprehensive exploration","self-supervised learning models","simclr","swav","moco","byol","the context","lyme disease detection","medical imaging","the effectiveness","performance","these models","standard metrics","f1 score","precision","recall","accuracy","we","the significance","progressive resizing","its implications","convolutional neural networks","cnns","medical image analysis","deep learning approaches","progressive resizing","self-supervised learning models","the challenges","lyme disease detection","this study","the application","our novel methodology","the execution","a comprehensive evaluation framework","invaluable insights","the development","more efficient and accurate diagnostic methods","lyme disease","it","our research","a catalyst","interdisciplinary collaborations","that","progress","the convergence","medicine","computing","technology","public health","nearly 60%"]},{"title":"Deep Learning Models for Skin Cancer Classification Across Diverse Color Spaces: Comprehensive Analysis","authors":["Anisha Paul","Asfak Ali","Sheli Sinha Chaudhuri"],"abstract":"Color space plays an important role in various aspects of imaging tasks. However, in deep learning-based computer vision, the RGB color model is predominantly employed. This research analyzes the impact of deep convolutional neural networks on cancer classification across different color spaces. The five most popular deep learning models undergo training and testing in eleven color spaces, revealing that YUV, LAB, and YIQ consistently outperform other color models in most cases. RGB images are frequently converted to alternative color spaces for enhanced representation in specific applications, like object detection and segmentation. This transformation induces alterations in the features of the color image due to variations in pixel intensity information across different color models. In this research, the aforementioned principle is applied to the classification of skin cancer using deep learning networks on images of skin lesions. The results exhibit diverse responses, with some networks achieving higher accuracy in alternative color spaces compared to RGB, while others do not. This study provides insights into the classification performance across RGB, HED, HSV, LAB, RGBCIE, XYZ, YCbCr, YDbDr, YIQ, YPbPr, and YUV color spaces. The research aims to illustrate how deep learning facilitates the analysis of skin cancer images in different color spaces.","doi":"10.1007\/s11831-024-10160-0","cleaned_title":"deep learning models for skin cancer classification across diverse color spaces: comprehensive analysis","cleaned_abstract":"color space plays an important role in various aspects of imaging tasks. however, in deep learning-based computer vision, the rgb color model is predominantly employed. this research analyzes the impact of deep convolutional neural networks on cancer classification across different color spaces. the five most popular deep learning models undergo training and testing in eleven color spaces, revealing that yuv, lab, and yiq consistently outperform other color models in most cases. rgb images are frequently converted to alternative color spaces for enhanced representation in specific applications, like object detection and segmentation. this transformation induces alterations in the features of the color image due to variations in pixel intensity information across different color models. in this research, the aforementioned principle is applied to the classification of skin cancer using deep learning networks on images of skin lesions. the results exhibit diverse responses, with some networks achieving higher accuracy in alternative color spaces compared to rgb, while others do not. this study provides insights into the classification performance across rgb, hed, hsv, lab, rgbcie, xyz, ycbcr, ydbdr, yiq, ypbpr, and yuv color spaces. the research aims to illustrate how deep learning facilitates the analysis of skin cancer images in different color spaces.","key_phrases":["color space","an important role","various aspects","imaging tasks","deep learning-based computer vision","the rgb color model","this research","the impact","deep convolutional neural networks","cancer classification","different color spaces","the five most popular deep learning models","training","testing","eleven color spaces","yuv","lab","yiq","other color models","most cases","rgb images","alternative color spaces","enhanced representation","specific applications","object detection","segmentation","this transformation","alterations","the features","the color image","variations","pixel intensity information","different color models","this research","the aforementioned principle","the classification","skin cancer","deep learning networks","images","skin lesions","the results","diverse responses","some networks","higher accuracy","alternative color spaces","rgb","others","this study","insights","the classification performance","rgb","the research","how deep learning","the analysis","skin cancer images","different color spaces","rgb","five","eleven","yuv","rgb","rgb","rgb"]},{"title":"Scaffolding cooperation in human groups with deep reinforcement learning","authors":["Kevin R. McKee","Andrea Tacchetti","Michiel A. Bakker","Jan Balaguer","Lucy Campbell-Gillingham","Richard Everett","Matthew Botvinick"],"abstract":"Effective approaches to encouraging group cooperation are still an open challenge. Here we apply recent advances in deep learning to structure networks of human participants playing a group cooperation game. We leverage deep reinforcement learning and simulation methods to train a \u2018social planner\u2019 capable of making recommendations to create or break connections between group members. The strategy that it develops succeeds at encouraging pro-sociality in networks of human participants (N\u2009=\u2009208 participants in 13 groups) playing for real monetary stakes. Under the social planner, groups finished the game with an average cooperation rate of 77.7%, compared with 42.8% in static networks (N\u2009=\u2009176 in 11 groups). In contrast to prior strategies that separate defectors from cooperators (tested here with N\u2009=\u2009384 in 24 groups), the social planner learns to take a conciliatory approach to defectors, encouraging them to act pro-socially by moving them to small highly cooperative neighbourhoods.","doi":"10.1038\/s41562-023-01686-7","cleaned_title":"scaffolding cooperation in human groups with deep reinforcement learning","cleaned_abstract":"effective approaches to encouraging group cooperation are still an open challenge. here we apply recent advances in deep learning to structure networks of human participants playing a group cooperation game. we leverage deep reinforcement learning and simulation methods to train a \u2018social planner\u2019 capable of making recommendations to create or break connections between group members. the strategy that it develops succeeds at encouraging pro-sociality in networks of human participants (n = 208 participants in 13 groups) playing for real monetary stakes. under the social planner, groups finished the game with an average cooperation rate of 77.7%, compared with 42.8% in static networks (n = 176 in 11 groups). in contrast to prior strategies that separate defectors from cooperators (tested here with n = 384 in 24 groups), the social planner learns to take a conciliatory approach to defectors, encouraging them to act pro-socially by moving them to small highly cooperative neighbourhoods.","key_phrases":["effective approaches","group cooperation","an open challenge","we","recent advances","deep learning","networks","human participants","a group cooperation game","we","deep reinforcement learning and simulation methods","a \u2018social planner","recommendations","connections","group members","the strategy","that","it","networks","human participants","= 208 participants","13 groups","real monetary stakes","the social planner","groups","the game","an average cooperation rate","77.7%","42.8%","static networks","11 groups","contrast","prior strategies","that","separate defectors","cooperators","n","24 groups","the social planner","a conciliatory approach","defectors","them","them","small highly cooperative neighbourhoods","208","13","77.7%","42.8%","176","11","384","24"]},{"title":"Deep-learning-enabled brain hemodynamic mapping using resting-state fMRI","authors":["Xirui Hou","Pengfei Guo","Puyang Wang","Peiying Liu","Doris D. M. Lin","Hongli Fan","Yang Li","Zhiliang Wei","Zixuan Lin","Dengrong Jiang","Jin Jin","Catherine Kelly","Jay J. Pillai","Judy Huang","Marco C. Pinho","Binu P. Thomas","Babu G. Welch","Denise C. Park","Vishal M. Patel","Argye E. Hillis","Hanzhang Lu"],"abstract":"Cerebrovascular disease is a leading cause of death globally. Prevention and early intervention are known to be the most effective forms of its management. Non-invasive imaging methods hold great promises for early stratification, but at present lack the sensitivity for personalized prognosis. Resting-state functional magnetic resonance imaging (rs-fMRI), a powerful tool previously used for mapping neural activity, is available in most hospitals. Here we show that rs-fMRI can be used to map cerebral hemodynamic function and delineate impairment. By exploiting time variations in breathing pattern during rs-fMRI, deep learning enables reproducible mapping of cerebrovascular reactivity (CVR) and bolus arrival time (BAT) of the human brain using resting-state CO2 fluctuations as a natural \u201ccontrast media\u201d. The deep-learning network is trained with CVR and BAT maps obtained with a reference method of CO2-inhalation MRI, which includes data from young and older healthy subjects and patients with Moyamoya disease and brain tumors. We demonstrate the performance of deep-learning cerebrovascular mapping in the detection of vascular abnormalities, evaluation of revascularization effects, and vascular alterations in normal aging. In addition, cerebrovascular maps obtained with the proposed method exhibit excellent reproducibility in both healthy volunteers and stroke patients. Deep-learning resting-state vascular imaging has the potential to become a useful tool in clinical cerebrovascular imaging.","doi":"10.1038\/s41746-023-00859-y","cleaned_title":"deep-learning-enabled brain hemodynamic mapping using resting-state fmri","cleaned_abstract":"cerebrovascular disease is a leading cause of death globally. prevention and early intervention are known to be the most effective forms of its management. non-invasive imaging methods hold great promises for early stratification, but at present lack the sensitivity for personalized prognosis. resting-state functional magnetic resonance imaging (rs-fmri), a powerful tool previously used for mapping neural activity, is available in most hospitals. here we show that rs-fmri can be used to map cerebral hemodynamic function and delineate impairment. by exploiting time variations in breathing pattern during rs-fmri, deep learning enables reproducible mapping of cerebrovascular reactivity (cvr) and bolus arrival time (bat) of the human brain using resting-state co2 fluctuations as a natural \u201ccontrast media\u201d. the deep-learning network is trained with cvr and bat maps obtained with a reference method of co2-inhalation mri, which includes data from young and older healthy subjects and patients with moyamoya disease and brain tumors. we demonstrate the performance of deep-learning cerebrovascular mapping in the detection of vascular abnormalities, evaluation of revascularization effects, and vascular alterations in normal aging. in addition, cerebrovascular maps obtained with the proposed method exhibit excellent reproducibility in both healthy volunteers and stroke patients. deep-learning resting-state vascular imaging has the potential to become a useful tool in clinical cerebrovascular imaging.","key_phrases":["cerebrovascular disease","a leading cause","death","prevention","early intervention","the most effective forms","its management","non-invasive imaging methods","great promises","early stratification","present lack","personalized prognosis","resting-state functional magnetic resonance imaging","rs-fmri","a powerful tool","neural activity","most hospitals","we","rs","fmri","cerebral hemodynamic function","delineate impairment","time variations","breathing pattern","rs-fmri","deep learning","reproducible mapping","cerebrovascular reactivity","cvr","bolus arrival time","bat","the human brain","resting-state co2 fluctuations","a natural \u201ccontrast media","the deep-learning network","cvr","bat maps","a reference method","co2-inhalation mri","which","data","young and older healthy subjects","patients","moyamoya disease","brain tumors","we","the performance","deep-learning cerebrovascular mapping","the detection","vascular abnormalities","evaluation","revascularization effects","vascular alterations","normal aging","addition","cerebrovascular maps","the proposed method","excellent reproducibility","both healthy volunteers","stroke patients","deep-learning resting-state vascular imaging","the potential","a useful tool","clinical cerebrovascular imaging"]},{"title":"An Alternative to Cognitivism: Computational Phenomenology for Deep Learning","authors":["Pierre Beckmann","Guillaume K\u00f6stner","In\u00eas Hip\u00f3lito"],"abstract":"We propose a non-representationalist framework for deep learning relying on a novel method computational phenomenology, a dialogue between the first-person perspective (relying on phenomenology) and the mechanisms of computational models. We thereby propose an alternative to the modern cognitivist interpretation of deep learning, according to which artificial neural networks encode representations of external entities. This interpretation mainly relies on neuro-representationalism, a position that combines a strong ontological commitment towards scientific theoretical entities and the idea that the brain operates on symbolic representations of these entities. We proceed as follows: after offering a review of cognitivism and neuro-representationalism in the field of deep learning, we first elaborate a phenomenological critique of these positions; we then sketch out computational phenomenology and distinguish it from existing alternatives; finally we apply this new method to deep learning models trained on specific tasks, in order to formulate a conceptual framework of deep-learning, that allows one to think of artificial neural networks\u2019 mechanisms in terms of lived experience.","doi":"10.1007\/s11023-023-09638-w","cleaned_title":"an alternative to cognitivism: computational phenomenology for deep learning","cleaned_abstract":"we propose a non-representationalist framework for deep learning relying on a novel method computational phenomenology, a dialogue between the first-person perspective (relying on phenomenology) and the mechanisms of computational models. we thereby propose an alternative to the modern cognitivist interpretation of deep learning, according to which artificial neural networks encode representations of external entities. this interpretation mainly relies on neuro-representationalism, a position that combines a strong ontological commitment towards scientific theoretical entities and the idea that the brain operates on symbolic representations of these entities. we proceed as follows: after offering a review of cognitivism and neuro-representationalism in the field of deep learning, we first elaborate a phenomenological critique of these positions; we then sketch out computational phenomenology and distinguish it from existing alternatives; finally we apply this new method to deep learning models trained on specific tasks, in order to formulate a conceptual framework of deep-learning, that allows one to think of artificial neural networks\u2019 mechanisms in terms of lived experience.","key_phrases":["we","a non-representationalist framework","deep learning","a novel method","computational phenomenology","a dialogue","the first-person perspective","phenomenology","the mechanisms","computational models","we","an alternative","the modern cognitivist interpretation","deep learning","which","artificial neural networks encode representations","external entities","this interpretation","neuro-representationalism","a position","that","a strong ontological commitment","scientific theoretical entities","the idea","the brain","symbolic representations","these entities","we","a review","cognitivism","neuro-representationalism","the field","deep learning","we","a phenomenological critique","these positions","we","computational phenomenology","it","existing alternatives","we","this new method","deep learning models","specific tasks","order","a conceptual framework","deep-learning","that","artificial neural networks\u2019 mechanisms","terms","lived experience","first","first"]},{"title":"Slideflow: deep learning for digital histopathology with real-time whole-slide visualization","authors":["James M. Dolezal","Sara Kochanny","Emma Dyer","Siddhi Ramesh","Andrew Srisuwananukorn","Matteo Sacco","Frederick M. Howard","Anran Li","Prajval Mohan","Alexander T. Pearson"],"abstract":"Deep learning methods have emerged as powerful tools for analyzing histopathological images, but current methods are often specialized for specific domains and software environments, and few open-source options exist for deploying models in an interactive interface. Experimenting with different deep learning approaches typically requires switching software libraries and reprocessing data, reducing the feasibility and practicality of experimenting with new architectures. We developed a flexible deep learning library for histopathology called Slideflow, a package which supports a broad array of deep learning methods for digital pathology and includes a fast whole-slide interface for deploying trained models. Slideflow includes unique tools for whole-slide image data processing, efficient stain normalization and augmentation, weakly-supervised whole-slide classification, uncertainty quantification, feature generation, feature space analysis, and explainability. Whole-slide image processing is highly optimized, enabling whole-slide tile extraction at 40x magnification in 2.5\u00a0s per slide. The framework-agnostic data processing pipeline enables rapid experimentation with new methods built with either Tensorflow or PyTorch, and the graphical user interface supports real-time visualization of slides, predictions, heatmaps, and feature space characteristics on a variety of hardware devices, including ARM-based devices such as the Raspberry Pi.","doi":"10.1186\/s12859-024-05758-x","cleaned_title":"slideflow: deep learning for digital histopathology with real-time whole-slide visualization","cleaned_abstract":"deep learning methods have emerged as powerful tools for analyzing histopathological images, but current methods are often specialized for specific domains and software environments, and few open-source options exist for deploying models in an interactive interface. experimenting with different deep learning approaches typically requires switching software libraries and reprocessing data, reducing the feasibility and practicality of experimenting with new architectures. we developed a flexible deep learning library for histopathology called slideflow, a package which supports a broad array of deep learning methods for digital pathology and includes a fast whole-slide interface for deploying trained models. slideflow includes unique tools for whole-slide image data processing, efficient stain normalization and augmentation, weakly-supervised whole-slide classification, uncertainty quantification, feature generation, feature space analysis, and explainability. whole-slide image processing is highly optimized, enabling whole-slide tile extraction at 40x magnification in 2.5 s per slide. the framework-agnostic data processing pipeline enables rapid experimentation with new methods built with either tensorflow or pytorch, and the graphical user interface supports real-time visualization of slides, predictions, heatmaps, and feature space characteristics on a variety of hardware devices, including arm-based devices such as the raspberry pi.","key_phrases":["deep learning methods","powerful tools","histopathological images","current methods","specific domains","software environments","few open-source options","models","an interactive interface","different deep learning approaches","software libraries","reprocessing data","the feasibility","practicality","new architectures","we","a flexible deep learning library","histopathology","a package","which","a broad array","deep learning methods","digital pathology","a fast whole-slide interface","trained models","slideflow","unique tools","whole-slide image data processing","efficient stain normalization","augmentation","weakly-supervised whole-slide classification","uncertainty quantification","feature generation","feature space analysis","explainability","whole-slide image processing","whole-slide tile extraction","40x magnification","2.5 s","slide","the framework-agnostic data processing pipeline","rapid experimentation","new methods","either tensorflow","pytorch","the graphical user interface","real-time visualization","slides","predictions","heatmaps","feature space characteristics","a variety","hardware devices","arm-based devices","the raspberry pi","40x","2.5"]},{"title":"Authorship attribution in twitter: a comparative study of machine learning and deep learning approaches","authors":["Rebeh Imane Ammar Aouchiche","Fatima Boumahdi","Mohamed Abdelkarim Remmide","Amina Madani"],"abstract":"As social media platforms gain popularity and influence, content integrity and user accountability issues become more critical. Authorship attribution (AA) is a powerful tool for tackling such issues by accurately determining the real author of online posts. This study proposes an AA approach using machine and deep learning algorithms to accurately predict the author of unknown posts on social media platforms. It introduces Temporal Convolutional Networks (TCN) for short texts, investigates the effectiveness of combining Long Short-Term Memory (LSTM) and Convolutional Neural Networks (CNN), and explores the use of an Autoencoder combined with Adaboost classifier. This approach was tested on a Twitter dataset, achieving 52.77% accuracy in AA through multiple experiments across various scenarios.","doi":"10.1007\/s41870-024-01788-z","cleaned_title":"authorship attribution in twitter: a comparative study of machine learning and deep learning approaches","cleaned_abstract":"as social media platforms gain popularity and influence, content integrity and user accountability issues become more critical. authorship attribution (aa) is a powerful tool for tackling such issues by accurately determining the real author of online posts. this study proposes an aa approach using machine and deep learning algorithms to accurately predict the author of unknown posts on social media platforms. it introduces temporal convolutional networks (tcn) for short texts, investigates the effectiveness of combining long short-term memory (lstm) and convolutional neural networks (cnn), and explores the use of an autoencoder combined with adaboost classifier. this approach was tested on a twitter dataset, achieving 52.77% accuracy in aa through multiple experiments across various scenarios.","key_phrases":["social media platforms","popularity","influence","content integrity","user accountability issues","authorship attribution","(aa","a powerful tool","such issues","the real author","online posts","this study","an aa approach","machine","algorithms","the author","unknown posts","social media platforms","it","temporal convolutional networks","tcn","short texts","the effectiveness","long short-term memory","lstm","convolutional neural networks","cnn","the use","an autoencoder","adaboost classifier","this approach","a twitter dataset","52.77% accuracy","aa","multiple experiments","various scenarios","cnn","52.77%"]},{"title":"Ultrasound-based deep learning radiomics nomogram for differentiating mass mastitis from invasive breast cancer","authors":["Linyong Wu","Songhua Li","Chaojun Wu","Shaofeng Wu","Yan Lin","Dayou Wei"],"abstract":"BackgroundThe purpose of this study is to develop and validate the potential value of the deep learning radiomics nomogram (DLRN) based on ultrasound to differentiate mass mastitis (MM) and invasive breast cancer (IBC).Methods50 cases of MM and 180 cases of IBC with ultrasound Breast Imaging Reporting and Data System 4 category were recruited (training cohort, n\u2009=\u2009161, validation cohort, n\u2009=\u200969). Based on PyRadiomics and ResNet50 extractors, radiomics and deep learning features were extracted, respectively. Based on supervised machine learning methods such as logistic regression, random forest, and support vector machine, as well as unsupervised machine learning methods using K-means clustering analysis, the differences in features between MM and IBC were analyzed to develop DLRN. The performance of DLRN had been evaluated by receiver operating characteristic curve, calibration, and clinical practicality.ResultsSupervised machine learning results showed that compared with radiomics models, especially random forest models, deep learning models were better at recognizing MM and IBC. The area under the curve (AUC) of the validation cohort was 0.84, the accuracy was 0.83, the sensitivity was 0.73, and the specificity was 0.83. Compared to radiomics or deep learning models, DLRN even further improved discrimination ability (AUC of 0.90 and 0.90, accuracy of 0.83 and 0.88 for training and validation cohorts), which had better clinical benefits and good calibratability. In addition, the information heterogeneity of deep learning features in MM and IBC was validated again through unsupervised machine learning clustering analysis, indicating that MM had a unique features phenotype.ConclusionThe DLRN developed based on radiomics and deep learning features of ultrasound images has potential clinical value in effectively distinguishing between MM and IBC. DLRN breaks through visual limitations and quantifies more image information related to MM based on computers, further utilizing machine learning to effectively utilize this information for clinical decision-making. As DLRN becomes an autonomous screening system, it will improve the recognition rate of MM in grassroots hospitals and reduce the possibility of incorrect treatment and overtreatment.","doi":"10.1186\/s12880-024-01353-x","cleaned_title":"ultrasound-based deep learning radiomics nomogram for differentiating mass mastitis from invasive breast cancer","cleaned_abstract":"backgroundthe purpose of this study is to develop and validate the potential value of the deep learning radiomics nomogram (dlrn) based on ultrasound to differentiate mass mastitis (mm) and invasive breast cancer (ibc).methods50 cases of mm and 180 cases of ibc with ultrasound breast imaging reporting and data system 4 category were recruited (training cohort, n = 161, validation cohort, n = 69). based on pyradiomics and resnet50 extractors, radiomics and deep learning features were extracted, respectively. based on supervised machine learning methods such as logistic regression, random forest, and support vector machine, as well as unsupervised machine learning methods using k-means clustering analysis, the differences in features between mm and ibc were analyzed to develop dlrn. the performance of dlrn had been evaluated by receiver operating characteristic curve, calibration, and clinical practicality.resultssupervised machine learning results showed that compared with radiomics models, especially random forest models, deep learning models were better at recognizing mm and ibc. the area under the curve (auc) of the validation cohort was 0.84, the accuracy was 0.83, the sensitivity was 0.73, and the specificity was 0.83. compared to radiomics or deep learning models, dlrn even further improved discrimination ability (auc of 0.90 and 0.90, accuracy of 0.83 and 0.88 for training and validation cohorts), which had better clinical benefits and good calibratability. in addition, the information heterogeneity of deep learning features in mm and ibc was validated again through unsupervised machine learning clustering analysis, indicating that mm had a unique features phenotype.conclusionthe dlrn developed based on radiomics and deep learning features of ultrasound images has potential clinical value in effectively distinguishing between mm and ibc. dlrn breaks through visual limitations and quantifies more image information related to mm based on computers, further utilizing machine learning to effectively utilize this information for clinical decision-making. as dlrn becomes an autonomous screening system, it will improve the recognition rate of mm in grassroots hospitals and reduce the possibility of incorrect treatment and overtreatment.","key_phrases":["backgroundthe purpose","this study","the potential value","the deep learning radiomics nomogram","dlrn","ultrasound","mm","mm","ibc","ultrasound breast imaging reporting","data system","pyradiomics","resnet50 extractors","radiomics","deep learning features","supervised machine learning methods","logistic regression","random forest","vector machine","unsupervised machine learning methods","k","clustering analysis","the differences","features","mm","ibc","dlrn","the performance","dlrn","receiver operating characteristic curve","calibration","clinical practicality.resultssupervised machine learning results","radiomics models","especially random forest models","deep learning models","mm","ibc","the area","the curve","auc","the validation cohort","the accuracy","the sensitivity","the specificity","radiomics","deep learning models","even further improved discrimination ability","auc","accuracy","training and validation cohorts","which","better clinical benefits","good calibratability","addition","the information heterogeneity","deep learning features","mm","ibc","unsupervised machine","clustering analysis","mm","a unique features","phenotype.conclusionthe dlrn","radiomics","deep learning features","ultrasound images","potential clinical value","mm","ibc","dlrn","visual limitations","quantifies","more image information","mm","computers","this information","clinical decision-making","dlrn","an autonomous screening system","it","the recognition rate","mm","grassroots hospitals","the possibility","incorrect treatment","overtreatment","mm","mm","180","4","161","69","resnet50","mm","0.84","0.83","0.73","0.83","0.90","0.90","0.83","0.88","mm","mm","dlrn","mm","mm","dlrn"]},{"title":"Two-layer Ensemble of Deep Learning Models for Medical Image Segmentation","authors":["Truong Dang","Tien Thanh Nguyen","John McCall","Eyad Elyan","Carlos Francisco Moreno-Garc\u00eda"],"abstract":"One of the most important areas in medical image analysis is segmentation, in which raw image data is partitioned into structured and meaningful regions to gain further insights. By using Deep Neural Networks (DNN), AI-based automated segmentation algorithms can potentially assist physicians with more effective imaging-based diagnoses. However, since it is difficult to acquire high-quality ground truths for medical images and DNN hyperparameters require significant manual tuning, the results by DNN-based medical models might be limited. A potential solution is to combine multiple DNN models using ensemble learning. We propose a two-layer ensemble of deep learning models in which the prediction of each training image pixel made by each model in the first layer is used as the augmented data of the training image for the second layer of the ensemble. The prediction of the second layer is then combined by using a weight-based scheme which is found by solving linear regression problems. To the best of our knowledge, our paper is the first work which proposes a two-layer ensemble of deep learning models with an augmented data technique in medical image segmentation. Experiments conducted on five different medical image datasets for diverse segmentation tasks show that proposed method achieves better results in terms of several performance metrics compared to some well-known benchmark algorithms. Our proposed two-layer ensemble of deep learning models for segmentation of medical images shows effectiveness compared to several benchmark algorithms. The research can be expanded in several directions like image classification.","doi":"10.1007\/s12559-024-10257-5","cleaned_title":"two-layer ensemble of deep learning models for medical image segmentation","cleaned_abstract":"one of the most important areas in medical image analysis is segmentation, in which raw image data is partitioned into structured and meaningful regions to gain further insights. by using deep neural networks (dnn), ai-based automated segmentation algorithms can potentially assist physicians with more effective imaging-based diagnoses. however, since it is difficult to acquire high-quality ground truths for medical images and dnn hyperparameters require significant manual tuning, the results by dnn-based medical models might be limited. a potential solution is to combine multiple dnn models using ensemble learning. we propose a two-layer ensemble of deep learning models in which the prediction of each training image pixel made by each model in the first layer is used as the augmented data of the training image for the second layer of the ensemble. the prediction of the second layer is then combined by using a weight-based scheme which is found by solving linear regression problems. to the best of our knowledge, our paper is the first work which proposes a two-layer ensemble of deep learning models with an augmented data technique in medical image segmentation. experiments conducted on five different medical image datasets for diverse segmentation tasks show that proposed method achieves better results in terms of several performance metrics compared to some well-known benchmark algorithms. our proposed two-layer ensemble of deep learning models for segmentation of medical images shows effectiveness compared to several benchmark algorithms. the research can be expanded in several directions like image classification.","key_phrases":["the most important areas","medical image analysis","segmentation","which","raw image data","structured and meaningful regions","further insights","deep neural networks","dnn","ai-based automated segmentation algorithms","physicians","more effective imaging-based diagnoses","it","high-quality ground truths","medical images","dnn hyperparameters","significant manual tuning","the results","dnn-based medical models","a potential solution","multiple dnn models","ensemble learning","we","a two-layer ensemble","deep learning models","which","the prediction","each training image pixel","each model","the first layer","the augmented data","the training image","the second layer","the ensemble","the prediction","the second layer","a weight-based scheme","which","linear regression problems","our knowledge","our paper","the first work","which","a two-layer ensemble","deep learning models","an augmented data technique","medical image segmentation","experiments","five different medical image datasets","diverse segmentation tasks","proposed method","better results","terms","several performance metrics","some well-known benchmark algorithms","our proposed two-layer ensemble","deep learning models","segmentation","medical images","effectiveness","several benchmark algorithms","the research","several directions","image classification","one","two","first","second","second","first","two","five","two"]},{"title":"A Diffusion Equation for Improving the Robustness of Deep Learning Speckle Removal Model","authors":["Li Cheng","Yuming Xing","Yao Li","Zhichang Guo"],"abstract":"Speckle removal aims to smooth noise while preserving image boundaries and texture information. In recent years, speckle removal models based on deep learning methods have attracted a lot of attention. However, it was found that these models are less robust to adversarial attacks. The adversarial attack makes the image recovery of deep learning methods significantly less effective when the speckle noise distribution is almost unchanged. In purpose of addressing the above problem, we propose a diffusion equation-based speckle removal model that can improve the robustness of deep learning algorithms in this paper. The model utilizes a deep learning image prior and an image grayscale detection operator together to construct the coefficient function of the diffusion equation. Among them, there is a high possibility that the deep learning image prior is inaccurate or even incorrect, but it will not affect the performance and the properties of the proposed diffusion equation model for noise removal. Moreover, we analyze the robustness of the proposed diffusion equation model in terms of theoretical and numerical properties. Experiments show that our proposed diffusion equation speckle removal model is not affected by adversarial attacks in any way and has stronger robustness.","doi":"10.1007\/s10851-024-01199-6","cleaned_title":"a diffusion equation for improving the robustness of deep learning speckle removal model","cleaned_abstract":"speckle removal aims to smooth noise while preserving image boundaries and texture information. in recent years, speckle removal models based on deep learning methods have attracted a lot of attention. however, it was found that these models are less robust to adversarial attacks. the adversarial attack makes the image recovery of deep learning methods significantly less effective when the speckle noise distribution is almost unchanged. in purpose of addressing the above problem, we propose a diffusion equation-based speckle removal model that can improve the robustness of deep learning algorithms in this paper. the model utilizes a deep learning image prior and an image grayscale detection operator together to construct the coefficient function of the diffusion equation. among them, there is a high possibility that the deep learning image prior is inaccurate or even incorrect, but it will not affect the performance and the properties of the proposed diffusion equation model for noise removal. moreover, we analyze the robustness of the proposed diffusion equation model in terms of theoretical and numerical properties. experiments show that our proposed diffusion equation speckle removal model is not affected by adversarial attacks in any way and has stronger robustness.","key_phrases":["speckle removal","noise","image boundaries and texture information","recent years","speckle removal models","deep learning methods","a lot","attention","it","these models","adversarial attacks","the adversarial attack","the image recovery","deep learning methods","the speckle noise distribution","purpose","the above problem","we","a diffusion equation-based speckle removal model","that","the robustness","deep learning algorithms","this paper","the model","a deep learning image","an image grayscale detection operator","the coefficient function","the diffusion equation","them","a high possibility","the deep learning image","it","the performance","the properties","the proposed diffusion equation model","noise removal","we","the robustness","the proposed diffusion equation model","terms","theoretical and numerical properties","experiments","our proposed diffusion equation speckle removal model","adversarial attacks","any way","stronger robustness","recent years"]},{"title":"Application of deep learning technique in next generation sequence experiments","authors":["Su \u00d6zg\u00fcr","Mehmet Orman"],"abstract":"In recent years, the widespread utilization of biological data processing technology has been driven by its cost-effectiveness. Consequently, next-generation sequencing (NGS) has become an integral component of biological research. NGS technologies enable the sequencing of billions of nucleotides in the entire genome, transcriptome, or specific target regions. This sequencing generates vast data matrices. Consequently, there is a growing demand for deep learning (DL) approaches, which employ multilayer artificial neural networks and systems capable of extracting meaningful information from these extensive data structures. In this study, the aim was to obtain optimized parameters and assess the prediction performance of deep learning and machine learning (ML) algorithms for binary classification in real and simulated whole genome data using a cloud-based system. The ART-simulated data and paired-end NGS (whole genome) data of Ch22, which includes ethnicity information, were evaluated using XGBoost, LightGBM, and DL algorithms. When the learning rate was set to 0.01 and 0.001, and the epoch values were updated to 500, 1000, and 2000 in the deep learning model for the ART simulated dataset, the median accuracy values of the ART models were as follows: 0.6320, 0.6800, and 0.7340 for epoch 0.01; and 0.6920, 0.7220, and 0.8020 for epoch 0.001, respectively. In comparison, the median accuracy values of the XGBoost and LightGBM models were 0.6990 and 0.6250 respectively. When the same process is repeated for Chr 22, the results are as follows: the median accuracy values of the DL models were 0.5290, 0.5420 and 0.5820 for epoch 0.01; and 0.5510, 0.5830 and 0.6040 for epoch 0.001, respectively. Additionally, the median accuracy values of the XGBoost and LightGBM models were 0.5760 and 0.5250, respectively. While the best classification estimates were obtained at 2000 epochs and a learning rate (LR) value of 0.001 for both real and simulated data, the XGBoost algorithm showed higher performance when the epoch value was 500 and the LR was 0.01. When dealing with class imbalance, the DL algorithm yielded similar and high Recall and Precision values. Conclusively, this study serves as a timely resource for genomic scientists, providing guidance on why, when, and how to effectively utilize deep learning\/machine learning methods for the analysis of human genomic data.","doi":"10.1186\/s40537-023-00838-w","cleaned_title":"application of deep learning technique in next generation sequence experiments","cleaned_abstract":"in recent years, the widespread utilization of biological data processing technology has been driven by its cost-effectiveness. consequently, next-generation sequencing (ngs) has become an integral component of biological research. ngs technologies enable the sequencing of billions of nucleotides in the entire genome, transcriptome, or specific target regions. this sequencing generates vast data matrices. consequently, there is a growing demand for deep learning (dl) approaches, which employ multilayer artificial neural networks and systems capable of extracting meaningful information from these extensive data structures. in this study, the aim was to obtain optimized parameters and assess the prediction performance of deep learning and machine learning (ml) algorithms for binary classification in real and simulated whole genome data using a cloud-based system. the art-simulated data and paired-end ngs (whole genome) data of ch22, which includes ethnicity information, were evaluated using xgboost, lightgbm, and dl algorithms. when the learning rate was set to 0.01 and 0.001, and the epoch values were updated to 500, 1000, and 2000 in the deep learning model for the art simulated dataset, the median accuracy values of the art models were as follows: 0.6320, 0.6800, and 0.7340 for epoch 0.01; and 0.6920, 0.7220, and 0.8020 for epoch 0.001, respectively. in comparison, the median accuracy values of the xgboost and lightgbm models were 0.6990 and 0.6250 respectively. when the same process is repeated for chr 22, the results are as follows: the median accuracy values of the dl models were 0.5290, 0.5420 and 0.5820 for epoch 0.01; and 0.5510, 0.5830 and 0.6040 for epoch 0.001, respectively. additionally, the median accuracy values of the xgboost and lightgbm models were 0.5760 and 0.5250, respectively. while the best classification estimates were obtained at 2000 epochs and a learning rate (lr) value of 0.001 for both real and simulated data, the xgboost algorithm showed higher performance when the epoch value was 500 and the lr was 0.01. when dealing with class imbalance, the dl algorithm yielded similar and high recall and precision values. conclusively, this study serves as a timely resource for genomic scientists, providing guidance on why, when, and how to effectively utilize deep learning\/machine learning methods for the analysis of human genomic data.","key_phrases":["recent years","the widespread utilization","biological data processing technology","its cost-effectiveness","(ngs","an integral component","biological research","ngs technologies","the sequencing","billions","nucleotides","the entire genome","specific target regions","vast data matrices","a growing demand","deep learning (dl) approaches","which","multilayer artificial neural networks","systems","meaningful information","these extensive data structures","this study","the aim","optimized parameters","the prediction performance","deep learning","machine learning","ml","algorithms","binary classification","real and simulated whole genome data","a cloud-based system","the art-simulated data","paired-end ngs","whole genome) data","ch22","which","ethnicity information","xgboost","lightgbm","the learning rate","the epoch values","the deep learning model","the art","the median accuracy values","the art models","epoch","epoch","comparison","the median accuracy values","the xgboost and lightgbm models","the same process","chr","the results","the median accuracy values","the dl models","epoch","epoch","the median accuracy values","the xgboost and lightgbm models","the best classification estimates","2000 epochs","a learning rate (lr) value","both real and simulated data","the xgboost algorithm","higher performance","the epoch value","the lr","class imbalance","the dl algorithm","similar and high recall","precision values","this study","a timely resource","genomic scientists","guidance","deep learning\/machine learning methods","the analysis","human genomic data","recent years","billions","ch22","0.01","0.001","500, 1000","2000","0.6320","0.6800","0.7340","0.01","0.6920","0.7220","0.8020","0.001","0.6990","0.6250","chr","22","0.5290","0.5420","0.5820","0.01","0.5510","0.5830","0.6040","0.001","0.5760","0.5250","2000","0.001","500","0.01"]},{"title":"Unsupervised deep learning for geometric feature detection and multilevel-multimodal image registration","authors":["Mohamed Lajili","Zakaria Belhachmi","Maher Moakher","Anis Theljani"],"abstract":"Medical image registration is a crucial step in computer-assisted medical diagnosis, and has seen significant progress with the adoption of deep learning methods like convolutional neural networks (CNN). Creating a deep learning network for image registration is complex because humans can\u2019t easily prepare or supervise the training data unless it\u2019s very basic. This article presents an innovative approach to unsupervised deep learning-based multilevel image registration approach. We propose to develop a CNN to detect the geometric features, such as edges and thin structures, from images using a loss function derived from the Blake-Zisserman energy. This method enables the detection of discontinuities at different scales without relying on labeled data. Subsequently, we use this geometric information extracted from the input images, to define a second loss function and to perform our multimodal image registration process. Furthermore, we introduce a novel deep neural network architecture for multilevel image registration, offering enhanced precision and efficiency compared to traditional methods. Numerical simulations are employed to demonstrate the accuracy and relevance of our approach. We perform some numerical simulations to show the accuracy and the relevance of our approach for multimodal registration and its multilevel implementation.","doi":"10.1007\/s10489-024-05585-w","cleaned_title":"unsupervised deep learning for geometric feature detection and multilevel-multimodal image registration","cleaned_abstract":"medical image registration is a crucial step in computer-assisted medical diagnosis, and has seen significant progress with the adoption of deep learning methods like convolutional neural networks (cnn). creating a deep learning network for image registration is complex because humans can\u2019t easily prepare or supervise the training data unless it\u2019s very basic. this article presents an innovative approach to unsupervised deep learning-based multilevel image registration approach. we propose to develop a cnn to detect the geometric features, such as edges and thin structures, from images using a loss function derived from the blake-zisserman energy. this method enables the detection of discontinuities at different scales without relying on labeled data. subsequently, we use this geometric information extracted from the input images, to define a second loss function and to perform our multimodal image registration process. furthermore, we introduce a novel deep neural network architecture for multilevel image registration, offering enhanced precision and efficiency compared to traditional methods. numerical simulations are employed to demonstrate the accuracy and relevance of our approach. we perform some numerical simulations to show the accuracy and the relevance of our approach for multimodal registration and its multilevel implementation.","key_phrases":["medical image registration","a crucial step","computer-assisted medical diagnosis","significant progress","the adoption","deep learning methods","convolutional neural networks","cnn","a deep learning network","image registration","humans","the training data","it","this article","an innovative approach","unsupervised deep learning-based multilevel image registration approach","we","a cnn","the geometric features","edges","thin structures","images","a loss function","the blake-zisserman energy","this method","the detection","discontinuities","different scales","labeled data","we","this geometric information","the input images","a second loss function","our multimodal image registration process","we","a novel deep neural network architecture","multilevel image registration","enhanced precision","efficiency","traditional methods","numerical simulations","the accuracy","relevance","our approach","we","some numerical simulations","the accuracy","the relevance","our approach","multimodal registration","its multilevel implementation","cnn","cnn","second"]},{"title":"Generative deep learning for data generation in natural hazard analysis: motivations, advances, challenges, and opportunities","authors":["Zhengjing Ma","Gang Mei","Nengxiong Xu"],"abstract":"Data mining and analysis are critical for preventing or mitigating natural hazards. However, data availability in natural hazard analysis is experiencing unprecedented challenges due to economic, technical, and environmental constraints. Recently, generative deep learning has become an increasingly attractive solution to these challenges, which can augment, impute, or synthesize data based on these learned complex, high-dimensional probability distributions of data. Over the last several years, much research has demonstrated the remarkable capabilities of generative deep learning for addressing data-related problems in natural hazards analysis. Data processed by deep generative models can be utilized to describe the evolution or occurrence of natural hazards and contribute to subsequent natural hazard modeling. Here we present a comprehensive review concerning generative deep learning for data generation in natural hazard analysis. (1) We summarized the limitations associated with data availability in natural hazards analysis and identified the fundamental motivations for employing generative deep learning as a critical response to these challenges. (2) We discuss several deep generative models that have been applied to overcome the problems caused by limited data availability in natural hazards analysis. (3) We analyze advances in utilizing generative deep learning for data generation in natural hazard analysis. (4) We discuss challenges associated with leveraging generative deep learning in natural hazard analysis. (5) We explore further opportunities for leveraging generative deep learning in natural hazard analysis. This comprehensive review provides a detailed roadmap for scholars interested in applying generative models for data generation in natural hazard analysis.","doi":"10.1007\/s10462-024-10764-9","cleaned_title":"generative deep learning for data generation in natural hazard analysis: motivations, advances, challenges, and opportunities","cleaned_abstract":"data mining and analysis are critical for preventing or mitigating natural hazards. however, data availability in natural hazard analysis is experiencing unprecedented challenges due to economic, technical, and environmental constraints. recently, generative deep learning has become an increasingly attractive solution to these challenges, which can augment, impute, or synthesize data based on these learned complex, high-dimensional probability distributions of data. over the last several years, much research has demonstrated the remarkable capabilities of generative deep learning for addressing data-related problems in natural hazards analysis. data processed by deep generative models can be utilized to describe the evolution or occurrence of natural hazards and contribute to subsequent natural hazard modeling. here we present a comprehensive review concerning generative deep learning for data generation in natural hazard analysis. (1) we summarized the limitations associated with data availability in natural hazards analysis and identified the fundamental motivations for employing generative deep learning as a critical response to these challenges. (2) we discuss several deep generative models that have been applied to overcome the problems caused by limited data availability in natural hazards analysis. (3) we analyze advances in utilizing generative deep learning for data generation in natural hazard analysis. (4) we discuss challenges associated with leveraging generative deep learning in natural hazard analysis. (5) we explore further opportunities for leveraging generative deep learning in natural hazard analysis. this comprehensive review provides a detailed roadmap for scholars interested in applying generative models for data generation in natural hazard analysis.","key_phrases":["data mining","analysis","natural hazards","data availability","natural hazard analysis","unprecedented challenges","economic, technical, and environmental constraints","generative deep learning","an increasingly attractive solution","these challenges","which","data","these learned complex, high-dimensional probability distributions","data","the last several years","much research","the remarkable capabilities","generative deep learning","data-related problems","natural hazards analysis","data","deep generative models","the evolution","occurrence","natural hazards","subsequent natural hazard modeling","we","a comprehensive review","generative deep learning","data generation","natural hazard analysis","we","the limitations","data availability","natural hazards analysis","the fundamental motivations","generative deep learning","a critical response","these challenges","we","several deep generative models","that","the problems","limited data availability","natural hazards analysis","we","advances","generative deep learning","data generation","natural hazard analysis","we","challenges","generative deep learning","natural hazard analysis","we","further opportunities","generative deep learning","natural hazard analysis","this comprehensive review","a detailed roadmap","scholars","generative models","data generation","natural hazard analysis","the last several years","1","2","3","4","5"]},{"title":"A Novel Discrete Deep Learning\u2013Based Cancer Classification Methodology","authors":["Marzieh Soltani","Mehdi Khashei","Negar Bakhtiarvand"],"abstract":"Classification is one of the most well-known data mining branches used in diverse domains and fields. In the literature, many different classification techniques, such as statistical\/intelligent, linear\/nonlinear, fuzzy\/crisp, shallow\/deep, and single\/hybrid, have been developed to cover data and systems with different characteristics. Intelligent classification approaches, especially deep learning classifiers, due to their unique features to provide accurate and efficient results, have recently attracted a lot of attention. However, in the learning process of the intelligent classifiers, a continuous distance-based cost function is used to estimate the connection weights, though the goal function in classification problems is discrete and using a continuous cost function in their learning process is unreasonable and inefficient. In this paper, a novel discrete learning\u2013based methodology is proposed to estimate the connection weights of intelligent classifiers more accurately. In the proposed learning process, they are discretely adjusted and at once jumped to the target. This is in contrast to conventional continuous learning algorithms in which the connection weights are continuously adjusted and step by step near the target. In the present research, the proposed methodology is exemplarily applied to the deep neural network (DNN), which is one of the most recognized deep classification approaches, with a solid mathematical foundation and strong practical results in complex problems. Although the proposed methodology is just implemented on the DNN, it is a general methodology that can be similarly applied to other shallow and deep intelligent classification models. It can be generally demonstrated that the performance of the proposed discrete learning\u2013based DNN (DIDNN) model, due to its consistency property, will not be worse than the conventional ones. The proposed DIDNN model is exemplarily evaluated on some well-known cancer classification benchmarks to illustrate the efficiency of the proposed model. The empirical results indicate that the proposed model outperforms the conventional versions of the selected deep approach in all data sets. Based on the performance analysis, the DIDNN model can improve the performance of the classic version by approximately 3.39%. Therefore, using this technique is an appropriate and effective alternative to conventional DNN-based models for classification purposes.","doi":"10.1007\/s12559-023-10170-3","cleaned_title":"a novel discrete deep learning\u2013based cancer classification methodology","cleaned_abstract":"classification is one of the most well-known data mining branches used in diverse domains and fields. in the literature, many different classification techniques, such as statistical\/intelligent, linear\/nonlinear, fuzzy\/crisp, shallow\/deep, and single\/hybrid, have been developed to cover data and systems with different characteristics. intelligent classification approaches, especially deep learning classifiers, due to their unique features to provide accurate and efficient results, have recently attracted a lot of attention. however, in the learning process of the intelligent classifiers, a continuous distance-based cost function is used to estimate the connection weights, though the goal function in classification problems is discrete and using a continuous cost function in their learning process is unreasonable and inefficient. in this paper, a novel discrete learning\u2013based methodology is proposed to estimate the connection weights of intelligent classifiers more accurately. in the proposed learning process, they are discretely adjusted and at once jumped to the target. this is in contrast to conventional continuous learning algorithms in which the connection weights are continuously adjusted and step by step near the target. in the present research, the proposed methodology is exemplarily applied to the deep neural network (dnn), which is one of the most recognized deep classification approaches, with a solid mathematical foundation and strong practical results in complex problems. although the proposed methodology is just implemented on the dnn, it is a general methodology that can be similarly applied to other shallow and deep intelligent classification models. it can be generally demonstrated that the performance of the proposed discrete learning\u2013based dnn (didnn) model, due to its consistency property, will not be worse than the conventional ones. the proposed didnn model is exemplarily evaluated on some well-known cancer classification benchmarks to illustrate the efficiency of the proposed model. the empirical results indicate that the proposed model outperforms the conventional versions of the selected deep approach in all data sets. based on the performance analysis, the didnn model can improve the performance of the classic version by approximately 3.39%. therefore, using this technique is an appropriate and effective alternative to conventional dnn-based models for classification purposes.","key_phrases":["classification","the most well-known data mining branches","diverse domains","fields","the literature","many different classification techniques","single\/hybrid","data","systems","different characteristics","intelligent classification approaches","especially deep learning classifiers","their unique features","accurate and efficient results","a lot","attention","the learning process","the intelligent classifiers","a continuous distance-based cost function","the connection weights","the goal function","classification problems","a continuous cost function","their learning process","this paper","a novel discrete learning","based methodology","the connection weights","intelligent classifiers","the proposed learning process","they","the target","this","contrast","conventional continuous learning algorithms","which","the connection weights","step","the target","the present research","the proposed methodology","the deep neural network","dnn","which","the most recognized deep classification approaches","a solid mathematical foundation","strong practical results","complex problems","the proposed methodology","the dnn","it","a general methodology","that","other shallow and deep intelligent classification models","it","the performance","the proposed discrete learning","based dnn (didnn) model","its consistency property","the conventional ones","the proposed didnn model","some well-known cancer classification benchmarks","the efficiency","the proposed model","the empirical results","the proposed model","the conventional versions","the selected deep approach","all data sets","the performance analysis","the didnn model","the performance","the classic version","approximately 3.39%","this technique","an appropriate and effective alternative","conventional dnn-based models","classification purposes","linear","approximately 3.39%"]},{"title":"Deep reinforcement learning for inverting earthquake focal mechanism and its potential application to marine earthquakes","authors":["Wenhuan Kuang","Zhihui Zou","Junhui Xing","Wei Wei"],"abstract":"Earthquake data are one of the key means by which to explore our planet. At a large scale, the layered structure of the Earth is revealed by the seismic waves of natural earthquakes that go deep into its inner core. At a local scale, seismology for exploration has successfully been employed to discover massive fossil energies. As the volume of recorded seismic data becomes greater, intelligent methods for processing such a volume of data are eagerly anticipated. In particular, earthquake focal mechanisms are important for assessing the severity of tsunamis, characterizing seismogenic faults, and investigating the stress perturbations that follow a major earthquake. Here, we report a novel deep reinforcement learning method for inverting the earthquake focal mechanism. Unlike more typical deep learning applications, which require a large training dataset, a deep reinforcement learning system learns by itself. We demonstrate the validity and efficacy of the proposed deep reinforcement learning method by applying it to the Mw 7.1 mainshock of the Ridgecrest earthquakes in southern California. In the foreseeable future, deep learning technologies may greatly contribute to our understanding of the oceanographic process. The proposed method may help us understand the mechanism of marine earthquakes.","doi":"10.1007\/s44295-024-00031-6","cleaned_title":"deep reinforcement learning for inverting earthquake focal mechanism and its potential application to marine earthquakes","cleaned_abstract":"earthquake data are one of the key means by which to explore our planet. at a large scale, the layered structure of the earth is revealed by the seismic waves of natural earthquakes that go deep into its inner core. at a local scale, seismology for exploration has successfully been employed to discover massive fossil energies. as the volume of recorded seismic data becomes greater, intelligent methods for processing such a volume of data are eagerly anticipated. in particular, earthquake focal mechanisms are important for assessing the severity of tsunamis, characterizing seismogenic faults, and investigating the stress perturbations that follow a major earthquake. here, we report a novel deep reinforcement learning method for inverting the earthquake focal mechanism. unlike more typical deep learning applications, which require a large training dataset, a deep reinforcement learning system learns by itself. we demonstrate the validity and efficacy of the proposed deep reinforcement learning method by applying it to the mw 7.1 mainshock of the ridgecrest earthquakes in southern california. in the foreseeable future, deep learning technologies may greatly contribute to our understanding of the oceanographic process. the proposed method may help us understand the mechanism of marine earthquakes.","key_phrases":["earthquake data","the key means","which","our planet","a large scale","the layered structure","the earth","the seismic waves","natural earthquakes","that","its inner core","a local scale","seismology","exploration","massive fossil energies","the volume","recorded seismic data","greater, intelligent methods","such a volume","data","earthquake focal mechanisms","the severity","tsunamis","seismogenic faults","the stress perturbations","that","a major earthquake","we","a novel deep reinforcement learning method","the earthquake focal mechanism","more typical deep learning applications","which","a large training dataset","a deep reinforcement learning system","itself","we","the validity","efficacy","the proposed deep reinforcement learning method","it","the mw 7.1 mainshock","the ridgecrest earthquakes","southern california","the foreseeable future","deep learning technologies","our understanding","the oceanographic process","the proposed method","us","the mechanism","marine earthquakes","one","7.1","southern california"]},{"title":"Advancements in hybrid approaches for brain tumor segmentation in MRI: a comprehensive review of machine learning and deep learning techniques","authors":["Ravikumar Sajjanar","Umesh D. Dixit","Vittalkumar K Vagga"],"abstract":"Magnetic resonance imaging (MRI) brain tumour segmentation is essential for the diagnosis, planning, and follow-up of patients with brain tumours. In an effort to increase efficiency and accuracy, a number of machine learning and deep learning algorithms have been developed over time to automate the segmentation process. Hybrid strategies, which include the advantages of both machine learning and deep learning, have become more and more popular as viable options. This in-depth analysis covers the developments in hybrid techniques for MRI segmentation of brain tumours. The essential ideas of machine learning and deep learning approaches are then covered, with an emphasis on their individual advantages and disadvantages. After that, the review explores the numerous hybrid strategies put out in the literature. In hybrid approaches, various phases of the segmentation pipeline are combined with machine learning and deep learning techniques. Pre-processing, feature extraction, and post-processing are examples of these phases. The paper examines at various combinations of methods utilised at these phases, such as segmentation using deep learning models and feature extraction utilising conventional machine learning algorithms. The implementation of ensemble approaches, which integrate forecasts from various models to improve segmentation accuracy, is also explored. The research study also examines the properties of freely accessible brain tumour datasets, which are essential for developing and testing hybrid models. To address the difficulties of generalisation and robustness in brain tumour segmentation, it emphasises the necessity of vast, varied, and annotated datasets. Additionally, by contrasting them with conventional machine learning and deep learning techniques, the review analyses the effectiveness of hybrid approaches reported in the literature. This comprehensive research provides information on recent advancements in hybrid techniques for MRI segmenting brain tumours. It emphasises the potential for merging deep learning and machine learning methods to enhance the precision and effectiveness of brain tumour segmentation, ultimately assisting in improving patient diagnosis and treatment planning.","doi":"10.1007\/s11042-023-16654-6","cleaned_title":"advancements in hybrid approaches for brain tumor segmentation in mri: a comprehensive review of machine learning and deep learning techniques","cleaned_abstract":"magnetic resonance imaging (mri) brain tumour segmentation is essential for the diagnosis, planning, and follow-up of patients with brain tumours. in an effort to increase efficiency and accuracy, a number of machine learning and deep learning algorithms have been developed over time to automate the segmentation process. hybrid strategies, which include the advantages of both machine learning and deep learning, have become more and more popular as viable options. this in-depth analysis covers the developments in hybrid techniques for mri segmentation of brain tumours. the essential ideas of machine learning and deep learning approaches are then covered, with an emphasis on their individual advantages and disadvantages. after that, the review explores the numerous hybrid strategies put out in the literature. in hybrid approaches, various phases of the segmentation pipeline are combined with machine learning and deep learning techniques. pre-processing, feature extraction, and post-processing are examples of these phases. the paper examines at various combinations of methods utilised at these phases, such as segmentation using deep learning models and feature extraction utilising conventional machine learning algorithms. the implementation of ensemble approaches, which integrate forecasts from various models to improve segmentation accuracy, is also explored. the research study also examines the properties of freely accessible brain tumour datasets, which are essential for developing and testing hybrid models. to address the difficulties of generalisation and robustness in brain tumour segmentation, it emphasises the necessity of vast, varied, and annotated datasets. additionally, by contrasting them with conventional machine learning and deep learning techniques, the review analyses the effectiveness of hybrid approaches reported in the literature. this comprehensive research provides information on recent advancements in hybrid techniques for mri segmenting brain tumours. it emphasises the potential for merging deep learning and machine learning methods to enhance the precision and effectiveness of brain tumour segmentation, ultimately assisting in improving patient diagnosis and treatment planning.","key_phrases":["mri","the diagnosis","planning","follow-up","patients","brain tumours","an effort","efficiency","accuracy","a number","machine learning","deep learning algorithms","time","the segmentation process","hybrid strategies","which","the advantages","both machine learning","deep learning","viable options","-depth","the developments","hybrid techniques","mri segmentation","brain tumours","the essential ideas","machine learning","deep learning approaches","an emphasis","their individual advantages","disadvantages","that","the review","the numerous hybrid strategies","the literature","hybrid approaches","various phases","the segmentation pipeline","machine learning","deep learning techniques","pre-processing, feature extraction","post-processing","examples","these phases","various combinations","methods","these phases","segmentation","deep learning models","extraction","conventional machine learning algorithms","the implementation","ensemble approaches","which","forecasts","various models","segmentation accuracy","the research study","the properties","freely accessible brain tumour datasets","which","hybrid models","the difficulties","generalisation","robustness","brain tumour segmentation","it","the necessity","datasets","them","conventional machine learning","deep learning techniques","the review","the effectiveness","hybrid approaches","the literature","this comprehensive research","information","recent advancements","hybrid techniques","mri segmenting brain tumours","it","the potential","deep learning and machine learning methods","the precision","effectiveness","brain tumour segmentation","patient diagnosis and treatment planning"]},{"title":"Analysis of English Classroom Teaching Behavior Mode in Environmental Protection Field Based on Deep Learning","authors":["Yuanyuan Li"],"abstract":"Learning is to use algorithms to enable machines to learn rules from a large amount of historical data, so as to intelligently identify new samples or predict the future. Deep learning can promote students\u2019 understanding of knowledge, conduct in-depth processing of new knowledge, integrate it with the original knowledge, and apply it to new situations, solve intelligent audio\u2013visual listening from the perspective of deep learning, and focus on cultivating students\u2019 in-depth learning ability and individual differences in innovative thinking. As the main position of ecological education, schools should effectively strengthen the publicity and education of ecological ideas and low-carbon concepts, and integrate them into education and teaching to effectively improve students\u2019 awareness of environmental protection. This study aims to explore the effectiveness of flipped classroom teaching model based on deep learning. Therefore, from the perspective of deep learning, this paper combs the theory of deep learning, constructs a new model of smart classroom, and provides ideas and directions for model reform. In this study, the flipped classroom teaching model based on deep learning was applied to English teaching, and an 8-week teaching experiment was conducted. In addition, this paper believes that it is of great practical significance to carry out environmental protection education with the help of English teaching.","doi":"10.1007\/s44196-024-00457-0","cleaned_title":"analysis of english classroom teaching behavior mode in environmental protection field based on deep learning","cleaned_abstract":"learning is to use algorithms to enable machines to learn rules from a large amount of historical data, so as to intelligently identify new samples or predict the future. deep learning can promote students\u2019 understanding of knowledge, conduct in-depth processing of new knowledge, integrate it with the original knowledge, and apply it to new situations, solve intelligent audio\u2013visual listening from the perspective of deep learning, and focus on cultivating students\u2019 in-depth learning ability and individual differences in innovative thinking. as the main position of ecological education, schools should effectively strengthen the publicity and education of ecological ideas and low-carbon concepts, and integrate them into education and teaching to effectively improve students\u2019 awareness of environmental protection. this study aims to explore the effectiveness of flipped classroom teaching model based on deep learning. therefore, from the perspective of deep learning, this paper combs the theory of deep learning, constructs a new model of smart classroom, and provides ideas and directions for model reform. in this study, the flipped classroom teaching model based on deep learning was applied to english teaching, and an 8-week teaching experiment was conducted. in addition, this paper believes that it is of great practical significance to carry out environmental protection education with the help of english teaching.","key_phrases":["algorithms","machines","rules","a large amount","historical data","new samples","the future","deep learning","students\u2019 understanding","knowledge","depth","new knowledge","it","the original knowledge","it","new situations","intelligent audio\u2013visual listening","the perspective","deep learning","students","depth","individual differences","innovative thinking","the main position","ecological education","schools","the publicity","education","ecological ideas","low-carbon concepts","them","education","teaching","students\u2019 awareness","environmental protection","this study","the effectiveness","flipped classroom teaching model","deep learning","the perspective","deep learning","this paper","the theory","deep learning","a new model","smart classroom","ideas","directions","model reform","this study","the flipped classroom teaching model","deep learning","english teaching","an 8-week teaching experiment","addition","this paper","it","great practical significance","environmental protection education","the help","english teaching","english","8-week","english"]},{"title":"Deep learning-based point cloud upsampling: a review of recent trends","authors":["Soonjo Kwon","Ji-Hyeon Hur","Hyungki Kim"],"abstract":"Point clouds are acquired primarily using 3D scanners and are used for product inspection and reverse engineering. The quality of the point cloud varies depending on the scanning environment and scanner specifications. The quality of the point cloud has a significant impact on the accuracy of automatic or manual modeling. In response, various point cloud post-processing technologies are being developed. Point cloud upsampling is a technique to improve the resolution of point clouds, and the purpose of upsampling is to generate additional points to express the target object more accurately and in higher detail. This technology is important in areas where high-resolution 3D representation is required, and approaches based on deep learning have been recently gaining attention. Deep learning-based point cloud upsampling research can be classified as surface consolidation or edge consolidation research depending on the target regions to be consolidated, and as supervised or self-supervised learning depending on the type of learning approaches. This study examines the latest research trends in deep learning-based point cloud sampling, analyzes the issues and limitations of each research category, and proposes future research directions.Graphical abstract","doi":"10.1007\/s42791-023-00058-6","cleaned_title":"deep learning-based point cloud upsampling: a review of recent trends","cleaned_abstract":"point clouds are acquired primarily using 3d scanners and are used for product inspection and reverse engineering. the quality of the point cloud varies depending on the scanning environment and scanner specifications. the quality of the point cloud has a significant impact on the accuracy of automatic or manual modeling. in response, various point cloud post-processing technologies are being developed. point cloud upsampling is a technique to improve the resolution of point clouds, and the purpose of upsampling is to generate additional points to express the target object more accurately and in higher detail. this technology is important in areas where high-resolution 3d representation is required, and approaches based on deep learning have been recently gaining attention. deep learning-based point cloud upsampling research can be classified as surface consolidation or edge consolidation research depending on the target regions to be consolidated, and as supervised or self-supervised learning depending on the type of learning approaches. this study examines the latest research trends in deep learning-based point cloud sampling, analyzes the issues and limitations of each research category, and proposes future research directions.graphical abstract","key_phrases":["point clouds","3d scanners","product inspection","engineering","the quality","the point cloud","the scanning environment","scanner specifications","the quality","the point cloud","a significant impact","the accuracy","automatic or manual modeling","response","various point cloud post-processing technologies","point cloud upsampling","a technique","the resolution","point clouds","the purpose","upsampling","additional points","the target object","higher detail","this technology","areas","high-resolution 3d representation","approaches","deep learning","attention","deep learning-based point cloud upsampling research","surface consolidation","edge consolidation research","the target regions","supervised or self-supervised learning","the type","approaches","this study","the latest research trends","deep learning-based point cloud sampling","the issues","limitations","each research category","future research","directions.graphical abstract","3d","3d"]},{"title":"Diagnosis of EV Gearbox Bearing Fault Using Deep Learning-Based Signal Processing","authors":["Kicheol Jeong","Chulwoo Moon"],"abstract":"The gearbox of an electric vehicle operates under the high load torque and axial load of electric vehicles. In particular, the bearings that support the shaft of the gearbox are subjected to several tons of axial load, and as the mileage increases, fault occurs on bearing rolling elements frequently. Such bearing fault has a serious impact on driving comfort and vehicle safety, however, bearing faults are diagnosed by human experts nowadays, and algorithm-based electric vehicle bearing fault diagnosis has not been implemented. Therefore, in this paper, a deep learning-based bearing vibration signal processing method to diagnose bearing fault in electric vehicle gearboxes is proposed. The proposed method consists of a deep neural network learning stage and an application stage of the pre-trained neural network. In the deep neural network learning stage, supervised learning is carried out based on two acceleration sensors. In the neural network application stage, signal processing of a single accelerometer signal is performed through a pre-trained neural network. In conclusion, the pre-trained neural network makes bearing fault signals stand out and can utilize these signals to extract frequency characteristics of bearing fault.","doi":"10.1007\/s12239-024-00094-8","cleaned_title":"diagnosis of ev gearbox bearing fault using deep learning-based signal processing","cleaned_abstract":"the gearbox of an electric vehicle operates under the high load torque and axial load of electric vehicles. in particular, the bearings that support the shaft of the gearbox are subjected to several tons of axial load, and as the mileage increases, fault occurs on bearing rolling elements frequently. such bearing fault has a serious impact on driving comfort and vehicle safety, however, bearing faults are diagnosed by human experts nowadays, and algorithm-based electric vehicle bearing fault diagnosis has not been implemented. therefore, in this paper, a deep learning-based bearing vibration signal processing method to diagnose bearing fault in electric vehicle gearboxes is proposed. the proposed method consists of a deep neural network learning stage and an application stage of the pre-trained neural network. in the deep neural network learning stage, supervised learning is carried out based on two acceleration sensors. in the neural network application stage, signal processing of a single accelerometer signal is performed through a pre-trained neural network. in conclusion, the pre-trained neural network makes bearing fault signals stand out and can utilize these signals to extract frequency characteristics of bearing fault.","key_phrases":["the gearbox","an electric vehicle","the high load torque","axial load","electric vehicles","the bearings","that","the shaft","the gearbox","several tons","axial load","the mileage increases","fault","rolling elements","fault","a serious impact","comfort","vehicle safety","faults","human experts","algorithm-based electric vehicle","fault diagnosis","this paper","a deep learning-based bearing vibration signal processing method","fault","electric vehicle gearboxes","the proposed method","a deep neural network learning stage","an application stage","the pre-trained neural network","the deep neural network learning stage","supervised learning","two acceleration sensors","the neural network application stage","signal processing","a single accelerometer signal","a pre-trained neural network","conclusion","the pre-trained neural network","fault signals","these signals","frequency characteristics","fault","several tons","two"]},{"title":"Sentiment analysis using a deep ensemble learning model","authors":["Muhammet Sinan Ba\u015farslan","Fatih Kayaalp"],"abstract":"The coronavirus pandemic has kept people away from social life and this has led to an increase in the use of social media over the past two years. Thanks to social media, people can now instantly share their thoughts on various topics such as their favourite movies, restaurants, hotels, etc. This has created a huge amount of data and many researchers from different sciences have focused on analysing this data. Natural Language Processing (NLP) is one of these areas of computer science that uses artificial technologies. Sentiment analysis is also one of the tasks of NLP, which is based on extracting emotions from huge post data. In this study, sentiment analysis was performed on two datasets of tweets about coronavirus and TripAdvisor hotel reviews. A frequency-based word representation method (Term Frequency-Inverse Document Frequency (TF-IDF)) and a prediction-based Word2Vec word embedding method were used to vectorise the datasets. Sentiment analysis models were then built using single machine learning methods (Decision Trees-DT, K-Nearest Neighbour-KNN, Naive Bayes-NB and Support Vector Machine-SVM), single deep learning methods (Long Short Term Memory-LSTM, Recurrent Neural Network-RNN) and heterogeneous ensemble learning methods (Stacking and Majority Voting) based on these single machine learning and deep learning methods. Accuracy was used as a performance measure. The heterogeneous model with stacking (LSTM-RNN) has outperformed the other models with accuracy values of 0.864 on the coronavirus dataset and 0.898 on the Trip Advisor dataset and they have been evaluated as promising results when compared to the literature. It has been observed that the use of single methods as an ensemble gives better results, which is consistent with the literature, which is a step forward in the detection of sentiments through posts. Investigating the performance of heterogeneous ensemble learning models based on different algorithms in sentiment analysis tasks is planned as future work.","doi":"10.1007\/s11042-023-17278-6","cleaned_title":"sentiment analysis using a deep ensemble learning model","cleaned_abstract":"the coronavirus pandemic has kept people away from social life and this has led to an increase in the use of social media over the past two years. thanks to social media, people can now instantly share their thoughts on various topics such as their favourite movies, restaurants, hotels, etc. this has created a huge amount of data and many researchers from different sciences have focused on analysing this data. natural language processing (nlp) is one of these areas of computer science that uses artificial technologies. sentiment analysis is also one of the tasks of nlp, which is based on extracting emotions from huge post data. in this study, sentiment analysis was performed on two datasets of tweets about coronavirus and tripadvisor hotel reviews. a frequency-based word representation method (term frequency-inverse document frequency (tf-idf)) and a prediction-based word2vec word embedding method were used to vectorise the datasets. sentiment analysis models were then built using single machine learning methods (decision trees-dt, k-nearest neighbour-knn, naive bayes-nb and support vector machine-svm), single deep learning methods (long short term memory-lstm, recurrent neural network-rnn) and heterogeneous ensemble learning methods (stacking and majority voting) based on these single machine learning and deep learning methods. accuracy was used as a performance measure. the heterogeneous model with stacking (lstm-rnn) has outperformed the other models with accuracy values of 0.864 on the coronavirus dataset and 0.898 on the trip advisor dataset and they have been evaluated as promising results when compared to the literature. it has been observed that the use of single methods as an ensemble gives better results, which is consistent with the literature, which is a step forward in the detection of sentiments through posts. investigating the performance of heterogeneous ensemble learning models based on different algorithms in sentiment analysis tasks is planned as future work.","key_phrases":["people","social life","this","an increase","the use","social media","the past two years","social media","people","their thoughts","various topics","their favourite movies","restaurants","hotels","this","a huge amount","data","many researchers","different sciences","this data","natural language processing","nlp","these areas","computer science","that","artificial technologies","sentiment analysis","the tasks","nlp","which","emotions","huge post data","this study","sentiment analysis","two datasets","tweets","coronavirus and tripadvisor hotel reviews","a frequency-based word representation method","term frequency-inverse document frequency","tf-idf","a prediction-based word2vec word","embedding method","the datasets","sentiment analysis models","single machine learning methods","decision trees","dt","k-nearest neighbour-knn, naive bayes","vector machine-svm","single deep learning methods","long short term memory-lstm","recurrent neural network-rnn","heterogeneous ensemble learning methods","stacking and majority voting","these single machine learning","deep learning methods","accuracy","a performance measure","the heterogeneous model","lstm-rnn","the other models","accuracy values","the coronavirus dataset","the trip advisor dataset","they","results","the literature","it","the use","single methods","an ensemble","better results","which","the literature","which","a step","the detection","sentiments","posts","the performance","heterogeneous ensemble learning models","different algorithms","sentiment analysis tasks","future work","the past two years","two","accuracy","0.864","0.898"]},{"title":"Deep Q-learning with hybrid quantum neural network on solving maze problems","authors":["Hao-Yuan Chen","Yen-Jui Chang","Shih-Wei Liao","Ching-Ray Chang"],"abstract":"Quantum computing holds great potential for advancing the limitations of machine learning algorithms to handle higher dimensions of data and reduce overall training parameters in deep learning (DL) models. This study uses a trainable variational quantum circuit (VQC) on a gate-based quantum computing model to investigate the potential for quantum benefit in a model-free reinforcement learning problem. Through a comprehensive investigation and evaluation of the current model and capabilities of quantum computers, we designed and trained a novel hybrid quantum neural network based on the latest Qiskit and PyTorch framework. We compared its performance with a full-classical CNN with and without an incorporated VQC. Our research provides insights into the potential of deep quantum learning to solve a maze problem and, potentially, other reinforcement learning problems. We conclude that reinforcement learning problems can be practical with reasonable training epochs. Moreover, a comparative study of full-classical and hybrid quantum neural networks is discussed to understand these two approaches\u2019 performance, advantages, and disadvantages to deep Q-learning problems, especially on larger-scale maze problems larger than 4\\(\\times \\)4.","doi":"10.1007\/s42484-023-00137-w","cleaned_title":"deep q-learning with hybrid quantum neural network on solving maze problems","cleaned_abstract":"quantum computing holds great potential for advancing the limitations of machine learning algorithms to handle higher dimensions of data and reduce overall training parameters in deep learning (dl) models. this study uses a trainable variational quantum circuit (vqc) on a gate-based quantum computing model to investigate the potential for quantum benefit in a model-free reinforcement learning problem. through a comprehensive investigation and evaluation of the current model and capabilities of quantum computers, we designed and trained a novel hybrid quantum neural network based on the latest qiskit and pytorch framework. we compared its performance with a full-classical cnn with and without an incorporated vqc. our research provides insights into the potential of deep quantum learning to solve a maze problem and, potentially, other reinforcement learning problems. we conclude that reinforcement learning problems can be practical with reasonable training epochs. moreover, a comparative study of full-classical and hybrid quantum neural networks is discussed to understand these two approaches\u2019 performance, advantages, and disadvantages to deep q-learning problems, especially on larger-scale maze problems larger than 4\\(\\times \\)4.","key_phrases":["quantum computing","great potential","the limitations","machine learning algorithms","higher dimensions","data","overall training parameters","deep learning","(dl) models","this study","a trainable variational quantum circuit","vqc","a gate-based quantum computing model","the potential","quantum benefit","a model-free reinforcement learning problem","a comprehensive investigation","evaluation","the current model","capabilities","quantum computers","we","a novel hybrid quantum neural network","the latest qiskit and pytorch framework","we","its performance","a full-classical cnn","an incorporated vqc","our research","insights","the potential","deep quantum","a maze problem","potentially, other reinforcement learning problems","we","reinforcement learning problems","reasonable training epochs","a comparative study","full-classical and hybrid quantum neural networks","these two approaches\u2019 performance","advantages","disadvantages","deep q-learning problems","larger-scale maze problems","4\\(\\times","quantum","quantum","quantum","quantum","quantum","two"]},{"title":"Deep learning algorithms for hedging with frictions","authors":["Xiaofei Shi","Daran Xu","Zhanhao Zhang"],"abstract":"This work studies the deep learning-based numerical algorithms for optimal hedging problems in markets with general convex transaction costs. Our main focus is on how these algorithms scale with the length of the trading time horizon. Based on the comparison results of the FBSDE solver by Han, Jentzen, and E (2018) and the Deep Hedging algorithm by Buehler, Gonon, Teichmann, and Wood (2019), we propose a Stable-Transfer Hedging (ST-Hedging) algorithm, to aggregate the convenience of the leading-order approximation formulas and the accuracy of the deep learning-based algorithms. Our ST-Hedging algorithm achieves the same state-of-the-art performance in short and moderately long time horizon as FBSDE solver and Deep Hedging, and generalize well to long time horizon when previous algorithms become suboptimal. With the transfer learning technique, ST-Hedging drastically reduce the training time, and shows great scalability to high-dimensional settings. This opens up new possibilities in model-based deep learning algorithms in economics, finance, and operational research, which takes advantage of the domain expert knowledge and the accuracy of the learning-based methods.","doi":"10.1007\/s42521-023-00075-z","cleaned_title":"deep learning algorithms for hedging with frictions","cleaned_abstract":"this work studies the deep learning-based numerical algorithms for optimal hedging problems in markets with general convex transaction costs. our main focus is on how these algorithms scale with the length of the trading time horizon. based on the comparison results of the fbsde solver by han, jentzen, and e (2018) and the deep hedging algorithm by buehler, gonon, teichmann, and wood (2019), we propose a stable-transfer hedging (st-hedging) algorithm, to aggregate the convenience of the leading-order approximation formulas and the accuracy of the deep learning-based algorithms. our st-hedging algorithm achieves the same state-of-the-art performance in short and moderately long time horizon as fbsde solver and deep hedging, and generalize well to long time horizon when previous algorithms become suboptimal. with the transfer learning technique, st-hedging drastically reduce the training time, and shows great scalability to high-dimensional settings. this opens up new possibilities in model-based deep learning algorithms in economics, finance, and operational research, which takes advantage of the domain expert knowledge and the accuracy of the learning-based methods.","key_phrases":["this work","the deep learning-based numerical algorithms","optimal hedging problems","markets","general convex transaction costs","our main focus","these algorithms","the length","the trading time horizon","the comparison results","the fbsde","han","jentzen","e","buehler","gonon","teichmann","wood","we","a stable-transfer hedging","st-hedging","the convenience","the leading-order approximation formulas","the accuracy","the deep learning-based algorithms","our st-hedging algorithm","the-art","short and moderately long time horizon","fbsde","long time horizon","previous algorithms","the transfer learning technique","st-hedging","the training time","great scalability","high-dimensional settings","this","new possibilities","model-based deep learning algorithms","economics","finance","operational research","which","advantage","the domain expert knowledge","the accuracy","the learning-based methods","han","2018","teichmann","2019"]},{"title":"Comparative study and analysis on skin cancer detection using machine learning\u00a0and\u00a0deep learning algorithms","authors":["V. Auxilia Osvin Nancy","P. Prabhavathy","Meenakshi S. Arya","B. Shamreen Ahamed"],"abstract":"Exposure to UV rays due to global warming can lead to sunburn and skin damage, ultimately resulting in skin cancer. Early prediction of this type of cancer is crucial. A detailed review in this paper explores various algorithms, including machine learning (ML) techniques as well as deep learning (DL) techniques. While deep learning strategies, particularly CNNs, are commonly employed for skin cancer identification and classification, there is also some usage of machine learning and hybrid approaches. These techniques have proven to be effective classifiers of skin lesions, offering promising results for early detection. The paper analyzes various researchers\u2019 reviews on skin cancer diagnosis to identify a suitable methodology for improving diagnostic accuracy. A publicly available dataset of dermoscopic images retrieved from the ISIC archive has been trained and evaluated. Performance analysis is done, considering metrics such as test and validation accuracy. The results indicate that the RF(random forest) algorithm outperforms other machine learning algorithms in both scenarios, with accuracies of 58.57% without augmentation and 87.32% with augmentation. MobileNetv2, ensemble of Dense Net and Inceptionv3 exhibit superior performance. During training without augmentation, MobileNetv2 achieves an accuracy of 88.81%, while the ensemble model achieves an accuracy of 88.80%. With augmentation techniques applied, the accuracies improved to 97.58% and 97.50%, respectively. Furthermore, experiment with a customized convolutional neural network (CNN) model was also conducted, varying the number of layers and applying various hyperparameter tuning methodologies. Suitable architectures, including a CNN with 7 layers and batch normalization, a CNN with 5 layers, and a CNN with 3 layers were identified. These models achieved accuracies of 77.92%, 97.72%, and 98.02% on the raw data and augmentation datasets, respectively. The experimental results suggest that these techniques hold promise for integration into clinical settings, and further research and validation are necessary. The results highlight the effectiveness of transfer learning models, in achieving high accuracy rates. The findings support the future adoption of these techniques in clinical practice, pending further research and validation.","doi":"10.1007\/s11042-023-16422-6","cleaned_title":"comparative study and analysis on skin cancer detection using machine learning and deep learning algorithms","cleaned_abstract":"exposure to uv rays due to global warming can lead to sunburn and skin damage, ultimately resulting in skin cancer. early prediction of this type of cancer is crucial. a detailed review in this paper explores various algorithms, including machine learning (ml) techniques as well as deep learning (dl) techniques. while deep learning strategies, particularly cnns, are commonly employed for skin cancer identification and classification, there is also some usage of machine learning and hybrid approaches. these techniques have proven to be effective classifiers of skin lesions, offering promising results for early detection. the paper analyzes various researchers\u2019 reviews on skin cancer diagnosis to identify a suitable methodology for improving diagnostic accuracy. a publicly available dataset of dermoscopic images retrieved from the isic archive has been trained and evaluated. performance analysis is done, considering metrics such as test and validation accuracy. the results indicate that the rf(random forest) algorithm outperforms other machine learning algorithms in both scenarios, with accuracies of 58.57% without augmentation and 87.32% with augmentation. mobilenetv2, ensemble of dense net and inceptionv3 exhibit superior performance. during training without augmentation, mobilenetv2 achieves an accuracy of 88.81%, while the ensemble model achieves an accuracy of 88.80%. with augmentation techniques applied, the accuracies improved to 97.58% and 97.50%, respectively. furthermore, experiment with a customized convolutional neural network (cnn) model was also conducted, varying the number of layers and applying various hyperparameter tuning methodologies. suitable architectures, including a cnn with 7 layers and batch normalization, a cnn with 5 layers, and a cnn with 3 layers were identified. these models achieved accuracies of 77.92%, 97.72%, and 98.02% on the raw data and augmentation datasets, respectively. the experimental results suggest that these techniques hold promise for integration into clinical settings, and further research and validation are necessary. the results highlight the effectiveness of transfer learning models, in achieving high accuracy rates. the findings support the future adoption of these techniques in clinical practice, pending further research and validation.","key_phrases":["exposure","uv rays","global warming","sunburn","skin damage","skin cancer","early prediction","this type","cancer","a detailed review","this paper","various algorithms","machine learning",") techniques","deep learning","(dl) techniques","deep learning strategies","particularly cnns","skin cancer identification","classification","some usage","machine learning","hybrid approaches","these techniques","effective classifiers","skin lesions","promising results","early detection","the paper","various researchers\u2019 reviews","skin cancer diagnosis","a suitable methodology","diagnostic accuracy","a publicly available dataset","dermoscopic images","the isic archive","performance analysis","metrics","test","validation accuracy","the results","algorithm","other machine learning algorithms","both scenarios","accuracies","58.57%","augmentation","87.32%","augmentation","mobilenetv2","dense net","inceptionv3","superior performance","training","augmentation","mobilenetv2","an accuracy","88.81%","the ensemble model","an accuracy","88.80%","augmentation techniques","the accuracies","97.58%","97.50%","experiment","a customized convolutional neural network (cnn) model","the number","layers","various hyperparameter","methodologies","suitable architectures","a cnn","7 layers","batch normalization","a cnn","5 layers","a cnn","3 layers","these models","accuracies","77.92%","97.72%","98.02%","the raw data","augmentation datasets","the experimental results","these techniques","promise","integration","clinical settings","further research","validation","the results","the effectiveness","transfer learning models","high accuracy rates","the findings","the future adoption","these techniques","clinical practice","further research","validation","58.57%","87.32%","mobilenetv2","inceptionv3","mobilenetv2","88.81%","88.80%","97.58%","97.50%","cnn","cnn","7","cnn","5","cnn","3","77.92%","97.72%","98.02%"]},{"title":"Enhancing trash classification in smart cities using federated deep learning","authors":["Haroon Ahmed Khan","Syed Saud Naqvi","Abeer A. K. Alharbi","Salihah Alotaibi","Mohammed Alkhathami"],"abstract":"Efficient Waste management plays a crucial role to ensure clean and green environment in the smart cities. This study investigates the critical role of efficient trash classification in achieving sustainable solid waste management within smart city environments. We conduct a comparative analysis of various trash classification methods utilizing deep learning models built on convolutional neural networks (CNNs). Leveraging the PyTorch open-source framework and the TrashBox dataset, we perform experiments involving ten unique deep neural network models. Our approach aims to maximize training accuracy. Through extensive experimentation, we observe the consistent superiority of the ResNext-101 model compared to others, achieving exceptional training, validation, and test accuracies. These findings illuminate the potential of CNN-based techniques in significantly advancing trash classification for optimized solid waste management within smart city initiatives. Lastly, this study presents a distributed framework based on federated learning that can be used to optimize the performance of a combination of CNN models for trash detection.","doi":"10.1038\/s41598-024-62003-4","cleaned_title":"enhancing trash classification in smart cities using federated deep learning","cleaned_abstract":"efficient waste management plays a crucial role to ensure clean and green environment in the smart cities. this study investigates the critical role of efficient trash classification in achieving sustainable solid waste management within smart city environments. we conduct a comparative analysis of various trash classification methods utilizing deep learning models built on convolutional neural networks (cnns). leveraging the pytorch open-source framework and the trashbox dataset, we perform experiments involving ten unique deep neural network models. our approach aims to maximize training accuracy. through extensive experimentation, we observe the consistent superiority of the resnext-101 model compared to others, achieving exceptional training, validation, and test accuracies. these findings illuminate the potential of cnn-based techniques in significantly advancing trash classification for optimized solid waste management within smart city initiatives. lastly, this study presents a distributed framework based on federated learning that can be used to optimize the performance of a combination of cnn models for trash detection.","key_phrases":["efficient waste management","a crucial role","clean and green environment","the smart cities","this study","the critical role","efficient trash classification","sustainable solid waste management","smart city environments","we","a comparative analysis","various trash classification methods","deep learning models","convolutional neural networks","cnns","the pytorch open-source framework","the trashbox dataset","we","experiments","ten unique deep neural network models","our approach","training accuracy","extensive experimentation","we","the consistent superiority","the resnext-101 model","others","exceptional training","validation","test accuracies","these findings","the potential","cnn-based techniques","significantly advancing trash classification","optimized solid waste management","smart city initiatives","this study","a distributed framework","federated learning","that","the performance","a combination","cnn models","trash detection","smart city","cnn","smart city","cnn"]},{"title":"Interdisciplinary approach to identify language markers for post-traumatic stress disorder using machine learning and deep learning","authors":["Robin Quillivic","Fr\u00e9d\u00e9rique Gayraud","Yann Aux\u00e9m\u00e9ry","Laurent Vanni","Denis Peschanski","Francis Eustache","Jacques Dayan","Salma Mesmoudi"],"abstract":"Post-traumatic stress disorder (PTSD) lacks clear biomarkers in clinical practice. Language as a potential diagnostic biomarker for PTSD is investigated in this study. We analyze an original cohort of 148 individuals exposed to the November 13, 2015, terrorist attacks in Paris. The interviews, conducted 5\u201311\u00a0months after the event, include individuals from similar socioeconomic backgrounds exposed to the same incident, responding to identical questions and using uniform PTSD measures. Using this dataset to collect nuanced insights that might be clinically relevant, we propose a three-step interdisciplinary methodology that integrates expertise from psychiatry, linguistics, and the Natural Language Processing (NLP) community to examine the relationship between language and PTSD. The first step assesses a clinical psychiatrist's ability to diagnose PTSD using interview transcription alone. The second step uses statistical analysis and machine learning models to create language features based on psycholinguistic hypotheses and evaluate their predictive strength. The third step is the application of a hypothesis-free deep learning approach to the classification of PTSD in our cohort. Results show that the clinical psychiatrist achieved a diagnosis of PTSD with an AUC of 0.72. This is comparable to a gold standard questionnaire (Area Under Curve (AUC)\u00a0\u2248\u00a00.80). The machine learning model achieved a diagnostic AUC of 0.69. The deep learning approach achieved an AUC of 0.64. An examination of model error informs our discussion. Importantly, the study controls for confounding factors, establishes associations between language and DSM-5 subsymptoms, and integrates automated methods with qualitative analysis. This study provides a direct and methodologically robust description of the relationship between PTSD and language. Our work lays the groundwork for advancing early and accurate diagnosis and using linguistic markers to assess the effectiveness of pharmacological treatments and psychotherapies.","doi":"10.1038\/s41598-024-61557-7","cleaned_title":"interdisciplinary approach to identify language markers for post-traumatic stress disorder using machine learning and deep learning","cleaned_abstract":"post-traumatic stress disorder (ptsd) lacks clear biomarkers in clinical practice. language as a potential diagnostic biomarker for ptsd is investigated in this study. we analyze an original cohort of 148 individuals exposed to the november 13, 2015, terrorist attacks in paris. the interviews, conducted 5\u201311 months after the event, include individuals from similar socioeconomic backgrounds exposed to the same incident, responding to identical questions and using uniform ptsd measures. using this dataset to collect nuanced insights that might be clinically relevant, we propose a three-step interdisciplinary methodology that integrates expertise from psychiatry, linguistics, and the natural language processing (nlp) community to examine the relationship between language and ptsd. the first step assesses a clinical psychiatrist's ability to diagnose ptsd using interview transcription alone. the second step uses statistical analysis and machine learning models to create language features based on psycholinguistic hypotheses and evaluate their predictive strength. the third step is the application of a hypothesis-free deep learning approach to the classification of ptsd in our cohort. results show that the clinical psychiatrist achieved a diagnosis of ptsd with an auc of 0.72. this is comparable to a gold standard questionnaire (area under curve (auc) \u2248 0.80). the machine learning model achieved a diagnostic auc of 0.69. the deep learning approach achieved an auc of 0.64. an examination of model error informs our discussion. importantly, the study controls for confounding factors, establishes associations between language and dsm-5 subsymptoms, and integrates automated methods with qualitative analysis. this study provides a direct and methodologically robust description of the relationship between ptsd and language. our work lays the groundwork for advancing early and accurate diagnosis and using linguistic markers to assess the effectiveness of pharmacological treatments and psychotherapies.","key_phrases":["post-traumatic stress disorder","ptsd","clear biomarkers","clinical practice","language","a potential diagnostic biomarker","ptsd","this study","we","an original cohort","148 individuals","the november","terrorist attacks","paris","the interviews","the event","individuals","similar socioeconomic backgrounds","the same incident","identical questions","uniform ptsd measures","this dataset","nuanced insights","that","we","a three-step interdisciplinary methodology","that","expertise","psychiatry","linguistics","the natural language processing (nlp) community","the relationship","language","ptsd","the first step","a clinical psychiatrist's ability","ptsd","interview transcription","the second step","statistical analysis","machine learning models","language features","psycholinguistic hypotheses","their predictive strength","the third step","the application","a hypothesis-free deep learning approach","the classification","ptsd","our cohort","results","the clinical psychiatrist","a diagnosis","ptsd","an auc","this","a gold standard questionnaire","area","curve","auc","\u2248","the machine learning model","a diagnostic auc","the deep learning approach","an auc","an examination","model error","our discussion","the study","factors","associations","language","dsm-5","subsymptoms","automated methods","qualitative analysis","this study","a direct and methodologically robust description","the relationship","ptsd","language","our work","the groundwork","early and accurate diagnosis","linguistic markers","the effectiveness","pharmacological treatments","psychotherapies","148","november 13, 2015","paris","5\u201311 months","three","first","second","third","0.72","0.69","0.64"]},{"title":"Study of Q-learning and deep Q-network learning control for a rotary inverted pendulum system","authors":["Zied Ben Hazem"],"abstract":"The rotary inverted pendulum system (RIPS) is an underactuated mechanical system with highly nonlinear dynamics and it is difficult to control a RIPS using the classic control models. In the last few years, reinforcement learning (RL) has become a popular nonlinear control method. RL has a powerful potential to control systems with high non-linearity and complex dynamics, such as RIPS. Nevertheless, RL control for RIPS has not been well studied and there is limited research on the development and evaluation of this control method. In this paper, RL control algorithms are developed for the swing-up and stabilization control of a single-link rotary inverted pendulum (SLRIP) and compared with classic control methods such as PID and LQR. A physical model of the SLRIP system is created using the MATLAB\/Simscape Toolbox, the model is used as a dynamic simulation in MATLAB\/Simulink to train the RL agents. An agent trainer system with Q-learning (QL) and deep Q-network learning (DQNL) is proposed for the data training. Furthermore, agent actions are actuating the horizontal arm of the system and states are the angles and velocities of the pendulum and the horizontal arm. The reward is computed according to the angles of the pendulum and horizontal arm. The reward is zero when the pendulum attends the upright position. The RL algorithms are used without a deep understanding of the classical controllers and are used to implement the agent. Finally, the outcome indicates the effectiveness of the QL and DQNL algorithms compared to the conventional PID and LQR controllers.","doi":"10.1007\/s42452-024-05690-y","cleaned_title":"study of q-learning and deep q-network learning control for a rotary inverted pendulum system","cleaned_abstract":"the rotary inverted pendulum system (rips) is an underactuated mechanical system with highly nonlinear dynamics and it is difficult to control a rips using the classic control models. in the last few years, reinforcement learning (rl) has become a popular nonlinear control method. rl has a powerful potential to control systems with high non-linearity and complex dynamics, such as rips. nevertheless, rl control for rips has not been well studied and there is limited research on the development and evaluation of this control method. in this paper, rl control algorithms are developed for the swing-up and stabilization control of a single-link rotary inverted pendulum (slrip) and compared with classic control methods such as pid and lqr. a physical model of the slrip system is created using the matlab\/simscape toolbox, the model is used as a dynamic simulation in matlab\/simulink to train the rl agents. an agent trainer system with q-learning (ql) and deep q-network learning (dqnl) is proposed for the data training. furthermore, agent actions are actuating the horizontal arm of the system and states are the angles and velocities of the pendulum and the horizontal arm. the reward is computed according to the angles of the pendulum and horizontal arm. the reward is zero when the pendulum attends the upright position. the rl algorithms are used without a deep understanding of the classical controllers and are used to implement the agent. finally, the outcome indicates the effectiveness of the ql and dqnl algorithms compared to the conventional pid and lqr controllers.","key_phrases":["the rotary inverted pendulum system","rips","an underactuated mechanical system","highly nonlinear dynamics","it","a rips","the classic control models","the last few years","reinforcement learning","(rl","a popular nonlinear control method","rl","a powerful potential","systems","high non-linearity and complex dynamics","rips","rl control","rips","limited research","the development","evaluation","this control method","this paper","rl control algorithms","the swing-up and stabilization control","a single-link rotary inverted pendulum","slrip","classic control methods","pid","lqr","a physical model","the slrip system","the matlab\/simscape toolbox","the model","a dynamic simulation","matlab\/simulink","the rl agents","an agent trainer system","q-learning","(ql","deep q-network learning","dqnl","the data training","agent actions","the horizontal arm","the system","states","the angles","velocities","the pendulum","the horizontal arm","the reward","the angles","the pendulum","horizontal arm","the reward","the pendulum","the upright position","the rl algorithms","a deep understanding","the classical controllers","the agent","the outcome","the effectiveness","the ql and dqnl algorithms","the conventional pid and lqr controllers","the last few years","zero"]},{"title":"Air combat maneuver decision based on deep reinforcement learning with auxiliary reward","authors":["Tingyu Zhang","Yongshuai Wang","Mingwei Sun","Zengqiang Chen"],"abstract":"For air combat maneuvering decision, the sparse reward during the application of deep reinforcement learning limits the exploration efficiency of the agents. To address this challenge, we propose an auxiliary reward function considering the impact of angle, range, and altitude. Furthermore, we investigate the influences of the network nodes, layers, and the learning rate on decision system, and reasonable parameter ranges are provided, which can serve as a guideline. Finally, four typical air combat scenarios demonstrate good adaptability and effectiveness of the proposed scheme, and the auxiliary reward significantly improves the learning ability of deep Q network (DQN) by leading the agents to explore more intently. Compared with the original deep deterministic policy gradient and soft actor critic algorithm, the proposed method exhibits superior exploration capability with higher reward, indicating that the trained agent can adapt to different air combats with good performance.","doi":"10.1007\/s00521-024-09720-z","cleaned_title":"air combat maneuver decision based on deep reinforcement learning with auxiliary reward","cleaned_abstract":"for air combat maneuvering decision, the sparse reward during the application of deep reinforcement learning limits the exploration efficiency of the agents. to address this challenge, we propose an auxiliary reward function considering the impact of angle, range, and altitude. furthermore, we investigate the influences of the network nodes, layers, and the learning rate on decision system, and reasonable parameter ranges are provided, which can serve as a guideline. finally, four typical air combat scenarios demonstrate good adaptability and effectiveness of the proposed scheme, and the auxiliary reward significantly improves the learning ability of deep q network (dqn) by leading the agents to explore more intently. compared with the original deep deterministic policy gradient and soft actor critic algorithm, the proposed method exhibits superior exploration capability with higher reward, indicating that the trained agent can adapt to different air combats with good performance.","key_phrases":["air combat maneuvering decision","the sparse","the application","deep reinforcement learning","the exploration efficiency","the agents","this challenge","we","an auxiliary reward function","the impact","angle","range","altitude","we","the influences","the network nodes","layers","the learning rate","decision system","reasonable parameter ranges","which","a guideline","four typical air combat scenarios","good adaptability","effectiveness","the proposed scheme","the auxiliary reward","the learning ability","deep q network","dqn","the agents","the original deep deterministic policy gradient","soft actor critic","algorithm","the proposed method","superior exploration capability","higher reward","the trained agent","different air combats","good performance","four"]},{"title":"Deep learning prediction of renal anomalies for prenatal ultrasound diagnosis","authors":["Olivier X. Miguel","Emily Kaczmarek","Inok Lee","Robin Ducharme","Alysha L. J. Dingwall-Harvey","Ruth Rennicks White","Brigitte Bonin","Richard I. Aviv","Steven Hawken","Christine M. Armour","Kevin Dick","Mark C. Walker"],"abstract":"Deep learning algorithms have demonstrated remarkable potential in clinical diagnostics, particularly in the field of medical imaging. In this study, we investigated the application of deep learning models in early detection of fetal kidney anomalies. To provide an enhanced interpretation of those models\u2019 predictions, we proposed an adapted two-class representation and developed a multi-class model interpretation approach for problems with more than two labels and variable hierarchical grouping of labels. Additionally, we employed the explainable AI (XAI) visualization tools Grad-CAM and HiResCAM, to gain insights into model predictions and identify reasons for misclassifications. The study dataset consisted of 969 ultrasound images from unique patients; 646 control images and 323 cases of kidney anomalies, including 259 cases of unilateral urinary tract dilation and 64 cases of unilateral multicystic dysplastic kidney. The best performing model achieved a cross-validated area under the ROC curve of 91.28%\u2009\u00b1\u20090.52%, with an overall accuracy of 84.03%\u2009\u00b1\u20090.76%, sensitivity of 77.39%\u2009\u00b1\u20091.99%, and specificity of 87.35%\u2009\u00b1\u20091.28%. Our findings emphasize the potential of deep learning models in predicting kidney anomalies from limited prenatal ultrasound imagery. The proposed adaptations in model representation and interpretation represent a novel solution to multi-class prediction problems.","doi":"10.1038\/s41598-024-59248-4","cleaned_title":"deep learning prediction of renal anomalies for prenatal ultrasound diagnosis","cleaned_abstract":"deep learning algorithms have demonstrated remarkable potential in clinical diagnostics, particularly in the field of medical imaging. in this study, we investigated the application of deep learning models in early detection of fetal kidney anomalies. to provide an enhanced interpretation of those models\u2019 predictions, we proposed an adapted two-class representation and developed a multi-class model interpretation approach for problems with more than two labels and variable hierarchical grouping of labels. additionally, we employed the explainable ai (xai) visualization tools grad-cam and hirescam, to gain insights into model predictions and identify reasons for misclassifications. the study dataset consisted of 969 ultrasound images from unique patients; 646 control images and 323 cases of kidney anomalies, including 259 cases of unilateral urinary tract dilation and 64 cases of unilateral multicystic dysplastic kidney. the best performing model achieved a cross-validated area under the roc curve of 91.28% \u00b1 0.52%, with an overall accuracy of 84.03% \u00b1 0.76%, sensitivity of 77.39% \u00b1 1.99%, and specificity of 87.35% \u00b1 1.28%. our findings emphasize the potential of deep learning models in predicting kidney anomalies from limited prenatal ultrasound imagery. the proposed adaptations in model representation and interpretation represent a novel solution to multi-class prediction problems.","key_phrases":["deep learning algorithms","remarkable potential","clinical diagnostics","the field","medical imaging","this study","we","the application","deep learning models","early detection","fetal kidney anomalies","an enhanced interpretation","those models\u2019 predictions","we","an adapted two-class representation","a multi-class model interpretation approach","problems","more than two labels","variable hierarchical grouping","labels","we","the explainable ai (xai) visualization tools","grad-cam","hirescam","insights","model predictions","reasons","misclassifications","the study dataset","969 ultrasound images","unique patients","646 control images","323 cases","kidney anomalies","259 cases","unilateral urinary tract dilation","64 cases","unilateral multicystic dysplastic kidney","the best performing model","a cross-validated area","the roc curve","91.28% \u00b1","0.52%","an overall accuracy","84.03%","\u00b1","0.76%","sensitivity","77.39% \u00b1","1.99%","specificity","87.35%","\u00b1","1.28%","our findings","the potential","deep learning models","kidney anomalies","limited prenatal ultrasound imagery","the proposed adaptations","model representation","interpretation","a novel solution","multi-class prediction problems","two","more than two","969","646","323","259","64","roc","91.28%","0.52%","84.03%","0.76%","77.39%","1.99%","87.35%","1.28%"]},{"title":"A review of deep learning and Generative Adversarial Networks applications in medical image analysis","authors":["D. N. Sindhura","Radhika M. Pai","Shyamasunder N. Bhat","Manohara M. M. Pai"],"abstract":"Nowadays, computer-aided decision support systems (CADs) for the analysis of images have been a perennial technique in the medical imaging field. In CADs, deep learning algorithms are widely used to perform tasks like classification, identification of patterns, detection, etc. Deep learning models learn feature representations from images rather than handcrafted features. Hence, deep learning models are quickly becoming the state-of-the-art method to achieve good performances in different computer-aided decision-support systems in medical applications. Similarly, deep learning-based generative models called Generative Adversarial Networks (GANs) have recently been developed as a novel method to produce realistic-looking synthetic data. GANs are used in different domains, including medical imaging generation. The common problems, like class imbalance and a small dataset, in healthcare are well addressed by GANs, and it is a leading area of research. Segmentation, reconstruction, detection, denoising, registration, etc. are the important applications of GANs. So in this work, the successes of deep learning methods in segmentation, classification, cell structure and fracture detection, computer-aided identification, and GANs in synthetic medical image generation, segmentation, reconstruction, detection, denoising, and registration in recent times are reviewed. Lately, the review article concludes by raising research directions for DL models and GANs in medical applications.","doi":"10.1007\/s00530-024-01349-1","cleaned_title":"a review of deep learning and generative adversarial networks applications in medical image analysis","cleaned_abstract":"nowadays, computer-aided decision support systems (cads) for the analysis of images have been a perennial technique in the medical imaging field. in cads, deep learning algorithms are widely used to perform tasks like classification, identification of patterns, detection, etc. deep learning models learn feature representations from images rather than handcrafted features. hence, deep learning models are quickly becoming the state-of-the-art method to achieve good performances in different computer-aided decision-support systems in medical applications. similarly, deep learning-based generative models called generative adversarial networks (gans) have recently been developed as a novel method to produce realistic-looking synthetic data. gans are used in different domains, including medical imaging generation. the common problems, like class imbalance and a small dataset, in healthcare are well addressed by gans, and it is a leading area of research. segmentation, reconstruction, detection, denoising, registration, etc. are the important applications of gans. so in this work, the successes of deep learning methods in segmentation, classification, cell structure and fracture detection, computer-aided identification, and gans in synthetic medical image generation, segmentation, reconstruction, detection, denoising, and registration in recent times are reviewed. lately, the review article concludes by raising research directions for dl models and gans in medical applications.","key_phrases":[", computer-aided decision support systems","cads","the analysis","images","a perennial technique","the medical imaging field","cads","deep learning algorithms","tasks","classification","identification","patterns","detection","deep learning models","feature representations","images","handcrafted features","deep learning models","the-art","good performances","different computer-aided decision-support systems","medical applications","deep learning-based generative models","generative adversarial networks","gans","a novel method","realistic-looking synthetic data","gans","different domains","medical imaging generation","the common problems","class imbalance","a small dataset","healthcare","gans","it","a leading area","research","segmentation","reconstruction","detection","denoising","registration","the important applications","gans","this work","the successes","deep learning methods","segmentation","classification","cell structure","fracture detection","computer-aided identification","gans","synthetic medical image generation","segmentation","reconstruction","detection","denoising","registration","recent times","the review article","research directions","dl models","gans","medical applications"]},{"title":"Deep learning segmentation of fibrous cap in intravascular optical coherence tomography images","authors":["Juhwan Lee","Justin N. Kim","Luis A. P. Dallan","Vladislav N. Zimin","Ammar Hoori","Neda S. Hassani","Mohamed H. E. Makhlouf","Giulio Guagliumi","Hiram G. Bezerra","David L. Wilson"],"abstract":"Thin-cap fibroatheroma (TCFA) is a prominent risk factor for plaque rupture. Intravascular optical coherence tomography (IVOCT) enables identification of fibrous cap (FC), measurement of FC thicknesses, and assessment of plaque vulnerability. We developed a fully-automated deep learning method for FC segmentation. This study included 32,531 images across 227 pullbacks from two registries (TRANSFORM-OCT and UHCMC). Images were semi-automatically labeled using our OCTOPUS with expert editing using established guidelines. We employed preprocessing including guidewire shadow detection, lumen segmentation, pixel-shifting, and Gaussian filtering on raw IVOCT (r,\u03b8) images. Data were augmented in a natural way by changing \u03b8 in spiral acquisitions and by changing intensity and noise values. We used a modified SegResNet and comparison networks to segment FCs. We employed transfer learning from our existing much larger, fully-labeled calcification IVOCT dataset to reduce deep-learning training. Postprocessing with a morphological operation enhanced segmentation performance. Overall, our method consistently delivered better FC segmentation results (Dice: 0.837\u2009\u00b1\u20090.012) than other deep-learning methods. Transfer learning reduced training time by 84% and reduced the need for more training samples. Our method showed a high level of generalizability, evidenced by highly-consistent segmentations across five-fold cross-validation (sensitivity: 85.0\u2009\u00b1\u20090.3%, Dice: 0.846\u2009\u00b1\u20090.011) and the held-out test (sensitivity: 84.9%, Dice: 0.816) sets. In addition, we found excellent agreement of FC thickness with ground truth (2.95\u2009\u00b1\u200920.73\u00a0\u00b5m), giving clinically insignificant bias. There was excellent reproducibility in pre- and post-stenting pullbacks (average FC angle: 200.9\u2009\u00b1\u2009128.0\u00b0\/202.0\u2009\u00b1\u2009121.1\u00b0). Our fully automated, deep-learning FC segmentation method demonstrated excellent performance, generalizability, and reproducibility on multi-center datasets. It will be useful for multiple research purposes and potentially for planning stent deployments that avoid placing a stent edge over an FC.","doi":"10.1038\/s41598-024-55120-7","cleaned_title":"deep learning segmentation of fibrous cap in intravascular optical coherence tomography images","cleaned_abstract":"thin-cap fibroatheroma (tcfa) is a prominent risk factor for plaque rupture. intravascular optical coherence tomography (ivoct) enables identification of fibrous cap (fc), measurement of fc thicknesses, and assessment of plaque vulnerability. we developed a fully-automated deep learning method for fc segmentation. this study included 32,531 images across 227 pullbacks from two registries (transform-oct and uhcmc). images were semi-automatically labeled using our octopus with expert editing using established guidelines. we employed preprocessing including guidewire shadow detection, lumen segmentation, pixel-shifting, and gaussian filtering on raw ivoct (r,\u03b8) images. data were augmented in a natural way by changing \u03b8 in spiral acquisitions and by changing intensity and noise values. we used a modified segresnet and comparison networks to segment fcs. we employed transfer learning from our existing much larger, fully-labeled calcification ivoct dataset to reduce deep-learning training. postprocessing with a morphological operation enhanced segmentation performance. overall, our method consistently delivered better fc segmentation results (dice: 0.837 \u00b1 0.012) than other deep-learning methods. transfer learning reduced training time by 84% and reduced the need for more training samples. our method showed a high level of generalizability, evidenced by highly-consistent segmentations across five-fold cross-validation (sensitivity: 85.0 \u00b1 0.3%, dice: 0.846 \u00b1 0.011) and the held-out test (sensitivity: 84.9%, dice: 0.816) sets. in addition, we found excellent agreement of fc thickness with ground truth (2.95 \u00b1 20.73 \u00b5m), giving clinically insignificant bias. there was excellent reproducibility in pre- and post-stenting pullbacks (average fc angle: 200.9 \u00b1 128.0\u00b0\/202.0 \u00b1 121.1\u00b0). our fully automated, deep-learning fc segmentation method demonstrated excellent performance, generalizability, and reproducibility on multi-center datasets. it will be useful for multiple research purposes and potentially for planning stent deployments that avoid placing a stent edge over an fc.","key_phrases":["thin-cap fibroatheroma","(tcfa","a prominent risk factor","plaque rupture","intravascular optical coherence tomography","ivoct","identification","fibrous cap","fc","measurement","fc thicknesses","assessment","plaque vulnerability","we","a fully-automated deep learning method","fc segmentation","this study","32,531 images","227 pullbacks","two registries","transform","oct","images","our octopus","expert editing","established guidelines","we","preprocessing","guidewire shadow detection","lumen segmentation","gaussian filtering","raw ivoct (r,\u03b8) images","data","a natural way","\u03b8","spiral acquisitions","intensity","noise values","we","a modified segresnet","comparison networks","we","transfer","our existing much larger, fully-labeled calcification ivoct dataset","deep-learning training","a morphological operation enhanced segmentation performance","our method","better fc segmentation results","dice","0.837 \u00b1","other deep-learning methods","transfer","reduced training time","84%","the need","more training samples","our method","a high level","generalizability","highly-consistent segmentations","five-fold cross","validation (sensitivity","85.0 \u00b1","0.3%","dice","0.846 \u00b1","the held-out test","84.9%","dice","addition","we","excellent agreement","fc thickness","ground truth","2.95 \u00b1","\u00b5m","clinically insignificant bias","excellent reproducibility","pre- and post-stenting pullbacks","average fc angle","200.9 \u00b1 128.0\u00b0","\u00b1","our fully automated, deep-learning fc segmentation method","excellent performance","generalizability","reproducibility","multi-center datasets","it","multiple research purposes","stent deployments","that","a stent edge","an fc","32,531","227","two","0.837 \u00b1","84%","five-fold","85.0","0.3%","0.846","0.011","84.9%","0.816","2.95","20.73","200.9","128.0","121.1"]},{"title":"Deep Q-learning with hybrid quantum neural network on solving maze problems","authors":["Hao-Yuan Chen","Yen-Jui Chang","Shih-Wei Liao","Ching-Ray Chang"],"abstract":"Quantum computing holds great potential for advancing the limitations of machine learning algorithms to handle higher dimensions of data and reduce overall training parameters in deep learning (DL) models. This study uses a trainable variational quantum circuit (VQC) on a gate-based quantum computing model to investigate the potential for quantum benefit in a model-free reinforcement learning problem. Through a comprehensive investigation and evaluation of the current model and capabilities of quantum computers, we designed and trained a novel hybrid quantum neural network based on the latest Qiskit and PyTorch framework. We compared its performance with a full-classical CNN with and without an incorporated VQC. Our research provides insights into the potential of deep quantum learning to solve a maze problem and, potentially, other reinforcement learning problems. We conclude that reinforcement learning problems can be practical with reasonable training epochs. Moreover, a comparative study of full-classical and hybrid quantum neural networks is discussed to understand these two approaches\u2019 performance, advantages, and disadvantages to deep Q-learning problems, especially on larger-scale maze problems larger than 4\\(\\times \\)4.","doi":"10.1007\/s42484-023-00137-w","cleaned_title":"deep q-learning with hybrid quantum neural network on solving maze problems","cleaned_abstract":"quantum computing holds great potential for advancing the limitations of machine learning algorithms to handle higher dimensions of data and reduce overall training parameters in deep learning (dl) models. this study uses a trainable variational quantum circuit (vqc) on a gate-based quantum computing model to investigate the potential for quantum benefit in a model-free reinforcement learning problem. through a comprehensive investigation and evaluation of the current model and capabilities of quantum computers, we designed and trained a novel hybrid quantum neural network based on the latest qiskit and pytorch framework. we compared its performance with a full-classical cnn with and without an incorporated vqc. our research provides insights into the potential of deep quantum learning to solve a maze problem and, potentially, other reinforcement learning problems. we conclude that reinforcement learning problems can be practical with reasonable training epochs. moreover, a comparative study of full-classical and hybrid quantum neural networks is discussed to understand these two approaches\u2019 performance, advantages, and disadvantages to deep q-learning problems, especially on larger-scale maze problems larger than 4\\(\\times \\)4.","key_phrases":["quantum computing","great potential","the limitations","machine learning algorithms","higher dimensions","data","overall training parameters","deep learning","(dl) models","this study","a trainable variational quantum circuit","vqc","a gate-based quantum computing model","the potential","quantum benefit","a model-free reinforcement learning problem","a comprehensive investigation","evaluation","the current model","capabilities","quantum computers","we","a novel hybrid quantum neural network","the latest qiskit and pytorch framework","we","its performance","a full-classical cnn","an incorporated vqc","our research","insights","the potential","deep quantum","a maze problem","potentially, other reinforcement learning problems","we","reinforcement learning problems","reasonable training epochs","a comparative study","full-classical and hybrid quantum neural networks","these two approaches\u2019 performance","advantages","disadvantages","deep q-learning problems","larger-scale maze problems","4\\(\\times","quantum","quantum","quantum","quantum","quantum","two"]},{"title":"Diagnosis of EV Gearbox Bearing Fault Using Deep Learning-Based Signal Processing","authors":["Kicheol Jeong","Chulwoo Moon"],"abstract":"The gearbox of an electric vehicle operates under the high load torque and axial load of electric vehicles. In particular, the bearings that support the shaft of the gearbox are subjected to several tons of axial load, and as the mileage increases, fault occurs on bearing rolling elements frequently. Such bearing fault has a serious impact on driving comfort and vehicle safety, however, bearing faults are diagnosed by human experts nowadays, and algorithm-based electric vehicle bearing fault diagnosis has not been implemented. Therefore, in this paper, a deep learning-based bearing vibration signal processing method to diagnose bearing fault in electric vehicle gearboxes is proposed. The proposed method consists of a deep neural network learning stage and an application stage of the pre-trained neural network. In the deep neural network learning stage, supervised learning is carried out based on two acceleration sensors. In the neural network application stage, signal processing of a single accelerometer signal is performed through a pre-trained neural network. In conclusion, the pre-trained neural network makes bearing fault signals stand out and can utilize these signals to extract frequency characteristics of bearing fault.","doi":"10.1007\/s12239-024-00094-8","cleaned_title":"diagnosis of ev gearbox bearing fault using deep learning-based signal processing","cleaned_abstract":"the gearbox of an electric vehicle operates under the high load torque and axial load of electric vehicles. in particular, the bearings that support the shaft of the gearbox are subjected to several tons of axial load, and as the mileage increases, fault occurs on bearing rolling elements frequently. such bearing fault has a serious impact on driving comfort and vehicle safety, however, bearing faults are diagnosed by human experts nowadays, and algorithm-based electric vehicle bearing fault diagnosis has not been implemented. therefore, in this paper, a deep learning-based bearing vibration signal processing method to diagnose bearing fault in electric vehicle gearboxes is proposed. the proposed method consists of a deep neural network learning stage and an application stage of the pre-trained neural network. in the deep neural network learning stage, supervised learning is carried out based on two acceleration sensors. in the neural network application stage, signal processing of a single accelerometer signal is performed through a pre-trained neural network. in conclusion, the pre-trained neural network makes bearing fault signals stand out and can utilize these signals to extract frequency characteristics of bearing fault.","key_phrases":["the gearbox","an electric vehicle","the high load torque","axial load","electric vehicles","the bearings","that","the shaft","the gearbox","several tons","axial load","the mileage increases","fault","rolling elements","fault","a serious impact","comfort","vehicle safety","faults","human experts","algorithm-based electric vehicle","fault diagnosis","this paper","a deep learning-based bearing vibration signal processing method","fault","electric vehicle gearboxes","the proposed method","a deep neural network learning stage","an application stage","the pre-trained neural network","the deep neural network learning stage","supervised learning","two acceleration sensors","the neural network application stage","signal processing","a single accelerometer signal","a pre-trained neural network","conclusion","the pre-trained neural network","fault signals","these signals","frequency characteristics","fault","several tons","two"]},{"title":"Comparative study and analysis on skin cancer detection using machine learning\u00a0and\u00a0deep learning algorithms","authors":["V. Auxilia Osvin Nancy","P. Prabhavathy","Meenakshi S. Arya","B. Shamreen Ahamed"],"abstract":"Exposure to UV rays due to global warming can lead to sunburn and skin damage, ultimately resulting in skin cancer. Early prediction of this type of cancer is crucial. A detailed review in this paper explores various algorithms, including machine learning (ML) techniques as well as deep learning (DL) techniques. While deep learning strategies, particularly CNNs, are commonly employed for skin cancer identification and classification, there is also some usage of machine learning and hybrid approaches. These techniques have proven to be effective classifiers of skin lesions, offering promising results for early detection. The paper analyzes various researchers\u2019 reviews on skin cancer diagnosis to identify a suitable methodology for improving diagnostic accuracy. A publicly available dataset of dermoscopic images retrieved from the ISIC archive has been trained and evaluated. Performance analysis is done, considering metrics such as test and validation accuracy. The results indicate that the RF(random forest) algorithm outperforms other machine learning algorithms in both scenarios, with accuracies of 58.57% without augmentation and 87.32% with augmentation. MobileNetv2, ensemble of Dense Net and Inceptionv3 exhibit superior performance. During training without augmentation, MobileNetv2 achieves an accuracy of 88.81%, while the ensemble model achieves an accuracy of 88.80%. With augmentation techniques applied, the accuracies improved to 97.58% and 97.50%, respectively. Furthermore, experiment with a customized convolutional neural network (CNN) model was also conducted, varying the number of layers and applying various hyperparameter tuning methodologies. Suitable architectures, including a CNN with 7 layers and batch normalization, a CNN with 5 layers, and a CNN with 3 layers were identified. These models achieved accuracies of 77.92%, 97.72%, and 98.02% on the raw data and augmentation datasets, respectively. The experimental results suggest that these techniques hold promise for integration into clinical settings, and further research and validation are necessary. The results highlight the effectiveness of transfer learning models, in achieving high accuracy rates. The findings support the future adoption of these techniques in clinical practice, pending further research and validation.","doi":"10.1007\/s11042-023-16422-6","cleaned_title":"comparative study and analysis on skin cancer detection using machine learning and deep learning algorithms","cleaned_abstract":"exposure to uv rays due to global warming can lead to sunburn and skin damage, ultimately resulting in skin cancer. early prediction of this type of cancer is crucial. a detailed review in this paper explores various algorithms, including machine learning (ml) techniques as well as deep learning (dl) techniques. while deep learning strategies, particularly cnns, are commonly employed for skin cancer identification and classification, there is also some usage of machine learning and hybrid approaches. these techniques have proven to be effective classifiers of skin lesions, offering promising results for early detection. the paper analyzes various researchers\u2019 reviews on skin cancer diagnosis to identify a suitable methodology for improving diagnostic accuracy. a publicly available dataset of dermoscopic images retrieved from the isic archive has been trained and evaluated. performance analysis is done, considering metrics such as test and validation accuracy. the results indicate that the rf(random forest) algorithm outperforms other machine learning algorithms in both scenarios, with accuracies of 58.57% without augmentation and 87.32% with augmentation. mobilenetv2, ensemble of dense net and inceptionv3 exhibit superior performance. during training without augmentation, mobilenetv2 achieves an accuracy of 88.81%, while the ensemble model achieves an accuracy of 88.80%. with augmentation techniques applied, the accuracies improved to 97.58% and 97.50%, respectively. furthermore, experiment with a customized convolutional neural network (cnn) model was also conducted, varying the number of layers and applying various hyperparameter tuning methodologies. suitable architectures, including a cnn with 7 layers and batch normalization, a cnn with 5 layers, and a cnn with 3 layers were identified. these models achieved accuracies of 77.92%, 97.72%, and 98.02% on the raw data and augmentation datasets, respectively. the experimental results suggest that these techniques hold promise for integration into clinical settings, and further research and validation are necessary. the results highlight the effectiveness of transfer learning models, in achieving high accuracy rates. the findings support the future adoption of these techniques in clinical practice, pending further research and validation.","key_phrases":["exposure","uv rays","global warming","sunburn","skin damage","skin cancer","early prediction","this type","cancer","a detailed review","this paper","various algorithms","machine learning",") techniques","deep learning","(dl) techniques","deep learning strategies","particularly cnns","skin cancer identification","classification","some usage","machine learning","hybrid approaches","these techniques","effective classifiers","skin lesions","promising results","early detection","the paper","various researchers\u2019 reviews","skin cancer diagnosis","a suitable methodology","diagnostic accuracy","a publicly available dataset","dermoscopic images","the isic archive","performance analysis","metrics","test","validation accuracy","the results","algorithm","other machine learning algorithms","both scenarios","accuracies","58.57%","augmentation","87.32%","augmentation","mobilenetv2","dense net","inceptionv3","superior performance","training","augmentation","mobilenetv2","an accuracy","88.81%","the ensemble model","an accuracy","88.80%","augmentation techniques","the accuracies","97.58%","97.50%","experiment","a customized convolutional neural network (cnn) model","the number","layers","various hyperparameter","methodologies","suitable architectures","a cnn","7 layers","batch normalization","a cnn","5 layers","a cnn","3 layers","these models","accuracies","77.92%","97.72%","98.02%","the raw data","augmentation datasets","the experimental results","these techniques","promise","integration","clinical settings","further research","validation","the results","the effectiveness","transfer learning models","high accuracy rates","the findings","the future adoption","these techniques","clinical practice","further research","validation","58.57%","87.32%","mobilenetv2","inceptionv3","mobilenetv2","88.81%","88.80%","97.58%","97.50%","cnn","cnn","7","cnn","5","cnn","3","77.92%","97.72%","98.02%"]},{"title":"Deep palmprint recognition algorithm based on self-supervised learning and uncertainty loss","authors":["Rui Fan","Xiaohong Han"],"abstract":"With the rapid development of deep learning technology, an increasing number of people are adopting palmprint recognition algorithms based on deep learning for identity authentication. However, these algorithms are susceptible to factors such as palm placement, light source, and insufficient data sampling, resulting in poor recognition accuracy. To address these issues, this paper proposes a new end-to-end deep palmprint recognition algorithm (SSLAUL), which introduces self-supervised representation learning based on contextual prediction, utilizing unlabeled palmprint data for pre-training before introducing the trained parameters into the downstream model for fine-tuning. An uncertainty loss function is introduced into the downstream model, using the homoskedastic uncertainty as a benchmark to do adaptive weight adjustment for different loss functions dynamically. Channel and spatial attention mechanisms are also introduced to extract highly discriminative local features. In this paper, the algorithm is validated on publicly available IITD, CASIA, and PolyU palmprint datasets. The method always achieves the best recognition performance compared to other state-of-the-art algorithms.","doi":"10.1007\/s11760-024-03104-5","cleaned_title":"deep palmprint recognition algorithm based on self-supervised learning and uncertainty loss","cleaned_abstract":"with the rapid development of deep learning technology, an increasing number of people are adopting palmprint recognition algorithms based on deep learning for identity authentication. however, these algorithms are susceptible to factors such as palm placement, light source, and insufficient data sampling, resulting in poor recognition accuracy. to address these issues, this paper proposes a new end-to-end deep palmprint recognition algorithm (sslaul), which introduces self-supervised representation learning based on contextual prediction, utilizing unlabeled palmprint data for pre-training before introducing the trained parameters into the downstream model for fine-tuning. an uncertainty loss function is introduced into the downstream model, using the homoskedastic uncertainty as a benchmark to do adaptive weight adjustment for different loss functions dynamically. channel and spatial attention mechanisms are also introduced to extract highly discriminative local features. in this paper, the algorithm is validated on publicly available iitd, casia, and polyu palmprint datasets. the method always achieves the best recognition performance compared to other state-of-the-art algorithms.","key_phrases":["the rapid development","deep learning technology","an increasing number","people","palmprint recognition algorithms","deep learning","identity authentication","these algorithms","factors","palm placement","light source","insufficient data sampling","poor recognition accuracy","these issues","this paper","end","recognition algorithm","sslaul","which","self-supervised representation learning","contextual prediction","unlabeled palmprint data","pre","-","training","the trained parameters","the downstream model","fine-tuning","an uncertainty loss function","the downstream model","the homoskedastic uncertainty","a benchmark","adaptive weight adjustment","different loss functions","channel","spatial attention mechanisms","highly discriminative local features","this paper","the algorithm","publicly available iitd","casia","polyu","datasets","the method","the best recognition performance","the-art","casia"]},{"title":"Challenges and practices of deep learning model reengineering: A case study on computer vision","authors":["Wenxin Jiang","Vishnu Banna","Naveen Vivek","Abhinav Goel","Nicholas Synovic","George K. Thiruvathukal","James C. Davis"],"abstract":"ContextMany engineering organizations are reimplementing and extending deep neural networks from the research community. We describe this process as deep learning model reengineering. Deep learning model reengineering \u2014 reusing, replicating, adapting, and enhancing state-of-the-art deep learning approaches \u2014 is challenging for reasons including under-documented reference models, changing requirements, and the cost of implementation and testing.ObjectivePrior work has characterized the challenges of deep learning model development, but as yet we know little about the deep learning model reengineering process and its common challenges. Prior work has examined DL systems from a \u201cproduct\u201d view, examining defects from projects regardless of the engineers\u2019 purpose. Our study is focused on reengineering activities from a \u201cprocess\u201d view, and focuses on engineers specifically engaged in the reengineering process.MethodOur goal is to understand the characteristics and challenges of deep learning model reengineering. We conducted a mixed-methods case study of this phenomenon, focusing on the context of computer vision. Our results draw from two data sources: defects reported in open-source reeengineering projects, and interviews conducted with practitioners and the leaders of a reengineering team. From the defect data source, we analyzed 348 defects from 27 open-source deep learning projects. Meanwhile, our reengineering team replicated 7 deep learning models over two years; we interviewed 2 open-source contributors, 4 practitioners, and 6 reengineering team leaders to understand their experiences.ResultsOur results describe how deep learning-based computer vision techniques are reengineered, quantitatively analyze the distribution of defects in this process, and qualitatively discuss challenges and practices. We found that most defects (58%) are reported by re-users, and that reproducibility-related defects tend to be discovered during training (68% of them are). Our analysis shows that most environment defects (88%) are interface defects, and most environment defects (46%) are caused by API defects. We found that training defects have diverse symptoms and root causes. We identified four main challenges in the DL reengineering process: model operationalization, performance debugging, portability of DL operations, and customized data pipeline. Integrating our quantitative and qualitative data, we propose a novel reengineering workflow.ConclusionsOur findings inform several conclusion, including: standardizing model reengineering practices, developing validation tools to support model reengineering, automated support beyond manual model reengineering, and measuring additional unknown aspects of model reengineering.","doi":"10.1007\/s10664-024-10521-0","cleaned_title":"challenges and practices of deep learning model reengineering: a case study on computer vision","cleaned_abstract":"contextmany engineering organizations are reimplementing and extending deep neural networks from the research community. we describe this process as deep learning model reengineering. deep learning model reengineering \u2014 reusing, replicating, adapting, and enhancing state-of-the-art deep learning approaches \u2014 is challenging for reasons including under-documented reference models, changing requirements, and the cost of implementation and testing.objectiveprior work has characterized the challenges of deep learning model development, but as yet we know little about the deep learning model reengineering process and its common challenges. prior work has examined dl systems from a \u201cproduct\u201d view, examining defects from projects regardless of the engineers\u2019 purpose. our study is focused on reengineering activities from a \u201cprocess\u201d view, and focuses on engineers specifically engaged in the reengineering process.methodour goal is to understand the characteristics and challenges of deep learning model reengineering. we conducted a mixed-methods case study of this phenomenon, focusing on the context of computer vision. our results draw from two data sources: defects reported in open-source reeengineering projects, and interviews conducted with practitioners and the leaders of a reengineering team. from the defect data source, we analyzed 348 defects from 27 open-source deep learning projects. meanwhile, our reengineering team replicated 7 deep learning models over two years; we interviewed 2 open-source contributors, 4 practitioners, and 6 reengineering team leaders to understand their experiences.resultsour results describe how deep learning-based computer vision techniques are reengineered, quantitatively analyze the distribution of defects in this process, and qualitatively discuss challenges and practices. we found that most defects (58%) are reported by re-users, and that reproducibility-related defects tend to be discovered during training (68% of them are). our analysis shows that most environment defects (88%) are interface defects, and most environment defects (46%) are caused by api defects. we found that training defects have diverse symptoms and root causes. we identified four main challenges in the dl reengineering process: model operationalization, performance debugging, portability of dl operations, and customized data pipeline. integrating our quantitative and qualitative data, we propose a novel reengineering workflow.conclusionsour findings inform several conclusion, including: standardizing model reengineering practices, developing validation tools to support model reengineering, automated support beyond manual model reengineering, and measuring additional unknown aspects of model reengineering.","key_phrases":["contextmany engineering organizations","deep neural networks","the research community","we","this process","deep learning model reengineering","deep learning model reengineering","the-art","reasons","under-documented reference models","changing requirements","the cost","implementation","testing.objectiveprior work","the challenges","deep learning model development","we","the deep learning model reengineering process","its common challenges","prior work","dl systems","a \u201cproduct\u201d view","defects","projects","the engineers\u2019 purpose","our study","activities","a \u201cprocess\u201d view","engineers","the reengineering","process.methodour goal","the characteristics","challenges","deep learning model reengineering","we","a mixed-methods case study","this phenomenon","the context","computer vision","our results","two data sources","defects","open-source reeengineering projects","interviews","practitioners","the leaders","a reengineering team","the defect data source","we","348 defects","27 open-source deep learning projects","our reengineering team","7 deep learning models","two years","we","2 open-source contributors","4 practitioners","6 reengineering team leaders","their experiences.resultsour results","how deep learning-based computer vision techniques","the distribution","defects","this process","challenges","practices","we","most defects","58%","re","-","users","reproducibility-related defects","training","68%","them","our analysis","most environment defects","88%","interface defects","most environment defects","46%","api defects","we","training defects","diverse symptoms","root causes","we","four main challenges","the dl reengineering process","model operationalization","performance debugging","portability","dl operations","customized data pipeline","our quantitative and qualitative data","we","a novel","workflow.conclusionsour findings","several conclusion","standardizing model reengineering practices","validation tools","model reengineering","automated support","manual model reengineering","additional unknown aspects","model reengineering","two","348","27","7","two years","2","4","6","58%","68%","88%","46%","four"]},{"title":"Study of Q-learning and deep Q-network learning control for a rotary inverted pendulum system","authors":["Zied Ben Hazem"],"abstract":"The rotary inverted pendulum system (RIPS) is an underactuated mechanical system with highly nonlinear dynamics and it is difficult to control a RIPS using the classic control models. In the last few years, reinforcement learning (RL) has become a popular nonlinear control method. RL has a powerful potential to control systems with high non-linearity and complex dynamics, such as RIPS. Nevertheless, RL control for RIPS has not been well studied and there is limited research on the development and evaluation of this control method. In this paper, RL control algorithms are developed for the swing-up and stabilization control of a single-link rotary inverted pendulum (SLRIP) and compared with classic control methods such as PID and LQR. A physical model of the SLRIP system is created using the MATLAB\/Simscape Toolbox, the model is used as a dynamic simulation in MATLAB\/Simulink to train the RL agents. An agent trainer system with Q-learning (QL) and deep Q-network learning (DQNL) is proposed for the data training. Furthermore, agent actions are actuating the horizontal arm of the system and states are the angles and velocities of the pendulum and the horizontal arm. The reward is computed according to the angles of the pendulum and horizontal arm. The reward is zero when the pendulum attends the upright position. The RL algorithms are used without a deep understanding of the classical controllers and are used to implement the agent. Finally, the outcome indicates the effectiveness of the QL and DQNL algorithms compared to the conventional PID and LQR controllers.","doi":"10.1007\/s42452-024-05690-y","cleaned_title":"study of q-learning and deep q-network learning control for a rotary inverted pendulum system","cleaned_abstract":"the rotary inverted pendulum system (rips) is an underactuated mechanical system with highly nonlinear dynamics and it is difficult to control a rips using the classic control models. in the last few years, reinforcement learning (rl) has become a popular nonlinear control method. rl has a powerful potential to control systems with high non-linearity and complex dynamics, such as rips. nevertheless, rl control for rips has not been well studied and there is limited research on the development and evaluation of this control method. in this paper, rl control algorithms are developed for the swing-up and stabilization control of a single-link rotary inverted pendulum (slrip) and compared with classic control methods such as pid and lqr. a physical model of the slrip system is created using the matlab\/simscape toolbox, the model is used as a dynamic simulation in matlab\/simulink to train the rl agents. an agent trainer system with q-learning (ql) and deep q-network learning (dqnl) is proposed for the data training. furthermore, agent actions are actuating the horizontal arm of the system and states are the angles and velocities of the pendulum and the horizontal arm. the reward is computed according to the angles of the pendulum and horizontal arm. the reward is zero when the pendulum attends the upright position. the rl algorithms are used without a deep understanding of the classical controllers and are used to implement the agent. finally, the outcome indicates the effectiveness of the ql and dqnl algorithms compared to the conventional pid and lqr controllers.","key_phrases":["the rotary inverted pendulum system","rips","an underactuated mechanical system","highly nonlinear dynamics","it","a rips","the classic control models","the last few years","reinforcement learning","(rl","a popular nonlinear control method","rl","a powerful potential","systems","high non-linearity and complex dynamics","rips","rl control","rips","limited research","the development","evaluation","this control method","this paper","rl control algorithms","the swing-up and stabilization control","a single-link rotary inverted pendulum","slrip","classic control methods","pid","lqr","a physical model","the slrip system","the matlab\/simscape toolbox","the model","a dynamic simulation","matlab\/simulink","the rl agents","an agent trainer system","q-learning","(ql","deep q-network learning","dqnl","the data training","agent actions","the horizontal arm","the system","states","the angles","velocities","the pendulum","the horizontal arm","the reward","the angles","the pendulum","horizontal arm","the reward","the pendulum","the upright position","the rl algorithms","a deep understanding","the classical controllers","the agent","the outcome","the effectiveness","the ql and dqnl algorithms","the conventional pid and lqr controllers","the last few years","zero"]},{"title":"Hybrid deep learning approach to improve classification of low-volume high-dimensional data","authors":["Pegah Mavaie","Lawrence Holder","Michael K. Skinner"],"abstract":"BackgroundThe performance of machine learning classification methods relies heavily on the choice of features. In many domains, feature generation can be labor-intensive and require domain knowledge, and feature selection methods do not scale well in high-dimensional datasets. Deep learning has shown success in feature generation but requires large datasets to achieve high classification accuracy. Biology domains typically exhibit these challenges with numerous handcrafted features (high-dimensional) and small amounts of training data (low volume).MethodA hybrid learning approach is proposed that first trains a deep network on the training data, extracts features from the deep network, and then uses these features to re-express the data for input to a non-deep learning method, which is trained to perform the final classification.ResultsThe approach is systematically evaluated to determine the best layer of the deep learning network from which to extract features and the threshold on training data volume that prefers this approach. Results from several domains show that this hybrid approach outperforms standalone deep and non-deep learning methods, especially on low-volume, high-dimensional datasets. The diverse collection of datasets further supports the robustness of the approach across different domains.ConclusionsThe hybrid approach combines the strengths of deep and non-deep learning paradigms to achieve high performance on high-dimensional, low volume learning tasks that are typical in biology domains.","doi":"10.1186\/s12859-023-05557-w","cleaned_title":"hybrid deep learning approach to improve classification of low-volume high-dimensional data","cleaned_abstract":"backgroundthe performance of machine learning classification methods relies heavily on the choice of features. in many domains, feature generation can be labor-intensive and require domain knowledge, and feature selection methods do not scale well in high-dimensional datasets. deep learning has shown success in feature generation but requires large datasets to achieve high classification accuracy. biology domains typically exhibit these challenges with numerous handcrafted features (high-dimensional) and small amounts of training data (low volume).methoda hybrid learning approach is proposed that first trains a deep network on the training data, extracts features from the deep network, and then uses these features to re-express the data for input to a non-deep learning method, which is trained to perform the final classification.resultsthe approach is systematically evaluated to determine the best layer of the deep learning network from which to extract features and the threshold on training data volume that prefers this approach. results from several domains show that this hybrid approach outperforms standalone deep and non-deep learning methods, especially on low-volume, high-dimensional datasets. the diverse collection of datasets further supports the robustness of the approach across different domains.conclusionsthe hybrid approach combines the strengths of deep and non-deep learning paradigms to achieve high performance on high-dimensional, low volume learning tasks that are typical in biology domains.","key_phrases":["backgroundthe performance","classification methods","the choice","features","many domains","feature generation","domain knowledge","feature selection methods","high-dimensional datasets","deep learning","success","feature generation","large datasets","high classification accuracy","biology domains","these challenges","numerous handcrafted features","small amounts","training data","low volume).methoda hybrid learning approach","a deep network","the training data","features","the deep network","these features","the data","input","a non-deep learning method","which","the final classification.resultsthe approach","the best layer","the deep learning network","which","features","the threshold","training data volume","that","this approach","results","several domains","this hybrid approach","deep and non-deep learning methods","low-volume, high-dimensional datasets","the diverse collection","datasets","the robustness","the approach","different domains.conclusionsthe hybrid approach","the strengths","deep and non-deep learning paradigms","high performance","high-dimensional, low volume learning tasks","that","biology domains","first"]},{"title":"Air combat maneuver decision based on deep reinforcement learning with auxiliary reward","authors":["Tingyu Zhang","Yongshuai Wang","Mingwei Sun","Zengqiang Chen"],"abstract":"For air combat maneuvering decision, the sparse reward during the application of deep reinforcement learning limits the exploration efficiency of the agents. To address this challenge, we propose an auxiliary reward function considering the impact of angle, range, and altitude. Furthermore, we investigate the influences of the network nodes, layers, and the learning rate on decision system, and reasonable parameter ranges are provided, which can serve as a guideline. Finally, four typical air combat scenarios demonstrate good adaptability and effectiveness of the proposed scheme, and the auxiliary reward significantly improves the learning ability of deep Q network (DQN) by leading the agents to explore more intently. Compared with the original deep deterministic policy gradient and soft actor critic algorithm, the proposed method exhibits superior exploration capability with higher reward, indicating that the trained agent can adapt to different air combats with good performance.","doi":"10.1007\/s00521-024-09720-z","cleaned_title":"air combat maneuver decision based on deep reinforcement learning with auxiliary reward","cleaned_abstract":"for air combat maneuvering decision, the sparse reward during the application of deep reinforcement learning limits the exploration efficiency of the agents. to address this challenge, we propose an auxiliary reward function considering the impact of angle, range, and altitude. furthermore, we investigate the influences of the network nodes, layers, and the learning rate on decision system, and reasonable parameter ranges are provided, which can serve as a guideline. finally, four typical air combat scenarios demonstrate good adaptability and effectiveness of the proposed scheme, and the auxiliary reward significantly improves the learning ability of deep q network (dqn) by leading the agents to explore more intently. compared with the original deep deterministic policy gradient and soft actor critic algorithm, the proposed method exhibits superior exploration capability with higher reward, indicating that the trained agent can adapt to different air combats with good performance.","key_phrases":["air combat maneuvering decision","the sparse","the application","deep reinforcement learning","the exploration efficiency","the agents","this challenge","we","an auxiliary reward function","the impact","angle","range","altitude","we","the influences","the network nodes","layers","the learning rate","decision system","reasonable parameter ranges","which","a guideline","four typical air combat scenarios","good adaptability","effectiveness","the proposed scheme","the auxiliary reward","the learning ability","deep q network","dqn","the agents","the original deep deterministic policy gradient","soft actor critic","algorithm","the proposed method","superior exploration capability","higher reward","the trained agent","different air combats","good performance","four"]},{"title":"Deep learning-based classification and application test of multiple crop leaf diseases using transfer learning and the attention mechanism","authors":["Yifu Zhang","Qian Sun","Ji Chen","Huini Zhou"],"abstract":"Crop diseases are among the major natural disasters in agricultural production that seriously restrict the growth and development of crops, threatening food security. Timely classification, accurate identification, and the application of methods suitable for the situation can effectively prevent and control crop diseases, improving the quality of agricultural products. Considering the huge variety of crops, diseases, and differences in the characteristics of diseases during each stage, the current convolutional neural network models based on deep learning need to meet the higher requirement of classifying crop diseases accurately. It is necessary to introduce a new architecture scheme to improve the recognition effect. Therefore, in this study, we optimized the deep learning-based classification model for multiple crop leaf diseases using combined transfer learning and the attention mechanism, the modified model was deployed in the smartphone for testing. Dataset that containing 10 types of crops, 61 types of diseases, and different degrees was established, the algorithm structure based on ResNet50 was designed using transfer learning and the SE attention mechanism. The classification performances of different improvement methods were compared by model training. Result indicates that the average accuracy of the proposed TL-SE-ResNet50 model is increased by 7.7%, reaching 96.32%. The model was also integrated and implemented in the smartphone and the test result of the application reaches 94.8%, and the average response time is 882\u00a0ms. The improved model proposed has a good effect on the identification of diseases and their condition of multiple crops, and the application can meet the portable usage needs of farmers. This study can provide reference for more crop disease management research in agricultural production.","doi":"10.1007\/s00607-024-01308-8","cleaned_title":"deep learning-based classification and application test of multiple crop leaf diseases using transfer learning and the attention mechanism","cleaned_abstract":"crop diseases are among the major natural disasters in agricultural production that seriously restrict the growth and development of crops, threatening food security. timely classification, accurate identification, and the application of methods suitable for the situation can effectively prevent and control crop diseases, improving the quality of agricultural products. considering the huge variety of crops, diseases, and differences in the characteristics of diseases during each stage, the current convolutional neural network models based on deep learning need to meet the higher requirement of classifying crop diseases accurately. it is necessary to introduce a new architecture scheme to improve the recognition effect. therefore, in this study, we optimized the deep learning-based classification model for multiple crop leaf diseases using combined transfer learning and the attention mechanism, the modified model was deployed in the smartphone for testing. dataset that containing 10 types of crops, 61 types of diseases, and different degrees was established, the algorithm structure based on resnet50 was designed using transfer learning and the se attention mechanism. the classification performances of different improvement methods were compared by model training. result indicates that the average accuracy of the proposed tl-se-resnet50 model is increased by 7.7%, reaching 96.32%. the model was also integrated and implemented in the smartphone and the test result of the application reaches 94.8%, and the average response time is 882 ms. the improved model proposed has a good effect on the identification of diseases and their condition of multiple crops, and the application can meet the portable usage needs of farmers. this study can provide reference for more crop disease management research in agricultural production.","key_phrases":["crop diseases","the major natural disasters","agricultural production","that","the growth","development","crops","food security","timely classification","accurate identification","the application","methods","the situation","crop diseases","the quality","agricultural products","the huge variety","crops","diseases","differences","the characteristics","diseases","each stage","the current convolutional neural network models","deep learning","the higher requirement","crop diseases","it","a new architecture scheme","the recognition effect","this study","we","the deep learning-based classification model","multiple crop leaf diseases","combined transfer learning","the attention mechanism","the modified model","the smartphone","testing","10 types","crops","61 types","diseases","different degrees","the algorithm structure","resnet50","transfer learning","the se attention mechanism","the classification performances","different improvement methods","model training","result","the average accuracy","the proposed tl-se-resnet50 model","7.7%","96.32%","the model","the smartphone","the test result","the application","94.8%","the average response time","882 ms","the improved model","a good effect","the identification","diseases","their condition","multiple crops","the application","the portable usage needs","farmers","this study","reference","more crop disease management research","agricultural production","10","61","resnet50","7.7%","96.32%","94.8%","882"]},{"title":"Sparse subspace clustering incorporated deep convolutional transform learning for hyperspectral band selection","authors":["Anurag Goel","Angshul Majumdar"],"abstract":"This work delves into a research area with a limited number of studies, that of convolutional filter-learning based clustering. Since clustering is an unsupervised formulation, one cannot formulate it around a conventional convolutional neural network since the latter is inherently supervised. This work, therefore, builds upon the recently developed framework of deep convolutional transform learning. The sparse subspace clustering formulation is embedded into deep convolutional transform learning to get an end-to-end clustering architecture. The proposed approach has been applied to the problem of hyperspectral band selection. Comparison with a variety of state-of-the-art techniques in this domain shows that our proposed approach improves over them in terms of the classification metrics.","doi":"10.1007\/s12145-024-01312-8","cleaned_title":"sparse subspace clustering incorporated deep convolutional transform learning for hyperspectral band selection","cleaned_abstract":"this work delves into a research area with a limited number of studies, that of convolutional filter-learning based clustering. since clustering is an unsupervised formulation, one cannot formulate it around a conventional convolutional neural network since the latter is inherently supervised. this work, therefore, builds upon the recently developed framework of deep convolutional transform learning. the sparse subspace clustering formulation is embedded into deep convolutional transform learning to get an end-to-end clustering architecture. the proposed approach has been applied to the problem of hyperspectral band selection. comparison with a variety of state-of-the-art techniques in this domain shows that our proposed approach improves over them in terms of the classification metrics.","key_phrases":["this work","a research area","a limited number","studies","that","convolutional filter-learning based clustering","an unsupervised formulation","one","it","a conventional convolutional neural network","the latter","this work","the recently developed framework","the sparse","clustering formulation","deep convolutional transform","end","the proposed approach","the problem","hyperspectral band selection","comparison","a variety","the-art","this domain","our proposed approach","them","terms","the classification metrics"]},{"title":"Effectiveness evaluation of sprint sports techniques and tactics based on deep learning","authors":["Jiankui Yan"],"abstract":"The assessment of sprint-like sports techniques and tactics involves evaluating their underlying running principles, structures, functions, methods, and real-world applications using existing technology. Particularly in the context of the fast-evolving sprint domain, a deficiency in evaluation expertise can result in lagging behind. Using deep learning algorithms, useful features, and information can be automatically extracted from extensive training and competition data, providing a more accurate and objective basis for evaluating sprint sports techniques and tactics. We can reduce manual intervention and tedious workloads by applying deep learning algorithms in automated processing and analysis, enabling intelligent training and competition. Deep learning algorithms can uncover and analyze patterns and trends in data, assisting coaches and athletes in identifying potential problems and areas for improvement and providing a foundation for athletes to develop targeted training plans and competition strategies. Therefore, studying the role of Deep Learning (DL) in the technological process is of great significance. This paper aims to effectively combine DL algorithms with sprint skill evaluation, utilizing DL algorithms for mining and leveraging them to achieve optimal sprint skill performance through effective tactical evaluation. This study trains a new model based on the classic VGGNet-16 model and compares different CNN models. The experiment reveals that the CaffeNet model is more susceptible to objects that do not entirely contain the target, resulting in an enlarging search box and eventually losing track of the target. The experimental results demonstrate that maintaining a step frequency of approximately 95% is crucial for achieving the highest speed. The SDP algorithm was tested on the VOC2007 dataset and achieved good results, with a mean average precision value of 69.4%, significantly surpassing the benchmark Fast R-CNN algorithm. Additionally, we have implemented a traditional tracking algorithm, the CamShift algorithm, based on color features.","doi":"10.1007\/s11761-024-00411-0","cleaned_title":"effectiveness evaluation of sprint sports techniques and tactics based on deep learning","cleaned_abstract":"the assessment of sprint-like sports techniques and tactics involves evaluating their underlying running principles, structures, functions, methods, and real-world applications using existing technology. particularly in the context of the fast-evolving sprint domain, a deficiency in evaluation expertise can result in lagging behind. using deep learning algorithms, useful features, and information can be automatically extracted from extensive training and competition data, providing a more accurate and objective basis for evaluating sprint sports techniques and tactics. we can reduce manual intervention and tedious workloads by applying deep learning algorithms in automated processing and analysis, enabling intelligent training and competition. deep learning algorithms can uncover and analyze patterns and trends in data, assisting coaches and athletes in identifying potential problems and areas for improvement and providing a foundation for athletes to develop targeted training plans and competition strategies. therefore, studying the role of deep learning (dl) in the technological process is of great significance. this paper aims to effectively combine dl algorithms with sprint skill evaluation, utilizing dl algorithms for mining and leveraging them to achieve optimal sprint skill performance through effective tactical evaluation. this study trains a new model based on the classic vggnet-16 model and compares different cnn models. the experiment reveals that the caffenet model is more susceptible to objects that do not entirely contain the target, resulting in an enlarging search box and eventually losing track of the target. the experimental results demonstrate that maintaining a step frequency of approximately 95% is crucial for achieving the highest speed. the sdp algorithm was tested on the voc2007 dataset and achieved good results, with a mean average precision value of 69.4%, significantly surpassing the benchmark fast r-cnn algorithm. additionally, we have implemented a traditional tracking algorithm, the camshift algorithm, based on color features.","key_phrases":["the assessment","sprint-like sports techniques","tactics","their underlying running principles","structures","functions","methods","real-world applications","existing technology","the context","the fast-evolving sprint domain","a deficiency","evaluation expertise","deep learning algorithms","useful features","information","extensive training","competition data","a more accurate and objective basis","sprint sports techniques","tactics","we","manual intervention","tedious workloads","deep learning algorithms","automated processing","analysis","intelligent training","competition","deep learning algorithms","patterns","trends","data","coaches","athletes","potential problems","areas","improvement","a foundation","athletes","targeted training plans","competition strategies","the role","deep learning","dl","the technological process","great significance","this paper","dl algorithms","sprint skill evaluation","dl algorithms","them","optimal sprint skill performance","effective tactical evaluation","this study","a new model","the classic vggnet-16 model","different cnn models","the experiment","the caffenet model","objects","that","the target","an enlarging search box","track","the target","the experimental results","a step frequency","approximately 95%","the highest speed","the sdp algorithm","the voc2007 dataset","good results","a mean average precision value","69.4%","the benchmark fast r-cnn algorithm","we","a traditional tracking algorithm","the camshift algorithm","color features","vggnet-16","cnn","approximately 95%","69.4%"]},{"title":"Image super-resolution reconstruction based on deep dictionary learning and A+","authors":["Yi Huang","Weixin Bian","Biao Jie","Zhiqiang Zhu","Wenhu Li"],"abstract":"The method of image super-resolution reconstruction through the dictionary usually only uses a single-layer dictionary, which not only cannot extract the deep features of the image but also requires a large trained dictionary if the reconstruction effect is to be better. This paper proposes a new deep dictionary learning model. Firstly, after preprocessing the images of the training set, the dictionary is trained by the deep dictionary learning method, and the adjusted anchored neighborhood regression method is used for image super-resolution reconstruction. The proposed algorithm is compared with several classical algorithms on Set5 dataset and Set14 dataset. The visualization and quantification results show that the proposed method improves PSNR and SSIM, effectively reduces the dictionary size and saves reconstruction time compared with traditional super-resolution algorithms.","doi":"10.1007\/s11760-023-02936-x","cleaned_title":"image super-resolution reconstruction based on deep dictionary learning and a+","cleaned_abstract":"the method of image super-resolution reconstruction through the dictionary usually only uses a single-layer dictionary, which not only cannot extract the deep features of the image but also requires a large trained dictionary if the reconstruction effect is to be better. this paper proposes a new deep dictionary learning model. firstly, after preprocessing the images of the training set, the dictionary is trained by the deep dictionary learning method, and the adjusted anchored neighborhood regression method is used for image super-resolution reconstruction. the proposed algorithm is compared with several classical algorithms on set5 dataset and set14 dataset. the visualization and quantification results show that the proposed method improves psnr and ssim, effectively reduces the dictionary size and saves reconstruction time compared with traditional super-resolution algorithms.","key_phrases":["the method","image super-resolution reconstruction","the dictionary","a single-layer dictionary","which","the deep features","the image","a large trained dictionary","the reconstruction effect","this paper","a new deep dictionary learning model","the images","the training set","the dictionary","the deep dictionary learning method","the adjusted anchored neighborhood regression method","image super-resolution reconstruction","the proposed algorithm","several classical algorithms","set5 dataset","the visualization and quantification results","the proposed method","psnr","ssim","the dictionary size","reconstruction time","traditional super-resolution algorithms","firstly","set5"]},{"title":"Efficient deep learning-based approach for malaria detection using red blood cell smears","authors":["Muhammad Mujahid","Furqan Rustam","Rahman Shafique","Elizabeth Caro Montero","Eduardo Silva Alvarado","Isabel de la Torre Diez","Imran Ashraf"],"abstract":"Malaria is an extremely malignant disease and is caused by the bites of infected female mosquitoes. This disease is not only infectious among humans, but among animals as well. Malaria causes mild symptoms like fever, headache, sweating and vomiting, and muscle discomfort; severe symptoms include coma, seizures, and kidney failure. The timely identification of malaria parasites is a challenging and chaotic endeavor for health staff. An expert technician examines the schematic blood smears of infected red blood cells through a microscope. The conventional methods for identifying malaria are not efficient. Machine learning approaches are effective for simple classification challenges but not for complex tasks. Furthermore, machine learning involves rigorous feature engineering to train the model and detect patterns in the features. On the other hand, deep learning works well with complex tasks and automatically extracts low and high-level features from the images to detect disease. In this paper, EfficientNet, a deep learning-based approach for detecting Malaria, is proposed that uses red blood cell images. Experiments are carried out and performance comparison is made with pre-trained deep learning models. In addition, k-fold cross-validation is also used to substantiate the results of the proposed approach. Experiments show that the proposed approach is 97.57% accurate in detecting Malaria from red blood cell images and can be beneficial practically for medical healthcare staff.","doi":"10.1038\/s41598-024-63831-0","cleaned_title":"efficient deep learning-based approach for malaria detection using red blood cell smears","cleaned_abstract":"malaria is an extremely malignant disease and is caused by the bites of infected female mosquitoes. this disease is not only infectious among humans, but among animals as well. malaria causes mild symptoms like fever, headache, sweating and vomiting, and muscle discomfort; severe symptoms include coma, seizures, and kidney failure. the timely identification of malaria parasites is a challenging and chaotic endeavor for health staff. an expert technician examines the schematic blood smears of infected red blood cells through a microscope. the conventional methods for identifying malaria are not efficient. machine learning approaches are effective for simple classification challenges but not for complex tasks. furthermore, machine learning involves rigorous feature engineering to train the model and detect patterns in the features. on the other hand, deep learning works well with complex tasks and automatically extracts low and high-level features from the images to detect disease. in this paper, efficientnet, a deep learning-based approach for detecting malaria, is proposed that uses red blood cell images. experiments are carried out and performance comparison is made with pre-trained deep learning models. in addition, k-fold cross-validation is also used to substantiate the results of the proposed approach. experiments show that the proposed approach is 97.57% accurate in detecting malaria from red blood cell images and can be beneficial practically for medical healthcare staff.","key_phrases":["malaria","an extremely malignant disease","the bites","infected female mosquitoes","this disease","humans","animals","malaria","mild symptoms","fever","headache","sweating","vomiting","muscle discomfort","severe symptoms","coma","seizures","kidney failure","the timely identification","malaria parasites","a challenging and chaotic endeavor","health staff","an expert technician","the schematic blood smears","infected red blood cells","a microscope","the conventional methods","malaria","machine learning approaches","simple classification challenges","complex tasks","machine learning","rigorous feature engineering","the model","patterns","the features","the other hand","deep learning","complex tasks","low and high-level features","the images","disease","this paper","efficientnet","a deep learning-based approach","malaria","that","red blood cell images","experiments","performance comparison","pre-trained deep learning models","addition","cross","-","validation","the results","the proposed approach","experiments","the proposed approach","97.57%","malaria","red blood cell images","medical healthcare staff","97.57%"]},{"title":"Optical electrocardiogram based heart disease prediction using hybrid deep learning","authors":["Avinash L. Golande","T. Pavankumar"],"abstract":"The diagnosis and categorization of cardiac disease using the low-cost tool electrocardiogram (ECG) becomes an intriguing study topic when contemplating intelligent healthcare applications. An ECG-based cardiac disease prediction system must be automated, accurate, and lightweight. The deep learning methods recently achieved automation and accuracy across multiple domains. However, applying deep learning for automatic ECG-based heart disease classification is a challenging research problem. Because using solely deep learning approaches failed to detect all of the important beats from the input ECG signal, a hybrid strategy is necessary to improve detection efficiency. The main objective of the proposed model is to enhance the ECG-based heart disease classification efficiency using a hybrid feature engineering approach. The proposed model consists of pre-processing, hybrid feature engineering, and classification. Pre-processing an ECG aims to eliminate powerline and baseline interference without disrupting the heartbeat. To efficiently classify data, we design a hybrid approach using a conventional ECG beats extraction algorithm and Convolutional Neural Network (CNN)-based features. For heart disease prediction, the hybrid feature vector is fed successively into the deep learning classifier Long Term Short Memory (LSTM). The results of the simulations show that the proposed model reduces both the number of diagnostic errors and the amount of time spent on each one when compared to the existing methods.","doi":"10.1186\/s40537-023-00820-6","cleaned_title":"optical electrocardiogram based heart disease prediction using hybrid deep learning","cleaned_abstract":"the diagnosis and categorization of cardiac disease using the low-cost tool electrocardiogram (ecg) becomes an intriguing study topic when contemplating intelligent healthcare applications. an ecg-based cardiac disease prediction system must be automated, accurate, and lightweight. the deep learning methods recently achieved automation and accuracy across multiple domains. however, applying deep learning for automatic ecg-based heart disease classification is a challenging research problem. because using solely deep learning approaches failed to detect all of the important beats from the input ecg signal, a hybrid strategy is necessary to improve detection efficiency. the main objective of the proposed model is to enhance the ecg-based heart disease classification efficiency using a hybrid feature engineering approach. the proposed model consists of pre-processing, hybrid feature engineering, and classification. pre-processing an ecg aims to eliminate powerline and baseline interference without disrupting the heartbeat. to efficiently classify data, we design a hybrid approach using a conventional ecg beats extraction algorithm and convolutional neural network (cnn)-based features. for heart disease prediction, the hybrid feature vector is fed successively into the deep learning classifier long term short memory (lstm). the results of the simulations show that the proposed model reduces both the number of diagnostic errors and the amount of time spent on each one when compared to the existing methods.","key_phrases":["the diagnosis","categorization","cardiac disease","the low-cost tool electrocardiogram","ecg","an intriguing study topic","intelligent healthcare applications","an ecg-based cardiac disease prediction system","the deep learning methods","automation","accuracy","multiple domains","deep learning","automatic ecg-based heart disease classification","a challenging research problem","solely deep learning approaches","all","the important beats","the input ecg signal","a hybrid strategy","detection efficiency","the main objective","the proposed model","the ecg-based heart disease classification efficiency","a hybrid feature engineering approach","the proposed model","pre-processing, hybrid feature engineering","classification","an ecg","powerline","baseline interference","the heartbeat","data","we","a hybrid approach","a conventional ecg beats extraction algorithm","convolutional neural network","cnn)-based features","heart disease prediction","the hybrid feature vector","the deep learning classifier long term short memory","lstm","the results","the simulations","the proposed model","both the number","diagnostic errors","the amount","time","each one","the existing methods","fed"]},{"title":"Deep learning-based automated assessment of canine hip dysplasia","authors":["C\u00e1tia Loureiro","Lio Gon\u00e7alves","Pedro Leite","Pedro Franco-Gon\u00e7alo","Ana In\u00eas Pereira","Bruno Cola\u00e7o","Sofia Alves-Pimenta","Fintan McEvoy","M\u00e1rio Ginja","V\u00edtor Filipe"],"abstract":"Radiographic canine hip dysplasia (CHD) diagnosis is crucial for breeding selection and disease management, delaying progression and alleviating the associated pain. Radiography is the primary imaging modality for CHD diagnosis, and visual assessment of radiographic features is sometimes used for accurate diagnosis. Specifically, alterations in femoral neck shape are crucial radiographic signs, with existing literature suggesting that dysplastic hips have a greater femoral neck thickness (FNT). In this study we aimed to develop a three-stage deep learning-based system that can automatically identify and quantify a femoral neck thickness index (FNTi) as a key metric to improve CHD diagnosis. Our system trained a keypoint detection model and a segmentation model to determine landmark and boundary coordinates of the femur and acetabulum, respectively. We then executed a series of mathematical operations to calculate the FNTi. The keypoint detection model achieved a mean absolute error (MAE) of 0.013 during training, while the femur segmentation results achieved a dice score (DS) of 0.978. Our three-stage deep learning-based system achieved an intraclass correlation coefficient of 0.86 (95% confidence interval) and showed no significant differences in paired t-test compared to a specialist (p > 0.05). As far as we know, this is the initial study to thoroughly measure FNTi by applying computer vision and deep learning-based approaches, which can provide reliable support in CHD diagnosis.","doi":"10.1007\/s11042-024-20072-7","cleaned_title":"deep learning-based automated assessment of canine hip dysplasia","cleaned_abstract":"radiographic canine hip dysplasia (chd) diagnosis is crucial for breeding selection and disease management, delaying progression and alleviating the associated pain. radiography is the primary imaging modality for chd diagnosis, and visual assessment of radiographic features is sometimes used for accurate diagnosis. specifically, alterations in femoral neck shape are crucial radiographic signs, with existing literature suggesting that dysplastic hips have a greater femoral neck thickness (fnt). in this study we aimed to develop a three-stage deep learning-based system that can automatically identify and quantify a femoral neck thickness index (fnti) as a key metric to improve chd diagnosis. our system trained a keypoint detection model and a segmentation model to determine landmark and boundary coordinates of the femur and acetabulum, respectively. we then executed a series of mathematical operations to calculate the fnti. the keypoint detection model achieved a mean absolute error (mae) of 0.013 during training, while the femur segmentation results achieved a dice score (ds) of 0.978. our three-stage deep learning-based system achieved an intraclass correlation coefficient of 0.86 (95% confidence interval) and showed no significant differences in paired t-test compared to a specialist (p > 0.05). as far as we know, this is the initial study to thoroughly measure fnti by applying computer vision and deep learning-based approaches, which can provide reliable support in chd diagnosis.","key_phrases":["radiographic canine hip dysplasia (chd) diagnosis","selection","disease management","progression","the associated pain","radiography","the primary imaging modality","chd diagnosis","visual assessment","radiographic features","accurate diagnosis","alterations","femoral neck shape","crucial radiographic signs","existing literature","dysplastic hips","a greater femoral neck thickness","fnt","this study","we","a three-stage deep learning-based system","that","a femoral neck","thickness index","a key metric","chd diagnosis","our system","a keypoint detection model","a segmentation model","landmark","boundary coordinates","the femur","acetabulum","we","a series","mathematical operations","the keypoint detection model","a mean absolute error","mae","training","the femur segmentation results","a dice score","ds","our three-stage deep learning-based system","an intraclass correlation","coefficient","0.86 (95% confidence interval","no significant differences","paired t-test","a specialist","(p","we","this","the initial study","computer vision","deep learning-based approaches","which","reliable support","chd diagnosis","three","0.013","0.978","three","0.86","95%","0.05"]},{"title":"An Explainable Deep Learning Approach for Oral Cancer Detection","authors":["P. Ashok Babu","Anjani Kumar Rai","Janjhyam Venkata Naga Ramesh","A. Nithyasri","S. Sangeetha","Pravin R. Kshirsagar","A. Rajendran","A. Rajaram","S. Dilipkumar"],"abstract":"With a high death rate, oral cancer is a major worldwide health problem, particularly in low- and middle-income nations. Timely detection and diagnosis are crucial for effective prevention and treatment. To address this challenge, there is a growing need for automated detection systems to aid healthcare professionals. Regular dental examinations play a vital role in early detection. Transfer learning, which leverages knowledge from related domains, can enhance performance in target categories. This study presents a unique approach to the early detection and diagnosis of oral cancer that makes use of the exceptional sensory capabilities of the mouth. Deep neural networks, particularly those based on automated systems, are employed to identify intricate patterns associated with the disease. By combining various transfer learning approaches and conducting comparative analyses, an optimal learning rate is achieved. The categorization analysis of the reference results is presented in detail. Our preliminary findings demonstrate that deep learning effectively addresses this challenging problem, with the Inception-V3 algorithm exhibiting superior accuracy compared to other algorithms.","doi":"10.1007\/s42835-023-01654-1","cleaned_title":"an explainable deep learning approach for oral cancer detection","cleaned_abstract":"with a high death rate, oral cancer is a major worldwide health problem, particularly in low- and middle-income nations. timely detection and diagnosis are crucial for effective prevention and treatment. to address this challenge, there is a growing need for automated detection systems to aid healthcare professionals. regular dental examinations play a vital role in early detection. transfer learning, which leverages knowledge from related domains, can enhance performance in target categories. this study presents a unique approach to the early detection and diagnosis of oral cancer that makes use of the exceptional sensory capabilities of the mouth. deep neural networks, particularly those based on automated systems, are employed to identify intricate patterns associated with the disease. by combining various transfer learning approaches and conducting comparative analyses, an optimal learning rate is achieved. the categorization analysis of the reference results is presented in detail. our preliminary findings demonstrate that deep learning effectively addresses this challenging problem, with the inception-v3 algorithm exhibiting superior accuracy compared to other algorithms.","key_phrases":["a high death rate","oral cancer","a major worldwide health problem","low- and middle-income nations","timely detection","diagnosis","effective prevention","treatment","this challenge","a growing need","automated detection systems","healthcare professionals","regular dental examinations","a vital role","early detection","transfer learning","which","knowledge","related domains","performance","target categories","this study","a unique approach","the early detection","diagnosis","oral cancer","that","use","the exceptional sensory capabilities","the mouth","deep neural networks","particularly those","automated systems","intricate patterns","the disease","various transfer","approaches","comparative analyses","an optimal learning rate","the categorization analysis","the reference results","detail","our preliminary findings","this challenging problem","the inception-v3 algorithm","superior accuracy","other algorithms"]},{"title":"Semantic speech analysis using machine learning and deep learning techniques: a comprehensive review","authors":["Suryakant Tyagi","S\u00e1ndor Sz\u00e9n\u00e1si"],"abstract":"Human cognitive functions such as perception, attention, learning, memory, reasoning, and problem-solving are all significantly influenced by emotion. Emotion has a particularly potent impact on attention, modifying its selectivity in particular and influencing behavior and action motivation. Artificial Emotional Intelligence (AEI) technologies enable computers to understand a user's emotional state and respond appropriately. These systems enable a realistic dialogue between people and machines. The current generation of adaptive user interference technologies is built on techniques from data analytics and machine learning (ML), namely deep learning (DL) artificial neural networks (ANN) from multimodal data, such as videos of facial expressions, stance, and gesture, voice, and bio-physiological data (such as eye movement, ECG, respiration, EEG, FMRT, EMG, eye tracking). In this study, we reviewed existing literature based on ML and data analytics techniques being used to detect emotions in speech. The efficacy of data analytics and ML techniques in this unique area of multimodal data processing and extracting emotions from speech. This study analyzes how emotional chatbots, facial expressions, images, and social media texts can be effective in detecting emotions. PRISMA methodology is used to review the existing survey. Support Vector Machines (SVM), Na\u00efve Bayes (NB), Random Forests (RF), Recurrent Neural Networks (RNN), Logistic Regression (LR), etc., are commonly used ML techniques for emotion extraction purposes. This study provides a new taxonomy about the application of ML in SER. The result shows that Long-Short Term Memory (LSTM) and Convolutional Neural Networks (CNN) are found to be the most useful methodology for this purpose.","doi":"10.1007\/s11042-023-17769-6","cleaned_title":"semantic speech analysis using machine learning and deep learning techniques: a comprehensive review","cleaned_abstract":"human cognitive functions such as perception, attention, learning, memory, reasoning, and problem-solving are all significantly influenced by emotion. emotion has a particularly potent impact on attention, modifying its selectivity in particular and influencing behavior and action motivation. artificial emotional intelligence (aei) technologies enable computers to understand a user's emotional state and respond appropriately. these systems enable a realistic dialogue between people and machines. the current generation of adaptive user interference technologies is built on techniques from data analytics and machine learning (ml), namely deep learning (dl) artificial neural networks (ann) from multimodal data, such as videos of facial expressions, stance, and gesture, voice, and bio-physiological data (such as eye movement, ecg, respiration, eeg, fmrt, emg, eye tracking). in this study, we reviewed existing literature based on ml and data analytics techniques being used to detect emotions in speech. the efficacy of data analytics and ml techniques in this unique area of multimodal data processing and extracting emotions from speech. this study analyzes how emotional chatbots, facial expressions, images, and social media texts can be effective in detecting emotions. prisma methodology is used to review the existing survey. support vector machines (svm), na\u00efve bayes (nb), random forests (rf), recurrent neural networks (rnn), logistic regression (lr), etc., are commonly used ml techniques for emotion extraction purposes. this study provides a new taxonomy about the application of ml in ser. the result shows that long-short term memory (lstm) and convolutional neural networks (cnn) are found to be the most useful methodology for this purpose.","key_phrases":["human cognitive functions","perception","attention","learning","memory","reasoning","problem-solving","emotion","emotion","a particularly potent impact","attention","its selectivity","behavior","action motivation","artificial emotional intelligence (aei) technologies","computers","a user's emotional state","these systems","a realistic dialogue","people","machines","the current generation","adaptive user","interference technologies","techniques","data analytics","machine learning","ml","namely deep learning","(dl) artificial neural networks","ann","multimodal data","videos","facial expressions","stance","gesture","voice","bio-physiological data","eye movement","ecg","respiration","fmrt","emg","eye tracking","this study","we","existing literature","ml and data analytics techniques","emotions","speech","the efficacy","data analytics","ml","techniques","this unique area","multimodal data processing","emotions","speech","this study","emotional chatbots","facial expressions","images","social media texts","emotions","prisma methodology","the existing survey","support vector machines","svm","na\u00efve bayes","random forests","neural networks","rnn","logistic regression","lr","ml techniques","emotion extraction purposes","this study","a new taxonomy","the application","ml","ser","the result","long-short term memory","lstm","convolutional neural networks","cnn","the most useful methodology","this purpose","na\u00efve bayes (nb","cnn"]},{"title":"An Explainable Deep Learning Approach for Oral Cancer Detection","authors":["P. Ashok Babu","Anjani Kumar Rai","Janjhyam Venkata Naga Ramesh","A. Nithyasri","S. Sangeetha","Pravin R. Kshirsagar","A. Rajendran","A. Rajaram","S. Dilipkumar"],"abstract":"With a high death rate, oral cancer is a major worldwide health problem, particularly in low- and middle-income nations. Timely detection and diagnosis are crucial for effective prevention and treatment. To address this challenge, there is a growing need for automated detection systems to aid healthcare professionals. Regular dental examinations play a vital role in early detection. Transfer learning, which leverages knowledge from related domains, can enhance performance in target categories. This study presents a unique approach to the early detection and diagnosis of oral cancer that makes use of the exceptional sensory capabilities of the mouth. Deep neural networks, particularly those based on automated systems, are employed to identify intricate patterns associated with the disease. By combining various transfer learning approaches and conducting comparative analyses, an optimal learning rate is achieved. The categorization analysis of the reference results is presented in detail. Our preliminary findings demonstrate that deep learning effectively addresses this challenging problem, with the Inception-V3 algorithm exhibiting superior accuracy compared to other algorithms.","doi":"10.1007\/s42835-023-01654-1","cleaned_title":"an explainable deep learning approach for oral cancer detection","cleaned_abstract":"with a high death rate, oral cancer is a major worldwide health problem, particularly in low- and middle-income nations. timely detection and diagnosis are crucial for effective prevention and treatment. to address this challenge, there is a growing need for automated detection systems to aid healthcare professionals. regular dental examinations play a vital role in early detection. transfer learning, which leverages knowledge from related domains, can enhance performance in target categories. this study presents a unique approach to the early detection and diagnosis of oral cancer that makes use of the exceptional sensory capabilities of the mouth. deep neural networks, particularly those based on automated systems, are employed to identify intricate patterns associated with the disease. by combining various transfer learning approaches and conducting comparative analyses, an optimal learning rate is achieved. the categorization analysis of the reference results is presented in detail. our preliminary findings demonstrate that deep learning effectively addresses this challenging problem, with the inception-v3 algorithm exhibiting superior accuracy compared to other algorithms.","key_phrases":["a high death rate","oral cancer","a major worldwide health problem","low- and middle-income nations","timely detection","diagnosis","effective prevention","treatment","this challenge","a growing need","automated detection systems","healthcare professionals","regular dental examinations","a vital role","early detection","transfer learning","which","knowledge","related domains","performance","target categories","this study","a unique approach","the early detection","diagnosis","oral cancer","that","use","the exceptional sensory capabilities","the mouth","deep neural networks","particularly those","automated systems","intricate patterns","the disease","various transfer","approaches","comparative analyses","an optimal learning rate","the categorization analysis","the reference results","detail","our preliminary findings","this challenging problem","the inception-v3 algorithm","superior accuracy","other algorithms"]},{"title":"Deep Learning Based Video Compression Techniques with Future Research Issues","authors":["Helen K. Joy","Manjunath R. Kounte","Arunkumar Chandrasekhar","Manoranjan Paul"],"abstract":"The advancements in the domain of video coding technologies are tremendously fluctuating in recent years. As the public got acquainted with the creation and availability of videos through internet boom and video acquisition devices including mobile phones, camera etc., the necessity of video compression become crucial. The resolution variance (4\u00a0K, 2\u00a0K etc.), framerate, display is some of the features that glorifies the importance of compression. Improving compression ratio with better efficiency and quality was the focus and it has many stumbling blocks to achieve it. The era of artificial intelligence, neural network, and especially deep learning provided light in the path of video processing area, particularly in compression. The paper mainly focuses on a precise, organized, meticulous review of the impact of deep learning on video compression. The content adaptivity quality of deep learning marks its importance in video compression to traditional signal processing. The development of intelligent and self-trained steps in video compression with deep learning is reviewed in detail. The relevant and noteworthy work that arose in each step of compression is inculcated in this paper. A detailed survey in the development of intra- prediction, inter-prediction, in-loop filtering, quantization, and entropy coding in hand with deep learning techniques are pointed along with envisages ideas in each field. The future scope of enhancement in various stages of compression and relevant research scope to explore with Deep Learning is emphasized.","doi":"10.1007\/s11277-023-10558-2","cleaned_title":"deep learning based video compression techniques with future research issues","cleaned_abstract":"the advancements in the domain of video coding technologies are tremendously fluctuating in recent years. as the public got acquainted with the creation and availability of videos through internet boom and video acquisition devices including mobile phones, camera etc., the necessity of video compression become crucial. the resolution variance (4 k, 2 k etc.), framerate, display is some of the features that glorifies the importance of compression. improving compression ratio with better efficiency and quality was the focus and it has many stumbling blocks to achieve it. the era of artificial intelligence, neural network, and especially deep learning provided light in the path of video processing area, particularly in compression. the paper mainly focuses on a precise, organized, meticulous review of the impact of deep learning on video compression. the content adaptivity quality of deep learning marks its importance in video compression to traditional signal processing. the development of intelligent and self-trained steps in video compression with deep learning is reviewed in detail. the relevant and noteworthy work that arose in each step of compression is inculcated in this paper. a detailed survey in the development of intra- prediction, inter-prediction, in-loop filtering, quantization, and entropy coding in hand with deep learning techniques are pointed along with envisages ideas in each field. the future scope of enhancement in various stages of compression and relevant research scope to explore with deep learning is emphasized.","key_phrases":["the advancements","the domain","video coding technologies","recent years","the public","the creation","availability","videos","internet boom and video acquisition devices","mobile phones","the necessity","video compression","the resolution variance","4 k","framerate, display","some","the features","that","the importance","compression","compression ratio","better efficiency","quality","the focus","it","many stumbling blocks","it","the era","artificial intelligence","neural network","especially deep learning","light","the path","video processing area","compression","the paper","a precise, organized, meticulous review","the impact","deep learning","video compression","the content adaptivity quality","deep learning","its importance","video compression","traditional signal processing","the development","intelligent and self-trained steps","video compression","deep learning","detail","the relevant and noteworthy work","that","each step","compression","this paper","a detailed survey","the development","intra- prediction","inter","-","loop","quantization","hand","deep learning techniques","envisages ideas","each field","the future scope","enhancement","various stages","compression","relevant research scope","deep learning","recent years","4","2"]},{"title":"Dermatological disease prediction and diagnosis system using deep learning","authors":["Neda Fatima","Syed Afzal Murtaza Rizvi","Major Syed Bilal Abbas Rizvi"],"abstract":"The prevalence of skin illnesses is higher than that of other diseases. Fungal infection, bacteria, allergies, viruses, genetic factors, and environmental factors are among important causative factors that have continuously escalated the degree and incidence of skin diseases. Medical technology based on lasers and photonics has made it possible to identify skin illnesses considerably more rapidly and correctly. However, the cost of such a diagnosis is currently limited and prohibitively high and restricted to developed areas. The present paper develops a holistic, critical, and important skin disease prediction system that utilizes machine learning and deep learning algorithms to accurately identify up to 20 different skin diseases with a high F1 score and efficiency. Deep learning algorithms like Xception, Inception-v3, Resnet50, DenseNet121, and Inception-ResNet-v2 were employed to accurately classify diseases based on the images. The training and testing have been performed on an enlarged dataset, and classification was performed for 20 diseases. The algorithm developed was free from any inherent bias and treated all classes equally. The present model, which was trained using the Xception algorithm, is highly efficient and accurate for 20 different skin conditions, with a dataset of over 10,000 photos. The developed system was able to classify 20 different dermatological diseases with high accuracy and precision.","doi":"10.1007\/s11845-023-03578-1","cleaned_title":"dermatological disease prediction and diagnosis system using deep learning","cleaned_abstract":"the prevalence of skin illnesses is higher than that of other diseases. fungal infection, bacteria, allergies, viruses, genetic factors, and environmental factors are among important causative factors that have continuously escalated the degree and incidence of skin diseases. medical technology based on lasers and photonics has made it possible to identify skin illnesses considerably more rapidly and correctly. however, the cost of such a diagnosis is currently limited and prohibitively high and restricted to developed areas. the present paper develops a holistic, critical, and important skin disease prediction system that utilizes machine learning and deep learning algorithms to accurately identify up to 20 different skin diseases with a high f1 score and efficiency. deep learning algorithms like xception, inception-v3, resnet50, densenet121, and inception-resnet-v2 were employed to accurately classify diseases based on the images. the training and testing have been performed on an enlarged dataset, and classification was performed for 20 diseases. the algorithm developed was free from any inherent bias and treated all classes equally. the present model, which was trained using the xception algorithm, is highly efficient and accurate for 20 different skin conditions, with a dataset of over 10,000 photos. the developed system was able to classify 20 different dermatological diseases with high accuracy and precision.","key_phrases":["the prevalence","skin illnesses","that","other diseases","fungal infection","bacteria","allergies","viruses","genetic factors","environmental factors","important causative factors","that","the degree","incidence","skin diseases","medical technology","lasers","photonics","it","skin illnesses","the cost","such a diagnosis","developed areas","the present paper","a holistic, critical, and important skin disease prediction system","that","machine learning","deep learning algorithms","up to 20 different skin diseases","a high f1 score","efficiency","algorithms","xception","inception-v3","resnet50","densenet121","inception-resnet-v2","diseases","the images","the training","testing","an enlarged dataset","classification","20 diseases","the algorithm","any inherent bias","all classes","the present model","which","the xception","algorithm","20 different skin conditions","a dataset","over 10,000 photos","the developed system","20 different dermatological diseases","high accuracy","precision","up to 20","resnet50","20","20","over 10,000","20"]},{"title":"Deep SqueezeNet learning model for diagnosis and prediction of maize leaf diseases","authors":["Prasannavenkatesan Theerthagiri","A. Usha Ruby","J. George Chellin Chandran","Tanvir Habib Sardar","Ahamed Shafeeq B. M."],"abstract":"The maize leaf diseases create severe yield reductions and critical problems. The maize leaf disease should be discovered early, perfectly identified, and precisely diagnosed to make greater yield. This work studies three main leaf diseases: common rust, blight, and grey leaf spot. This approach involves pre-processing, including sampling and labelling, while ensuring class balance and preventing overfitting via the SMOTE algorithm. The maize leaf dataset with augmentation was used to classify these diseases using several deep-learning pre-trained networks, including VGG16, Resnet34, Resnet50, and SqueezeNet. The model was evaluated using a maize leaf dataset that included various leaf classes, mini-batch sizes, and input sizes. Performance measures, recall, precision, accuracy, F1-score, and confusion matrix were computed for each network. The SqueezeNet learning model produces an accuracy of 97% in classifying four different classes of plant leaf datasets. Comparatively, the SqueezeNet learning model has improved accuracy by 2\u20135% and reduced the mean square error by 4\u201311% over VGG16, Resnet34, and Resnet50 deep learning models.","doi":"10.1186\/s40537-024-00972-z","cleaned_title":"deep squeezenet learning model for diagnosis and prediction of maize leaf diseases","cleaned_abstract":"the maize leaf diseases create severe yield reductions and critical problems. the maize leaf disease should be discovered early, perfectly identified, and precisely diagnosed to make greater yield. this work studies three main leaf diseases: common rust, blight, and grey leaf spot. this approach involves pre-processing, including sampling and labelling, while ensuring class balance and preventing overfitting via the smote algorithm. the maize leaf dataset with augmentation was used to classify these diseases using several deep-learning pre-trained networks, including vgg16, resnet34, resnet50, and squeezenet. the model was evaluated using a maize leaf dataset that included various leaf classes, mini-batch sizes, and input sizes. performance measures, recall, precision, accuracy, f1-score, and confusion matrix were computed for each network. the squeezenet learning model produces an accuracy of 97% in classifying four different classes of plant leaf datasets. comparatively, the squeezenet learning model has improved accuracy by 2\u20135% and reduced the mean square error by 4\u201311% over vgg16, resnet34, and resnet50 deep learning models.","key_phrases":["the maize leaf diseases","severe yield reductions","critical problems","the maize leaf disease","greater yield","this work","three main leaf diseases","common rust","blight","grey leaf spot","this approach","labelling","class balance","the smote algorithm","the maize leaf dataset","augmentation","these diseases","several deep-learning pre-trained networks","vgg16","resnet34","resnet50","squeezenet","the model","a maize leaf dataset","that","various leaf classes","mini-batch sizes","input sizes","performance measures","recall","precision","accuracy","f1-score","confusion matrix","each network","the squeezenet learning model","an accuracy","97%","four different classes","plant leaf datasets","the squeezenet learning model","accuracy","2\u20135%","the mean square error","4\u201311%","vgg16","resnet34","deep learning models","three","grey leaf spot","resnet34","resnet50","97%","four","2\u20135%","4\u201311%","resnet34","resnet50"]},{"title":"Review and perspective on bioinformatics tools using machine learning and deep learning for predicting antiviral peptides","authors":["Nicol\u00e1s Lefin","Lisandra Herrera-Bel\u00e9n","Jorge G. Farias","Jorge F. Beltr\u00e1n"],"abstract":"Viruses constitute a constant threat to global health and have caused millions of human and animal deaths throughout human history. Despite advances in the discovery of antiviral compounds that help fight these pathogens, finding a solution to this problem continues to be a task that consumes time and financial resources. Currently, artificial intelligence (AI) has revolutionized many areas of the biological sciences, making it possible to decipher patterns in amino acid sequences that encode different functions and activities. Within the field of AI, machine learning, and deep learning algorithms have been used to discover antimicrobial peptides. Due to their effectiveness and specificity, antimicrobial peptides (AMPs) hold excellent promise for treating various infections caused by pathogens. Antiviral peptides (AVPs) are a specific type of AMPs that have activity against certain viruses. Unlike the research focused on the development of tools and methods for the prediction of antimicrobial peptides, those related to the prediction of AVPs are still scarce. Given the significance of AVPs as potential pharmaceutical options for human and animal health and the ongoing AI revolution, we have reviewed and summarized the current machine learning and deep learning-based tools and methods available for predicting these types of peptides.","doi":"10.1007\/s11030-023-10718-3","cleaned_title":"review and perspective on bioinformatics tools using machine learning and deep learning for predicting antiviral peptides","cleaned_abstract":"viruses constitute a constant threat to global health and have caused millions of human and animal deaths throughout human history. despite advances in the discovery of antiviral compounds that help fight these pathogens, finding a solution to this problem continues to be a task that consumes time and financial resources. currently, artificial intelligence (ai) has revolutionized many areas of the biological sciences, making it possible to decipher patterns in amino acid sequences that encode different functions and activities. within the field of ai, machine learning, and deep learning algorithms have been used to discover antimicrobial peptides. due to their effectiveness and specificity, antimicrobial peptides (amps) hold excellent promise for treating various infections caused by pathogens. antiviral peptides (avps) are a specific type of amps that have activity against certain viruses. unlike the research focused on the development of tools and methods for the prediction of antimicrobial peptides, those related to the prediction of avps are still scarce. given the significance of avps as potential pharmaceutical options for human and animal health and the ongoing ai revolution, we have reviewed and summarized the current machine learning and deep learning-based tools and methods available for predicting these types of peptides.","key_phrases":["viruses","a constant threat","global health","millions","human and animal deaths","human history","advances","the discovery","antiviral compounds","that","these pathogens","a solution","this problem","a task","that","time","financial resources","artificial intelligence","ai","many areas","the biological sciences","it","patterns","amino acid sequences","that","different functions","activities","the field","ai",", machine learning","deep learning algorithms","antimicrobial peptides","their effectiveness","specificity","antimicrobial peptides","amps","excellent promise","various infections","pathogens","antiviral peptides","avps","a specific type","amps","that","activity","certain viruses","the research","the development","tools","methods","the prediction","antimicrobial peptides","those","the prediction","avps","the significance","avps","potential pharmaceutical options","human and animal health","the ongoing ai revolution","we","the current machine learning","deep learning-based tools","methods","these types","peptides","millions","avps"]},{"title":"A review of deep learning and Generative Adversarial Networks applications in medical image analysis","authors":["D. N. Sindhura","Radhika M. Pai","Shyamasunder N. Bhat","Manohara M. M. Pai"],"abstract":"Nowadays, computer-aided decision support systems (CADs) for the analysis of images have been a perennial technique in the medical imaging field. In CADs, deep learning algorithms are widely used to perform tasks like classification, identification of patterns, detection, etc. Deep learning models learn feature representations from images rather than handcrafted features. Hence, deep learning models are quickly becoming the state-of-the-art method to achieve good performances in different computer-aided decision-support systems in medical applications. Similarly, deep learning-based generative models called Generative Adversarial Networks (GANs) have recently been developed as a novel method to produce realistic-looking synthetic data. GANs are used in different domains, including medical imaging generation. The common problems, like class imbalance and a small dataset, in healthcare are well addressed by GANs, and it is a leading area of research. Segmentation, reconstruction, detection, denoising, registration, etc. are the important applications of GANs. So in this work, the successes of deep learning methods in segmentation, classification, cell structure and fracture detection, computer-aided identification, and GANs in synthetic medical image generation, segmentation, reconstruction, detection, denoising, and registration in recent times are reviewed. Lately, the review article concludes by raising research directions for DL models and GANs in medical applications.","doi":"10.1007\/s00530-024-01349-1","cleaned_title":"a review of deep learning and generative adversarial networks applications in medical image analysis","cleaned_abstract":"nowadays, computer-aided decision support systems (cads) for the analysis of images have been a perennial technique in the medical imaging field. in cads, deep learning algorithms are widely used to perform tasks like classification, identification of patterns, detection, etc. deep learning models learn feature representations from images rather than handcrafted features. hence, deep learning models are quickly becoming the state-of-the-art method to achieve good performances in different computer-aided decision-support systems in medical applications. similarly, deep learning-based generative models called generative adversarial networks (gans) have recently been developed as a novel method to produce realistic-looking synthetic data. gans are used in different domains, including medical imaging generation. the common problems, like class imbalance and a small dataset, in healthcare are well addressed by gans, and it is a leading area of research. segmentation, reconstruction, detection, denoising, registration, etc. are the important applications of gans. so in this work, the successes of deep learning methods in segmentation, classification, cell structure and fracture detection, computer-aided identification, and gans in synthetic medical image generation, segmentation, reconstruction, detection, denoising, and registration in recent times are reviewed. lately, the review article concludes by raising research directions for dl models and gans in medical applications.","key_phrases":[", computer-aided decision support systems","cads","the analysis","images","a perennial technique","the medical imaging field","cads","deep learning algorithms","tasks","classification","identification","patterns","detection","deep learning models","feature representations","images","handcrafted features","deep learning models","the-art","good performances","different computer-aided decision-support systems","medical applications","deep learning-based generative models","generative adversarial networks","gans","a novel method","realistic-looking synthetic data","gans","different domains","medical imaging generation","the common problems","class imbalance","a small dataset","healthcare","gans","it","a leading area","research","segmentation","reconstruction","detection","denoising","registration","the important applications","gans","this work","the successes","deep learning methods","segmentation","classification","cell structure","fracture detection","computer-aided identification","gans","synthetic medical image generation","segmentation","reconstruction","detection","denoising","registration","recent times","the review article","research directions","dl models","gans","medical applications"]},{"title":"Deep learning-based fishing ground prediction with multiple environmental factors","authors":["Mingyang Xie","Bin Liu","Xinjun Chen"],"abstract":"Improving the accuracy of fishing ground prediction for oceanic economic species has always been one of the most concerning issues in fisheries research. Recent studies have confirmed that deep learning has achieved superior results over traditional methods in the era of big data. However, the deep learning-based fishing ground prediction model with a single environment suffers from the problem that the area of the fishing ground is too large and not concentrated. In this study, we developed a deep learning-based fishing ground prediction model with multiple environmental factors using neon flying squid (Ommastrephes bartramii) in Northwest Pacific Ocean as an example. Based on the modified U-Net model, the approach involves the sea surface temperature, sea surface height, sea surface salinity, and chlorophyll a as inputs, and the center fishing ground as the output. The model is trained with data from July to November in 2002\u20132019, and tested with data of 2020. We considered and compared five temporal scales (3, 6, 10, 15, and 30\u00a0days) and seven multiple environmental factor combinations. By comparing different cases, we found that the optimal temporal scale is 30\u00a0days, and the optimal multiple environmental factor combination contained SST and Chl a. The inclusion of multiple factors in the model greatly improved the concentration of the center fishing ground. The selection of a suitable combination of multiple environmental factors is beneficial to the precise spatial distribution of fishing grounds. This study deepens the understanding of the mechanism of environmental field influence on fishing grounds from the perspective of artificial intelligence and fishery science.","doi":"10.1007\/s42995-024-00222-4","cleaned_title":"deep learning-based fishing ground prediction with multiple environmental factors","cleaned_abstract":"improving the accuracy of fishing ground prediction for oceanic economic species has always been one of the most concerning issues in fisheries research. recent studies have confirmed that deep learning has achieved superior results over traditional methods in the era of big data. however, the deep learning-based fishing ground prediction model with a single environment suffers from the problem that the area of the fishing ground is too large and not concentrated. in this study, we developed a deep learning-based fishing ground prediction model with multiple environmental factors using neon flying squid (ommastrephes bartramii) in northwest pacific ocean as an example. based on the modified u-net model, the approach involves the sea surface temperature, sea surface height, sea surface salinity, and chlorophyll a as inputs, and the center fishing ground as the output. the model is trained with data from july to november in 2002\u20132019, and tested with data of 2020. we considered and compared five temporal scales (3, 6, 10, 15, and 30 days) and seven multiple environmental factor combinations. by comparing different cases, we found that the optimal temporal scale is 30 days, and the optimal multiple environmental factor combination contained sst and chl a. the inclusion of multiple factors in the model greatly improved the concentration of the center fishing ground. the selection of a suitable combination of multiple environmental factors is beneficial to the precise spatial distribution of fishing grounds. this study deepens the understanding of the mechanism of environmental field influence on fishing grounds from the perspective of artificial intelligence and fishery science.","key_phrases":["the accuracy","fishing ground prediction","oceanic economic species","the most concerning issues","fisheries research","recent studies","deep learning","superior results","traditional methods","the era","big data","the deep learning-based fishing ground prediction model","a single environment","the problem","the area","the fishing ground","this study","we","a deep learning-based fishing ground prediction model","multiple environmental factors","neon flying squid","ommastrephes","northwest pacific ocean","an example","the modified u-net model","the approach","the sea surface temperature","sea surface height","sea surface salinity","a","inputs","the center fishing ground","the output","the model","data","july","november","data","we","five temporal scales","30 days","seven multiple environmental factor combinations","different cases","we","the optimal temporal scale","30 days","the optimal multiple environmental factor combination","sst and chl a.","the inclusion","multiple factors","the model","the concentration","the center fishing ground","the selection","a suitable combination","multiple environmental factors","the precise spatial distribution","fishing grounds","this study","the understanding","the mechanism","environmental field influence","fishing grounds","the perspective","artificial intelligence","fishery science","july","november","2002\u20132019","2020","five","3","6","10","15","30 days","seven","30 days","chl a."]},{"title":"DEEPBIN: Deep Learning Based Garbage Classification for Households Using Sustainable Natural Technologies","authors":["Yu Song","Xin He","Xiwang Tang","Bo Yin","Jie Du","Jiali Liu","Zhongbao Zhao","Shigang Geng"],"abstract":"Today, things that are accessible worldwide are upgrading to innovative technology. In this research, an intelligent garbage system will be designed with State-of-the-art methods using deep learning technologies. Garbage is highly produced due to urbanization and the rising population in urban areas. It is essential to manage daily trash from homes and living environments. This research aims to provide an intelligent IoT-based garbage bin system, and classification is done using Deep learning techniques. This smart bin is capable of sensing more varieties of garbage from home. Though there are more technologies successfully implemented with IoT and machine learning, there is still a need for sustainable natural technologies to manage daily waste. The innovative IoT-based garbage system uses various sensors like humidity, temperature, gas, and liquid sensors to identify the garbage condition. Initially, the Smart Garbage Bin system is designed, and then the data are collected using a garbage annotation application. Next, the deep learning method is used for object detection and classification of garbage images. Arithmetic Optimization Algorithm (AOA) with Improved RefineDet (IRD) is used for object detection. Next, the EfficientNet-B0 model is used for the classification of garbage images. The garbage content is identified, and the content is prepared to train the deep learning model to perform efficient classification tasks. For result evaluation, smart bins are deployed in real-time, and accuracy is estimated. Furthermore, fine-tuning region-specific litter photos led to enhanced categorization.","doi":"10.1007\/s10723-023-09722-6","cleaned_title":"deepbin: deep learning based garbage classification for households using sustainable natural technologies","cleaned_abstract":"today, things that are accessible worldwide are upgrading to innovative technology. in this research, an intelligent garbage system will be designed with state-of-the-art methods using deep learning technologies. garbage is highly produced due to urbanization and the rising population in urban areas. it is essential to manage daily trash from homes and living environments. this research aims to provide an intelligent iot-based garbage bin system, and classification is done using deep learning techniques. this smart bin is capable of sensing more varieties of garbage from home. though there are more technologies successfully implemented with iot and machine learning, there is still a need for sustainable natural technologies to manage daily waste. the innovative iot-based garbage system uses various sensors like humidity, temperature, gas, and liquid sensors to identify the garbage condition. initially, the smart garbage bin system is designed, and then the data are collected using a garbage annotation application. next, the deep learning method is used for object detection and classification of garbage images. arithmetic optimization algorithm (aoa) with improved refinedet (ird) is used for object detection. next, the efficientnet-b0 model is used for the classification of garbage images. the garbage content is identified, and the content is prepared to train the deep learning model to perform efficient classification tasks. for result evaluation, smart bins are deployed in real-time, and accuracy is estimated. furthermore, fine-tuning region-specific litter photos led to enhanced categorization.","key_phrases":["things","that","innovative technology","this research","an intelligent garbage system","the-art","deep learning technologies","garbage","urbanization","the rising population","urban areas","it","daily trash","homes","living environments","this research","an intelligent iot-based garbage bin system","classification","deep learning techniques","this smart bin","more varieties","garbage","home","more technologies","iot and machine learning","a need","sustainable natural technologies","daily waste","the innovative iot-based garbage system","various sensors","humidity","temperature","gas","liquid sensors","the garbage condition","the smart garbage bin system","the data","a garbage annotation application","the deep learning method","object detection","classification","garbage images","arithmetic optimization algorithm","aoa","improved refinedet","ird","object detection","the efficientnet-b0 model","the classification","garbage images","the garbage content","the content","the deep learning model","efficient classification tasks","result evaluation","smart bins","real-time","accuracy","fine-tuning region-specific litter photos","enhanced categorization","today","daily","bin system","daily","bin system"]},{"title":"Epileptic Seizures Detection Using iEEG Signals and Deep Learning Models","authors":["Nourane Abderrahim","Amira Echtioui","Rafik Khemakhem","Wassim Zouch","Mohamed Ghorbel","Ahmed Ben Hamida"],"abstract":"Epilepsy is a common neurological disorder that affects millions of people worldwide, and many patients do not respond well to traditional anti-epileptic drugs. To improve the lives of these patients, there is a need to develop accurate methods for predicting epileptic seizures. Seizure prediction involves classifying preictal and interictal states, which is a challenging classification problem. Deep learning techniques, such as convolutional neural networks (CNNs), have shown great promise in analyzing and classifying EEG signals related to epilepsy. In this study, we proposed four deep learning models (S-CNN, Modif-CNN, CNN-SVM, and Comb-2CNN) to classify epilepsy states, which we evaluated on an iEEG dataset from the American Epilepsy Society database. Our models achieved high accuracy rates, with the S-CNN and Comb-2CNN models achieving 96.53%, CNN-SVM achieving 96.99%, and the Modif-CNN model achieving 97.96% in our experiments. These findings suggest that deep learning models could be an effective approach for classifying epilepsy states and could potentially improve seizure prediction methods, ultimately enhancing the quality of life for people with epilepsy.","doi":"10.1007\/s00034-023-02527-8","cleaned_title":"epileptic seizures detection using ieeg signals and deep learning models","cleaned_abstract":"epilepsy is a common neurological disorder that affects millions of people worldwide, and many patients do not respond well to traditional anti-epileptic drugs. to improve the lives of these patients, there is a need to develop accurate methods for predicting epileptic seizures. seizure prediction involves classifying preictal and interictal states, which is a challenging classification problem. deep learning techniques, such as convolutional neural networks (cnns), have shown great promise in analyzing and classifying eeg signals related to epilepsy. in this study, we proposed four deep learning models (s-cnn, modif-cnn, cnn-svm, and comb-2cnn) to classify epilepsy states, which we evaluated on an ieeg dataset from the american epilepsy society database. our models achieved high accuracy rates, with the s-cnn and comb-2cnn models achieving 96.53%, cnn-svm achieving 96.99%, and the modif-cnn model achieving 97.96% in our experiments. these findings suggest that deep learning models could be an effective approach for classifying epilepsy states and could potentially improve seizure prediction methods, ultimately enhancing the quality of life for people with epilepsy.","key_phrases":["epilepsy","a common neurological disorder","that","millions","people","many patients","traditional anti-epileptic drugs","the lives","these patients","a need","accurate methods","epileptic seizures","seizure prediction","preictal and interictal states","which","a challenging classification problem","deep learning techniques","convolutional neural networks","cnns","great promise","eeg signals","epilepsy","this study","we","four deep learning models","s","cnn","modif-cnn, cnn-svm","comb-2cnn","epilepsy states","which","we","an ieeg dataset","the american epilepsy society database","our models","high accuracy rates","the s-cnn and comb-2cnn models","96.53%","cnn-svm","96.99%","the modif-cnn model","97.96%","our experiments","these findings","deep learning models","an effective approach","epilepsy states","seizure prediction methods","the quality","life","people","epilepsy","millions","four","modif-cnn","cnn","american","comb-2cnn","96.53%","cnn","96.99%","modif","97.96%"]},{"title":"Real-time deep learning-based model predictive control of a 3-DOF biped robot leg","authors":["Haitham El-Hussieny"],"abstract":"Our research utilized deep learning to enhance the control of a 3 Degrees of Freedom biped robot leg. We created a dynamic model based on a detailed joint angles and actuator torques dataset. This model was then integrated into a Model Predictive Control (MPC) framework, allowing for precise trajectory tracking without the need for traditional analytical dynamic models. By incorporating specific constraints within the MPC, we met operational and safety standards. The experimental results demonstrate the effectiveness of deep learning models in improving robotic control, leading to precise trajectory tracking and suggesting potential for further integration of deep learning into robotic system control. This approach not only outperforms traditional control methods in accuracy and efficiency but also opens the way for new research in robotics, highlighting the potential of utilizing deep learning models in predictive control techniques.","doi":"10.1038\/s41598-024-66104-y","cleaned_title":"real-time deep learning-based model predictive control of a 3-dof biped robot leg","cleaned_abstract":"our research utilized deep learning to enhance the control of a 3 degrees of freedom biped robot leg. we created a dynamic model based on a detailed joint angles and actuator torques dataset. this model was then integrated into a model predictive control (mpc) framework, allowing for precise trajectory tracking without the need for traditional analytical dynamic models. by incorporating specific constraints within the mpc, we met operational and safety standards. the experimental results demonstrate the effectiveness of deep learning models in improving robotic control, leading to precise trajectory tracking and suggesting potential for further integration of deep learning into robotic system control. this approach not only outperforms traditional control methods in accuracy and efficiency but also opens the way for new research in robotics, highlighting the potential of utilizing deep learning models in predictive control techniques.","key_phrases":["our research","deep learning","the control","a 3 degrees","freedom biped robot leg","we","a dynamic model","a detailed joint angles","actuator torques","this model","a model predictive control","mpc","framework","precise trajectory tracking","the need","traditional analytical dynamic models","specific constraints","the mpc","we","operational and safety standards","the experimental results","the effectiveness","deep learning models","robotic control","precise trajectory tracking","potential","further integration","deep learning","robotic system control","this approach","traditional control methods","accuracy","efficiency","the way","new research","robotics","the potential","deep learning models","predictive control techniques","3 degrees"]},{"title":"Deep learning for survival analysis: a review","authors":["Simon Wiegrebe","Philipp Kopper","Raphael Sonabend","Bernd Bischl","Andreas Bender"],"abstract":"The influx of deep learning (DL) techniques into the field of survival analysis in recent years has led to substantial methodological progress; for instance, learning from unstructured or high-dimensional data such as images, text or omics data. In this work, we conduct a comprehensive systematic review of DL-based methods for time-to-event analysis, characterizing them according to both survival- and DL-related attributes. In summary, the reviewed methods often address only a small subset of tasks relevant to time-to-event data\u2014e.g., single-risk right-censored data\u2014and neglect to incorporate more complex settings. Our findings are summarized in an editable, open-source, interactive table: https:\/\/survival-org.github.io\/DL4Survival. As this research area is advancing rapidly, we encourage community contribution in order to keep this database up to date.","doi":"10.1007\/s10462-023-10681-3","cleaned_title":"deep learning for survival analysis: a review","cleaned_abstract":"the influx of deep learning (dl) techniques into the field of survival analysis in recent years has led to substantial methodological progress; for instance, learning from unstructured or high-dimensional data such as images, text or omics data. in this work, we conduct a comprehensive systematic review of dl-based methods for time-to-event analysis, characterizing them according to both survival- and dl-related attributes. in summary, the reviewed methods often address only a small subset of tasks relevant to time-to-event data\u2014e.g., single-risk right-censored data\u2014and neglect to incorporate more complex settings. our findings are summarized in an editable, open-source, interactive table: https:\/\/survival-org.github.io\/dl4survival. as this research area is advancing rapidly, we encourage community contribution in order to keep this database up to date.","key_phrases":["the influx","deep learning","(dl) techniques","the field","survival analysis","recent years","substantial methodological progress","instance","unstructured or high-dimensional data","images","text","omics","data","this work","we","a comprehensive systematic review","dl-based methods","event","them","both survival- and dl-related attributes","summary","the reviewed methods","only a small subset","tasks","event","e.g., single-risk right-censored data","neglect","more complex settings","our findings","an editable, open-source, interactive table","https:\/\/survival-org.github.io\/dl4survival","this research area","we","community contribution","order","this database","date","recent years"]},{"title":"A review on emotion detection by using deep learning techniques","authors":["Tulika Chutia","Nomi Baruah"],"abstract":"Along with the growth of Internet with its numerous potential applications and diverse fields, artificial intelligence (AI) and sentiment analysis (SA) have become significant and popular research areas. Additionally, it was a key technology that contributed to the Fourth Industrial Revolution (IR 4.0). The subset of AI known as emotion recognition systems facilitates communication between IR 4.0 and IR 5.0. Nowadays users of social media, digital marketing, and e-commerce sites are increasing day by day resulting in massive amounts of unstructured data. Medical, marketing, public safety, education, human resources, business, and other industries also use the emotion recognition system widely. Hence it provides a large amount of textual data to extract the emotions from them. The paper presents a systematic literature review of the existing literature published between 2013 to 2023 in text-based emotion detection. This review scrupulously summarized 330 research papers from different conferences, journals, workshops, and dissertations. This paper explores different approaches, methods, different deep learning models, key aspects, description of datasets, evaluation techniques, Future prospects of deep learning, challenges in existing studies and presents limitations and practical implications.","doi":"10.1007\/s10462-024-10831-1","cleaned_title":"a review on emotion detection by using deep learning techniques","cleaned_abstract":"along with the growth of internet with its numerous potential applications and diverse fields, artificial intelligence (ai) and sentiment analysis (sa) have become significant and popular research areas. additionally, it was a key technology that contributed to the fourth industrial revolution (ir 4.0). the subset of ai known as emotion recognition systems facilitates communication between ir 4.0 and ir 5.0. nowadays users of social media, digital marketing, and e-commerce sites are increasing day by day resulting in massive amounts of unstructured data. medical, marketing, public safety, education, human resources, business, and other industries also use the emotion recognition system widely. hence it provides a large amount of textual data to extract the emotions from them. the paper presents a systematic literature review of the existing literature published between 2013 to 2023 in text-based emotion detection. this review scrupulously summarized 330 research papers from different conferences, journals, workshops, and dissertations. this paper explores different approaches, methods, different deep learning models, key aspects, description of datasets, evaluation techniques, future prospects of deep learning, challenges in existing studies and presents limitations and practical implications.","key_phrases":["the growth","internet","its numerous potential applications","diverse fields","artificial intelligence","analysis","sa","significant and popular research areas","it","a key technology","that","the fourth industrial revolution","the subset","emotion recognition systems","communication","users","social media","digital marketing","e-commerce sites","day","massive amounts","unstructured data","marketing","public safety","education","human resources","business","other industries","the emotion recognition system","it","a large amount","textual data","the emotions","them","the paper","a systematic literature review","the existing literature","text-based emotion detection","this review","330 research papers","different conferences","journals","workshops","dissertations","this paper","different approaches","methods","different deep learning models","key aspects","description","datasets","evaluation techniques","future prospects","deep learning","challenges","existing studies","limitations","practical implications","fourth","4.0","4.0","5.0","between 2013 to 2023","330"]},{"title":"Theoretical Assessment for Weather Nowcasting Using Deep Learning Methods","authors":["Abhay B. Upadhyay","Saurin R. Shah","Rajesh A. Thakkar"],"abstract":"Weather is influenced by various factors such as temperature, pressure, air movement, moisture\/water vapor, and the Earth\u2019s rotating motion. Accurate weather forecasting at a high geographical resolution is a complex and computationally expensive task. This study employs a nowcasting approach using meteorological radar images. Building upon the principles of unsupervised representation in deep learning, we delve into the emerging field of next-frame prediction in computer vision. This research focuses on predicting future images based on prior image data, with applications ranging from robot decision-making to autonomous driving. We present the latest advancements in next-frame prediction networks, categorizing them into two approaches: Machine Learners and deep learners. We discuss the merits and limitations of each approach, comparing them based on various parameters. Finally, we outline potential directions for future research in this field, aiming to make weather forecasting more precise and accessible.","doi":"10.1007\/s11831-024-10096-5","cleaned_title":"theoretical assessment for weather nowcasting using deep learning methods","cleaned_abstract":"weather is influenced by various factors such as temperature, pressure, air movement, moisture\/water vapor, and the earth\u2019s rotating motion. accurate weather forecasting at a high geographical resolution is a complex and computationally expensive task. this study employs a nowcasting approach using meteorological radar images. building upon the principles of unsupervised representation in deep learning, we delve into the emerging field of next-frame prediction in computer vision. this research focuses on predicting future images based on prior image data, with applications ranging from robot decision-making to autonomous driving. we present the latest advancements in next-frame prediction networks, categorizing them into two approaches: machine learners and deep learners. we discuss the merits and limitations of each approach, comparing them based on various parameters. finally, we outline potential directions for future research in this field, aiming to make weather forecasting more precise and accessible.","key_phrases":["weather","various factors","temperature","pressure","air movement","moisture\/water vapor","the earth","motion","accurate weather forecasting","a high geographical resolution","a complex and computationally expensive task","this study","a nowcasting approach","meteorological radar images","the principles","unsupervised representation","deep learning","we","the emerging field","next-frame prediction","computer vision","this research","future images","prior image data","applications","robot decision-making","autonomous driving","we","the latest advancements","next-frame prediction networks","them","two approaches","machine learners","deep learners","we","the merits","limitations","each approach","them","various parameters","we","potential directions","future research","this field","earth","two"]},{"title":"Deep Learning for Stock Market Prediction Using Sentiment and Technical Analysis","authors":["Georgios-Markos Chatziloizos","Dimitrios Gunopulos","Konstantinos Konstantinou"],"abstract":"Machine learning and deep learning techniques are applied by researchers with a background in both economics and computer science, to predict stock prices and trends. These techniques are particularly attractive as an alternative to existing models and methodologies because of their ability to extract abstract features from data.\u00a0Most existing research approaches are based on using either numerical\/economical data or textual\/sentimental data.\u00a0In this article,\u00a0we use cutting-edge deep learning\/machine learning approaches on both numerical\/economical data and textual\/sentimental data in order not only to predict stock market prices and trends based on combined data but also to understand how a stock's Technical Analysis can be strengthened by using Sentiment Analysis. Using the four tickers AAPL, GOOG, NVDA and S&P 500 Information Technology, we collected historical financial data and historical textual data and we used each type of data individually and in unison, to display in which case the results were more accurate and more profitable. We describe in detail how we analyzed each type of data, and how we used it to come up with our results.","doi":"10.1007\/s42979-024-02651-5","cleaned_title":"deep learning for stock market prediction using sentiment and technical analysis","cleaned_abstract":"machine learning and deep learning techniques are applied by researchers with a background in both economics and computer science, to predict stock prices and trends. these techniques are particularly attractive as an alternative to existing models and methodologies because of their ability to extract abstract features from data. most existing research approaches are based on using either numerical\/economical data or textual\/sentimental data. in this article, we use cutting-edge deep learning\/machine learning approaches on both numerical\/economical data and textual\/sentimental data in order not only to predict stock market prices and trends based on combined data but also to understand how a stock's technical analysis can be strengthened by using sentiment analysis. using the four tickers aapl, goog, nvda and s&p 500 information technology, we collected historical financial data and historical textual data and we used each type of data individually and in unison, to display in which case the results were more accurate and more profitable. we describe in detail how we analyzed each type of data, and how we used it to come up with our results.","key_phrases":["machine learning","deep learning techniques","researchers","a background","both economics","computer science","stock prices","trends","these techniques","an alternative","existing models","methodologies","their ability","abstract features","data","most existing research approaches","either numerical\/economical data","textual\/sentimental data","this article","we","approaches","both numerical\/economical data","textual\/sentimental data","order","stock market prices","trends","combined data","a stock's technical analysis","sentiment analysis","the four tickers","aapl","goog","nvda","s&p","information technology","we","historical financial data","historical textual data","we","each type","data","unison","which case","the results","we","detail","we","each type","data","we","it","our results","four"]},{"title":"IoT-enabled healthcare transformation leveraging deep learning for advanced patient monitoring and diagnosis","authors":["Nawaf Alharbe","Manal Almalki"],"abstract":"Deep learning and the Internet of Things (IoT) are revolutionizing the healthcare industry. This study explores the potential commercial transformation resulting from IoT-enabled healthcare systems that use deep learning for patient monitoring and diagnosis. Wearables, smart sensors, and internet-connected medical devices allow doctors to monitor patients' vital signs, activities, and physiological traits in real time. However, these devices generate vast and complex data, making analysis and diagnosis challenging. Deep learning models are well-suited to analyze this growing volume of medical data. Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) networks can automatically recognize complex patterns and relationships in sensor data, electronic health records, and patient-reported information. This capability aids clinical professionals in diagnosing illnesses, identifying warning signs, and tailoring treatments. This paper describes a Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) -based IoT-enabled healthcare system that performs feature extraction, classification, prediction, and data preparation. Additionally, it addresses interpretability issues, privacy concerns, and resource limitations of deep learning models in real-time healthcare settings. The study demonstrates the effectiveness of Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) -powered IoT-based healthcare solutions, such as real-time patient monitoring, disease detection, risk prediction, and therapy optimization. These techniques can improve the quality, cost, and outcomes of healthcare. Combining Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) with IoT can significantly enhance healthcare by improving disease detection, personalized treatment, and patient monitoring through connected devices and powerful analytics.","doi":"10.1007\/s11042-024-19919-w","cleaned_title":"iot-enabled healthcare transformation leveraging deep learning for advanced patient monitoring and diagnosis","cleaned_abstract":"deep learning and the internet of things (iot) are revolutionizing the healthcare industry. this study explores the potential commercial transformation resulting from iot-enabled healthcare systems that use deep learning for patient monitoring and diagnosis. wearables, smart sensors, and internet-connected medical devices allow doctors to monitor patients' vital signs, activities, and physiological traits in real time. however, these devices generate vast and complex data, making analysis and diagnosis challenging. deep learning models are well-suited to analyze this growing volume of medical data. convolutional neural networks (cnns) and long short-term memory (lstm) networks can automatically recognize complex patterns and relationships in sensor data, electronic health records, and patient-reported information. this capability aids clinical professionals in diagnosing illnesses, identifying warning signs, and tailoring treatments. this paper describes a convolutional neural networks (cnns) and long short-term memory (lstm) -based iot-enabled healthcare system that performs feature extraction, classification, prediction, and data preparation. additionally, it addresses interpretability issues, privacy concerns, and resource limitations of deep learning models in real-time healthcare settings. the study demonstrates the effectiveness of convolutional neural networks (cnns) and long short-term memory (lstm) -powered iot-based healthcare solutions, such as real-time patient monitoring, disease detection, risk prediction, and therapy optimization. these techniques can improve the quality, cost, and outcomes of healthcare. combining convolutional neural networks (cnns) and long short-term memory (lstm) with iot can significantly enhance healthcare by improving disease detection, personalized treatment, and patient monitoring through connected devices and powerful analytics.","key_phrases":["deep learning","the internet","things","iot","the healthcare industry","this study","the potential commercial transformation","iot-enabled healthcare systems","that","deep learning","patient monitoring","diagnosis","wearables","smart sensors","internet-connected medical devices","doctors","patients' vital signs","activities","physiological traits","real time","these devices","vast and complex data","analysis","diagnosis","challenging","deep learning models","this growing volume","medical data","convolutional neural networks","cnns","long short-term memory","(lstm) networks","complex patterns","relationships","sensor data","electronic health records","patient-reported information","this capability","clinical professionals","illnesses","warning signs","tailoring treatments","this paper","a convolutional neural networks","cnns","long short-term memory","lstm","that","feature extraction","classification","prediction","data preparation","it","interpretability issues","privacy concerns","resource limitations","deep learning models","real-time healthcare settings","the study","the effectiveness","convolutional neural networks","cnns","long short-term memory","lstm","real-time patient monitoring","disease detection","risk prediction","therapy optimization","these techniques","the quality","cost","outcomes","healthcare","convolutional neural networks","cnns","long short-term memory","lstm","iot","healthcare","disease detection","personalized treatment","patient monitoring","connected devices","powerful analytics"]},{"title":"An improved ensemble deep learning framework for glaucoma detection","authors":["K. J. Subha","R. Rajavel","B. Paulchamy"],"abstract":"In the realm of biomedical engineering, a significant challenge involves detecting physiological changes within the human body. Currently, these irregularities are assessed manually, which proves to be both laborious and time-consuming due to the intricate nature of the methods used for identification. To address the need for early disease detection, there is growing interest in employing computer-assisted diagnostics. The core goal of this proposed work is to develop a computer-aided system (CAD) aimed at the prompt identification, screening, and treatment of glaucoma. A fundus camera is used to extract structural characteristics from segmented optic discs and optic cups, which assists in the characterization of glaucoma and the evaluation of its severity.. Here, a novel ensemble-based deep learning model is used which consists of three modified pre-trained convolutional neural networks such as ResNet50, VGGNet19, and the Inception-V3. To assess the proposed algorithm's performance, a combined four distinct publicly available datasets which includes RIM-ONE, DRISHTI-GS, DRIONS-DB and HRF were utilized. An accuracy of 98.58%, specificity of 98.17%, sensitivity of 98.80%, Precision of 98.86%, F1 Score of 98.83% and AUC of 0.98 was obtained for the combined dataset and also an outstanding accuracy rates were attained across various datasets: 98.67%, 97.71%, 97.22% and 97.50% for RIM-ONE, DRIONS-DB, HRF, DRISHTI-GS respectively. The ensemble framework exhibits a significant advantage over state-of-the-art techniques, showcasing a performance improvement of up to 1.58% in glaucoma classification than the existing deep learning techniques.","doi":"10.1007\/s11042-024-20088-z","cleaned_title":"an improved ensemble deep learning framework for glaucoma detection","cleaned_abstract":"in the realm of biomedical engineering, a significant challenge involves detecting physiological changes within the human body. currently, these irregularities are assessed manually, which proves to be both laborious and time-consuming due to the intricate nature of the methods used for identification. to address the need for early disease detection, there is growing interest in employing computer-assisted diagnostics. the core goal of this proposed work is to develop a computer-aided system (cad) aimed at the prompt identification, screening, and treatment of glaucoma. a fundus camera is used to extract structural characteristics from segmented optic discs and optic cups, which assists in the characterization of glaucoma and the evaluation of its severity.. here, a novel ensemble-based deep learning model is used which consists of three modified pre-trained convolutional neural networks such as resnet50, vggnet19, and the inception-v3. to assess the proposed algorithm's performance, a combined four distinct publicly available datasets which includes rim-one, drishti-gs, drions-db and hrf were utilized. an accuracy of 98.58%, specificity of 98.17%, sensitivity of 98.80%, precision of 98.86%, f1 score of 98.83% and auc of 0.98 was obtained for the combined dataset and also an outstanding accuracy rates were attained across various datasets: 98.67%, 97.71%, 97.22% and 97.50% for rim-one, drions-db, hrf, drishti-gs respectively. the ensemble framework exhibits a significant advantage over state-of-the-art techniques, showcasing a performance improvement of up to 1.58% in glaucoma classification than the existing deep learning techniques.","key_phrases":["the realm","biomedical engineering","a significant challenge","physiological changes","the human body","these irregularities","which","the intricate nature","the methods","identification","the need","early disease detection","interest","computer-assisted diagnostics","the core goal","this proposed work","a computer-aided system","cad","the prompt identification","screening","treatment","glaucoma","a fundus camera","structural characteristics","segmented optic discs and optic cups","which","the characterization","glaucoma","the evaluation","its severity","a novel ensemble-based deep learning model","which","three modified pre-trained convolutional neural networks","resnet50","vggnet19","the inception","v3","the proposed algorithm's performance","a combined four distinct publicly available datasets","which","drishti-gs, drions","db","an accuracy","98.58%","specificity","98.17%","sensitivity","98.80%","precision","98.86%","f1 score","98.83%","auc","the combined dataset","an outstanding accuracy rates","various datasets","98.67%","97.71%","97.22%","97.50%","drions","db","drishti-gs","the ensemble framework","a significant advantage","the-art","a performance improvement","up to 1.58%","glaucoma classification","the existing deep learning techniques","three","resnet50","vggnet19","four","98.58%","98.17%","98.80%","98.86%","98.83%","0.98","98.67%","97.71%","97.22%","97.50%","up to 1.58%"]},{"title":"Deep palmprint recognition algorithm based on self-supervised learning and uncertainty loss","authors":["Rui Fan","Xiaohong Han"],"abstract":"With the rapid development of deep learning technology, an increasing number of people are adopting palmprint recognition algorithms based on deep learning for identity authentication. However, these algorithms are susceptible to factors such as palm placement, light source, and insufficient data sampling, resulting in poor recognition accuracy. To address these issues, this paper proposes a new end-to-end deep palmprint recognition algorithm (SSLAUL), which introduces self-supervised representation learning based on contextual prediction, utilizing unlabeled palmprint data for pre-training before introducing the trained parameters into the downstream model for fine-tuning. An uncertainty loss function is introduced into the downstream model, using the homoskedastic uncertainty as a benchmark to do adaptive weight adjustment for different loss functions dynamically. Channel and spatial attention mechanisms are also introduced to extract highly discriminative local features. In this paper, the algorithm is validated on publicly available IITD, CASIA, and PolyU palmprint datasets. The method always achieves the best recognition performance compared to other state-of-the-art algorithms.","doi":"10.1007\/s11760-024-03104-5","cleaned_title":"deep palmprint recognition algorithm based on self-supervised learning and uncertainty loss","cleaned_abstract":"with the rapid development of deep learning technology, an increasing number of people are adopting palmprint recognition algorithms based on deep learning for identity authentication. however, these algorithms are susceptible to factors such as palm placement, light source, and insufficient data sampling, resulting in poor recognition accuracy. to address these issues, this paper proposes a new end-to-end deep palmprint recognition algorithm (sslaul), which introduces self-supervised representation learning based on contextual prediction, utilizing unlabeled palmprint data for pre-training before introducing the trained parameters into the downstream model for fine-tuning. an uncertainty loss function is introduced into the downstream model, using the homoskedastic uncertainty as a benchmark to do adaptive weight adjustment for different loss functions dynamically. channel and spatial attention mechanisms are also introduced to extract highly discriminative local features. in this paper, the algorithm is validated on publicly available iitd, casia, and polyu palmprint datasets. the method always achieves the best recognition performance compared to other state-of-the-art algorithms.","key_phrases":["the rapid development","deep learning technology","an increasing number","people","palmprint recognition algorithms","deep learning","identity authentication","these algorithms","factors","palm placement","light source","insufficient data sampling","poor recognition accuracy","these issues","this paper","end","recognition algorithm","sslaul","which","self-supervised representation learning","contextual prediction","unlabeled palmprint data","pre","-","training","the trained parameters","the downstream model","fine-tuning","an uncertainty loss function","the downstream model","the homoskedastic uncertainty","a benchmark","adaptive weight adjustment","different loss functions","channel","spatial attention mechanisms","highly discriminative local features","this paper","the algorithm","publicly available iitd","casia","polyu","datasets","the method","the best recognition performance","the-art","casia"]},{"title":"On the use of deep learning for phase recovery","authors":["Kaiqiang Wang","Li Song","Chutian Wang","Zhenbo Ren","Guangyuan Zhao","Jiazhen Dou","Jianglei Di","George Barbastathis","Renjie Zhou","Jianlin Zhao","Edmund Y. Lam"],"abstract":"Phase recovery (PR) refers to calculating the phase of the light field from its intensity measurements. As exemplified from quantitative phase imaging and coherent diffraction imaging to adaptive optics, PR is essential for reconstructing the refractive index distribution or topography of an object and correcting the aberration of an imaging system. In recent years, deep learning (DL), often implemented through deep neural networks, has provided unprecedented support for computational imaging, leading to more efficient solutions for various PR problems. In this review, we first briefly introduce conventional methods for PR. Then, we review how DL provides support for PR from the following three stages, namely, pre-processing, in-processing, and post-processing. We also review how DL is used in phase image processing. Finally, we summarize the work in DL for PR and provide an outlook on how to better use DL to improve the reliability and efficiency of PR. Furthermore, we present a live-updating resource (https:\/\/github.com\/kqwang\/phase-recovery) for readers to learn more about PR.","doi":"10.1038\/s41377-023-01340-x","cleaned_title":"on the use of deep learning for phase recovery","cleaned_abstract":"phase recovery (pr) refers to calculating the phase of the light field from its intensity measurements. as exemplified from quantitative phase imaging and coherent diffraction imaging to adaptive optics, pr is essential for reconstructing the refractive index distribution or topography of an object and correcting the aberration of an imaging system. in recent years, deep learning (dl), often implemented through deep neural networks, has provided unprecedented support for computational imaging, leading to more efficient solutions for various pr problems. in this review, we first briefly introduce conventional methods for pr. then, we review how dl provides support for pr from the following three stages, namely, pre-processing, in-processing, and post-processing. we also review how dl is used in phase image processing. finally, we summarize the work in dl for pr and provide an outlook on how to better use dl to improve the reliability and efficiency of pr. furthermore, we present a live-updating resource (https:\/\/github.com\/kqwang\/phase-recovery) for readers to learn more about pr.","key_phrases":["phase recovery","(pr","the phase","the light field","its intensity measurements","quantitative phase imaging","coherent diffraction","adaptive optics","the refractive index distribution","topography","an object","the aberration","an imaging system","recent years","deep learning","dl","deep neural networks","unprecedented support","computational imaging","more efficient solutions","various pr problems","this review","we","conventional methods","we","dl","support","pr","the following three stages","-processing","processing","we","dl","phase image processing","we","the work","dl","pr","an outlook","dl","the reliability","efficiency","pr","we","a live-updating resource","https:\/\/github.com\/kqwang\/phase-recovery","readers","pr","recent years","first","three"]},{"title":"Efficient Generation of Pretraining Samples for Developing a Deep Learning Brain Injury Model via Transfer Learning","authors":["Nan Lin","Shaoju Wu","Zheyang Wu","Songbai Ji"],"abstract":"The large amount of training samples required to develop a deep learning brain injury model demands enormous computational resources. Here, we study how a transformer neural network (TNN) of high accuracy can be used to efficiently generate pretraining samples for a convolutional neural network (CNN) brain injury model to reduce computational cost. The samples use synthetic impacts emulating real-world events or augmented impacts generated from limited measured impacts. First, we verify that the TNN remains highly accurate for the two impact types (N\u2009=\u2009100 each; \\({R}^{2}\\) of 0.948\u20130.967 with root mean squared error, RMSE,\u2009~\u20090.01, for voxelized peak strains). The TNN-estimated samples (1000\u20135000 for each data type) are then used to pretrain a CNN, which is further finetuned using directly simulated training samples (250\u20135000). An independent measured impact dataset considered of complete capture of impact event is used to assess estimation accuracy (N\u2009=\u2009191). We find that pretraining can significantly improve CNN accuracy via transfer learning compared to a baseline CNN without pretraining. It is most effective when the finetuning dataset is relatively small (e.g., 2000\u20134000 pretraining synthetic or augmented samples improves success rate from 0.72 to 0.81 with 500 finetuning samples). When finetuning samples reach 3000 or more, no obvious improvement occurs from pretraining. These results support using the TNN to rapidly generate pretraining samples to facilitate a more efficient training strategy for future deep learning brain models, by limiting the number of costly direct simulations from an alternative baseline model. This study could contribute to a wider adoption of deep learning brain injury models for large-scale predictive modeling and ultimately, enhancing safety protocols and protective equipment.","doi":"10.1007\/s10439-023-03354-3","cleaned_title":"efficient generation of pretraining samples for developing a deep learning brain injury model via transfer learning","cleaned_abstract":"the large amount of training samples required to develop a deep learning brain injury model demands enormous computational resources. here, we study how a transformer neural network (tnn) of high accuracy can be used to efficiently generate pretraining samples for a convolutional neural network (cnn) brain injury model to reduce computational cost. the samples use synthetic impacts emulating real-world events or augmented impacts generated from limited measured impacts. first, we verify that the tnn remains highly accurate for the two impact types (n = 100 each; \\({r}^{2}\\) of 0.948\u20130.967 with root mean squared error, rmse, ~ 0.01, for voxelized peak strains). the tnn-estimated samples (1000\u20135000 for each data type) are then used to pretrain a cnn, which is further finetuned using directly simulated training samples (250\u20135000). an independent measured impact dataset considered of complete capture of impact event is used to assess estimation accuracy (n = 191). we find that pretraining can significantly improve cnn accuracy via transfer learning compared to a baseline cnn without pretraining. it is most effective when the finetuning dataset is relatively small (e.g., 2000\u20134000 pretraining synthetic or augmented samples improves success rate from 0.72 to 0.81 with 500 finetuning samples). when finetuning samples reach 3000 or more, no obvious improvement occurs from pretraining. these results support using the tnn to rapidly generate pretraining samples to facilitate a more efficient training strategy for future deep learning brain models, by limiting the number of costly direct simulations from an alternative baseline model. this study could contribute to a wider adoption of deep learning brain injury models for large-scale predictive modeling and ultimately, enhancing safety protocols and protective equipment.","key_phrases":["the large amount","training samples","a deep learning brain injury model","enormous computational resources","we","a transformer neural network","tnn","high accuracy","pretraining samples","a convolutional neural network (cnn) brain injury model","computational cost","the samples","synthetic impacts","real-world events","augmented impacts","limited measured impacts","we","the tnn","the two impact types","root mean squared error","rmse","voxelized peak strains","the tnn-estimated samples","each data type","a cnn","which","directly simulated training samples","an independent measured impact dataset","complete capture","impact event","estimation accuracy","we","pretraining","cnn accuracy","transfer learning","a baseline cnn","it","the finetuning dataset","2000\u20134000 pretraining synthetic or augmented samples","success rate","500 finetuning samples","samples","no obvious improvement","these results","the tnn","pretraining samples","a more efficient training strategy","future deep learning brain models","the number","costly direct simulations","an alternative baseline model","this study","a wider adoption","deep learning brain injury models","large-scale predictive modeling","safety protocols","protective equipment","cnn","first","two","100","0.948\u20130.967","0.01","1000\u20135000","cnn","191","cnn","cnn","2000\u20134000","0.72","0.81","500","3000"]},{"title":"Curriculum learning for ab initio deep learned refractive optics","authors":["Xinge Yang","Qiang Fu","Wolfgang Heidrich"],"abstract":"Deep optical optimization has recently emerged as a new paradigm for designing computational imaging systems using only the output image as the objective. However, it has been limited to either simple optical systems consisting of a single element such as a diffractive optical element or metalens, or the fine-tuning of compound lenses from good initial designs. Here we present a DeepLens design method based on curriculum learning, which is able to learn optical designs of compound lenses ab initio from randomly initialized surfaces without human intervention, therefore overcoming the need for a good initial design. We demonstrate the effectiveness of our approach by fully automatically designing both classical imaging lenses and a large field-of-view extended depth-of-field computational lens in a cellphone-style form factor, with highly aspheric surfaces and a short back focal length.","doi":"10.1038\/s41467-024-50835-7","cleaned_title":"curriculum learning for ab initio deep learned refractive optics","cleaned_abstract":"deep optical optimization has recently emerged as a new paradigm for designing computational imaging systems using only the output image as the objective. however, it has been limited to either simple optical systems consisting of a single element such as a diffractive optical element or metalens, or the fine-tuning of compound lenses from good initial designs. here we present a deeplens design method based on curriculum learning, which is able to learn optical designs of compound lenses ab initio from randomly initialized surfaces without human intervention, therefore overcoming the need for a good initial design. we demonstrate the effectiveness of our approach by fully automatically designing both classical imaging lenses and a large field-of-view extended depth-of-field computational lens in a cellphone-style form factor, with highly aspheric surfaces and a short back focal length.","key_phrases":["deep optical optimization","a new paradigm","computational imaging systems","only the output image","the objective","it","either simple optical systems","a single element","a diffractive optical element","metalens","the fine-tuning","compound lenses","good initial designs","we","a deeplens design method","curriculum learning","which","optical designs","compound lenses","ab initio","randomly initialized surfaces","human intervention","the need","a good initial design","we","the effectiveness","our approach","both classical imaging lenses","view","field","a cellphone-style form factor","highly aspheric surfaces","a short back focal length","deeplens"]},{"title":"Verifying the Generalization of Deep Learning to Out-of-Distribution Domains","authors":["Guy Amir","Osher Maayan","Tom Zelazny","Guy Katz","Michael Schapira"],"abstract":"Deep neural networks (DNNs) play a crucial role in the field of machine learning, demonstrating state-of-the-art performance across various application domains. However, despite their success, DNN-based models may occasionally exhibit challenges with generalization, i.e., may fail to handle inputs that were not encountered during training. This limitation is a significant challenge when it comes to deploying deep learning for safety-critical tasks, as well as in real-world settings characterized by substantial variability. We introduce a novel approach for harnessing DNN verification technology to identify DNN-driven decision rules that exhibit robust generalization to previously unencountered input domains. Our method assesses generalization within an input domain by measuring the level of agreement between independently trained deep neural networks for inputs in this domain. We also efficiently realize our approach by using off-the-shelf DNN verification engines, and extensively evaluate it on both supervised and unsupervised DNN benchmarks, including a deep reinforcement learning (DRL) system for Internet congestion control\u2014demonstrating the applicability of our approach for real-world settings. Moreover, our research introduces a fresh objective for formal verification, offering the prospect of mitigating the challenges linked to deploying DNN-driven systems in real-world scenarios.","doi":"10.1007\/s10817-024-09704-7","cleaned_title":"verifying the generalization of deep learning to out-of-distribution domains","cleaned_abstract":"deep neural networks (dnns) play a crucial role in the field of machine learning, demonstrating state-of-the-art performance across various application domains. however, despite their success, dnn-based models may occasionally exhibit challenges with generalization, i.e., may fail to handle inputs that were not encountered during training. this limitation is a significant challenge when it comes to deploying deep learning for safety-critical tasks, as well as in real-world settings characterized by substantial variability. we introduce a novel approach for harnessing dnn verification technology to identify dnn-driven decision rules that exhibit robust generalization to previously unencountered input domains. our method assesses generalization within an input domain by measuring the level of agreement between independently trained deep neural networks for inputs in this domain. we also efficiently realize our approach by using off-the-shelf dnn verification engines, and extensively evaluate it on both supervised and unsupervised dnn benchmarks, including a deep reinforcement learning (drl) system for internet congestion control\u2014demonstrating the applicability of our approach for real-world settings. moreover, our research introduces a fresh objective for formal verification, offering the prospect of mitigating the challenges linked to deploying dnn-driven systems in real-world scenarios.","key_phrases":["deep neural networks","dnns","a crucial role","the field","machine learning","the-art","various application domains","their success","dnn-based models","challenges","generalization","inputs","that","training","this limitation","a significant challenge","it","deep learning","safety-critical tasks","real-world settings","substantial variability","we","a novel approach","dnn verification technology","dnn-driven decision rules","that","robust generalization","input domains","our method","generalization","an input domain","the level","agreement","independently trained deep neural networks","inputs","this domain","we","our approach","the-shelf","dnn verification engines","it","dnn benchmarks","a deep reinforcement learning (drl) system","internet congestion control","the applicability","our approach","real-world settings","our research","a fresh objective","formal verification","the prospect","the challenges","dnn-driven systems","real-world scenarios"]},{"title":"A multimodal deep learning model for predicting severe hemorrhage in placenta previa","authors":["Munetoshi Akazawa","Kazunori Hashimoto"],"abstract":"Placenta previa causes life-threatening bleeding and accurate prediction of severe hemorrhage leads to risk stratification and optimum allocation of interventions. We aimed to use a multimodal deep learning model to predict severe hemorrhage. Using MRI T2-weighted image of the placenta and tabular data consisting of patient demographics and preoperative blood examination data, a multimodal deep learning model was constructed to predict cases of intraoperative blood loss\u2009>\u20092000\u00a0ml. We evaluated the prediction performance of the model by comparing it with that of two machine learning methods using only tabular data and MRI images, as well as with that of two human expert obstetricians. Among the enrolled 48 patients, 26 (54.2%) lost\u2009>\u20092000\u00a0ml of blood and 22 (45.8%) lost\u2009<\u20092000\u00a0ml of blood. Multimodal deep learning model showed the best accuracy of 0.68 and AUC of 0.74, whereas the machine learning model using tabular data and MRI images had a class accuracy of 0.61 and 0.53, respectively. The human experts had median accuracies of 0.61. Multimodal deep learning models could integrate the two types of information and predict severe hemorrhage cases. The model might assist human expert in the prediction of intraoperative hemorrhage in the case of placenta previa.","doi":"10.1038\/s41598-023-44634-1","cleaned_title":"a multimodal deep learning model for predicting severe hemorrhage in placenta previa","cleaned_abstract":"placenta previa causes life-threatening bleeding and accurate prediction of severe hemorrhage leads to risk stratification and optimum allocation of interventions. we aimed to use a multimodal deep learning model to predict severe hemorrhage. using mri t2-weighted image of the placenta and tabular data consisting of patient demographics and preoperative blood examination data, a multimodal deep learning model was constructed to predict cases of intraoperative blood loss > 2000 ml. we evaluated the prediction performance of the model by comparing it with that of two machine learning methods using only tabular data and mri images, as well as with that of two human expert obstetricians. among the enrolled 48 patients, 26 (54.2%) lost > 2000 ml of blood and 22 (45.8%) lost < 2000 ml of blood. multimodal deep learning model showed the best accuracy of 0.68 and auc of 0.74, whereas the machine learning model using tabular data and mri images had a class accuracy of 0.61 and 0.53, respectively. the human experts had median accuracies of 0.61. multimodal deep learning models could integrate the two types of information and predict severe hemorrhage cases. the model might assist human expert in the prediction of intraoperative hemorrhage in the case of placenta previa.","key_phrases":["placenta previa","life-threatening bleeding","accurate prediction","severe hemorrhage","risk stratification","optimum allocation","interventions","we","a multimodal deep learning model","severe hemorrhage","mri t2-weighted image","the placenta and tabular data","patient demographics","preoperative blood examination data","a multimodal deep learning model","cases","intraoperative blood loss","we","the prediction performance","the model","it","that","two machine learning methods","only tabular data and mri images","that","two human expert obstetricians","the enrolled 48 patients","54.2%","2000 ml","blood","22 (45.8%","2000 ml","blood","multimodal deep learning model","the best accuracy","auc","the machine learning model","tabular data and mri images","a class accuracy","the human experts","median accuracies","multimodal deep learning models","the two types","information","severe hemorrhage cases","the model","human expert","the prediction","intraoperative hemorrhage","the case","placenta previa","placenta previa","2000 ml","two","two","48","26","54.2%","2000 ml","22","45.8%","2000 ml","0.68","0.74","0.61","0.53","0.61","two"]},{"title":"Deep Reinforcement Learning for Mineral  Prospectivity  Mapping","authors":["Zixian Shi","Renguang Zuo","Bao Zhou"],"abstract":"Machine learning algorithms, including supervised and unsupervised learning ones, have been widely used in mineral prospectivity mapping. Supervised learning algorithms require the use of numerous known mineral deposits to ensure the reliability of the training results. Unsupervised learning algorithms can be applied to areas with rare or no known deposits. Reinforcement learning (RL) is a type of machine learning algorithm that differs from supervised and unsupervised learning models in that the learning process is performed interactively by the agent and environment. The environment feeds the agent with reward signals and states, and the agent synthetically evaluates the mineralization potential of each state based on these rewards. In this study, a deep RL framework was constructed for mineral prospectivity mapping, and a case study for mapping gold prospectivity in northwest Hubei Province, China, was used to test the framework. The deep RL agent extracted the information of known mineralization by automatically interacting with the environment while simultaneously mining potential mineralization information from the unlabeled dataset. Its comparison with random forest and isolation forest models demonstrates that deep RL performs better regardless of the number of known mineral deposits because of its unique reward and feedback mechanism. The delineated high-potential areas show a strong spatial correlation with known gold deposits and can therefore provide significant clues for future prospecting in the study area.","doi":"10.1007\/s11004-023-10059-9","cleaned_title":"deep reinforcement learning for mineral prospectivity mapping","cleaned_abstract":"machine learning algorithms, including supervised and unsupervised learning ones, have been widely used in mineral prospectivity mapping. supervised learning algorithms require the use of numerous known mineral deposits to ensure the reliability of the training results. unsupervised learning algorithms can be applied to areas with rare or no known deposits. reinforcement learning (rl) is a type of machine learning algorithm that differs from supervised and unsupervised learning models in that the learning process is performed interactively by the agent and environment. the environment feeds the agent with reward signals and states, and the agent synthetically evaluates the mineralization potential of each state based on these rewards. in this study, a deep rl framework was constructed for mineral prospectivity mapping, and a case study for mapping gold prospectivity in northwest hubei province, china, was used to test the framework. the deep rl agent extracted the information of known mineralization by automatically interacting with the environment while simultaneously mining potential mineralization information from the unlabeled dataset. its comparison with random forest and isolation forest models demonstrates that deep rl performs better regardless of the number of known mineral deposits because of its unique reward and feedback mechanism. the delineated high-potential areas show a strong spatial correlation with known gold deposits and can therefore provide significant clues for future prospecting in the study area.","key_phrases":["machine learning algorithms","supervised and unsupervised learning ones","mineral prospectivity mapping","supervised learning algorithms","the use","numerous known mineral deposits","the reliability","the training results","unsupervised learning algorithms","areas","rare or no known deposits","reinforcement learning","rl","a type","machine learning algorithm","that","supervised and unsupervised learning models","the learning process","the agent","environment","the environment","the agent","reward signals","states","the agent","the mineralization potential","each state","these rewards","this study","a deep rl framework","mineral prospectivity mapping","a case study","mapping","gold prospectivity","northwest hubei province","china","the framework","the deep rl agent","the information","known mineralization","the environment","potential mineralization information","the unlabeled dataset","its comparison","random forest and isolation forest models","deep rl","the number","known mineral deposits","its unique reward","feedback mechanism","the delineated high-potential areas","a strong spatial correlation","known gold deposits","significant clues","future prospecting","the study area","china"]},{"title":"A deep learning based approach for image retrieval extraction in mobile edge computing","authors":["Jamal Alasadi","Ghassan F. Bati","Ahmed Al Hilli"],"abstract":"Deep learning has been widely explored in 5G applications, including computer vision, the Internet of Things (IoT), and intermedia classification. However, applying the deep learning approach in limited-resource mobile devices is one of the most challenging issues. At the same time, users\u2019 experience in terms of Quality of Service (QoS) (e.g., service latency, outcome accuracy, and achievable data rate) performs poorly while interacting with machine learning applications. Mobile edge computing (MEC) has been introduced as a cooperative approach to bring computation resources in proximity to end-user devices to overcome these limitations. This article aims to design a novel image reiterative extraction algorithm based on convolution neural network (CNN) learning and computational task offloading to support machine learning-based mobile applications in resource-limited and uncertain environments. Accordingly, we leverage the framework of image retrieval extraction and introduce three approaches. First, privacy preservation is strict and aims to protect personal data. Second, network traffic reduction. Third, minimizing feature matching time. Our simulation results associated with real-time experiments on a small-scale MEC server have shown the effectiveness of the proposed deep learning-based approach over existing schemes. The source code is available here: https:\/\/github.com\/jamalalasadi\/CNN_Image_retrieval.","doi":"10.1007\/s43995-024-00060-6","cleaned_title":"a deep learning based approach for image retrieval extraction in mobile edge computing","cleaned_abstract":"deep learning has been widely explored in 5g applications, including computer vision, the internet of things (iot), and intermedia classification. however, applying the deep learning approach in limited-resource mobile devices is one of the most challenging issues. at the same time, users\u2019 experience in terms of quality of service (qos) (e.g., service latency, outcome accuracy, and achievable data rate) performs poorly while interacting with machine learning applications. mobile edge computing (mec) has been introduced as a cooperative approach to bring computation resources in proximity to end-user devices to overcome these limitations. this article aims to design a novel image reiterative extraction algorithm based on convolution neural network (cnn) learning and computational task offloading to support machine learning-based mobile applications in resource-limited and uncertain environments. accordingly, we leverage the framework of image retrieval extraction and introduce three approaches. first, privacy preservation is strict and aims to protect personal data. second, network traffic reduction. third, minimizing feature matching time. our simulation results associated with real-time experiments on a small-scale mec server have shown the effectiveness of the proposed deep learning-based approach over existing schemes. the source code is available here: https:\/\/github.com\/jamalalasadi\/cnn_image_retrieval.","key_phrases":["deep learning","5g applications","computer vision","the internet","things","iot","intermedia classification","the deep learning approach","limited-resource mobile devices","the most challenging issues","the same time","users\u2019 experience","terms","quality","service","qos","machine learning applications","mobile edge computing","(mec","a cooperative approach","computation resources","proximity","end-user devices","these limitations","this article","a novel image reiterative extraction algorithm","convolution neural network","(cnn) learning","computational task","machine learning-based mobile applications","resource-limited and uncertain environments","we","the framework","image retrieval extraction","three approaches","privacy preservation","personal data","feature matching time","our simulation results","real-time experiments","a small-scale mec server","the effectiveness","the proposed deep learning-based approach","existing schemes","the source code","https:\/\/github.com\/jamalalasadi\/cnn_image_retrieval","5","cnn","three","first","second","third"]},{"title":"Deep learning-based automated angle measurement for flatfoot diagnosis in weight-bearing lateral radiographs","authors":["Won-Jun Noh","Mu Sook Lee","Byoung-Dai Lee"],"abstract":"This study aimed to develop and evaluate a deep learning-based system for the automatic measurement of angles (specifically, Meary\u2019s angle and calcaneal pitch) in weight-bearing lateral radiographs of the foot for flatfoot diagnosis. We utilized 3960 lateral radiographs, either from the left or right foot, sourced from a pool of 4000 patients to construct and evaluate a deep learning-based model. These radiographs were captured between June and November 2021, and patients who had undergone total ankle replacement surgery or ankle arthrodesis surgery were excluded. Various methods, including correlation analysis, Bland\u2013Altman plots, and paired T-tests, were employed to assess the concordance between the angles automatically measured using the system and those assessed by clinical experts. The evaluation dataset comprised 150 weight-bearing radiographs from 150 patients. In all test cases, the angles automatically computed using the deep learning-based system were in good agreement with the reference standards (Meary\u2019s angle: Pearson correlation coefficient (PCC)\u2009=\u20090.964, intraclass correlation coefficient (ICC)\u2009=\u20090.963, concordance correlation coefficient (CCC)\u2009=\u20090.963, p-value\u2009=\u20090.632, mean absolute error (MAE)\u2009=\u20091.59\u00b0; calcaneal pitch: PCC\u2009=\u20090.988, ICC\u2009=\u20090.987, CCC\u2009=\u20090.987, p-value\u2009=\u20090.055, MAE\u2009=\u20090.63\u00b0). The average time required for angle measurement using only the CPU to execute the deep learning-based system was 11\u2009\u00b1\u20091\u00a0s. The deep learning-based automatic angle measurement system, a tool for diagnosing flatfoot, demonstrated comparable accuracy and reliability with the results obtained by medical professionals for patients without internal fixation devices.","doi":"10.1038\/s41598-024-69549-3","cleaned_title":"deep learning-based automated angle measurement for flatfoot diagnosis in weight-bearing lateral radiographs","cleaned_abstract":"this study aimed to develop and evaluate a deep learning-based system for the automatic measurement of angles (specifically, meary\u2019s angle and calcaneal pitch) in weight-bearing lateral radiographs of the foot for flatfoot diagnosis. we utilized 3960 lateral radiographs, either from the left or right foot, sourced from a pool of 4000 patients to construct and evaluate a deep learning-based model. these radiographs were captured between june and november 2021, and patients who had undergone total ankle replacement surgery or ankle arthrodesis surgery were excluded. various methods, including correlation analysis, bland\u2013altman plots, and paired t-tests, were employed to assess the concordance between the angles automatically measured using the system and those assessed by clinical experts. the evaluation dataset comprised 150 weight-bearing radiographs from 150 patients. in all test cases, the angles automatically computed using the deep learning-based system were in good agreement with the reference standards (meary\u2019s angle: pearson correlation coefficient (pcc) = 0.964, intraclass correlation coefficient (icc) = 0.963, concordance correlation coefficient (ccc) = 0.963, p-value = 0.632, mean absolute error (mae) = 1.59\u00b0; calcaneal pitch: pcc = 0.988, icc = 0.987, ccc = 0.987, p-value = 0.055, mae = 0.63\u00b0). the average time required for angle measurement using only the cpu to execute the deep learning-based system was 11 \u00b1 1 s. the deep learning-based automatic angle measurement system, a tool for diagnosing flatfoot, demonstrated comparable accuracy and reliability with the results obtained by medical professionals for patients without internal fixation devices.","key_phrases":["this study","a deep learning-based system","the automatic measurement","angles","(specifically, meary\u2019s angle","calcaneal pitch","weight-bearing lateral radiographs","the foot","flatfoot diagnosis","we","3960 lateral radiographs","the left or right foot","a pool","4000 patients","a deep learning-based model","these radiographs","june","november","patients","who","total ankle replacement surgery","ankle arthrodesis surgery","various methods","correlation analysis","bland\u2013altman plots","t-tests","the concordance","the angles","the system","those","clinical experts","the evaluation dataset","150 weight-bearing","150 patients","all test cases","the angles","the deep learning-based system","good agreement","the reference standards","meary\u2019s angle","pcc","intraclass correlation coefficient","icc","concordance correlation coefficient","ccc","absolute error","mae","1.59\u00b0","calcaneal pitch","the average time","angle measurement","only the cpu","the deep learning-based system","11 \u00b1","the deep learning-based automatic angle measurement system","a tool","comparable accuracy","reliability","the results","medical professionals","patients","internal fixation devices","3960","4000","between june and november 2021","150","150","0.964","0.963","0.963","0.632","1.59","0.988","0.987","0.987","0.055","0.63","11","1"]},{"title":"Visual sentiment analysis using data-augmented deep transfer learning techniques","authors":["Zhiguo Jiang","Waneeza Zaheer","Aamir Wali","S. A. M. Gilani"],"abstract":"There has been a growing trend among users of social media platforms to express their emotions using visual content. Visual sentiment analysis is the process of understanding the emotional polarity of images or videos and is still considered a challenging problem in artificial intelligence. Most of the existing models are based on robust machine learning or deep learning techniques. The idea of using deep transfer learning techniques for visual sentiment analysis is fairly new. In this paper, we propose a new approach using data-augmented-transfer learning architecture consisting of a pre-trained VGG16 model that is fine-tuned using SVM with augmented training data. For fine-tuning and evaluation, we initially use two Twitter image datasets. We further validated the proposed model on a third dataset. The commonly used geometric augmentation methods such as rotation, zoom range, width shift, height shift, shear range and horizontal flip were are used. We compare our proposed VGG16-SVM model with 3 other state-of-the-art deep models commonly used for transfer learning and 4 machine learning models (besides SVM) used for fine-tuning. The results show that VGG16-SVM produces the overall best accuracy (94%) and recall (96%) among all transfer learning and machine learning pairs. We also show that our proposed model outperforms all previous studies that use the same dataset.","doi":"10.1007\/s11042-023-16262-4","cleaned_title":"visual sentiment analysis using data-augmented deep transfer learning techniques","cleaned_abstract":"there has been a growing trend among users of social media platforms to express their emotions using visual content. visual sentiment analysis is the process of understanding the emotional polarity of images or videos and is still considered a challenging problem in artificial intelligence. most of the existing models are based on robust machine learning or deep learning techniques. the idea of using deep transfer learning techniques for visual sentiment analysis is fairly new. in this paper, we propose a new approach using data-augmented-transfer learning architecture consisting of a pre-trained vgg16 model that is fine-tuned using svm with augmented training data. for fine-tuning and evaluation, we initially use two twitter image datasets. we further validated the proposed model on a third dataset. the commonly used geometric augmentation methods such as rotation, zoom range, width shift, height shift, shear range and horizontal flip were are used. we compare our proposed vgg16-svm model with 3 other state-of-the-art deep models commonly used for transfer learning and 4 machine learning models (besides svm) used for fine-tuning. the results show that vgg16-svm produces the overall best accuracy (94%) and recall (96%) among all transfer learning and machine learning pairs. we also show that our proposed model outperforms all previous studies that use the same dataset.","key_phrases":["a growing trend","users","social media platforms","their emotions","visual content","visual sentiment analysis","the process","the emotional polarity","images","videos","a challenging problem","artificial intelligence","the existing models","robust machine learning","deep learning techniques","the idea","deep transfer","techniques","visual sentiment analysis","this paper","we","a new approach","data-augmented-transfer learning architecture","a pre-trained vgg16 model","that","svm","augmented training data","fine-tuning","evaluation","we","two twitter image datasets","we","the proposed model","a third dataset","the commonly used geometric augmentation methods","rotation","zoom","range",", width shift","height shift","shear range","horizontal flip","we","our proposed vgg16-svm model","the-art","transfer learning","4 machine learning models","svm","fine-tuning","the results","vgg16-svm","the overall best accuracy","94%","96%","all transfer learning and machine learning pairs","we","our proposed model","all previous studies","that","the same dataset","two","third","3","4","94%","96%"]},{"title":"Deep learning-based electrocardiographic screening for chronic kidney disease","authors":["Lauri Holmstrom","Matthew Christensen","Neal Yuan","J. Weston Hughes","John Theurer","Melvin Jujjavarapu","Pedram Fatehi","Alan Kwan","Roopinder K. Sandhu","Joseph Ebinger","Susan Cheng","James Zou","Sumeet S. Chugh","David Ouyang"],"abstract":"BackgroundUndiagnosed chronic kidney disease (CKD) is a common and usually asymptomatic disorder that causes a high burden of morbidity and early mortality worldwide. We developed a deep learning model for CKD screening from routinely acquired ECGs.MethodsWe collected data from a primary cohort with 111,370 patients which had 247,655 ECGs between 2005 and 2019. Using this data, we developed, trained, validated, and tested a deep learning model to predict whether an ECG was taken within one year of the patient receiving a CKD diagnosis. The model was additionally validated using an external cohort from another healthcare system which had 312,145 patients with 896,620 ECGs between 2005 and 2018.ResultsUsing 12-lead ECG waveforms, our deep learning algorithm achieves discrimination for CKD of any stage with an AUC of 0.767 (95% CI 0.760\u20130.773) in a held-out test set and an AUC of 0.709 (0.708\u20130.710) in the external cohort. Our 12-lead ECG-based model performance is consistent across the severity of CKD, with an AUC of 0.753 (0.735\u20130.770) for mild CKD, AUC of 0.759 (0.750\u20130.767) for moderate-severe CKD, and an AUC of 0.783 (0.773\u20130.793) for ESRD. In patients under 60 years old, our model achieves high performance in detecting any stage CKD with both 12-lead (AUC 0.843 [0.836\u20130.852]) and 1-lead ECG waveform (0.824 [0.815\u20130.832]).ConclusionsOur deep learning algorithm is able to detect CKD using ECG waveforms, with stronger performance in younger patients and more severe CKD stages. This ECG algorithm has the potential to augment screening for CKD.","doi":"10.1038\/s43856-023-00278-w","cleaned_title":"deep learning-based electrocardiographic screening for chronic kidney disease","cleaned_abstract":"backgroundundiagnosed chronic kidney disease (ckd) is a common and usually asymptomatic disorder that causes a high burden of morbidity and early mortality worldwide. we developed a deep learning model for ckd screening from routinely acquired ecgs.methodswe collected data from a primary cohort with 111,370 patients which had 247,655 ecgs between 2005 and 2019. using this data, we developed, trained, validated, and tested a deep learning model to predict whether an ecg was taken within one year of the patient receiving a ckd diagnosis. the model was additionally validated using an external cohort from another healthcare system which had 312,145 patients with 896,620 ecgs between 2005 and 2018.resultsusing 12-lead ecg waveforms, our deep learning algorithm achieves discrimination for ckd of any stage with an auc of 0.767 (95% ci 0.760\u20130.773) in a held-out test set and an auc of 0.709 (0.708\u20130.710) in the external cohort. our 12-lead ecg-based model performance is consistent across the severity of ckd, with an auc of 0.753 (0.735\u20130.770) for mild ckd, auc of 0.759 (0.750\u20130.767) for moderate-severe ckd, and an auc of 0.783 (0.773\u20130.793) for esrd. in patients under 60 years old, our model achieves high performance in detecting any stage ckd with both 12-lead (auc 0.843 [0.836\u20130.852]) and 1-lead ecg waveform (0.824 [0.815\u20130.832]).conclusionsour deep learning algorithm is able to detect ckd using ecg waveforms, with stronger performance in younger patients and more severe ckd stages. this ecg algorithm has the potential to augment screening for ckd.","key_phrases":["backgroundundiagnosed chronic kidney disease","ckd","a common and usually asymptomatic disorder","that","a high burden","morbidity","early mortality","we","a deep learning model","ckd","routinely acquired ecgs.methodswe collected data","a primary cohort","111,370 patients","which","247,655 ecgs","this data","we","a deep learning model","an ecg","one year","the patient","a ckd diagnosis","the model","an external cohort","another healthcare system","which","312,145 patients","896,620 ecgs","our deep learning","algorithm","discrimination","ckd","any stage","an auc","95%","ci 0.760\u20130.773","an auc","(0.708\u20130.710","the external cohort","our 12-lead ecg-based model performance","the severity","ckd","an auc","(0.735\u20130.770","mild ckd","auc","0.750\u20130.767","moderate-severe ckd","an auc","esrd","patients","our model","high performance","any stage ckd","both 12-lead","auc","deep learning algorithm","ckd","ecg waveforms","stronger performance","younger patients","more severe ckd stages","this ecg algorithm","the potential","augment","ckd","111,370","247,655","between 2005 and 2019","one year","312,145","896,620","between 2005","2018.resultsusing","12","0.767","95%","0.760\u20130.773","0.709","12","0.753","0.735\u20130.770","0.759","0.750\u20130.767","0.783","0.773\u20130.793","under 60 years old","12","0.843","1","0.824"]},{"title":"Deep-learning based supervisory monitoring of robotized DE-GMAW process through learning from human welders","authors":["Rui Yu","Yue Cao","Jennifer Martin","Otto Chiang","YuMing Zhang"],"abstract":"Double-electrode gas metal arc welding (DE-GMAW) modifies GMAW by adding a second electrode to bypass a portion of the current flowing from the wire. This reduces the current to, and the heat input on, the workpiece. Successful bypassing depends on the relative position of the bypass electrode to the continuously varying wire tip. To ensure proper operation, we propose robotizing the system using a follower robot to carry and adaptively adjust the bypass electrode. The primary information for monitoring this process is the arc image, which directly shows desired and undesired modes. However, developing a robust algorithm for processing the complex arc image is time-consuming and challenging. Employing a deep learning approach requires labeling numerous arc images for the corresponding DE-GMAW modes, which is not practically feasible. To introduce alternative labels, we analyze arc phenomena in various DE-GMAW modes and correlate them with distinct arc systems having varying voltages. These voltages serve as automatically derived labels to train the deep-learning network. The results demonstrated reliable process monitoring.","doi":"10.1007\/s40194-023-01635-y","cleaned_title":"deep-learning based supervisory monitoring of robotized de-gmaw process through learning from human welders","cleaned_abstract":"double-electrode gas metal arc welding (de-gmaw) modifies gmaw by adding a second electrode to bypass a portion of the current flowing from the wire. this reduces the current to, and the heat input on, the workpiece. successful bypassing depends on the relative position of the bypass electrode to the continuously varying wire tip. to ensure proper operation, we propose robotizing the system using a follower robot to carry and adaptively adjust the bypass electrode. the primary information for monitoring this process is the arc image, which directly shows desired and undesired modes. however, developing a robust algorithm for processing the complex arc image is time-consuming and challenging. employing a deep learning approach requires labeling numerous arc images for the corresponding de-gmaw modes, which is not practically feasible. to introduce alternative labels, we analyze arc phenomena in various de-gmaw modes and correlate them with distinct arc systems having varying voltages. these voltages serve as automatically derived labels to train the deep-learning network. the results demonstrated reliable process monitoring.","key_phrases":["double-electrode gas metal arc welding","de","-","gmaw","gmaw","a second electrode","a portion","the current","the wire","this","the heat input","the workpiece","successful bypassing","the relative position","the bypass","the continuously varying wire tip","proper operation","we","the system","a follower robot","the bypass","the primary information","this process","the arc image","which","desired and undesired modes","a robust algorithm","the complex arc image","a deep learning approach","numerous arc images","the corresponding de-gmaw modes","which","alternative labels","we","arc phenomena","various de-gmaw modes","them","distinct arc systems","varying voltages","these voltages","automatically derived labels","the deep-learning network","the results","reliable process monitoring","second"]},{"title":"A Pragmatic Privacy-Preserving Deep Learning Framework Satisfying Differential Privacy","authors":["Tran Khanh Dang","Phat T. Tran-Truong"],"abstract":"With the increasing use of technology in our daily lives, data privacy has become a critical issue. It is essential to carefully design technologies to ensure the protection of people\u2019s personal information. In fact, what we need are privacy-enhancing technologies (PETs) rather than solely focusing on technologies themselves. Artificial intelligence (AI) and deep learning technologies, which are considered societal locomotives, are no exception. However, AI practitioners usually design and develop without considering privacy concerns. To address this gap, we propose a pragmatic privacy-preserving deep learning framework that is suitable for AI practitioners. Our proposed framework is designed to satisfy differential privacy, a rigorous standard for preserving privacy. It is based on a setting called Private Aggregation of Teacher Ensembles (PATE), in which we have made several improvements to achieve a better level of accuracy and privacy protection. Specifically, we use a differential private aggregation mechanism called sparse vector technique and combine it with several other improvements such as human-in-the-loop and pre-trained models. Our proposed solution demonstrates the possibility of producing privacy-preserving models that approximate ground-truth models with a fixed privacy budget. These models are capable of handling a large number of training requests, making them suitable for deep learning training processes. Furthermore, our framework can be deployed in both centralized and distributed training settings. We hope that our work will encourage AI practitioners to adopt PETs and build technologies with privacy in mind.","doi":"10.1007\/s42979-023-02437-1","cleaned_title":"a pragmatic privacy-preserving deep learning framework satisfying differential privacy","cleaned_abstract":"with the increasing use of technology in our daily lives, data privacy has become a critical issue. it is essential to carefully design technologies to ensure the protection of people\u2019s personal information. in fact, what we need are privacy-enhancing technologies (pets) rather than solely focusing on technologies themselves. artificial intelligence (ai) and deep learning technologies, which are considered societal locomotives, are no exception. however, ai practitioners usually design and develop without considering privacy concerns. to address this gap, we propose a pragmatic privacy-preserving deep learning framework that is suitable for ai practitioners. our proposed framework is designed to satisfy differential privacy, a rigorous standard for preserving privacy. it is based on a setting called private aggregation of teacher ensembles (pate), in which we have made several improvements to achieve a better level of accuracy and privacy protection. specifically, we use a differential private aggregation mechanism called sparse vector technique and combine it with several other improvements such as human-in-the-loop and pre-trained models. our proposed solution demonstrates the possibility of producing privacy-preserving models that approximate ground-truth models with a fixed privacy budget. these models are capable of handling a large number of training requests, making them suitable for deep learning training processes. furthermore, our framework can be deployed in both centralized and distributed training settings. we hope that our work will encourage ai practitioners to adopt pets and build technologies with privacy in mind.","key_phrases":["the increasing use","technology","our daily lives","data privacy","a critical issue","it","technologies","the protection","people\u2019s personal information","fact","what","we","privacy-enhancing technologies","pets","technologies","themselves","artificial intelligence","deep learning technologies","which","societal locomotives","no exception","practitioners","privacy concerns","this gap","we","a pragmatic privacy-preserving deep learning framework","that","ai practitioners","our proposed framework","differential privacy","a rigorous standard","privacy","it","a setting","private aggregation","teacher ensembles","which","we","several improvements","a better level","accuracy","privacy protection","we","a differential private aggregation mechanism","sparse vector technique","it","several other improvements","the-loop","our proposed solution","the possibility","privacy-preserving models","that","a fixed privacy budget","these models","a large number","training requests","them","deep learning training processes","our framework","training settings","we","our work","practitioners","pets","technologies","privacy","mind"]},{"title":"A stacked deep learning approach for multiclass classification of plant diseases","authors":["Aman Sharma","Raghav Dalmia","Aarush Saxena","Rajni Mohana"],"abstract":"PurposePlant diseases are one of the main factors affecting food production and reducing production losses, they must be swiftly identified and treated. Deep learning algorithms in association with computer vision techniques have recently found usage in the diagnosis of plant diseases, offering a potent tool with highly accurate results. The objective of this study is to identify a stacking ensemble-based solution by using several algorithms in the process of classifying and diagnosing plant diseases, describing trends, and emphasizing gaps.MethodThe stacking ensemble is made using top four performing deep learning algorithms and multi-layered perceptron as meta classifier. In this regard, we reviewed more than 15 studies from the previous three years that address problems with disease detection, dataset characteristics, researched crops, and pathogens in various ways.ResultsThe proposed ensemble model achieved a maximum accuracy of 98.13% compared to the conventional architectures. For comparing the results, various performance metrics are used such as accuracy, loss, precision etc. which outperformed the results of the deep learning algorithms run separately for the data as shown in Table\u00a05.ConclusionThe suggested framework can help identify the presence of disease in a sample of plant leaves as a preventative strategy as the results were quite promising compared to the results of existing literature.","doi":"10.1007\/s11104-024-06719-2","cleaned_title":"a stacked deep learning approach for multiclass classification of plant diseases","cleaned_abstract":"purposeplant diseases are one of the main factors affecting food production and reducing production losses, they must be swiftly identified and treated. deep learning algorithms in association with computer vision techniques have recently found usage in the diagnosis of plant diseases, offering a potent tool with highly accurate results. the objective of this study is to identify a stacking ensemble-based solution by using several algorithms in the process of classifying and diagnosing plant diseases, describing trends, and emphasizing gaps.methodthe stacking ensemble is made using top four performing deep learning algorithms and multi-layered perceptron as meta classifier. in this regard, we reviewed more than 15 studies from the previous three years that address problems with disease detection, dataset characteristics, researched crops, and pathogens in various ways.resultsthe proposed ensemble model achieved a maximum accuracy of 98.13% compared to the conventional architectures. for comparing the results, various performance metrics are used such as accuracy, loss, precision etc. which outperformed the results of the deep learning algorithms run separately for the data as shown in table 5.conclusionthe suggested framework can help identify the presence of disease in a sample of plant leaves as a preventative strategy as the results were quite promising compared to the results of existing literature.","key_phrases":["purposeplant diseases","the main factors","food production","production losses","they","deep learning algorithms","association","computer vision techniques","usage","the diagnosis","plant diseases","a potent tool","highly accurate results","the objective","this study","a stacking ensemble-based solution","several algorithms","the process","plant diseases","trends","gaps.methodthe stacking ensemble","deep learning algorithms","multi-layered perceptron","meta classifier","this regard","we","more than 15 studies","the previous three years","that","problems","disease detection","dataset characteristics","crops","pathogens","various ways.resultsthe proposed ensemble model","a maximum accuracy","98.13%","the conventional architectures","the results","various performance metrics","accuracy","loss","which","the results","the deep learning algorithms","the data","table","5.conclusionthe suggested framework","the presence","disease","a sample","plant leaves","a preventative strategy","the results","the results","existing literature","one","four","meta classifier","more than 15","the previous three years","98.13%"]},{"title":"Forecasting bitcoin volatility: exploring the potential of deep learning","authors":["Tiago E. Pratas","Filipe R. Ramos","Lihki Rubio"],"abstract":"This study aims to evaluate forecasting properties of classic methodologies (ARCH and GARCH models) in comparison with deep learning methodologies (MLP, RNN, and LSTM architectures) for predicting Bitcoin's volatility. As a new asset class with unique characteristics, Bitcoin's high volatility and structural breaks make forecasting challenging. Based on 2753 observations from 08-09-2014 to 01-05-2022, this study focuses on Bitcoin logarithmic returns. Results show that deep learning methodologies have advantages in terms of forecast quality, although significant computational costs are required. Although both MLP and RNN models produce smoother forecasts with less fluctuation, they fail to capture large spikes. The LSTM architecture, on the other hand, reacts strongly to such movements and tries to adjust its forecast accordingly. To compare forecasting accuracy at different horizons MAPE, MAE metrics are used. Diebold\u2013Mariano tests were conducted to compare the forecast, confirming the superiority of deep learning methodologies. Overall, this study suggests that deep learning methodologies could provide a promising tool for forecasting Bitcoin returns (and therefore volatility), especially for short-term horizons.","doi":"10.1007\/s40822-023-00232-0","cleaned_title":"forecasting bitcoin volatility: exploring the potential of deep learning","cleaned_abstract":"this study aims to evaluate forecasting properties of classic methodologies (arch and garch models) in comparison with deep learning methodologies (mlp, rnn, and lstm architectures) for predicting bitcoin's volatility. as a new asset class with unique characteristics, bitcoin's high volatility and structural breaks make forecasting challenging. based on 2753 observations from 08-09-2014 to 01-05-2022, this study focuses on bitcoin logarithmic returns. results show that deep learning methodologies have advantages in terms of forecast quality, although significant computational costs are required. although both mlp and rnn models produce smoother forecasts with less fluctuation, they fail to capture large spikes. the lstm architecture, on the other hand, reacts strongly to such movements and tries to adjust its forecast accordingly. to compare forecasting accuracy at different horizons mape, mae metrics are used. diebold\u2013mariano tests were conducted to compare the forecast, confirming the superiority of deep learning methodologies. overall, this study suggests that deep learning methodologies could provide a promising tool for forecasting bitcoin returns (and therefore volatility), especially for short-term horizons.","key_phrases":["this study","forecasting properties","classic methodologies","arch and garch models","comparison","deep learning methodologies","mlp","rnn","lstm architectures","bitcoin's volatility","a new asset class","unique characteristics","bitcoin's high volatility","structural breaks","forecasting","2753 observations","this study","bitcoin logarithmic returns","results","deep learning methodologies","advantages","terms","forecast quality","significant computational costs","both mlp","models","smoother forecasts","less fluctuation","they","large spikes","the lstm architecture","the other hand","such movements","its forecast","forecasting accuracy","different horizons mape","mae metrics","\u2013mariano tests","the forecast","the superiority","deep learning methodologies","this study","deep learning methodologies","a promising tool","bitcoin returns","therefore volatility","short-term horizons","2753","08","01","mae metrics"]},{"title":"Presenting a three layer stacking ensemble classifier of deep learning and machine learning for skin cancer classification","authors":["Bahman Jafari Tabaghsar","Reza Tavoli","Mohammad Mahdi Alizadeh Toosi"],"abstract":"One of the most common types of cancer in the world is skin cancer. Despite the different types of skin diseases with different shapes, the classification of skin diseases is a very difficult task. As a result, considering such a problem, a combination model of deep learning algorithms and machine has been proposed for skin disease classification.In this paper, a three-layer architecture based on ensemble learning is presented. In the first layer, the training input is given to convolutional neural network and EfficientNET. The output of the first layer is given to the classifiers of the second layer including machine learning classifiers. The output of the best decision of these classifiers is sent to the third layer classifier and the final prediction is made.The reason for using the three-layer architecture based on group learning was the lack of correct recognition of some classes by simple classifications. On the other hand, some diseases with different classes are classified in the same class. This model helps to correctly identify input samples with the correct combination of classifications in different layers.HAM10000 data set has been used to test and validate the proposed method. The mentioned dataset includes 10,015 images of skin lesions in seven different classes and includes different types of skin diseases. The accuracy is 99.97 on the testing set, which was much better than the previous heavy models.","doi":"10.1007\/s11042-024-19195-8","cleaned_title":"presenting a three layer stacking ensemble classifier of deep learning and machine learning for skin cancer classification","cleaned_abstract":"one of the most common types of cancer in the world is skin cancer. despite the different types of skin diseases with different shapes, the classification of skin diseases is a very difficult task. as a result, considering such a problem, a combination model of deep learning algorithms and machine has been proposed for skin disease classification.in this paper, a three-layer architecture based on ensemble learning is presented. in the first layer, the training input is given to convolutional neural network and efficientnet. the output of the first layer is given to the classifiers of the second layer including machine learning classifiers. the output of the best decision of these classifiers is sent to the third layer classifier and the final prediction is made.the reason for using the three-layer architecture based on group learning was the lack of correct recognition of some classes by simple classifications. on the other hand, some diseases with different classes are classified in the same class. this model helps to correctly identify input samples with the correct combination of classifications in different layers.ham10000 data set has been used to test and validate the proposed method. the mentioned dataset includes 10,015 images of skin lesions in seven different classes and includes different types of skin diseases. the accuracy is 99.97 on the testing set, which was much better than the previous heavy models.","key_phrases":["the most common types","cancer","the world","skin cancer","the different types","skin diseases","different shapes","the classification","skin diseases","a very difficult task","a result","such a problem","a combination model","deep learning algorithms","machine","a three-layer architecture","ensemble learning","the first layer","the training input","convolutional neural network","efficientnet","the output","the first layer","the classifiers","the second layer","machine learning classifiers","the output","the best decision","these classifiers","the third layer classifier","the final prediction","made.the reason","the three-layer architecture","group learning","the lack","correct recognition","some classes","simple classifications","the other hand","some diseases","different classes","the same class","this model","input samples","the correct combination","classifications","different layers.ham10000 data set","the proposed method","the mentioned dataset","10,015 images","skin lesions","seven different classes","different types","skin diseases","the accuracy","the testing set","which","the previous heavy models","one","three","first","first","second","third","three","10,015","seven","99.97"]},{"title":"RRS: Review-Based Recommendation System Using Deep Learning for Vietnamese","authors":["Minh Hoang Nguyen","Thuat Thien Nguyen","Minh Nhat Ta","Tien Minh Nguyen","Kiet Van Nguyen"],"abstract":"Tourism, which includes sightseeing, relaxation, and discovery, is a fundamental aspect of human life. One of the most critical considerations for traveling is accommodation, mainly hotels. To improve the travel experience, we have presented a solution for building a recommendation model using Vietnamese and user data to suggest travelers choose the ideal hotel. Our data was collected from two well-known websites, Traveloka and Ivivu, and includes information about hotels in Vietnam and users\u2019 feedback history, such as comments, ratings, and the names of users and hotels. We then preprocessed and labeled the inter-annotator agreement for various aspects, including service (0.89), infrastructure (0.84), sanitary (0.83), location (0.89), and attitude (0.83). Our recommendation model is built by using Collaborative Filtering and deep learning techniques. Furthermore, we suggest incorporating context vectors from tourists\u2019 Vietnamese comments in the recommendation process. The context model is developed by using deep learning techniques to extract topics and sentiments from the words effectively. The results of our proposed model, as measured by the MSE, were 0.027, which is significantly better than a context-free model using the same parameters, which had an MSE of 0.061. Additionally, our Deep Learning, created using PhoBERT embedding, had an accuracy of 81% for topic classification and 82% for sentiment classification. The FastText based model had 82% and 81% accuracy for topic and sentiment, respectively. Our research demonstrates that our approach improves the accuracy of the recommendation model and has the potential for further development in the future. This idea can introduce a new Recommendation System that can overcome existing limitations and apply to other areas.","doi":"10.1007\/s42979-024-02812-6","cleaned_title":"rrs: review-based recommendation system using deep learning for vietnamese","cleaned_abstract":"tourism, which includes sightseeing, relaxation, and discovery, is a fundamental aspect of human life. one of the most critical considerations for traveling is accommodation, mainly hotels. to improve the travel experience, we have presented a solution for building a recommendation model using vietnamese and user data to suggest travelers choose the ideal hotel. our data was collected from two well-known websites, traveloka and ivivu, and includes information about hotels in vietnam and users\u2019 feedback history, such as comments, ratings, and the names of users and hotels. we then preprocessed and labeled the inter-annotator agreement for various aspects, including service (0.89), infrastructure (0.84), sanitary (0.83), location (0.89), and attitude (0.83). our recommendation model is built by using collaborative filtering and deep learning techniques. furthermore, we suggest incorporating context vectors from tourists\u2019 vietnamese comments in the recommendation process. the context model is developed by using deep learning techniques to extract topics and sentiments from the words effectively. the results of our proposed model, as measured by the mse, were 0.027, which is significantly better than a context-free model using the same parameters, which had an mse of 0.061. additionally, our deep learning, created using phobert embedding, had an accuracy of 81% for topic classification and 82% for sentiment classification. the fasttext based model had 82% and 81% accuracy for topic and sentiment, respectively. our research demonstrates that our approach improves the accuracy of the recommendation model and has the potential for further development in the future. this idea can introduce a new recommendation system that can overcome existing limitations and apply to other areas.","key_phrases":["tourism","which","sightseeing","relaxation","discovery","a fundamental aspect","human life","the most critical considerations","accommodation",", mainly hotels","the travel experience","we","a solution","a recommendation model","user data","travelers","the ideal hotel","our data","two well-known websites","traveloka","ivivu","information","hotels","vietnam","users\u2019 feedback history","comments","ratings","the names","users","hotels","we","the inter-annotator agreement","various aspects","service","infrastructure","location","attitude","our recommendation model","collaborative filtering","deep learning techniques","we","context vectors","tourists\u2019 vietnamese comments","the recommendation process","the context model","deep learning techniques","topics","sentiments","the words","the results","our proposed model","the mse","which","a context-free model","the same parameters","which","an mse","our deep learning","phobert","an accuracy","81%","topic classification","82%","sentiment classification","the fasttext based model","82%","81% accuracy","topic","sentiment","our research","our approach","the accuracy","the recommendation model","the potential","further development","the future","this idea","a new recommendation system","that","existing limitations","other areas","one","vietnamese","two","vietnam","0.89","0.84","0.83","0.89","0.83","vietnamese","0.027","0.061","81%","82%","82% and","81%"]},{"title":"Basal Cell Carcinoma Diagnosis with Fusion of Deep Learning and Telangiectasia Features","authors":["Akanksha Maurya","R. Joe Stanley","Hemanth Y. Aradhyula","Norsang Lama","Anand K. Nambisan","Gehana Patel","Daniyal Saeed","Samantha Swinfard","Colin Smith","Sadhika Jagannathan","Jason R. Hagerty","William V. Stoecker"],"abstract":"In recent years, deep learning (DL) has been used extensively and successfully to diagnose different cancers in dermoscopic images. However, most approaches lack clinical inputs supported by dermatologists that could aid in higher accuracy and explainability. To dermatologists, the presence of telangiectasia, or narrow blood vessels that typically appear serpiginous or arborizing, is a critical indicator of basal cell carcinoma (BCC). Exploiting the feature information present in telangiectasia through a combination of DL-based techniques could create a pathway for both, improving DL results as well as aiding dermatologists in BCC diagnosis. This study demonstrates a novel \u201cfusion\u201d technique for BCC vs non-BCC classification using ensemble learning on a combination of (a) handcrafted features from semantically segmented telangiectasia (U-Net-based) and (b) deep learning features generated from whole lesion images (EfficientNet-B5-based). This fusion method achieves a binary classification accuracy of 97.2%, with a 1.3% improvement over the corresponding DL-only model, on a holdout test set of 395 images. An increase of 3.7% in sensitivity, 1.5% in specificity, and 1.5% in precision along with an AUC of 0.99 was also achieved. Metric improvements were demonstrated in three stages: (1) the addition of handcrafted telangiectasia features to deep learning features, (2) including areas near telangiectasia (surround areas), (3) discarding the noisy lower-importance features through feature importance. Another novel approach to feature finding with weak annotations through the examination of the surrounding areas of telangiectasia is offered in this study. The experimental results show state-of-the-art accuracy and precision in the diagnosis of BCC, compared to three benchmark techniques. Further exploration of deep learning techniques for individual dermoscopy feature detection is warranted.","doi":"10.1007\/s10278-024-00969-3","cleaned_title":"basal cell carcinoma diagnosis with fusion of deep learning and telangiectasia features","cleaned_abstract":"in recent years, deep learning (dl) has been used extensively and successfully to diagnose different cancers in dermoscopic images. however, most approaches lack clinical inputs supported by dermatologists that could aid in higher accuracy and explainability. to dermatologists, the presence of telangiectasia, or narrow blood vessels that typically appear serpiginous or arborizing, is a critical indicator of basal cell carcinoma (bcc). exploiting the feature information present in telangiectasia through a combination of dl-based techniques could create a pathway for both, improving dl results as well as aiding dermatologists in bcc diagnosis. this study demonstrates a novel \u201cfusion\u201d technique for bcc vs non-bcc classification using ensemble learning on a combination of (a) handcrafted features from semantically segmented telangiectasia (u-net-based) and (b) deep learning features generated from whole lesion images (efficientnet-b5-based). this fusion method achieves a binary classification accuracy of 97.2%, with a 1.3% improvement over the corresponding dl-only model, on a holdout test set of 395 images. an increase of 3.7% in sensitivity, 1.5% in specificity, and 1.5% in precision along with an auc of 0.99 was also achieved. metric improvements were demonstrated in three stages: (1) the addition of handcrafted telangiectasia features to deep learning features, (2) including areas near telangiectasia (surround areas), (3) discarding the noisy lower-importance features through feature importance. another novel approach to feature finding with weak annotations through the examination of the surrounding areas of telangiectasia is offered in this study. the experimental results show state-of-the-art accuracy and precision in the diagnosis of bcc, compared to three benchmark techniques. further exploration of deep learning techniques for individual dermoscopy feature detection is warranted.","key_phrases":["recent years","deep learning","dl","different cancers","dermoscopic images","most approaches","clinical inputs","dermatologists","that","higher accuracy","explainability","dermatologists","the presence","telangiectasia","narrow blood vessels","that","a critical indicator","basal cell carcinoma","(bcc","the feature information","telangiectasia","a combination","dl-based techniques","a pathway","both","dl results","dermatologists","bcc diagnosis","this study","a novel \u201cfusion\u201d technique","bcc","non-bcc classification","ensemble learning","a combination","(a) handcrafted features","semantically segmented telangiectasia","(b) deep learning features","whole lesion images","this fusion method","a binary classification accuracy","97.2%","a 1.3% improvement","the corresponding dl-only model","a holdout test","395 images","an increase","3.7%","sensitivity","1.5%","specificity","1.5%","precision","an auc","metric improvements","three stages","the addition","deep learning features","areas","telangiectasia","surround areas","the noisy lower-importance features","feature importance","another novel approach","weak annotations","the examination","the surrounding areas","telangiectasia","this study","the experimental results","the-art","precision","the diagnosis","bcc","three benchmark techniques","further exploration","deep learning techniques","individual dermoscopy feature detection","recent years","telangiectasia","bcc","telangiectasia","telangiectasia","97.2%","1.3%","395","3.7%","1.5%","1.5%","0.99","three","1","telangiectasia","2","telangiectasia","3","telangiectasia","bcc","three"]},{"title":"Semantic speech analysis using machine learning and deep learning techniques: a comprehensive review","authors":["Suryakant Tyagi","S\u00e1ndor Sz\u00e9n\u00e1si"],"abstract":"Human cognitive functions such as perception, attention, learning, memory, reasoning, and problem-solving are all significantly influenced by emotion. Emotion has a particularly potent impact on attention, modifying its selectivity in particular and influencing behavior and action motivation. Artificial Emotional Intelligence (AEI) technologies enable computers to understand a user's emotional state and respond appropriately. These systems enable a realistic dialogue between people and machines. The current generation of adaptive user interference technologies is built on techniques from data analytics and machine learning (ML), namely deep learning (DL) artificial neural networks (ANN) from multimodal data, such as videos of facial expressions, stance, and gesture, voice, and bio-physiological data (such as eye movement, ECG, respiration, EEG, FMRT, EMG, eye tracking). In this study, we reviewed existing literature based on ML and data analytics techniques being used to detect emotions in speech. The efficacy of data analytics and ML techniques in this unique area of multimodal data processing and extracting emotions from speech. This study analyzes how emotional chatbots, facial expressions, images, and social media texts can be effective in detecting emotions. PRISMA methodology is used to review the existing survey. Support Vector Machines (SVM), Na\u00efve Bayes (NB), Random Forests (RF), Recurrent Neural Networks (RNN), Logistic Regression (LR), etc., are commonly used ML techniques for emotion extraction purposes. This study provides a new taxonomy about the application of ML in SER. The result shows that Long-Short Term Memory (LSTM) and Convolutional Neural Networks (CNN) are found to be the most useful methodology for this purpose.","doi":"10.1007\/s11042-023-17769-6","cleaned_title":"semantic speech analysis using machine learning and deep learning techniques: a comprehensive review","cleaned_abstract":"human cognitive functions such as perception, attention, learning, memory, reasoning, and problem-solving are all significantly influenced by emotion. emotion has a particularly potent impact on attention, modifying its selectivity in particular and influencing behavior and action motivation. artificial emotional intelligence (aei) technologies enable computers to understand a user's emotional state and respond appropriately. these systems enable a realistic dialogue between people and machines. the current generation of adaptive user interference technologies is built on techniques from data analytics and machine learning (ml), namely deep learning (dl) artificial neural networks (ann) from multimodal data, such as videos of facial expressions, stance, and gesture, voice, and bio-physiological data (such as eye movement, ecg, respiration, eeg, fmrt, emg, eye tracking). in this study, we reviewed existing literature based on ml and data analytics techniques being used to detect emotions in speech. the efficacy of data analytics and ml techniques in this unique area of multimodal data processing and extracting emotions from speech. this study analyzes how emotional chatbots, facial expressions, images, and social media texts can be effective in detecting emotions. prisma methodology is used to review the existing survey. support vector machines (svm), na\u00efve bayes (nb), random forests (rf), recurrent neural networks (rnn), logistic regression (lr), etc., are commonly used ml techniques for emotion extraction purposes. this study provides a new taxonomy about the application of ml in ser. the result shows that long-short term memory (lstm) and convolutional neural networks (cnn) are found to be the most useful methodology for this purpose.","key_phrases":["human cognitive functions","perception","attention","learning","memory","reasoning","problem-solving","emotion","emotion","a particularly potent impact","attention","its selectivity","behavior","action motivation","artificial emotional intelligence (aei) technologies","computers","a user's emotional state","these systems","a realistic dialogue","people","machines","the current generation","adaptive user","interference technologies","techniques","data analytics","machine learning","ml","namely deep learning","(dl) artificial neural networks","ann","multimodal data","videos","facial expressions","stance","gesture","voice","bio-physiological data","eye movement","ecg","respiration","fmrt","emg","eye tracking","this study","we","existing literature","ml and data analytics techniques","emotions","speech","the efficacy","data analytics","ml","techniques","this unique area","multimodal data processing","emotions","speech","this study","emotional chatbots","facial expressions","images","social media texts","emotions","prisma methodology","the existing survey","support vector machines","svm","na\u00efve bayes","random forests","neural networks","rnn","logistic regression","lr","ml techniques","emotion extraction purposes","this study","a new taxonomy","the application","ml","ser","the result","long-short term memory","lstm","convolutional neural networks","cnn","the most useful methodology","this purpose","na\u00efve bayes (nb","cnn"]},{"title":"Detection of Image Tampering Using Deep Learning, Error Levels and Noise Residuals","authors":["Sunen Chakraborty","Kingshuk Chatterjee","Paramita Dey"],"abstract":"Images once were considered a reliable source of information. However, when photo-editing software started to get noticed it gave rise to illegal activities which is called image tampering. These days we can come across innumerable tampered images across the internet. Software such as Photoshop, GNU Image Manipulation Program, etc. are applied to form tampered images from real ones in just a few minutes. To discover hidden signs of tampering in an image deep learning models are an effective tool than any other methods. Models used in deep learning are capable of extracting intricate features from an image automatically. Here we proposed a combination of traditional handcrafted features along with a deep learning model to differentiate between authentic and tampered images. We have presented a dual-branch Convolutional Neural Network in conjunction with Error Level Analysis and noise residuals from Spatial Rich Model. For our experiment, we utilized the freely accessible CASIA dataset. After training the dual-branch network for 16 epochs, it generated an accuracy of 98.55%. We have also provided a comparative analysis with other previously proposed work in the field of image forgery detection. This hybrid approach proves that deep learning models along with some well-known traditional approaches can provide better results for detecting tampered images.","doi":"10.1007\/s11063-024-11448-9","cleaned_title":"detection of image tampering using deep learning, error levels and noise residuals","cleaned_abstract":"images once were considered a reliable source of information. however, when photo-editing software started to get noticed it gave rise to illegal activities which is called image tampering. these days we can come across innumerable tampered images across the internet. software such as photoshop, gnu image manipulation program, etc. are applied to form tampered images from real ones in just a few minutes. to discover hidden signs of tampering in an image deep learning models are an effective tool than any other methods. models used in deep learning are capable of extracting intricate features from an image automatically. here we proposed a combination of traditional handcrafted features along with a deep learning model to differentiate between authentic and tampered images. we have presented a dual-branch convolutional neural network in conjunction with error level analysis and noise residuals from spatial rich model. for our experiment, we utilized the freely accessible casia dataset. after training the dual-branch network for 16 epochs, it generated an accuracy of 98.55%. we have also provided a comparative analysis with other previously proposed work in the field of image forgery detection. this hybrid approach proves that deep learning models along with some well-known traditional approaches can provide better results for detecting tampered images.","key_phrases":["images","a reliable source","information","photo-editing software","it","rise","illegal activities","which","image tampering","we","innumerable tampered images","the internet","software","photoshop","gnu image manipulation program","tampered images","real ones","just a few minutes","hidden signs","an image","deep learning models","an effective tool","any other methods","models","deep learning","intricate features","an image","we","a combination","traditional handcrafted features","a deep learning model","authentic and tampered images","we","a dual-branch convolutional neural network","conjunction","error level analysis","noise residuals","spatial rich model","our experiment","we","the freely accessible casia dataset","the dual-branch network","16 epochs","it","an accuracy","98.55%","we","a comparative analysis","other previously proposed work","the field","image forgery detection","this hybrid approach","deep learning models","some well-known traditional approaches","better results","tampered images","these days","just a few minutes","16","98.55%"]},{"title":"Repetition Dynamics-based Deep Learning Model for Next Basket Recommendation","authors":["Kaushlendra Kumar Sinha","Somaraju Suvvari"],"abstract":"Next Basket Recommendation system analyzes users\u2019 past interactions to provide personalized recommendations. For a better understanding of the complex relations between users and objects and to handle large datasets, most recommendation system researchers present deep learning-based models. These deep learning based models considers users short-term preferences and long-term preferences for better performance, but these existing deep learning-based models neglected the repetition behavior of the user, as some researchers have highlighted the importance of repetition behavior in the literature. Towards this objective, we proposed a deep learning model that considers the user\u2019s repetition behavior, and also our model included the correlation dynamics at the embedding level. This means that the generated embeddings represent a combination of a user\u2019s repetition behavior and the correlation-based dynamics of items. Further, to extract the short-term preferences of the user, we fed these embeddings to an LSTM architecture, and finally, it generates suitable personalized recommendations. To evaluate the effectiveness of the proposed model, we tested our model on different real-world datasets by using extensive performance metrics, and the results are compared against state-of-the-art models. There was been considerable improvement in recall ranging from 42 to 99% over various values of k in Ta Feng and Dunnhumby dataset and considering precision there has been an improvement of 221\u2013400%.","doi":"10.1007\/s42979-023-02403-x","cleaned_title":"repetition dynamics-based deep learning model for next basket recommendation","cleaned_abstract":"next basket recommendation system analyzes users\u2019 past interactions to provide personalized recommendations. for a better understanding of the complex relations between users and objects and to handle large datasets, most recommendation system researchers present deep learning-based models. these deep learning based models considers users short-term preferences and long-term preferences for better performance, but these existing deep learning-based models neglected the repetition behavior of the user, as some researchers have highlighted the importance of repetition behavior in the literature. towards this objective, we proposed a deep learning model that considers the user\u2019s repetition behavior, and also our model included the correlation dynamics at the embedding level. this means that the generated embeddings represent a combination of a user\u2019s repetition behavior and the correlation-based dynamics of items. further, to extract the short-term preferences of the user, we fed these embeddings to an lstm architecture, and finally, it generates suitable personalized recommendations. to evaluate the effectiveness of the proposed model, we tested our model on different real-world datasets by using extensive performance metrics, and the results are compared against state-of-the-art models. there was been considerable improvement in recall ranging from 42 to 99% over various values of k in ta feng and dunnhumby dataset and considering precision there has been an improvement of 221\u2013400%.","key_phrases":["next basket recommendation system","users","interactions","personalized recommendations","a better understanding","the complex relations","users","objects","large datasets","most recommendation system researchers","deep learning-based models","these deep learning based models","users","better performance","these existing deep learning-based models","the repetition behavior","the user","some researchers","the importance","repetition behavior","the literature","this objective","we","a deep learning model","that","the user\u2019s repetition behavior","our model","the correlation dynamics","the embedding level","this","the generated embeddings","a combination","a user\u2019s repetition behavior","the correlation-based dynamics","items","the short-term preferences","the user","we","these embeddings","an lstm architecture","it","suitable personalized recommendations","the effectiveness","the proposed model","we","our model","different real-world datasets","extensive performance metrics","the results","the-art","considerable improvement","recall","42 to 99%","various values","k","ta feng","dunnhumby","precision","an improvement","221\u2013400%","42 to 99%","feng","221\u2013400%"]},{"title":"On the use of deep learning for phase recovery","authors":["Kaiqiang Wang","Li Song","Chutian Wang","Zhenbo Ren","Guangyuan Zhao","Jiazhen Dou","Jianglei Di","George Barbastathis","Renjie Zhou","Jianlin Zhao","Edmund Y. Lam"],"abstract":"Phase recovery (PR) refers to calculating the phase of the light field from its intensity measurements. As exemplified from quantitative phase imaging and coherent diffraction imaging to adaptive optics, PR is essential for reconstructing the refractive index distribution or topography of an object and correcting the aberration of an imaging system. In recent years, deep learning (DL), often implemented through deep neural networks, has provided unprecedented support for computational imaging, leading to more efficient solutions for various PR problems. In this review, we first briefly introduce conventional methods for PR. Then, we review how DL provides support for PR from the following three stages, namely, pre-processing, in-processing, and post-processing. We also review how DL is used in phase image processing. Finally, we summarize the work in DL for PR and provide an outlook on how to better use DL to improve the reliability and efficiency of PR. Furthermore, we present a live-updating resource (https:\/\/github.com\/kqwang\/phase-recovery) for readers to learn more about PR.","doi":"10.1038\/s41377-023-01340-x","cleaned_title":"on the use of deep learning for phase recovery","cleaned_abstract":"phase recovery (pr) refers to calculating the phase of the light field from its intensity measurements. as exemplified from quantitative phase imaging and coherent diffraction imaging to adaptive optics, pr is essential for reconstructing the refractive index distribution or topography of an object and correcting the aberration of an imaging system. in recent years, deep learning (dl), often implemented through deep neural networks, has provided unprecedented support for computational imaging, leading to more efficient solutions for various pr problems. in this review, we first briefly introduce conventional methods for pr. then, we review how dl provides support for pr from the following three stages, namely, pre-processing, in-processing, and post-processing. we also review how dl is used in phase image processing. finally, we summarize the work in dl for pr and provide an outlook on how to better use dl to improve the reliability and efficiency of pr. furthermore, we present a live-updating resource (https:\/\/github.com\/kqwang\/phase-recovery) for readers to learn more about pr.","key_phrases":["phase recovery","(pr","the phase","the light field","its intensity measurements","quantitative phase imaging","coherent diffraction","adaptive optics","the refractive index distribution","topography","an object","the aberration","an imaging system","recent years","deep learning","dl","deep neural networks","unprecedented support","computational imaging","more efficient solutions","various pr problems","this review","we","conventional methods","we","dl","support","pr","the following three stages","-processing","processing","we","dl","phase image processing","we","the work","dl","pr","an outlook","dl","the reliability","efficiency","pr","we","a live-updating resource","https:\/\/github.com\/kqwang\/phase-recovery","readers","pr","recent years","first","three"]},{"title":"Curriculum learning for ab initio deep learned refractive optics","authors":["Xinge Yang","Qiang Fu","Wolfgang Heidrich"],"abstract":"Deep optical optimization has recently emerged as a new paradigm for designing computational imaging systems using only the output image as the objective. However, it has been limited to either simple optical systems consisting of a single element such as a diffractive optical element or metalens, or the fine-tuning of compound lenses from good initial designs. Here we present a DeepLens design method based on curriculum learning, which is able to learn optical designs of compound lenses ab initio from randomly initialized surfaces without human intervention, therefore overcoming the need for a good initial design. We demonstrate the effectiveness of our approach by fully automatically designing both classical imaging lenses and a large field-of-view extended depth-of-field computational lens in a cellphone-style form factor, with highly aspheric surfaces and a short back focal length.","doi":"10.1038\/s41467-024-50835-7","cleaned_title":"curriculum learning for ab initio deep learned refractive optics","cleaned_abstract":"deep optical optimization has recently emerged as a new paradigm for designing computational imaging systems using only the output image as the objective. however, it has been limited to either simple optical systems consisting of a single element such as a diffractive optical element or metalens, or the fine-tuning of compound lenses from good initial designs. here we present a deeplens design method based on curriculum learning, which is able to learn optical designs of compound lenses ab initio from randomly initialized surfaces without human intervention, therefore overcoming the need for a good initial design. we demonstrate the effectiveness of our approach by fully automatically designing both classical imaging lenses and a large field-of-view extended depth-of-field computational lens in a cellphone-style form factor, with highly aspheric surfaces and a short back focal length.","key_phrases":["deep optical optimization","a new paradigm","computational imaging systems","only the output image","the objective","it","either simple optical systems","a single element","a diffractive optical element","metalens","the fine-tuning","compound lenses","good initial designs","we","a deeplens design method","curriculum learning","which","optical designs","compound lenses","ab initio","randomly initialized surfaces","human intervention","the need","a good initial design","we","the effectiveness","our approach","both classical imaging lenses","view","field","a cellphone-style form factor","highly aspheric surfaces","a short back focal length","deeplens"]},{"title":"A technique to forecast Pakistan\u2019s news using deep hybrid learning model","authors":["Rukhshanda Ihsan","Syed Khaldoon Khurshid","Muhammad Shoaib","Sadia Ali","Sana Mahnoor","Syed Muhammad Hamza"],"abstract":"Forecasting future events is a challenging task that can have a significant impact on decision-making and policy-making. In this research, we focus on forecasting news related to Pakistan. Despite the importance of accurate predictions in this field, there currently exists no dataset for forecasting Pakistani news, specifically with regards to politics. Unlike numerical time series data, textual data includes information about the event's potential causes in addition to its impact. Better forecasts are thus anticipated as a result of this greater information. In order to address this gap, our research aims to create a first Pakistani news dataset for forecasting of Pakistan news that is mostly related to politics of Pakistan. This dataset was collected from various sources, including Pakistani news websites and social media platforms, as well as frequently asked questions about Pakistani politics. We develop a forecasting model using this dataset and evaluate the effectiveness of cutting-edge deep hybrid learning techniques incorporating neural networks, random forest, Word2vec, Natural language processing (NLP), and Naive Bayes. To the best of our understanding, no research has been done on the application of a deep hybrid learning model\u2014a blend of deep learning and machine learning\u2014for news forecasting. The accuracy for forecasting model is 97%. According to our findings, the model's performance is adequate when compared to that of other forecasting models. Our research not only fills the gap in the current literature but also presents a new challenge for large language models and has the potential to bring significant practical advantages in the field of forecasting. The unique contribution of this study lies in the intelligent modeling of the prediction challenge, allowing for the utilization of text rich in content for forecasting objectives.","doi":"10.1007\/s41870-024-01781-6","cleaned_title":"a technique to forecast pakistan\u2019s news using deep hybrid learning model","cleaned_abstract":"forecasting future events is a challenging task that can have a significant impact on decision-making and policy-making. in this research, we focus on forecasting news related to pakistan. despite the importance of accurate predictions in this field, there currently exists no dataset for forecasting pakistani news, specifically with regards to politics. unlike numerical time series data, textual data includes information about the event's potential causes in addition to its impact. better forecasts are thus anticipated as a result of this greater information. in order to address this gap, our research aims to create a first pakistani news dataset for forecasting of pakistan news that is mostly related to politics of pakistan. this dataset was collected from various sources, including pakistani news websites and social media platforms, as well as frequently asked questions about pakistani politics. we develop a forecasting model using this dataset and evaluate the effectiveness of cutting-edge deep hybrid learning techniques incorporating neural networks, random forest, word2vec, natural language processing (nlp), and naive bayes. to the best of our understanding, no research has been done on the application of a deep hybrid learning model\u2014a blend of deep learning and machine learning\u2014for news forecasting. the accuracy for forecasting model is 97%. according to our findings, the model's performance is adequate when compared to that of other forecasting models. our research not only fills the gap in the current literature but also presents a new challenge for large language models and has the potential to bring significant practical advantages in the field of forecasting. the unique contribution of this study lies in the intelligent modeling of the prediction challenge, allowing for the utilization of text rich in content for forecasting objectives.","key_phrases":["future events","a challenging task","that","a significant impact","decision-making","policy-making","this research","we","news","pakistan","the importance","accurate predictions","this field","no dataset","pakistani news","regards","politics","numerical time series data","textual data","information","the event's potential causes","addition","its impact","better forecasts","a result","this greater information","order","this gap","our research","a first pakistani news dataset","forecasting","pakistan news","that","politics","pakistan","this dataset","various sources","pakistani news websites","social media platforms","questions","pakistani politics","we","a forecasting model","this dataset","the effectiveness","cutting-edge deep hybrid learning techniques","neural networks","random forest","word2vec","natural language processing","nlp","naive bayes","our understanding","no research","the application","a deep hybrid learning model","a blend","deep learning","machine learning","news forecasting","the accuracy","forecasting model","97%","our findings","the model's performance","that","other forecasting models","our research","the gap","the current literature","a new challenge","large language models","the potential","significant practical advantages","the field","forecasting","the unique contribution","this study","the intelligent modeling","the prediction challenge","the utilization","text","content","forecasting objectives","pakistan","pakistani","first","pakistani","pakistan","pakistan","pakistani","pakistani","97%"]},{"title":"A Transfer Learning-Based CNN Deep Learning Model for Unfavorable Driving State Recognition","authors":["Jichi Chen","Hong Wang","Enqiu He"],"abstract":"The detection of unfavorable driving states (UDS) of drivers based on electroencephalogram (EEG) measures has received continuous attention from extensive scholars on account of directly reflecting brain neural activity with high temporal resolution and low risk of being deceived. However, the existing EEG-based driver UDS detection methods involve limited exploration of the functional connectivity patterns and interaction relationships within the brain network. Therefore, there is still room for improvement in the accuracy of detection. In this project, we propose three pretrained convolutional neural network (CNN)-based automatic detection frameworks for UDS of drivers with 30-channel EEG signals. The frameworks are investigated by adjusting the learning rate and choosing the optimization solver, etc. Two different conditions of driving experiments are performed, collecting EEG signals from sixteen subjects. The acquired 1-dimensional 30-channel EEG signals are converted into 2-dimensional matrices by the Granger causality (GC) method to form the functional connectivity graphs of the brain (FCGB). Then, the FCGB are fed into pretrained deep learning models that employed transfer learning strategy for feature extraction and judgment of different EEG signal types. Furthermore, we adopt two visualization interpretability techniques, named, activation visualization and gradient-weighted class activation mapping (Grad-CAM) for better visualizing and understanding the predictions of the pretrained models after fine-tuning. The experimental outcomes show that Resnet 18 model yields the highest average recognition accuracy of 90% using the rmsprop optimizer with a learning rate of 1e\u2009\u2212\u20093. The overall outcomes suggest that cooperating of biologically inspired functional connectivity graphs of the brain and pretrained transfer learning algorithms is a prospective approach in reducing the rate of major traffic accidents caused by driver unfavorable driving states.","doi":"10.1007\/s12559-023-10196-7","cleaned_title":"a transfer learning-based cnn deep learning model for unfavorable driving state recognition","cleaned_abstract":"the detection of unfavorable driving states (uds) of drivers based on electroencephalogram (eeg) measures has received continuous attention from extensive scholars on account of directly reflecting brain neural activity with high temporal resolution and low risk of being deceived. however, the existing eeg-based driver uds detection methods involve limited exploration of the functional connectivity patterns and interaction relationships within the brain network. therefore, there is still room for improvement in the accuracy of detection. in this project, we propose three pretrained convolutional neural network (cnn)-based automatic detection frameworks for uds of drivers with 30-channel eeg signals. the frameworks are investigated by adjusting the learning rate and choosing the optimization solver, etc. two different conditions of driving experiments are performed, collecting eeg signals from sixteen subjects. the acquired 1-dimensional 30-channel eeg signals are converted into 2-dimensional matrices by the granger causality (gc) method to form the functional connectivity graphs of the brain (fcgb). then, the fcgb are fed into pretrained deep learning models that employed transfer learning strategy for feature extraction and judgment of different eeg signal types. furthermore, we adopt two visualization interpretability techniques, named, activation visualization and gradient-weighted class activation mapping (grad-cam) for better visualizing and understanding the predictions of the pretrained models after fine-tuning. the experimental outcomes show that resnet 18 model yields the highest average recognition accuracy of 90% using the rmsprop optimizer with a learning rate of 1e \u2212 3. the overall outcomes suggest that cooperating of biologically inspired functional connectivity graphs of the brain and pretrained transfer learning algorithms is a prospective approach in reducing the rate of major traffic accidents caused by driver unfavorable driving states.","key_phrases":["the detection","unfavorable driving states","uds","drivers","electroencephalogram (eeg) measures","continuous attention","extensive scholars","account","brain neural activity","high temporal resolution","low risk","the existing eeg-based driver uds detection methods","limited exploration","the functional connectivity patterns","interaction relationships","the brain network","room","improvement","the accuracy","detection","this project","we","three pretrained convolutional neural network","cnn)-based automatic detection frameworks","uds","drivers","30-channel eeg signals","the frameworks","the learning rate","the optimization","two different conditions","driving experiments","eeg signals","sixteen subjects","the acquired 1-dimensional 30-channel eeg signals","2-dimensional matrices","the granger causality","(gc) method","the functional connectivity graphs","the brain","fcgb","the fcgb","pretrained deep learning models","that","transfer learning strategy","feature extraction","judgment","different eeg signal types","we","two visualization interpretability techniques","grad-cam","better visualizing","the predictions","the pretrained models","fine-tuning","the experimental outcomes","resnet","18 model","the highest average recognition accuracy","90%","the rmsprop optimizer","a learning rate","the overall outcomes","biologically inspired functional connectivity graphs","the brain","pretrained transfer learning algorithms","a prospective approach","the rate","major traffic accidents","driver unfavorable driving states","three","30","two","1","30","2","fed","two","18","90%","1e \u2212","3"]},{"title":"Deep learning reconstruction for lumbar spine MRI acceleration: a prospective study","authors":["Hui Tang","Ming Hong","Lu Yu","Yang Song","Mengqiu Cao","Lei Xiang","Yan Zhou","Shiteng Suo"],"abstract":"BackgroundWe compared magnetic resonance imaging (MRI) turbo spin-echo images reconstructed using a deep learning technique (TSE-DL) with standard turbo spin-echo (TSE-SD) images of the lumbar spine regarding image quality and detection performance of common degenerative pathologies.MethodsThis prospective, single-center study included 31 patients (15 males and 16 females; aged 51\u2009\u00b1\u200916\u00a0years (mean\u2009\u00b1\u2009standard deviation)) who underwent lumbar spine exams with both TSE-SD and TSE-DL acquisitions for degenerative spine diseases. Images were analyzed by two radiologists and assessed for qualitative image quality using a 4-point Likert scale, quantitative signal-to-noise ratio (SNR) of anatomic landmarks, and detection of common pathologies. Paired-sample t, Wilcoxon, and McNemar tests, unweighted\/linearly weighted Cohen \u03ba statistics, and intraclass correlation coefficients were used.ResultsScan time for TSE-DL and TSE-SD protocols was 2:55 and 5:17\u00a0min:s, respectively. The overall image quality was either significantly higher for TSE-DL or not significantly different between TSE-SD and TSE-DL. TSE-DL demonstrated higher SNR and subject noise scores than TSE-SD. For pathology detection, the interreader agreement was substantial to almost perfect for TSE-DL, with \u03ba values ranging from 0.61 to 1.00; the interprotocol agreement was almost perfect for both readers, with \u03ba values ranging from 0.84 to 1.00. There was no significant difference in the diagnostic confidence or detection rate of common pathologies between the two sequences (p\u2009\u2265\u20090.081).ConclusionsTSE-DL allowed for a 45% reduction in scan time over TSE-SD in lumbar spine MRI without compromising the overall image quality and showed comparable detection performance of common pathologies in the evaluation of degenerative lumbar spine changes.Relevance statementDeep learning-reconstructed lumbar spine MRI protocol enabled a 45% reduction in scan time compared with conventional reconstruction, with comparable image quality and detection performance of common degenerative pathologies.Key points\u2022 Lumbar spine MRI with deep learning reconstruction has broad application prospects.\u2022 Deep learning reconstruction of lumbar spine MRI saved 45% scan time without compromising overall image quality.\u2022 When compared with standard sequences, deep learning reconstruction showed similar detection performance of common degenerative lumbar spine pathologies.Graphical Abstract","doi":"10.1186\/s41747-024-00470-0","cleaned_title":"deep learning reconstruction for lumbar spine mri acceleration: a prospective study","cleaned_abstract":"backgroundwe compared magnetic resonance imaging (mri) turbo spin-echo images reconstructed using a deep learning technique (tse-dl) with standard turbo spin-echo (tse-sd) images of the lumbar spine regarding image quality and detection performance of common degenerative pathologies.methodsthis prospective, single-center study included 31 patients (15 males and 16 females; aged 51 \u00b1 16 years (mean \u00b1 standard deviation)) who underwent lumbar spine exams with both tse-sd and tse-dl acquisitions for degenerative spine diseases. images were analyzed by two radiologists and assessed for qualitative image quality using a 4-point likert scale, quantitative signal-to-noise ratio (snr) of anatomic landmarks, and detection of common pathologies. paired-sample t, wilcoxon, and mcnemar tests, unweighted\/linearly weighted cohen \u03ba statistics, and intraclass correlation coefficients were used.resultsscan time for tse-dl and tse-sd protocols was 2:55 and 5:17 min:s, respectively. the overall image quality was either significantly higher for tse-dl or not significantly different between tse-sd and tse-dl. tse-dl demonstrated higher snr and subject noise scores than tse-sd. for pathology detection, the interreader agreement was substantial to almost perfect for tse-dl, with \u03ba values ranging from 0.61 to 1.00; the interprotocol agreement was almost perfect for both readers, with \u03ba values ranging from 0.84 to 1.00. there was no significant difference in the diagnostic confidence or detection rate of common pathologies between the two sequences (p \u2265 0.081).conclusionstse-dl allowed for a 45% reduction in scan time over tse-sd in lumbar spine mri without compromising the overall image quality and showed comparable detection performance of common pathologies in the evaluation of degenerative lumbar spine changes.relevance statementdeep learning-reconstructed lumbar spine mri protocol enabled a 45% reduction in scan time compared with conventional reconstruction, with comparable image quality and detection performance of common degenerative pathologies.key points\u2022 lumbar spine mri with deep learning reconstruction has broad application prospects.\u2022 deep learning reconstruction of lumbar spine mri saved 45% scan time without compromising overall image quality.\u2022 when compared with standard sequences, deep learning reconstruction showed similar detection performance of common degenerative lumbar spine pathologies.graphical abstract","key_phrases":["backgroundwe","mri","a deep learning technique","tse-dl","standard turbo spin-echo (tse-sd) images","the lumbar spine","image quality and detection performance","common degenerative pathologies.methodsthis prospective, single-center study","31 patients","15 males","16 females","51 \u00b1","\u00b1 standard deviation","who","lumbar spine exams","both tse-sd and tse-dl acquisitions","degenerative spine diseases","images","two radiologists","qualitative image quality","a 4-point likert scale","noise","snr","anatomic landmarks","detection","common pathologies","paired-sample t","wilcoxon","mcnemar tests","cohen \u03ba statistics","correlation coefficients","used.resultsscan time","tse-dl and tse-sd protocols","min",":","s","the overall image quality","tse-dl","tse-sd and tse-dl. tse-dl","higher snr and subject noise scores","tse-sd","pathology detection","the interreader agreement","tse-dl","\u03ba values","the interprotocol agreement","both readers","\u03ba values","no significant difference","the diagnostic confidence or detection rate","common pathologies","the two sequences","p \u2265","0.081).conclusionstse-dl","a 45% reduction","scan time","tse-sd","lumbar spine mri","the overall image quality","comparable detection performance","common pathologies","the evaluation","degenerative lumbar spine changes.relevance","learning-reconstructed lumbar spine mri protocol","a 45% reduction","scan time","conventional reconstruction","comparable image quality and detection performance","lumbar spine mri","deep learning reconstruction","broad application prospects.\u2022","reconstruction","lumbar spine mri","45% scan time","overall image","standard sequences","deep learning reconstruction","similar detection performance","common degenerative lumbar spine","31","15","16","51 \u00b1 16 years","two","4","mcnemar","used.resultsscan","2:55","5:17","0.61","1.00","0.84","1.00","two","45%","45%","45%"]},{"title":"A multimodal deep learning model for predicting severe hemorrhage in placenta previa","authors":["Munetoshi Akazawa","Kazunori Hashimoto"],"abstract":"Placenta previa causes life-threatening bleeding and accurate prediction of severe hemorrhage leads to risk stratification and optimum allocation of interventions. We aimed to use a multimodal deep learning model to predict severe hemorrhage. Using MRI T2-weighted image of the placenta and tabular data consisting of patient demographics and preoperative blood examination data, a multimodal deep learning model was constructed to predict cases of intraoperative blood loss\u2009>\u20092000\u00a0ml. We evaluated the prediction performance of the model by comparing it with that of two machine learning methods using only tabular data and MRI images, as well as with that of two human expert obstetricians. Among the enrolled 48 patients, 26 (54.2%) lost\u2009>\u20092000\u00a0ml of blood and 22 (45.8%) lost\u2009<\u20092000\u00a0ml of blood. Multimodal deep learning model showed the best accuracy of 0.68 and AUC of 0.74, whereas the machine learning model using tabular data and MRI images had a class accuracy of 0.61 and 0.53, respectively. The human experts had median accuracies of 0.61. Multimodal deep learning models could integrate the two types of information and predict severe hemorrhage cases. The model might assist human expert in the prediction of intraoperative hemorrhage in the case of placenta previa.","doi":"10.1038\/s41598-023-44634-1","cleaned_title":"a multimodal deep learning model for predicting severe hemorrhage in placenta previa","cleaned_abstract":"placenta previa causes life-threatening bleeding and accurate prediction of severe hemorrhage leads to risk stratification and optimum allocation of interventions. we aimed to use a multimodal deep learning model to predict severe hemorrhage. using mri t2-weighted image of the placenta and tabular data consisting of patient demographics and preoperative blood examination data, a multimodal deep learning model was constructed to predict cases of intraoperative blood loss > 2000 ml. we evaluated the prediction performance of the model by comparing it with that of two machine learning methods using only tabular data and mri images, as well as with that of two human expert obstetricians. among the enrolled 48 patients, 26 (54.2%) lost > 2000 ml of blood and 22 (45.8%) lost < 2000 ml of blood. multimodal deep learning model showed the best accuracy of 0.68 and auc of 0.74, whereas the machine learning model using tabular data and mri images had a class accuracy of 0.61 and 0.53, respectively. the human experts had median accuracies of 0.61. multimodal deep learning models could integrate the two types of information and predict severe hemorrhage cases. the model might assist human expert in the prediction of intraoperative hemorrhage in the case of placenta previa.","key_phrases":["placenta previa","life-threatening bleeding","accurate prediction","severe hemorrhage","risk stratification","optimum allocation","interventions","we","a multimodal deep learning model","severe hemorrhage","mri t2-weighted image","the placenta and tabular data","patient demographics","preoperative blood examination data","a multimodal deep learning model","cases","intraoperative blood loss","we","the prediction performance","the model","it","that","two machine learning methods","only tabular data and mri images","that","two human expert obstetricians","the enrolled 48 patients","54.2%","2000 ml","blood","22 (45.8%","2000 ml","blood","multimodal deep learning model","the best accuracy","auc","the machine learning model","tabular data and mri images","a class accuracy","the human experts","median accuracies","multimodal deep learning models","the two types","information","severe hemorrhage cases","the model","human expert","the prediction","intraoperative hemorrhage","the case","placenta previa","placenta previa","2000 ml","two","two","48","26","54.2%","2000 ml","22","45.8%","2000 ml","0.68","0.74","0.61","0.53","0.61","two"]},{"title":"IRADA: integrated reinforcement learning and deep learning algorithm for attack detection in wireless sensor networks","authors":["Vandana Shakya","Jaytrilok Choudhary","Dhirendra Pratap Singh"],"abstract":"Wireless Sensor Networks (WSNs) play a vital role in various applications, necessitating robust network security to protect sensitive data. Intrusion Detection Systems (IDSs) are crucial for preserving the integrity, availability, and confidentiality of WSNs by detecting and countering potential attacks. Despite significant research efforts, the existing IDS solutions still suffer from challenges related to detection accuracy and false alarms. To address these challenges, in this paper, we propose a Bayesian optimization-based Deep Learning (DL) model. However, the proposed optimized DL model, while showing promising results in enhancing security, encounters challenges such as data dependency, computational complexity, and the potential for overfitting. In the literature, researchers have employed Reinforcement Learning (RL) to address these issues. However, it also introduces its own concerns, including exploration, reward design, and prolonged training times. Consequently, to address these challenges, this paper proposes an Innovative Integrated RL-based Advanced DL Algorithm (IRADA) for attack detection in WSNs. IRADA leverages the convergence of DL and RL models to achieve superior intrusion detection performance. The performance analysis of IRADA reveals impressive results, including accuracy (99.50%), specificity (99.94%), sensitivity (99.48%), F1-Score (98.26%), Kappa statistics (99.42%), and area under the curve (99.38%). Additionally, we analyze IRADA\u2019s robustness against adversarial attacks, ensuring its applicability in real-world security scenarios.","doi":"10.1007\/s11042-024-18289-7","cleaned_title":"irada: integrated reinforcement learning and deep learning algorithm for attack detection in wireless sensor networks","cleaned_abstract":"wireless sensor networks (wsns) play a vital role in various applications, necessitating robust network security to protect sensitive data. intrusion detection systems (idss) are crucial for preserving the integrity, availability, and confidentiality of wsns by detecting and countering potential attacks. despite significant research efforts, the existing ids solutions still suffer from challenges related to detection accuracy and false alarms. to address these challenges, in this paper, we propose a bayesian optimization-based deep learning (dl) model. however, the proposed optimized dl model, while showing promising results in enhancing security, encounters challenges such as data dependency, computational complexity, and the potential for overfitting. in the literature, researchers have employed reinforcement learning (rl) to address these issues. however, it also introduces its own concerns, including exploration, reward design, and prolonged training times. consequently, to address these challenges, this paper proposes an innovative integrated rl-based advanced dl algorithm (irada) for attack detection in wsns. irada leverages the convergence of dl and rl models to achieve superior intrusion detection performance. the performance analysis of irada reveals impressive results, including accuracy (99.50%), specificity (99.94%), sensitivity (99.48%), f1-score (98.26%), kappa statistics (99.42%), and area under the curve (99.38%). additionally, we analyze irada\u2019s robustness against adversarial attacks, ensuring its applicability in real-world security scenarios.","key_phrases":["wireless sensor networks","wsns","a vital role","various applications","robust network security","sensitive data","intrusion detection systems","idss","the integrity","availability","confidentiality","wsns","potential attacks","significant research efforts","the existing ids solutions","challenges","detection accuracy","false alarms","these challenges","this paper","we","a bayesian optimization-based deep learning (dl) model","the proposed optimized dl model","promising results","security","challenges","data dependency","computational complexity","the potential","the literature","researchers","reinforcement learning","these issues","it","its own concerns","exploration","reward design","prolonged training times","these challenges","this paper","an innovative integrated rl-based advanced dl algorithm","irada","attack detection","wsns","irada","the convergence","dl and rl models","superior intrusion detection performance","the performance analysis","irada","impressive results","accuracy","99.50%","specificity","99.94%","sensitivity","99.48%","f1-score","98.26%","kappa statistics","99.42%","area","the curve","99.38%","we","irada\u2019s robustness","adversarial attacks","its applicability","real-world security scenarios","advanced dl algorithm","irada","irada","irada","99.50%","99.94%","99.48%","98.26%","99.42%","99.38%","irada"]},{"title":"A deep learning based approach for image retrieval extraction in mobile edge computing","authors":["Jamal Alasadi","Ghassan F. Bati","Ahmed Al Hilli"],"abstract":"Deep learning has been widely explored in 5G applications, including computer vision, the Internet of Things (IoT), and intermedia classification. However, applying the deep learning approach in limited-resource mobile devices is one of the most challenging issues. At the same time, users\u2019 experience in terms of Quality of Service (QoS) (e.g., service latency, outcome accuracy, and achievable data rate) performs poorly while interacting with machine learning applications. Mobile edge computing (MEC) has been introduced as a cooperative approach to bring computation resources in proximity to end-user devices to overcome these limitations. This article aims to design a novel image reiterative extraction algorithm based on convolution neural network (CNN) learning and computational task offloading to support machine learning-based mobile applications in resource-limited and uncertain environments. Accordingly, we leverage the framework of image retrieval extraction and introduce three approaches. First, privacy preservation is strict and aims to protect personal data. Second, network traffic reduction. Third, minimizing feature matching time. Our simulation results associated with real-time experiments on a small-scale MEC server have shown the effectiveness of the proposed deep learning-based approach over existing schemes. The source code is available here: https:\/\/github.com\/jamalalasadi\/CNN_Image_retrieval.","doi":"10.1007\/s43995-024-00060-6","cleaned_title":"a deep learning based approach for image retrieval extraction in mobile edge computing","cleaned_abstract":"deep learning has been widely explored in 5g applications, including computer vision, the internet of things (iot), and intermedia classification. however, applying the deep learning approach in limited-resource mobile devices is one of the most challenging issues. at the same time, users\u2019 experience in terms of quality of service (qos) (e.g., service latency, outcome accuracy, and achievable data rate) performs poorly while interacting with machine learning applications. mobile edge computing (mec) has been introduced as a cooperative approach to bring computation resources in proximity to end-user devices to overcome these limitations. this article aims to design a novel image reiterative extraction algorithm based on convolution neural network (cnn) learning and computational task offloading to support machine learning-based mobile applications in resource-limited and uncertain environments. accordingly, we leverage the framework of image retrieval extraction and introduce three approaches. first, privacy preservation is strict and aims to protect personal data. second, network traffic reduction. third, minimizing feature matching time. our simulation results associated with real-time experiments on a small-scale mec server have shown the effectiveness of the proposed deep learning-based approach over existing schemes. the source code is available here: https:\/\/github.com\/jamalalasadi\/cnn_image_retrieval.","key_phrases":["deep learning","5g applications","computer vision","the internet","things","iot","intermedia classification","the deep learning approach","limited-resource mobile devices","the most challenging issues","the same time","users\u2019 experience","terms","quality","service","qos","machine learning applications","mobile edge computing","(mec","a cooperative approach","computation resources","proximity","end-user devices","these limitations","this article","a novel image reiterative extraction algorithm","convolution neural network","(cnn) learning","computational task","machine learning-based mobile applications","resource-limited and uncertain environments","we","the framework","image retrieval extraction","three approaches","privacy preservation","personal data","feature matching time","our simulation results","real-time experiments","a small-scale mec server","the effectiveness","the proposed deep learning-based approach","existing schemes","the source code","https:\/\/github.com\/jamalalasadi\/cnn_image_retrieval","5","cnn","three","first","second","third"]},{"title":"Machine learning vs deep learning in stock market investment: an international evidence","authors":["Jing Hao","Feng He","Feng Ma","Shibo Zhang","Xiaotao Zhang"],"abstract":"Machine learning and deep learning are powerful tools for quantitative investment. To examine the effectiveness of the models in different markets, this paper applies random forest and DNN models to forecast stock prices and construct statistical arbitrage strategies in five stock markets, including mainland China, the United States, the United Kingdom, Canada and Japan. Each model is applied to the price of major stock indices constituting stocks in these markets from 2005 to 2020 to construct a long-short portfolio with 20 selected stocks by the model. The results show that the a particular model obtains significantly different profits in different markets, among which DNN has the best performance, especially in the Chinese stock market. We find that DNN models generally perform better than other machine learning models in all markets.","doi":"10.1007\/s10479-023-05286-6","cleaned_title":"machine learning vs deep learning in stock market investment: an international evidence","cleaned_abstract":"machine learning and deep learning are powerful tools for quantitative investment. to examine the effectiveness of the models in different markets, this paper applies random forest and dnn models to forecast stock prices and construct statistical arbitrage strategies in five stock markets, including mainland china, the united states, the united kingdom, canada and japan. each model is applied to the price of major stock indices constituting stocks in these markets from 2005 to 2020 to construct a long-short portfolio with 20 selected stocks by the model. the results show that the a particular model obtains significantly different profits in different markets, among which dnn has the best performance, especially in the chinese stock market. we find that dnn models generally perform better than other machine learning models in all markets.","key_phrases":["machine learning","deep learning","powerful tools","quantitative investment","the effectiveness","the models","different markets","this paper","random forest","dnn models","stock prices","statistical arbitrage strategies","five stock markets","mainland china","the united states","the united kingdom","canada","japan","each model","the price","major stock indices","stocks","these markets","a long-short portfolio","20 selected stocks","the model","the results","the a particular model","significantly different profits","different markets","which","dnn","the best performance","the chinese stock market","we","dnn models","other machine learning models","all markets","five","china","the united states","the united kingdom","canada","japan","2005","2020","20","chinese"]},{"title":"Deep learning-based automated angle measurement for flatfoot diagnosis in weight-bearing lateral radiographs","authors":["Won-Jun Noh","Mu Sook Lee","Byoung-Dai Lee"],"abstract":"This study aimed to develop and evaluate a deep learning-based system for the automatic measurement of angles (specifically, Meary\u2019s angle and calcaneal pitch) in weight-bearing lateral radiographs of the foot for flatfoot diagnosis. We utilized 3960 lateral radiographs, either from the left or right foot, sourced from a pool of 4000 patients to construct and evaluate a deep learning-based model. These radiographs were captured between June and November 2021, and patients who had undergone total ankle replacement surgery or ankle arthrodesis surgery were excluded. Various methods, including correlation analysis, Bland\u2013Altman plots, and paired T-tests, were employed to assess the concordance between the angles automatically measured using the system and those assessed by clinical experts. The evaluation dataset comprised 150 weight-bearing radiographs from 150 patients. In all test cases, the angles automatically computed using the deep learning-based system were in good agreement with the reference standards (Meary\u2019s angle: Pearson correlation coefficient (PCC)\u2009=\u20090.964, intraclass correlation coefficient (ICC)\u2009=\u20090.963, concordance correlation coefficient (CCC)\u2009=\u20090.963, p-value\u2009=\u20090.632, mean absolute error (MAE)\u2009=\u20091.59\u00b0; calcaneal pitch: PCC\u2009=\u20090.988, ICC\u2009=\u20090.987, CCC\u2009=\u20090.987, p-value\u2009=\u20090.055, MAE\u2009=\u20090.63\u00b0). The average time required for angle measurement using only the CPU to execute the deep learning-based system was 11\u2009\u00b1\u20091\u00a0s. The deep learning-based automatic angle measurement system, a tool for diagnosing flatfoot, demonstrated comparable accuracy and reliability with the results obtained by medical professionals for patients without internal fixation devices.","doi":"10.1038\/s41598-024-69549-3","cleaned_title":"deep learning-based automated angle measurement for flatfoot diagnosis in weight-bearing lateral radiographs","cleaned_abstract":"this study aimed to develop and evaluate a deep learning-based system for the automatic measurement of angles (specifically, meary\u2019s angle and calcaneal pitch) in weight-bearing lateral radiographs of the foot for flatfoot diagnosis. we utilized 3960 lateral radiographs, either from the left or right foot, sourced from a pool of 4000 patients to construct and evaluate a deep learning-based model. these radiographs were captured between june and november 2021, and patients who had undergone total ankle replacement surgery or ankle arthrodesis surgery were excluded. various methods, including correlation analysis, bland\u2013altman plots, and paired t-tests, were employed to assess the concordance between the angles automatically measured using the system and those assessed by clinical experts. the evaluation dataset comprised 150 weight-bearing radiographs from 150 patients. in all test cases, the angles automatically computed using the deep learning-based system were in good agreement with the reference standards (meary\u2019s angle: pearson correlation coefficient (pcc) = 0.964, intraclass correlation coefficient (icc) = 0.963, concordance correlation coefficient (ccc) = 0.963, p-value = 0.632, mean absolute error (mae) = 1.59\u00b0; calcaneal pitch: pcc = 0.988, icc = 0.987, ccc = 0.987, p-value = 0.055, mae = 0.63\u00b0). the average time required for angle measurement using only the cpu to execute the deep learning-based system was 11 \u00b1 1 s. the deep learning-based automatic angle measurement system, a tool for diagnosing flatfoot, demonstrated comparable accuracy and reliability with the results obtained by medical professionals for patients without internal fixation devices.","key_phrases":["this study","a deep learning-based system","the automatic measurement","angles","(specifically, meary\u2019s angle","calcaneal pitch","weight-bearing lateral radiographs","the foot","flatfoot diagnosis","we","3960 lateral radiographs","the left or right foot","a pool","4000 patients","a deep learning-based model","these radiographs","june","november","patients","who","total ankle replacement surgery","ankle arthrodesis surgery","various methods","correlation analysis","bland\u2013altman plots","t-tests","the concordance","the angles","the system","those","clinical experts","the evaluation dataset","150 weight-bearing","150 patients","all test cases","the angles","the deep learning-based system","good agreement","the reference standards","meary\u2019s angle","pcc","intraclass correlation coefficient","icc","concordance correlation coefficient","ccc","absolute error","mae","1.59\u00b0","calcaneal pitch","the average time","angle measurement","only the cpu","the deep learning-based system","11 \u00b1","the deep learning-based automatic angle measurement system","a tool","comparable accuracy","reliability","the results","medical professionals","patients","internal fixation devices","3960","4000","between june and november 2021","150","150","0.964","0.963","0.963","0.632","1.59","0.988","0.987","0.987","0.055","0.63","11","1"]},{"title":"Deep-learning based supervisory monitoring of robotized DE-GMAW process through learning from human welders","authors":["Rui Yu","Yue Cao","Jennifer Martin","Otto Chiang","YuMing Zhang"],"abstract":"Double-electrode gas metal arc welding (DE-GMAW) modifies GMAW by adding a second electrode to bypass a portion of the current flowing from the wire. This reduces the current to, and the heat input on, the workpiece. Successful bypassing depends on the relative position of the bypass electrode to the continuously varying wire tip. To ensure proper operation, we propose robotizing the system using a follower robot to carry and adaptively adjust the bypass electrode. The primary information for monitoring this process is the arc image, which directly shows desired and undesired modes. However, developing a robust algorithm for processing the complex arc image is time-consuming and challenging. Employing a deep learning approach requires labeling numerous arc images for the corresponding DE-GMAW modes, which is not practically feasible. To introduce alternative labels, we analyze arc phenomena in various DE-GMAW modes and correlate them with distinct arc systems having varying voltages. These voltages serve as automatically derived labels to train the deep-learning network. The results demonstrated reliable process monitoring.","doi":"10.1007\/s40194-023-01635-y","cleaned_title":"deep-learning based supervisory monitoring of robotized de-gmaw process through learning from human welders","cleaned_abstract":"double-electrode gas metal arc welding (de-gmaw) modifies gmaw by adding a second electrode to bypass a portion of the current flowing from the wire. this reduces the current to, and the heat input on, the workpiece. successful bypassing depends on the relative position of the bypass electrode to the continuously varying wire tip. to ensure proper operation, we propose robotizing the system using a follower robot to carry and adaptively adjust the bypass electrode. the primary information for monitoring this process is the arc image, which directly shows desired and undesired modes. however, developing a robust algorithm for processing the complex arc image is time-consuming and challenging. employing a deep learning approach requires labeling numerous arc images for the corresponding de-gmaw modes, which is not practically feasible. to introduce alternative labels, we analyze arc phenomena in various de-gmaw modes and correlate them with distinct arc systems having varying voltages. these voltages serve as automatically derived labels to train the deep-learning network. the results demonstrated reliable process monitoring.","key_phrases":["double-electrode gas metal arc welding","de","-","gmaw","gmaw","a second electrode","a portion","the current","the wire","this","the heat input","the workpiece","successful bypassing","the relative position","the bypass","the continuously varying wire tip","proper operation","we","the system","a follower robot","the bypass","the primary information","this process","the arc image","which","desired and undesired modes","a robust algorithm","the complex arc image","a deep learning approach","numerous arc images","the corresponding de-gmaw modes","which","alternative labels","we","arc phenomena","various de-gmaw modes","them","distinct arc systems","varying voltages","these voltages","automatically derived labels","the deep-learning network","the results","reliable process monitoring","second"]},{"title":"Surface wave inversion with unknown number of soil layers based on a hybrid learning procedure of deep learning and genetic algorithm","authors":["Zan Zhou","Thomas Man-Hoi Lok","Wan-Huan Zhou"],"abstract":"Surface wave inversion is a key step in the application of surface waves to soil velocity profiling. Currently, a common practice for the process of inversion is that the number of soil layers is assumed to be known before using heuristic search algorithms to compute the shear wave velocity profile or the number of soil layers is considered as an optimization variable. However, an improper selection of the number of layers may lead to an incorrect shear wave velocity profile. In this study, a deep learning and genetic algorithm hybrid learning procedure is proposed to perform the surface wave inversion without the need to assume the number of soil layers. First, a deep neural network is adapted to learn from a large number of synthetic dispersion curves for inferring the layer number. Then, the shear-wave velocity profile is determined by a genetic algorithm with the known layer number. By applying this procedure to both simulated and real-world cases, the results indicate that the proposed method is reliable and efficient for surface wave inversion.","doi":"10.1007\/s11803-024-2240-1","cleaned_title":"surface wave inversion with unknown number of soil layers based on a hybrid learning procedure of deep learning and genetic algorithm","cleaned_abstract":"surface wave inversion is a key step in the application of surface waves to soil velocity profiling. currently, a common practice for the process of inversion is that the number of soil layers is assumed to be known before using heuristic search algorithms to compute the shear wave velocity profile or the number of soil layers is considered as an optimization variable. however, an improper selection of the number of layers may lead to an incorrect shear wave velocity profile. in this study, a deep learning and genetic algorithm hybrid learning procedure is proposed to perform the surface wave inversion without the need to assume the number of soil layers. first, a deep neural network is adapted to learn from a large number of synthetic dispersion curves for inferring the layer number. then, the shear-wave velocity profile is determined by a genetic algorithm with the known layer number. by applying this procedure to both simulated and real-world cases, the results indicate that the proposed method is reliable and efficient for surface wave inversion.","key_phrases":["surface wave inversion","a key step","the application","surface waves","velocity profiling","a common practice","the process","inversion","the number","soil layers","heuristic search algorithms","the shear wave velocity profile","the number","soil layers","an optimization variable","an improper selection","the number","layers","an incorrect shear wave velocity profile","this study","a deep learning and genetic algorithm hybrid learning procedure","the surface wave inversion","the need","the number","soil layers","a deep neural network","a large number","synthetic dispersion curves","the layer number","the shear-wave velocity profile","a genetic algorithm","the known layer number","this procedure","both simulated and real-world cases","the results","the proposed method","surface wave inversion","first"]},{"title":"Presenting a three layer stacking ensemble classifier of deep learning and machine learning for skin cancer classification","authors":["Bahman Jafari Tabaghsar","Reza Tavoli","Mohammad Mahdi Alizadeh Toosi"],"abstract":"One of the most common types of cancer in the world is skin cancer. Despite the different types of skin diseases with different shapes, the classification of skin diseases is a very difficult task. As a result, considering such a problem, a combination model of deep learning algorithms and machine has been proposed for skin disease classification.In this paper, a three-layer architecture based on ensemble learning is presented. In the first layer, the training input is given to convolutional neural network and EfficientNET. The output of the first layer is given to the classifiers of the second layer including machine learning classifiers. The output of the best decision of these classifiers is sent to the third layer classifier and the final prediction is made.The reason for using the three-layer architecture based on group learning was the lack of correct recognition of some classes by simple classifications. On the other hand, some diseases with different classes are classified in the same class. This model helps to correctly identify input samples with the correct combination of classifications in different layers.HAM10000 data set has been used to test and validate the proposed method. The mentioned dataset includes 10,015 images of skin lesions in seven different classes and includes different types of skin diseases. The accuracy is 99.97 on the testing set, which was much better than the previous heavy models.","doi":"10.1007\/s11042-024-19195-8","cleaned_title":"presenting a three layer stacking ensemble classifier of deep learning and machine learning for skin cancer classification","cleaned_abstract":"one of the most common types of cancer in the world is skin cancer. despite the different types of skin diseases with different shapes, the classification of skin diseases is a very difficult task. as a result, considering such a problem, a combination model of deep learning algorithms and machine has been proposed for skin disease classification.in this paper, a three-layer architecture based on ensemble learning is presented. in the first layer, the training input is given to convolutional neural network and efficientnet. the output of the first layer is given to the classifiers of the second layer including machine learning classifiers. the output of the best decision of these classifiers is sent to the third layer classifier and the final prediction is made.the reason for using the three-layer architecture based on group learning was the lack of correct recognition of some classes by simple classifications. on the other hand, some diseases with different classes are classified in the same class. this model helps to correctly identify input samples with the correct combination of classifications in different layers.ham10000 data set has been used to test and validate the proposed method. the mentioned dataset includes 10,015 images of skin lesions in seven different classes and includes different types of skin diseases. the accuracy is 99.97 on the testing set, which was much better than the previous heavy models.","key_phrases":["the most common types","cancer","the world","skin cancer","the different types","skin diseases","different shapes","the classification","skin diseases","a very difficult task","a result","such a problem","a combination model","deep learning algorithms","machine","a three-layer architecture","ensemble learning","the first layer","the training input","convolutional neural network","efficientnet","the output","the first layer","the classifiers","the second layer","machine learning classifiers","the output","the best decision","these classifiers","the third layer classifier","the final prediction","made.the reason","the three-layer architecture","group learning","the lack","correct recognition","some classes","simple classifications","the other hand","some diseases","different classes","the same class","this model","input samples","the correct combination","classifications","different layers.ham10000 data set","the proposed method","the mentioned dataset","10,015 images","skin lesions","seven different classes","different types","skin diseases","the accuracy","the testing set","which","the previous heavy models","one","three","first","first","second","third","three","10,015","seven","99.97"]},{"title":"RRS: Review-Based Recommendation System Using Deep Learning for Vietnamese","authors":["Minh Hoang Nguyen","Thuat Thien Nguyen","Minh Nhat Ta","Tien Minh Nguyen","Kiet Van Nguyen"],"abstract":"Tourism, which includes sightseeing, relaxation, and discovery, is a fundamental aspect of human life. One of the most critical considerations for traveling is accommodation, mainly hotels. To improve the travel experience, we have presented a solution for building a recommendation model using Vietnamese and user data to suggest travelers choose the ideal hotel. Our data was collected from two well-known websites, Traveloka and Ivivu, and includes information about hotels in Vietnam and users\u2019 feedback history, such as comments, ratings, and the names of users and hotels. We then preprocessed and labeled the inter-annotator agreement for various aspects, including service (0.89), infrastructure (0.84), sanitary (0.83), location (0.89), and attitude (0.83). Our recommendation model is built by using Collaborative Filtering and deep learning techniques. Furthermore, we suggest incorporating context vectors from tourists\u2019 Vietnamese comments in the recommendation process. The context model is developed by using deep learning techniques to extract topics and sentiments from the words effectively. The results of our proposed model, as measured by the MSE, were 0.027, which is significantly better than a context-free model using the same parameters, which had an MSE of 0.061. Additionally, our Deep Learning, created using PhoBERT embedding, had an accuracy of 81% for topic classification and 82% for sentiment classification. The FastText based model had 82% and 81% accuracy for topic and sentiment, respectively. Our research demonstrates that our approach improves the accuracy of the recommendation model and has the potential for further development in the future. This idea can introduce a new Recommendation System that can overcome existing limitations and apply to other areas.","doi":"10.1007\/s42979-024-02812-6","cleaned_title":"rrs: review-based recommendation system using deep learning for vietnamese","cleaned_abstract":"tourism, which includes sightseeing, relaxation, and discovery, is a fundamental aspect of human life. one of the most critical considerations for traveling is accommodation, mainly hotels. to improve the travel experience, we have presented a solution for building a recommendation model using vietnamese and user data to suggest travelers choose the ideal hotel. our data was collected from two well-known websites, traveloka and ivivu, and includes information about hotels in vietnam and users\u2019 feedback history, such as comments, ratings, and the names of users and hotels. we then preprocessed and labeled the inter-annotator agreement for various aspects, including service (0.89), infrastructure (0.84), sanitary (0.83), location (0.89), and attitude (0.83). our recommendation model is built by using collaborative filtering and deep learning techniques. furthermore, we suggest incorporating context vectors from tourists\u2019 vietnamese comments in the recommendation process. the context model is developed by using deep learning techniques to extract topics and sentiments from the words effectively. the results of our proposed model, as measured by the mse, were 0.027, which is significantly better than a context-free model using the same parameters, which had an mse of 0.061. additionally, our deep learning, created using phobert embedding, had an accuracy of 81% for topic classification and 82% for sentiment classification. the fasttext based model had 82% and 81% accuracy for topic and sentiment, respectively. our research demonstrates that our approach improves the accuracy of the recommendation model and has the potential for further development in the future. this idea can introduce a new recommendation system that can overcome existing limitations and apply to other areas.","key_phrases":["tourism","which","sightseeing","relaxation","discovery","a fundamental aspect","human life","the most critical considerations","accommodation",", mainly hotels","the travel experience","we","a solution","a recommendation model","user data","travelers","the ideal hotel","our data","two well-known websites","traveloka","ivivu","information","hotels","vietnam","users\u2019 feedback history","comments","ratings","the names","users","hotels","we","the inter-annotator agreement","various aspects","service","infrastructure","location","attitude","our recommendation model","collaborative filtering","deep learning techniques","we","context vectors","tourists\u2019 vietnamese comments","the recommendation process","the context model","deep learning techniques","topics","sentiments","the words","the results","our proposed model","the mse","which","a context-free model","the same parameters","which","an mse","our deep learning","phobert","an accuracy","81%","topic classification","82%","sentiment classification","the fasttext based model","82%","81% accuracy","topic","sentiment","our research","our approach","the accuracy","the recommendation model","the potential","further development","the future","this idea","a new recommendation system","that","existing limitations","other areas","one","vietnamese","two","vietnam","0.89","0.84","0.83","0.89","0.83","vietnamese","0.027","0.061","81%","82%","82% and","81%"]},{"title":"An Extensive Review on Deep Learning and Machine Learning Intervention in Prediction and Classification of Types of Aneurysms","authors":["Renugadevi Ammapalayam Sinnaswamy","Natesan Palanisamy","Kavitha Subramaniam","Suresh Muthusamy","Ravita Lamba","Sreejith Sekaran"],"abstract":"Aneurysm (Rupture of blood vessels) may happen in the cerebrum, abdominal aorta and thoracic aorta of humans, which has a high fatal rate. The advancement of the artificial technologies specifically machine learning algorithms and deep learning models have attempted to predict the aneurysm, which may reduce the death rate. The main objective of this paper is to provide the review of various algorithms and models for the early prediction of the various types of aneurysms. The focused literature review was conducted from the preferred journals from 2007 to 2022 on various parameters such as way of collecting images, the techniques used, number of images used in data set, performance metrics and future work. The summarized overview of advances in prediction of aneurysms using the machine learning algorithms from non linear kernel support regression algorithm to 3D Unet architecture of deep learning models starting from CT scan images to final performance analysis in prediction. The range of sensitivity, specificity and area under receiving operating characteristic was from 0. 7 to 1 for the abdominal aortic aneurysm detection, intracranial aneurysm detection. The thoracic aortic aneurysm was not concentrated much in the literature review, so the prediction of thoracic aortic aneurysm using machine learning as well as deep learning model is recommended.","doi":"10.1007\/s11277-023-10532-y","cleaned_title":"an extensive review on deep learning and machine learning intervention in prediction and classification of types of aneurysms","cleaned_abstract":"aneurysm (rupture of blood vessels) may happen in the cerebrum, abdominal aorta and thoracic aorta of humans, which has a high fatal rate. the advancement of the artificial technologies specifically machine learning algorithms and deep learning models have attempted to predict the aneurysm, which may reduce the death rate. the main objective of this paper is to provide the review of various algorithms and models for the early prediction of the various types of aneurysms. the focused literature review was conducted from the preferred journals from 2007 to 2022 on various parameters such as way of collecting images, the techniques used, number of images used in data set, performance metrics and future work. the summarized overview of advances in prediction of aneurysms using the machine learning algorithms from non linear kernel support regression algorithm to 3d unet architecture of deep learning models starting from ct scan images to final performance analysis in prediction. the range of sensitivity, specificity and area under receiving operating characteristic was from 0. 7 to 1 for the abdominal aortic aneurysm detection, intracranial aneurysm detection. the thoracic aortic aneurysm was not concentrated much in the literature review, so the prediction of thoracic aortic aneurysm using machine learning as well as deep learning model is recommended.","key_phrases":["aneurysm","rupture","blood vessels","the cerebrum","abdominal aorta","aorta","humans","which","a high fatal rate","the advancement","the artificial technologies","specifically machine learning algorithms","deep learning models","the aneurysm","which","the death rate","the main objective","this paper","the review","various algorithms","models","the early prediction","the various types","aneurysms","the focused literature review","the preferred journals","various parameters","way","images","the techniques","images","data set","performance metrics","future work","the summarized overview","advances","prediction","aneurysms","the machine learning algorithms","non linear kernel support regression algorithm","3d unet architecture","deep learning models","ct scan images","final performance analysis","prediction","the range","sensitivity","specificity","area","operating characteristic","the abdominal aortic aneurysm detection","intracranial aneurysm detection","the thoracic aortic aneurysm","the literature review","the prediction","thoracic aortic aneurysm","machine learning","deep learning model","2007","2022","3d","scan","0","7 to 1","intracranial aneurysm detection","the thoracic aortic aneurysm"]},{"title":"Machine learning and deep learning techniques for breast cancer diagnosis and classification: a comprehensive review of medical imaging studies","authors":["Mehran Radak","Haider Yabr Lafta","Hossein Fallahi"],"abstract":"BackgroundBreast cancer is a major public health concern, and early diagnosis and classification are critical for effective treatment. Machine learning and deep learning techniques have shown great promise in the classification and diagnosis of breast cancer.PurposeIn this review, we examine studies that have used these techniques for breast cancer classification and diagnosis, focusing on five groups of medical images: mammography, ultrasound, MRI, histology, and thermography. We discuss the use of five popular machine learning techniques, including Nearest Neighbor, SVM, Naive Bayesian Network, DT, and ANN, as well as deep learning architectures and convolutional neural networks.ConclusionOur review finds that machine learning and deep learning techniques have achieved high accuracy rates in breast cancer classification and diagnosis across various medical imaging modalities. Furthermore, these techniques have the potential to improve clinical decision-making and ultimately lead to better patient outcomes.","doi":"10.1007\/s00432-023-04956-z","cleaned_title":"machine learning and deep learning techniques for breast cancer diagnosis and classification: a comprehensive review of medical imaging studies","cleaned_abstract":"backgroundbreast cancer is a major public health concern, and early diagnosis and classification are critical for effective treatment. machine learning and deep learning techniques have shown great promise in the classification and diagnosis of breast cancer.purposein this review, we examine studies that have used these techniques for breast cancer classification and diagnosis, focusing on five groups of medical images: mammography, ultrasound, mri, histology, and thermography. we discuss the use of five popular machine learning techniques, including nearest neighbor, svm, naive bayesian network, dt, and ann, as well as deep learning architectures and convolutional neural networks.conclusionour review finds that machine learning and deep learning techniques have achieved high accuracy rates in breast cancer classification and diagnosis across various medical imaging modalities. furthermore, these techniques have the potential to improve clinical decision-making and ultimately lead to better patient outcomes.","key_phrases":["backgroundbreast cancer","a major public health concern","early diagnosis","classification","effective treatment","machine learning","deep learning techniques","great promise","the classification","diagnosis","breast cancer.purposein","this review","we","studies","that","these techniques","breast cancer classification","diagnosis","five groups","medical images","mammography","ultrasound","mri","histology","thermography","we","the use","five popular machine learning techniques","neighbor","svm","naive bayesian network","dt","ann","deep learning architectures","convolutional neural networks.conclusionour review","machine learning","deep learning techniques","high accuracy rates","breast cancer classification","diagnosis","various medical imaging modalities","these techniques","the potential","clinical decision-making","better patient outcomes","five","five"]},{"title":"SAKMR: Industrial control anomaly detection based on semi-supervised hybrid deep learning","authors":["Shijie Tang","Yong Ding","Meng Zhao","Huiyong Wang"],"abstract":"With the advent of Industry 4.0, industrial control systems (ICS) are more and more closely connected with the Internet, leading to a rapid increase in the types and quantities of security threats that arise from ICS. Anomaly detection is an effective defense measure against attacks. At present, it is the main trend to use hybrid deep learning methods to realize ICS anomaly detection. However, we found that many ICS anomaly detection methods based on hybrid deep learning adopt phased learning, in which each phase is optimized separately with optimization goals deviating from the overall goal. In view of this issue, we propose an end-to-end anomaly detection method SAKMR based on hybrid deep learning. Our method uses radial basis function network (RBFN) to realize K-means clustering, and combines it with stacked auto-encoder (SAE), which is conducive to defining reconstruction error and clustering error into an objective function to ensure joint optimization of feature extraction and classification. Experiments were conducted on the commonly used KDDCUP99 and SWAT datasets. The results show that SAKMR is effective in detecting abnormal industrial control data and outperforms the baseline methods on multiple performance indicators such as F1-Measure.","doi":"10.1007\/s12083-023-01586-7","cleaned_title":"sakmr: industrial control anomaly detection based on semi-supervised hybrid deep learning","cleaned_abstract":"with the advent of industry 4.0, industrial control systems (ics) are more and more closely connected with the internet, leading to a rapid increase in the types and quantities of security threats that arise from ics. anomaly detection is an effective defense measure against attacks. at present, it is the main trend to use hybrid deep learning methods to realize ics anomaly detection. however, we found that many ics anomaly detection methods based on hybrid deep learning adopt phased learning, in which each phase is optimized separately with optimization goals deviating from the overall goal. in view of this issue, we propose an end-to-end anomaly detection method sakmr based on hybrid deep learning. our method uses radial basis function network (rbfn) to realize k-means clustering, and combines it with stacked auto-encoder (sae), which is conducive to defining reconstruction error and clustering error into an objective function to ensure joint optimization of feature extraction and classification. experiments were conducted on the commonly used kddcup99 and swat datasets. the results show that sakmr is effective in detecting abnormal industrial control data and outperforms the baseline methods on multiple performance indicators such as f1-measure.","key_phrases":["the advent","industry","industrial control systems","ics","the internet","a rapid increase","the types","quantities","security threats","that","ics","anomaly detection","an effective defense measure","attacks","present","it","the main trend","hybrid deep learning methods","ics anomaly detection","we","that many ics anomaly detection methods","hybrid deep learning adopt phased learning","which","each phase","optimization goals","the overall goal","view","this issue","we","end","hybrid deep learning","our method","radial basis function network","rbfn","k","it","stacked auto-encoder","sae","which","reconstruction error","error","an objective function","joint optimization","feature extraction","classification","experiments","the commonly used kddcup99","swat datasets","the results","sakmr","abnormal industrial control data","the baseline methods","multiple performance indicators","f1-measure","4.0","anomaly detection","anomaly detection method","kddcup99"]},{"title":"Deep learning for the harmonization of structural MRI scans: a survey","authors":["Soolmaz Abbasi","Haoyu Lan","Jeiran Choupan","Nasim Sheikh-Bahaei","Gaurav Pandey","Bino Varghese"],"abstract":"Medical imaging datasets for research are frequently collected from multiple imaging centers using different scanners, protocols, and settings. These variations affect data consistency and compatibility across different sources. Image harmonization is a critical step to mitigate the effects of factors like inherent differences between various vendors, hardware upgrades, protocol changes, and scanner calibration drift, as well as to ensure consistent data for medical image processing techniques. Given the critical importance and widespread relevance of this issue, a vast array of image harmonization methodologies have emerged, with deep learning-based approaches driving substantial advancements in recent times. The goal of this review paper is to examine the latest deep learning techniques employed for image harmonization by analyzing cutting-edge architectural approaches in the field of medical image harmonization, evaluating both their strengths and limitations. This paper begins by providing a comprehensive fundamental overview of image harmonization strategies, covering three critical aspects: established imaging datasets, commonly used evaluation metrics, and characteristics of different scanners. Subsequently, this paper analyzes recent structural MRI (Magnetic Resonance Imaging) harmonization techniques based on network architecture, network learning algorithm, network supervision strategy, and network output. The underlying architectures include U-Net, Generative Adversarial Networks (GANs), Variational Autoencoders (VAEs), flow-based generative models, transformer-based approaches, as well as custom-designed network architectures. This paper investigates the effectiveness of Disentangled Representation Learning (DRL) as a pivotal learning algorithm in harmonization. Lastly, the review highlights the primary limitations in harmonization techniques, specifically the lack of comprehensive quantitative comparisons across different methods. The overall aim of this review is to serve as a guide for researchers and practitioners to select appropriate architectures based on their specific conditions and requirements. It also aims to foster discussions around ongoing challenges in the field and shed light on promising future research directions with the potential for significant advancements.","doi":"10.1186\/s12938-024-01280-6","cleaned_title":"deep learning for the harmonization of structural mri scans: a survey","cleaned_abstract":"medical imaging datasets for research are frequently collected from multiple imaging centers using different scanners, protocols, and settings. these variations affect data consistency and compatibility across different sources. image harmonization is a critical step to mitigate the effects of factors like inherent differences between various vendors, hardware upgrades, protocol changes, and scanner calibration drift, as well as to ensure consistent data for medical image processing techniques. given the critical importance and widespread relevance of this issue, a vast array of image harmonization methodologies have emerged, with deep learning-based approaches driving substantial advancements in recent times. the goal of this review paper is to examine the latest deep learning techniques employed for image harmonization by analyzing cutting-edge architectural approaches in the field of medical image harmonization, evaluating both their strengths and limitations. this paper begins by providing a comprehensive fundamental overview of image harmonization strategies, covering three critical aspects: established imaging datasets, commonly used evaluation metrics, and characteristics of different scanners. subsequently, this paper analyzes recent structural mri (magnetic resonance imaging) harmonization techniques based on network architecture, network learning algorithm, network supervision strategy, and network output. the underlying architectures include u-net, generative adversarial networks (gans), variational autoencoders (vaes), flow-based generative models, transformer-based approaches, as well as custom-designed network architectures. this paper investigates the effectiveness of disentangled representation learning (drl) as a pivotal learning algorithm in harmonization. lastly, the review highlights the primary limitations in harmonization techniques, specifically the lack of comprehensive quantitative comparisons across different methods. the overall aim of this review is to serve as a guide for researchers and practitioners to select appropriate architectures based on their specific conditions and requirements. it also aims to foster discussions around ongoing challenges in the field and shed light on promising future research directions with the potential for significant advancements.","key_phrases":["medical imaging datasets","research","multiple imaging centers","different scanners","protocols","settings","these variations","data consistency","compatibility","different sources","image harmonization","a critical step","the effects","factors","inherent differences","various vendors","hardware upgrades","protocol changes","scanner calibration drift","consistent data","medical image processing techniques","the critical importance","widespread relevance","this issue","a vast array","image harmonization methodologies","deep learning-based approaches","substantial advancements","recent times","the goal","this review paper","the latest deep learning techniques","image harmonization","cutting-edge architectural approaches","the field","medical image harmonization","both their strengths","limitations","this paper","a comprehensive fundamental overview","image harmonization strategies","three critical aspects","imaging datasets","commonly used evaluation metrics","characteristics","different scanners","this paper","recent structural mri","magnetic resonance imaging","harmonization techniques","network architecture","network learning algorithm","network supervision strategy","network output","the underlying architectures","u","-","net","generative adversarial networks","gans","variational autoencoders","flow-based generative models","transformer-based approaches","custom-designed network architectures","this paper","the effectiveness","drl","a pivotal learning algorithm","harmonization","the review","the primary limitations","harmonization techniques","specifically the lack","comprehensive quantitative comparisons","different methods","the overall aim","this review","a guide","researchers","practitioners","appropriate architectures","their specific conditions","requirements","it","discussions","ongoing challenges","the field","light","future research directions","the potential","significant advancements","three"]},{"title":"PolySeg Plus: Polyp Segmentation Using Deep Learning with Cost Effective Active Learning","authors":["Abdelrahman I. Saad","Fahima A. Maghraby","Osama Badawy"],"abstract":"A deep convolution neural network image segmentation model based on a cost-effective active learning mechanism is proposed and named PolySeg Plus. It is intended to address polyp segmentation with a lack of labeled data and a high false-positive rate of polyp discovery. In addition to applying active learning, which assisted in labeling more image samples, a comprehensive polyp dataset formed of five benchmark datasets was generated to increase the number of images. To enhance the captured image features, the locally shared feature method is used, which utilizes the power of employing neighboring features together with one another to improve the quality of image features and overcome the drawbacks of the Conditional Random Features method. Medical image segmentation was performed using ResUNet++, ResUNet, UNet++, and UNet models. Gaussian noise was removed from the images using a gaussian filter, and the images were then augmented before being fed into the models. In addition to optimizing model performance through hyperparameter tuning, grid search is used to select the optimum parameters to maximize model performance. The results demonstrated a significant improvement and applicability of the proposed method in polyp segmentation when compared to state-of-the-art methods on the datasets CVC-ClinicDB, CVC-ColonDB, ETIS Larib Polyp DB, KVASIR-SEG, and Kvasir-Sessile, with Dice coefficients of 0.9558, 0.8947, 0.7547, 0.9476, and 0.6023, respectively. Not only did the suggested method improve the dice coefficients on the individual datasets, but it also produced better results on the comprehensive dataset, which will contribute to the development of computer-aided diagnosis systems.","doi":"10.1007\/s44196-023-00330-6","cleaned_title":"polyseg plus: polyp segmentation using deep learning with cost effective active learning","cleaned_abstract":"a deep convolution neural network image segmentation model based on a cost-effective active learning mechanism is proposed and named polyseg plus. it is intended to address polyp segmentation with a lack of labeled data and a high false-positive rate of polyp discovery. in addition to applying active learning, which assisted in labeling more image samples, a comprehensive polyp dataset formed of five benchmark datasets was generated to increase the number of images. to enhance the captured image features, the locally shared feature method is used, which utilizes the power of employing neighboring features together with one another to improve the quality of image features and overcome the drawbacks of the conditional random features method. medical image segmentation was performed using resunet++, resunet, unet++, and unet models. gaussian noise was removed from the images using a gaussian filter, and the images were then augmented before being fed into the models. in addition to optimizing model performance through hyperparameter tuning, grid search is used to select the optimum parameters to maximize model performance. the results demonstrated a significant improvement and applicability of the proposed method in polyp segmentation when compared to state-of-the-art methods on the datasets cvc-clinicdb, cvc-colondb, etis larib polyp db, kvasir-seg, and kvasir-sessile, with dice coefficients of 0.9558, 0.8947, 0.7547, 0.9476, and 0.6023, respectively. not only did the suggested method improve the dice coefficients on the individual datasets, but it also produced better results on the comprehensive dataset, which will contribute to the development of computer-aided diagnosis systems.","key_phrases":["a deep convolution neural network image segmentation model","a cost-effective active learning mechanism","polyseg","it","polyp segmentation","a lack","labeled data","a high false-positive rate","polyp discovery","addition","active learning","which","more image samples","a comprehensive polyp dataset","five benchmark datasets","the number","images","the captured image features","the locally shared feature method","which","the power","neighboring features","the quality","image features","the drawbacks","the conditional random features method","medical image segmentation","resunet++","resunet","unet++","unet models","gaussian noise","the images","a gaussian filter","the images","the models","addition","model performance","hyperparameter tuning","grid search","the optimum parameters","model performance","the results","a significant improvement","applicability","the proposed method","polyp segmentation","the-art","the datasets cvc-clinicdb","cvc-colondb","etis larib polyp","kvasir-seg","kvasir-sessile","dice coefficients","the suggested method","the dice coefficients","the individual datasets","it","better results","the comprehensive dataset","which","the development","computer-aided diagnosis systems","polyseg","five","gaussian noise","0.9558","0.8947","0.7547","0.9476","0.6023"]},{"title":"Deep learning large-scale drug discovery and repurposing","authors":["Min Yu","Weiming Li","Yunru Yu","Yu Zhao","Lizhi Xiao","Volker M. Lauschke","Yiyu Cheng","Xingcai Zhang","Yi Wang"],"abstract":"Large-scale drug discovery and repurposing is challenging. Identifying the mechanism of action (MOA) is crucial, yet current approaches are costly and low-throughput. Here we present an approach for MOA identification by profiling changes in mitochondrial phenotypes. By temporally imaging mitochondrial morphology and membrane potential, we established a pipeline for monitoring time-resolved mitochondrial images, resulting in a dataset comprising 570,096 single-cell images of cells exposed to 1,068 United States Food and Drug Administration-approved drugs. A deep learning model named MitoReID, using a re-identification (ReID) framework and an Inflated 3D ResNet backbone, was developed. It achieved 76.32% Rank-1 and 65.92% mean average precision on the testing set and successfully identified the MOAs for six untrained drugs on the basis of mitochondrial phenotype. Furthermore, MitoReID identified cyclooxygenase-2 inhibition as the MOA of the natural compound epicatechin in tea, which was successfully validated in vitro. Our approach thus provides an automated and cost-effective alternative for target identification that could accelerate large-scale drug discovery and repurposing.","doi":"10.1038\/s43588-024-00679-4","cleaned_title":"deep learning large-scale drug discovery and repurposing","cleaned_abstract":"large-scale drug discovery and repurposing is challenging. identifying the mechanism of action (moa) is crucial, yet current approaches are costly and low-throughput. here we present an approach for moa identification by profiling changes in mitochondrial phenotypes. by temporally imaging mitochondrial morphology and membrane potential, we established a pipeline for monitoring time-resolved mitochondrial images, resulting in a dataset comprising 570,096 single-cell images of cells exposed to 1,068 united states food and drug administration-approved drugs. a deep learning model named mitoreid, using a re-identification (reid) framework and an inflated 3d resnet backbone, was developed. it achieved 76.32% rank-1 and 65.92% mean average precision on the testing set and successfully identified the moas for six untrained drugs on the basis of mitochondrial phenotype. furthermore, mitoreid identified cyclooxygenase-2 inhibition as the moa of the natural compound epicatechin in tea, which was successfully validated in vitro. our approach thus provides an automated and cost-effective alternative for target identification that could accelerate large-scale drug discovery and repurposing.","key_phrases":["large-scale drug discovery","repurposing","the mechanism","action","(moa","current approaches","we","an approach","moa identification","changes","mitochondrial phenotypes","mitochondrial morphology","membrane potential","we","a pipeline","time-resolved mitochondrial images","a dataset","570,096 single-cell images","cells","1,068 united states food and drug administration-approved drugs","a deep learning model","mitoreid","a re","-","identification (reid) framework","an inflated 3d resnet backbone","it","76.32%","rank-1","65.92%","the moas","six untrained drugs","the basis","mitochondrial phenotype","mitoreid","inhibition","the moa","the natural compound epicatechin","tea","which","our approach","an automated and cost-effective alternative","target identification","that","large-scale drug discovery","repurposing","570,096","1,068","united states","food and drug administration","mitoreid","3d","76.32%","65.92%","six","mitoreid"]},{"title":"CoVSeverity-Net: an efficient deep learning model for COVID-19 severity estimation from Chest X-Ray images","authors":["Sagar Deep Deb","Rajib Kumar Jha","Rajnish Kumar","Prem S. Tripathi","Yash Talera","Manish Kumar"],"abstract":"PurposeCOVID-19 is not going anywhere and is slowly becoming a part of our life. The World Health Organization declared it a pandemic in 2020, and it has affected all of us in many ways. Several deep learning techniques have been developed to detect COVID-19 from Chest X-Ray images. COVID-19 infection severity scoring can aid in establishing the optimum course of treatment and care for a positive patient, as all COVID-19 positive patients do not require special medical attention. Still, very few works are reported to estimate the severity of the disease from the Chest X-Ray images. The unavailability of the large-scale dataset might be a reason.MethodsWe aim to propose CoVSeverity-Net, a deep learning-based architecture for predicting the severity of COVID-19 from Chest X-ray images. CoVSeverity-Net is trained on a public COVID-19 dataset, curated by experienced radiologists for severity estimation. For that, a large publicly available dataset is collected and divided into three levels of severity, namely Mild, Moderate, and Severe.ResultsAn accuracy of 85.71% is reported. Conducting 5-fold cross-validation, we have obtained an accuracy of 87.82 \u00b1 6.25%. Similarly, conducting 10-fold cross-validation we obtained accuracy of 91.26 \u00b1 3.42. The results were better when compared with other state-of-the-art architectures.ConclusionWe strongly believe that this study has a high chance of reducing the workload of overworked front-line radiologists, speeding up patient diagnosis and treatment, and easing pandemic control. Future work would be to train a novel deep learning-based architecture on a larger dataset for severity estimation.","doi":"10.1007\/s42600-022-00254-8","cleaned_title":"covseverity-net: an efficient deep learning model for covid-19 severity estimation from chest x-ray images","cleaned_abstract":"purposecovid-19 is not going anywhere and is slowly becoming a part of our life. the world health organization declared it a pandemic in 2020, and it has affected all of us in many ways. several deep learning techniques have been developed to detect covid-19 from chest x-ray images. covid-19 infection severity scoring can aid in establishing the optimum course of treatment and care for a positive patient, as all covid-19 positive patients do not require special medical attention. still, very few works are reported to estimate the severity of the disease from the chest x-ray images. the unavailability of the large-scale dataset might be a reason.methodswe aim to propose covseverity-net, a deep learning-based architecture for predicting the severity of covid-19 from chest x-ray images. covseverity-net is trained on a public covid-19 dataset, curated by experienced radiologists for severity estimation. for that, a large publicly available dataset is collected and divided into three levels of severity, namely mild, moderate, and severe.resultsan accuracy of 85.71% is reported. conducting 5-fold cross-validation, we have obtained an accuracy of 87.82 \u00b1 6.25%. similarly, conducting 10-fold cross-validation we obtained accuracy of 91.26 \u00b1 3.42. the results were better when compared with other state-of-the-art architectures.conclusionwe strongly believe that this study has a high chance of reducing the workload of overworked front-line radiologists, speeding up patient diagnosis and treatment, and easing pandemic control. future work would be to train a novel deep learning-based architecture on a larger dataset for severity estimation.","key_phrases":["purposecovid-19","a part","our life","the world health organization","it","it","all","us","many ways","several deep learning techniques","covid-19","chest x-ray images","covid-19 infection severity scoring","the optimum course","treatment","care","a positive patient","all covid-19 positive patients","special medical attention","very few works","the severity","the disease","the chest x-ray images","the unavailability","the large-scale dataset","a reason.methodswe aim","covseverity-net","a deep learning-based architecture","the severity","covid-19","chest x-ray images","covseverity-net","a public covid-19 dataset","experienced radiologists","severity estimation","that","a large publicly available dataset","three levels","severity","namely mild, moderate, and severe.resultsan accuracy","85.71%","5-fold cross","-","validation","we","an accuracy","87.82 \u00b1","6.25%","cross","validation","we","accuracy","91.26 \u00b1","the results","the-art","this study","a high chance","the workload","overworked front-line radiologists","patient diagnosis","treatment","pandemic control","future work","a novel deep learning-based architecture","a larger dataset","severity estimation","purposecovid-19","the world health organization","2020","covid-19","covid-19","covid-19","covid-19","covid-19","three","85.71%","5-fold","87.82","6.25%","10-fold","91.26","3.42"]},{"title":"A comparison between machine and deep learning models on high stationarity data","authors":["Domenico Santoro","Tiziana Ciano","Massimiliano Ferrara"],"abstract":"Advances in sensor, computing, and communication technologies are enabling big data analytics by providing time series data. However, conventional models struggle to identify sequence features and forecast accuracy. This paper investigates time series features and shows that some machine learning algorithms can outperform deep learning models. In particular, the problem analyzed concerned predicting the number of vehicles passing through an Italian tollbooth in 2021. The dataset, composed of 8766 rows and 6 columns relating to additional tollbooths, proved to have high stationarity and was treated through machine learning methods such as support vector machine, random forest, and eXtreme gradient boosting (XGBoost), as well as deep learning through recurrent neural networks with long short-term memory (RNN-LSTM) cells. From the comparison of these models, the prediction through the XGBoost algorithm outperforms competing algorithms, particularly in terms of MAE and MSE. The result highlights how a shallower algorithm than a neural network is, in this case, able to obtain a better adaptation to the time series instead of a much deeper model that tends to develop a smoother prediction.","doi":"10.1038\/s41598-024-70341-6","cleaned_title":"a comparison between machine and deep learning models on high stationarity data","cleaned_abstract":"advances in sensor, computing, and communication technologies are enabling big data analytics by providing time series data. however, conventional models struggle to identify sequence features and forecast accuracy. this paper investigates time series features and shows that some machine learning algorithms can outperform deep learning models. in particular, the problem analyzed concerned predicting the number of vehicles passing through an italian tollbooth in 2021. the dataset, composed of 8766 rows and 6 columns relating to additional tollbooths, proved to have high stationarity and was treated through machine learning methods such as support vector machine, random forest, and extreme gradient boosting (xgboost), as well as deep learning through recurrent neural networks with long short-term memory (rnn-lstm) cells. from the comparison of these models, the prediction through the xgboost algorithm outperforms competing algorithms, particularly in terms of mae and mse. the result highlights how a shallower algorithm than a neural network is, in this case, able to obtain a better adaptation to the time series instead of a much deeper model that tends to develop a smoother prediction.","key_phrases":["advances","sensor","computing","communication technologies","big data analytics","time series data","conventional models","sequence features","forecast accuracy","this paper investigates","some machine learning algorithms","deep learning models","the problem","the number","vehicles","an italian tollbooth","the dataset","8766 rows","6 columns","additional tollbooths","high stationarity","machine learning methods","support vector machine","random forest","deep learning","recurrent neural networks","long short-term memory","rnn-lstm) cells","the comparison","these models","the prediction","the xgboost algorithm","competing algorithms","terms","mae","mse","the result","a shallower algorithm","a neural network","this case","a better adaptation","the time series","a much deeper model","that","a smoother prediction","italian","2021","8766","6","mae"]},{"title":"Efficient deep learning-based automated diagnosis from echocardiography with contrastive self-supervised learning","authors":["Gregory Holste","Evangelos K. Oikonomou","Bobak J. Mortazavi","Zhangyang Wang","Rohan Khera"],"abstract":"BackgroundAdvances in self-supervised learning (SSL) have enabled state-of-the-art automated medical image diagnosis from small, labeled datasets. This label efficiency is often desirable, given the difficulty of obtaining expert labels for medical image recognition tasks. However, most efforts toward SSL in medical imaging are not adapted to video-based modalities, such as echocardiography.MethodsWe developed a self-supervised contrastive learning approach, EchoCLR, for echocardiogram videos with the goal of learning strong representations for efficient fine-tuning on downstream cardiac disease diagnosis. EchoCLR pretraining involves (i) contrastive learning, where the model is trained to identify distinct videos of the same patient, and (ii) frame reordering, where the model is trained to predict the correct of video frames after being randomly shuffled.ResultsWhen fine-tuned on small portions of labeled data, EchoCLR pretraining significantly improves classification performance for left ventricular hypertrophy (LVH) and aortic stenosis (AS) over other transfer learning and SSL approaches across internal and external test sets. When fine-tuning on 10% of available training data (519 studies), an EchoCLR-pretrained model achieves 0.72 AUROC (95% CI: [0.69, 0.75]) on LVH classification, compared to 0.61 AUROC (95% CI: [0.57, 0.64]) with a standard transfer learning approach. Similarly, using 1% of available training data (53 studies), EchoCLR pretraining achieves 0.82 AUROC (95% CI: [0.79, 0.84]) on severe AS classification, compared to 0.61 AUROC (95% CI: [0.58, 0.65]) with transfer learning.ConclusionsEchoCLR is unique in its ability to learn representations of echocardiogram videos and demonstrates that SSL can enable label-efficient disease classification from small amounts of labeled data.","doi":"10.1038\/s43856-024-00538-3","cleaned_title":"efficient deep learning-based automated diagnosis from echocardiography with contrastive self-supervised learning","cleaned_abstract":"backgroundadvances in self-supervised learning (ssl) have enabled state-of-the-art automated medical image diagnosis from small, labeled datasets. this label efficiency is often desirable, given the difficulty of obtaining expert labels for medical image recognition tasks. however, most efforts toward ssl in medical imaging are not adapted to video-based modalities, such as echocardiography.methodswe developed a self-supervised contrastive learning approach, echoclr, for echocardiogram videos with the goal of learning strong representations for efficient fine-tuning on downstream cardiac disease diagnosis. echoclr pretraining involves (i) contrastive learning, where the model is trained to identify distinct videos of the same patient, and (ii) frame reordering, where the model is trained to predict the correct of video frames after being randomly shuffled.resultswhen fine-tuned on small portions of labeled data, echoclr pretraining significantly improves classification performance for left ventricular hypertrophy (lvh) and aortic stenosis (as) over other transfer learning and ssl approaches across internal and external test sets. when fine-tuning on 10% of available training data (519 studies), an echoclr-pretrained model achieves 0.72 auroc (95% ci: [0.69, 0.75]) on lvh classification, compared to 0.61 auroc (95% ci: [0.57, 0.64]) with a standard transfer learning approach. similarly, using 1% of available training data (53 studies), echoclr pretraining achieves 0.82 auroc (95% ci: [0.79, 0.84]) on severe as classification, compared to 0.61 auroc (95% ci: [0.58, 0.65]) with transfer learning.conclusionsechoclr is unique in its ability to learn representations of echocardiogram videos and demonstrates that ssl can enable label-efficient disease classification from small amounts of labeled data.","key_phrases":["backgroundadvances","self-supervised learning","ssl","the-art","small, labeled datasets","this label efficiency","the difficulty","expert labels","medical image recognition tasks","most efforts","ssl","medical imaging","video-based modalities","echocardiography.methodswe","a self-supervised contrastive learning approach","echocardiogram videos","the goal","strong representations","efficient fine-tuning","downstream cardiac disease diagnosis","echoclr pretraining","(i) contrastive learning","the model","distinct videos","the same patient","(ii) frame","the model","the correct","video frames","small portions","labeled data","echoclr pretraining","classification performance","left ventricular hypertrophy","lvh","aortic stenosis","other transfer learning","approaches","internal and external test sets","10%","available training data","519 studies","an echoclr-pretrained model","0.72 auroc","95% ci","lvh classification","0.61 auroc","95% ci","a standard transfer learning approach","1%","available training data","53 studies","0.82 auroc","95% ci","classification","0.61 auroc","95% ci","transfer learning.conclusionsechoclr","its ability","representations","echocardiogram videos","ssl","label-efficient disease classification","small amounts","labeled data","10%","519","0.72","95%","0.69","0.75","0.61","95%","0.57","0.64","1%","53","0.82","95%","0.79","0.84","0.61","95%","0.58","0.65"]},{"title":"Detecting defects in PCB manufacturing: an exploration using Yolov8 deep learning","authors":["Weifeng LI"],"abstract":"Detecting defects in automated inspection systems for Printed Circuit Board (PCB) manufacturing stands as a critical endeavor for ensuring product quality. Despite numerous methodologies explored in the literature, recent advancements highlight the superiority of deep learning techniques in defect identification, driven by their remarkable accuracy surpassing conventional methods. However, the persistent challenge of achieving higher accuracy rates and real-time processing capabilities persists, motivating the need for innovative solutions. In response, this study introduces a novel approach utilizing the Yolov8 architecture for PCB defect detection, addressing the aforementioned research challenge. Leveraging a custom dataset tailored explicitly for this purpose, our method showcases remarkable performance, with Yolov8x emerging as the top-performing model, achieving a notable F1 Score of 98% and a high mean Average Precision (mAP) of 98.9%. This study\u2019s novelty lies in its comprehensive evaluation of deep learning models specifically tailored for PCB defect detection, offering promising prospects for enhancing precision, non-destructiveness, and real-time capabilities in manufacturing processes.","doi":"10.1007\/s12008-024-01986-w","cleaned_title":"detecting defects in pcb manufacturing: an exploration using yolov8 deep learning","cleaned_abstract":"detecting defects in automated inspection systems for printed circuit board (pcb) manufacturing stands as a critical endeavor for ensuring product quality. despite numerous methodologies explored in the literature, recent advancements highlight the superiority of deep learning techniques in defect identification, driven by their remarkable accuracy surpassing conventional methods. however, the persistent challenge of achieving higher accuracy rates and real-time processing capabilities persists, motivating the need for innovative solutions. in response, this study introduces a novel approach utilizing the yolov8 architecture for pcb defect detection, addressing the aforementioned research challenge. leveraging a custom dataset tailored explicitly for this purpose, our method showcases remarkable performance, with yolov8x emerging as the top-performing model, achieving a notable f1 score of 98% and a high mean average precision (map) of 98.9%. this study\u2019s novelty lies in its comprehensive evaluation of deep learning models specifically tailored for pcb defect detection, offering promising prospects for enhancing precision, non-destructiveness, and real-time capabilities in manufacturing processes.","key_phrases":["defects","automated inspection systems","printed circuit board (pcb) manufacturing","a critical endeavor","product quality","numerous methodologies","the literature","recent advancements","the superiority","deep learning techniques","defect identification","their remarkable accuracy","conventional methods","however, the persistent challenge","higher accuracy rates","real-time processing capabilities persists","the need","innovative solutions","response","this study","a novel approach","the yolov8 architecture","pcb defect detection","the aforementioned research challenge","a custom dataset","this purpose","our method","remarkable performance","yolov8x","the top-performing model","a notable f1 score","98%","a high mean average precision","(map","98.9%","this study\u2019s novelty","its comprehensive evaluation","deep learning models","pcb defect detection","promising prospects","precision","real-time capabilities","manufacturing processes","yolov8","98%","98.9%"]},{"title":"Deep learning in medical image super resolution: a review","authors":["Hujun Yang","Zhongyang Wang","Xinyao Liu","Chuangang Li","Junchang Xin","Zhiqiong Wang"],"abstract":"Super-resolution (SR) reconstruction is a hot topic in medical image processing. SR implies reconstructing corresponding high-resolution (HR) images from observed low-resolution (LR) images or image sequences. In recent years, significant breakthroughs in SR based on deep learning have been made, and many advanced results have been achieved. However, there is a lack of review literature that summarizes the field\u2019s current state and provides an outlook on future developments. Therefore, we provide a comprehensive summary of the literature on medical image SR (MedSR) based on deep learning since 2018 in five aspects: (1) The SR problem of medical images is described, and the methods of image degradation are summarized. (2) We divide the existing studies into three categories: two-dimensional image SR (2DISR), three-dimensional image SR (3DISR), and video SR (VSR). Each category is subdivided. We analyze the network structure and method characteristics of typical methods. (3) Existing SR reconstruction quality evaluation metrics are presented in detail. (4) The application of MedSR methods based on deep learning is discussed. (5) We discuss the challenges of this phase and point out valuable research directions.","doi":"10.1007\/s10489-023-04566-9","cleaned_title":"deep learning in medical image super resolution: a review","cleaned_abstract":"super-resolution (sr) reconstruction is a hot topic in medical image processing. sr implies reconstructing corresponding high-resolution (hr) images from observed low-resolution (lr) images or image sequences. in recent years, significant breakthroughs in sr based on deep learning have been made, and many advanced results have been achieved. however, there is a lack of review literature that summarizes the field\u2019s current state and provides an outlook on future developments. therefore, we provide a comprehensive summary of the literature on medical image sr (medsr) based on deep learning since 2018 in five aspects: (1) the sr problem of medical images is described, and the methods of image degradation are summarized. (2) we divide the existing studies into three categories: two-dimensional image sr (2disr), three-dimensional image sr (3disr), and video sr (vsr). each category is subdivided. we analyze the network structure and method characteristics of typical methods. (3) existing sr reconstruction quality evaluation metrics are presented in detail. (4) the application of medsr methods based on deep learning is discussed. (5) we discuss the challenges of this phase and point out valuable research directions.","key_phrases":["-","resolution (sr) reconstruction","a hot topic","medical image processing","sr","hr","observed low-resolution (lr) images","image sequences","recent years","significant breakthroughs","sr","deep learning","many advanced results","a lack","review literature","that","the field\u2019s current state","an outlook","future developments","we","a comprehensive summary","the literature","medical image sr","medsr","deep learning","five aspects","the sr problem","medical images","the methods","image degradation","we","the existing studies","three categories","two-dimensional image sr","three-dimensional image sr","video sr","vsr","each category","we","the network structure","method","characteristics","typical methods","existing sr reconstruction quality evaluation metrics","detail","the application","medsr methods","deep learning","we","the challenges","this phase","valuable research directions","recent years","2018","five","1","2","three","two","2disr","three","vsr","3","4","5"]},{"title":"Aspect-oriented extraction and sentiment analysis using optimized hybrid deep learning approaches","authors":["Srividya Kotagiri","A. Mary Sowjanya","B. Anilkumar","N Lakshmi Devi"],"abstract":"Aspect-oriented extraction involves the identification and extraction of specific aspects, features, or entities within a piece of text. Traditional methods often struggled with the complexity and variability of language, leading to the exploration of advanced deep learning approaches. In the realm of sentiment analysis, the conventional approaches often fall short when it comes to providing a nuanced understanding of sentiments expressed in textual data. Traditional sentiment analysis models often overlook the specific aspects or entities within the text that contribute to the overall sentiment. This limitation poses a significant challenge for businesses and organizations aiming to gain detailed insights into customer opinions, product reviews, and other forms of user-generated content.In this research, we propose an innovative approach for aspect-oriented extraction and sentiment analysis leveraging optimized hybrid deep learning techniques. Our methodology integrates the powerful capabilities of deep learning models with the efficiency of Reptile Search Optimization. Furthermore, we introduce an advanced sentiment analysis framework employing the state-of-the-art Extreme Gradient Boosting Algorithm. The fusion of these techniques aims to enhance the precision and interpretability of aspect-oriented sentiment analysis. The proposed approach first utilizes deep learning architectures to extract and comprehend diverse aspects within textual data. Through the incorporation of Reptile Search Optimization, we optimize the learning process, ensuring adaptability and improved model generalization across various datasets. Subsequently, the sentiment analysis phase employs the robust Extreme Gradient Boosting Algorithm, known for its effectiveness in handling complex relationships and patterns within data. Our experiments, conducted on diverse datasets, demonstrate the superior performance of the proposed methodology in comparison to traditional approaches. The optimized hybrid deep learning approach, coupled with the Reptile Search Optimization and Extreme Gradient Boosting Algorithm, showcases promising results in accurately capturing nuanced sentiments associated with different aspects. This research contributes to the advancement of aspect-oriented sentiment analysis techniques, offering a comprehensive and efficient solution for understanding sentiment nuances in textual data across various domains. The ResNet 50 and EfficientNet B7 architecture of the modified pre-trained model is proposed for the aspect extraction function. The Reptile Search Optimization based Extreme Gradient Boosting Algorithm (RSO-EGBA) is proposed to analyze and predict customer sentiments. The execution of this study is carried out using python software. It has been observed that the overall accuracy of our proposed method is 99.8%, while that of the other state-of-the-art. The overall accuracy of our proposed method shows an increment of 9\u201316% from that of the state-of-the-art methods.","doi":"10.1007\/s11042-024-18964-9","cleaned_title":"aspect-oriented extraction and sentiment analysis using optimized hybrid deep learning approaches","cleaned_abstract":"aspect-oriented extraction involves the identification and extraction of specific aspects, features, or entities within a piece of text. traditional methods often struggled with the complexity and variability of language, leading to the exploration of advanced deep learning approaches. in the realm of sentiment analysis, the conventional approaches often fall short when it comes to providing a nuanced understanding of sentiments expressed in textual data. traditional sentiment analysis models often overlook the specific aspects or entities within the text that contribute to the overall sentiment. this limitation poses a significant challenge for businesses and organizations aiming to gain detailed insights into customer opinions, product reviews, and other forms of user-generated content.in this research, we propose an innovative approach for aspect-oriented extraction and sentiment analysis leveraging optimized hybrid deep learning techniques. our methodology integrates the powerful capabilities of deep learning models with the efficiency of reptile search optimization. furthermore, we introduce an advanced sentiment analysis framework employing the state-of-the-art extreme gradient boosting algorithm. the fusion of these techniques aims to enhance the precision and interpretability of aspect-oriented sentiment analysis. the proposed approach first utilizes deep learning architectures to extract and comprehend diverse aspects within textual data. through the incorporation of reptile search optimization, we optimize the learning process, ensuring adaptability and improved model generalization across various datasets. subsequently, the sentiment analysis phase employs the robust extreme gradient boosting algorithm, known for its effectiveness in handling complex relationships and patterns within data. our experiments, conducted on diverse datasets, demonstrate the superior performance of the proposed methodology in comparison to traditional approaches. the optimized hybrid deep learning approach, coupled with the reptile search optimization and extreme gradient boosting algorithm, showcases promising results in accurately capturing nuanced sentiments associated with different aspects. this research contributes to the advancement of aspect-oriented sentiment analysis techniques, offering a comprehensive and efficient solution for understanding sentiment nuances in textual data across various domains. the resnet 50 and efficientnet b7 architecture of the modified pre-trained model is proposed for the aspect extraction function. the reptile search optimization based extreme gradient boosting algorithm (rso-egba) is proposed to analyze and predict customer sentiments. the execution of this study is carried out using python software. it has been observed that the overall accuracy of our proposed method is 99.8%, while that of the other state-of-the-art. the overall accuracy of our proposed method shows an increment of 9\u201316% from that of the state-of-the-art methods.","key_phrases":["aspect-oriented extraction","the identification","extraction","specific aspects","features","entities","a piece","text","traditional methods","the complexity","variability","language","the exploration","advanced deep learning approaches","the realm","sentiment analysis","the conventional approaches","it","a nuanced understanding","sentiments","textual data","traditional sentiment analysis models","the specific aspects","entities","the text","that","the overall sentiment","this limitation","a significant challenge","businesses","organizations","detailed insights","customer opinions","product reviews","other forms","we","an innovative approach","aspect-oriented extraction","sentiment analysis","optimized hybrid deep learning techniques","our methodology","the powerful capabilities","deep learning models","the efficiency","reptile search optimization","we","an advanced sentiment analysis framework","the-art","algorithm","the fusion","these techniques","the precision","interpretability","aspect-oriented sentiment analysis","the proposed approach","deep learning architectures","diverse aspects","textual data","the incorporation","reptile search optimization","we","the learning process","adaptability","improved model generalization","various datasets","the sentiment analysis phase","the robust extreme gradient","algorithm","its effectiveness","complex relationships","patterns","data","our experiments","diverse datasets","the superior performance","the proposed methodology","comparison","traditional approaches","the optimized hybrid deep learning approach","the reptile search optimization","extreme gradient","algorithm","promising results","nuanced sentiments","different aspects","this research","the advancement","aspect-oriented sentiment analysis techniques","a comprehensive and efficient solution","sentiment nuances","textual data","various domains","the resnet 50 and efficientnet b7 architecture","the modified pre-trained model","the aspect extraction function","the reptile search optimization","based extreme gradient","algorithm","rso","egba","customer sentiments","the execution","this study","python software","it","the overall accuracy","our proposed method","99.8%","the other state","the-art","the overall accuracy","our proposed method","an increment","9\u201316%","that","the-art","first","50","99.8%","9\u201316%"]},{"title":"A review of research on micro-expression recognition algorithms based on deep learning","authors":["Fan Zhang","Lin Chai"],"abstract":"Micro-expression is a special kind of human emotion. Due to its characteristics of short time, low intensity, and local region, micro-expression recognition is a difficult task. At the same time, it is a natural, spontaneous, and unconcealable emotion that can well convey a person's actual psychological state and, therefore, has certain research value and practical significance. This paper focuses on micro-expression recognition in the field of deep learning through the survey and understanding of existing micro-expression recognition research, as well as grasping the research trend, for the previous literature on micro-expression review ignored the handcrafted features as an important part of the micro-expression recognition framework, and at the same time lacked the analysis of the various enhancement processing, a new micro-expression recognition framework based on deep learning is proposed. The model is designed from the perspective of modularity and streaming data. On the other hand, unlike the previous process of feeding the data directly into the network for training and recognition, the handcrafted features are used as the initial encoding of the micro-expression recognition data, followed by the training and learning of the deep model and at the same time the modular embedding approach is used to incorporate the feature enhancement module, and finally the classification and recognition. The article provides a detailed summary and analysis of each part of the whole framework and a comprehensive introduction to the current problems, experimental protocols, evaluation metrics, and application areas. Finally, it summarizes and gives possible future research directions. Therefore, this paper provides a comprehensive summary and analysis of micro-expression recognition in deep learning so that the related personnel can have a new understanding of the development of this field. On the other hand, it proposes a new recognition framework that also provides a reference for the researchers' later research.","doi":"10.1007\/s00521-024-10262-7","cleaned_title":"a review of research on micro-expression recognition algorithms based on deep learning","cleaned_abstract":"micro-expression is a special kind of human emotion. due to its characteristics of short time, low intensity, and local region, micro-expression recognition is a difficult task. at the same time, it is a natural, spontaneous, and unconcealable emotion that can well convey a person's actual psychological state and, therefore, has certain research value and practical significance. this paper focuses on micro-expression recognition in the field of deep learning through the survey and understanding of existing micro-expression recognition research, as well as grasping the research trend, for the previous literature on micro-expression review ignored the handcrafted features as an important part of the micro-expression recognition framework, and at the same time lacked the analysis of the various enhancement processing, a new micro-expression recognition framework based on deep learning is proposed. the model is designed from the perspective of modularity and streaming data. on the other hand, unlike the previous process of feeding the data directly into the network for training and recognition, the handcrafted features are used as the initial encoding of the micro-expression recognition data, followed by the training and learning of the deep model and at the same time the modular embedding approach is used to incorporate the feature enhancement module, and finally the classification and recognition. the article provides a detailed summary and analysis of each part of the whole framework and a comprehensive introduction to the current problems, experimental protocols, evaluation metrics, and application areas. finally, it summarizes and gives possible future research directions. therefore, this paper provides a comprehensive summary and analysis of micro-expression recognition in deep learning so that the related personnel can have a new understanding of the development of this field. on the other hand, it proposes a new recognition framework that also provides a reference for the researchers' later research.","key_phrases":["micro","-","expression","a special kind","human emotion","its characteristics","short time","low intensity","local region","micro-expression recognition","a difficult task","the same time","it","a natural, spontaneous, and unconcealable emotion","that","a person's actual psychological state","certain research value","practical significance","this paper","micro-expression recognition","the field","deep learning","the survey","understanding","existing micro-expression recognition research","the research trend","the previous literature","micro-expression review","the handcrafted features","an important part","the micro-expression recognition framework","the same time","the analysis","the various enhancement processing","a new micro-expression recognition framework","deep learning","the model","the perspective","modularity and streaming data","the other hand","the previous process","the data","the network","training","recognition","the handcrafted features","the initial encoding","the micro-expression recognition data","the training","learning","the deep model","the same time","the modular embedding approach","the feature enhancement module","finally the classification","recognition","the article","a detailed summary","analysis","each part","the whole framework","a comprehensive introduction","the current problems","experimental protocols","evaluation metrics","application areas","it","possible future research directions","this paper","a comprehensive summary","analysis","micro-expression recognition","deep learning","the related personnel","a new understanding","the development","this field","the other hand","it","a new recognition framework","that","a reference","the researchers' later research"]},{"title":"On the use of deep learning for phase recovery","authors":["Kaiqiang Wang","Li Song","Chutian Wang","Zhenbo Ren","Guangyuan Zhao","Jiazhen Dou","Jianglei Di","George Barbastathis","Renjie Zhou","Jianlin Zhao","Edmund Y. Lam"],"abstract":"Phase recovery (PR) refers to calculating the phase of the light field from its intensity measurements. As exemplified from quantitative phase imaging and coherent diffraction imaging to adaptive optics, PR is essential for reconstructing the refractive index distribution or topography of an object and correcting the aberration of an imaging system. In recent years, deep learning (DL), often implemented through deep neural networks, has provided unprecedented support for computational imaging, leading to more efficient solutions for various PR problems. In this review, we first briefly introduce conventional methods for PR. Then, we review how DL provides support for PR from the following three stages, namely, pre-processing, in-processing, and post-processing. We also review how DL is used in phase image processing. Finally, we summarize the work in DL for PR and provide an outlook on how to better use DL to improve the reliability and efficiency of PR. Furthermore, we present a live-updating resource (https:\/\/github.com\/kqwang\/phase-recovery) for readers to learn more about PR.","doi":"10.1038\/s41377-023-01340-x","cleaned_title":"on the use of deep learning for phase recovery","cleaned_abstract":"phase recovery (pr) refers to calculating the phase of the light field from its intensity measurements. as exemplified from quantitative phase imaging and coherent diffraction imaging to adaptive optics, pr is essential for reconstructing the refractive index distribution or topography of an object and correcting the aberration of an imaging system. in recent years, deep learning (dl), often implemented through deep neural networks, has provided unprecedented support for computational imaging, leading to more efficient solutions for various pr problems. in this review, we first briefly introduce conventional methods for pr. then, we review how dl provides support for pr from the following three stages, namely, pre-processing, in-processing, and post-processing. we also review how dl is used in phase image processing. finally, we summarize the work in dl for pr and provide an outlook on how to better use dl to improve the reliability and efficiency of pr. furthermore, we present a live-updating resource (https:\/\/github.com\/kqwang\/phase-recovery) for readers to learn more about pr.","key_phrases":["phase recovery","(pr","the phase","the light field","its intensity measurements","quantitative phase imaging","coherent diffraction","adaptive optics","the refractive index distribution","topography","an object","the aberration","an imaging system","recent years","deep learning","dl","deep neural networks","unprecedented support","computational imaging","more efficient solutions","various pr problems","this review","we","conventional methods","we","dl","support","pr","the following three stages","-processing","processing","we","dl","phase image processing","we","the work","dl","pr","an outlook","dl","the reliability","efficiency","pr","we","a live-updating resource","https:\/\/github.com\/kqwang\/phase-recovery","readers","pr","recent years","first","three"]},{"title":"A multimodal deep learning model for predicting severe hemorrhage in placenta previa","authors":["Munetoshi Akazawa","Kazunori Hashimoto"],"abstract":"Placenta previa causes life-threatening bleeding and accurate prediction of severe hemorrhage leads to risk stratification and optimum allocation of interventions. We aimed to use a multimodal deep learning model to predict severe hemorrhage. Using MRI T2-weighted image of the placenta and tabular data consisting of patient demographics and preoperative blood examination data, a multimodal deep learning model was constructed to predict cases of intraoperative blood loss\u2009>\u20092000\u00a0ml. We evaluated the prediction performance of the model by comparing it with that of two machine learning methods using only tabular data and MRI images, as well as with that of two human expert obstetricians. Among the enrolled 48 patients, 26 (54.2%) lost\u2009>\u20092000\u00a0ml of blood and 22 (45.8%) lost\u2009<\u20092000\u00a0ml of blood. Multimodal deep learning model showed the best accuracy of 0.68 and AUC of 0.74, whereas the machine learning model using tabular data and MRI images had a class accuracy of 0.61 and 0.53, respectively. The human experts had median accuracies of 0.61. Multimodal deep learning models could integrate the two types of information and predict severe hemorrhage cases. The model might assist human expert in the prediction of intraoperative hemorrhage in the case of placenta previa.","doi":"10.1038\/s41598-023-44634-1","cleaned_title":"a multimodal deep learning model for predicting severe hemorrhage in placenta previa","cleaned_abstract":"placenta previa causes life-threatening bleeding and accurate prediction of severe hemorrhage leads to risk stratification and optimum allocation of interventions. we aimed to use a multimodal deep learning model to predict severe hemorrhage. using mri t2-weighted image of the placenta and tabular data consisting of patient demographics and preoperative blood examination data, a multimodal deep learning model was constructed to predict cases of intraoperative blood loss > 2000 ml. we evaluated the prediction performance of the model by comparing it with that of two machine learning methods using only tabular data and mri images, as well as with that of two human expert obstetricians. among the enrolled 48 patients, 26 (54.2%) lost > 2000 ml of blood and 22 (45.8%) lost < 2000 ml of blood. multimodal deep learning model showed the best accuracy of 0.68 and auc of 0.74, whereas the machine learning model using tabular data and mri images had a class accuracy of 0.61 and 0.53, respectively. the human experts had median accuracies of 0.61. multimodal deep learning models could integrate the two types of information and predict severe hemorrhage cases. the model might assist human expert in the prediction of intraoperative hemorrhage in the case of placenta previa.","key_phrases":["placenta previa","life-threatening bleeding","accurate prediction","severe hemorrhage","risk stratification","optimum allocation","interventions","we","a multimodal deep learning model","severe hemorrhage","mri t2-weighted image","the placenta and tabular data","patient demographics","preoperative blood examination data","a multimodal deep learning model","cases","intraoperative blood loss","we","the prediction performance","the model","it","that","two machine learning methods","only tabular data and mri images","that","two human expert obstetricians","the enrolled 48 patients","54.2%","2000 ml","blood","22 (45.8%","2000 ml","blood","multimodal deep learning model","the best accuracy","auc","the machine learning model","tabular data and mri images","a class accuracy","the human experts","median accuracies","multimodal deep learning models","the two types","information","severe hemorrhage cases","the model","human expert","the prediction","intraoperative hemorrhage","the case","placenta previa","placenta previa","2000 ml","two","two","48","26","54.2%","2000 ml","22","45.8%","2000 ml","0.68","0.74","0.61","0.53","0.61","two"]},{"title":"Curriculum learning for ab initio deep learned refractive optics","authors":["Xinge Yang","Qiang Fu","Wolfgang Heidrich"],"abstract":"Deep optical optimization has recently emerged as a new paradigm for designing computational imaging systems using only the output image as the objective. However, it has been limited to either simple optical systems consisting of a single element such as a diffractive optical element or metalens, or the fine-tuning of compound lenses from good initial designs. Here we present a DeepLens design method based on curriculum learning, which is able to learn optical designs of compound lenses ab initio from randomly initialized surfaces without human intervention, therefore overcoming the need for a good initial design. We demonstrate the effectiveness of our approach by fully automatically designing both classical imaging lenses and a large field-of-view extended depth-of-field computational lens in a cellphone-style form factor, with highly aspheric surfaces and a short back focal length.","doi":"10.1038\/s41467-024-50835-7","cleaned_title":"curriculum learning for ab initio deep learned refractive optics","cleaned_abstract":"deep optical optimization has recently emerged as a new paradigm for designing computational imaging systems using only the output image as the objective. however, it has been limited to either simple optical systems consisting of a single element such as a diffractive optical element or metalens, or the fine-tuning of compound lenses from good initial designs. here we present a deeplens design method based on curriculum learning, which is able to learn optical designs of compound lenses ab initio from randomly initialized surfaces without human intervention, therefore overcoming the need for a good initial design. we demonstrate the effectiveness of our approach by fully automatically designing both classical imaging lenses and a large field-of-view extended depth-of-field computational lens in a cellphone-style form factor, with highly aspheric surfaces and a short back focal length.","key_phrases":["deep optical optimization","a new paradigm","computational imaging systems","only the output image","the objective","it","either simple optical systems","a single element","a diffractive optical element","metalens","the fine-tuning","compound lenses","good initial designs","we","a deeplens design method","curriculum learning","which","optical designs","compound lenses","ab initio","randomly initialized surfaces","human intervention","the need","a good initial design","we","the effectiveness","our approach","both classical imaging lenses","view","field","a cellphone-style form factor","highly aspheric surfaces","a short back focal length","deeplens"]},{"title":"HDLP: air quality modeling with hybrid deep learning approaches and particle swam optimization","authors":["Elmenawy Osman","C. Banerjee","Ajeet Singh Poonia"],"abstract":"Predicting air pollution in cities has become an important tool for preventing its negative impacts. Therefore, citizens should be aware of air quality level, especially for individuals suffering from diseases caused by air pollutants. Collective efforts from researchers, environmental institutions, governments, industrial companies, and policy makers are shaping the future of the Air Quality Index (AQI) to effectively address severe air pollution in urban areas. Many air quality prediction models have been introduced in the literature; modern advances in deep learning techniques are promising more precise prediction results and data integration. The aim of this paper is to review methods, contributions, findings, limitations, and gaps in predicting air quality index and PM2.5 concentrations using a hybrid deep learning approach. A literature review led researchers to propose a Hybrid Deep Learning model with Particle Swarm Optimization (HDLP) which combines CNN, LSTM, and PSO. Algorithm, first Discrete Wavelet Transform (DWT) is used to solve the air pollution signal, then it is fed to CNN-LSTM neural network for PSO optimization to get the final prediction result. Optimized parameters input models are trained on the original data. In addition to the beneficial assessment cycle, to outperform current models.","doi":"10.1007\/s11334-024-00559-0","cleaned_title":"hdlp: air quality modeling with hybrid deep learning approaches and particle swam optimization","cleaned_abstract":"predicting air pollution in cities has become an important tool for preventing its negative impacts. therefore, citizens should be aware of air quality level, especially for individuals suffering from diseases caused by air pollutants. collective efforts from researchers, environmental institutions, governments, industrial companies, and policy makers are shaping the future of the air quality index (aqi) to effectively address severe air pollution in urban areas. many air quality prediction models have been introduced in the literature; modern advances in deep learning techniques are promising more precise prediction results and data integration. the aim of this paper is to review methods, contributions, findings, limitations, and gaps in predicting air quality index and pm2.5 concentrations using a hybrid deep learning approach. a literature review led researchers to propose a hybrid deep learning model with particle swarm optimization (hdlp) which combines cnn, lstm, and pso. algorithm, first discrete wavelet transform (dwt) is used to solve the air pollution signal, then it is fed to cnn-lstm neural network for pso optimization to get the final prediction result. optimized parameters input models are trained on the original data. in addition to the beneficial assessment cycle, to outperform current models.","key_phrases":["air pollution","cities","an important tool","its negative impacts","citizens","air quality level","individuals","diseases","air pollutants","collective efforts","researchers","environmental institutions","governments","industrial companies","policy makers","the future","the air quality index","aqi","severe air pollution","urban areas","many air quality prediction models","the literature","modern advances","deep learning techniques","more precise prediction results","data integration","the aim","this paper","methods","contributions","findings","limitations","gaps","air quality index","pm2.5 concentrations","a hybrid deep learning approach","a literature review","researchers","a hybrid deep learning model","particle swarm optimization","hdlp","which","cnn","lstm","algorithm","first discrete wavelet transform","dwt","the air pollution signal","it","cnn-lstm neural network","pso optimization","the final prediction result","optimized parameters input models","the original data","addition","the beneficial assessment cycle","current models","cnn","first","fed","cnn"]},{"title":"Research progress in water quality prediction based on deep learning technology: a review","authors":["Wenhao Li","Yin Zhao","Yining Zhu","Zhongtian Dong","Fenghe Wang","Fengliang Huang"],"abstract":"Water, an invaluable and non-renewable resource, plays an indispensable role in human survival and societal development. Accurate forecasting of water quality involves early identification of future pollutant concentrations and water quality indices, enabling evidence-based decision-making and targeted environmental interventions. The emergence of advanced computational technologies, particularly deep learning, has garnered considerable interest among researchers for applications in water quality prediction because of its robust data analytics capabilities. This article comprehensively reviews the deployment of deep learning methodologies in water quality forecasting, encompassing single-model and mixed-model approaches. Additionally, we delineate optimization strategies, data fusion techniques, and other factors influencing the efficacy of deep learning-based water quality prediction models, because understanding and mastering these factors are crucial for accurate water quality prediction. Although challenges such as data scarcity, long-term prediction accuracy, and limited deployments of large-scale models persist, future research aims to address these limitations by refining prediction algorithms, leveraging high-dimensional datasets, evaluating model performance, and broadening large-scale model application. These efforts contribute to precise water resource management and environmental conservation.","doi":"10.1007\/s11356-024-33058-7","cleaned_title":"research progress in water quality prediction based on deep learning technology: a review","cleaned_abstract":"water, an invaluable and non-renewable resource, plays an indispensable role in human survival and societal development. accurate forecasting of water quality involves early identification of future pollutant concentrations and water quality indices, enabling evidence-based decision-making and targeted environmental interventions. the emergence of advanced computational technologies, particularly deep learning, has garnered considerable interest among researchers for applications in water quality prediction because of its robust data analytics capabilities. this article comprehensively reviews the deployment of deep learning methodologies in water quality forecasting, encompassing single-model and mixed-model approaches. additionally, we delineate optimization strategies, data fusion techniques, and other factors influencing the efficacy of deep learning-based water quality prediction models, because understanding and mastering these factors are crucial for accurate water quality prediction. although challenges such as data scarcity, long-term prediction accuracy, and limited deployments of large-scale models persist, future research aims to address these limitations by refining prediction algorithms, leveraging high-dimensional datasets, evaluating model performance, and broadening large-scale model application. these efforts contribute to precise water resource management and environmental conservation.","key_phrases":["water","an invaluable and non-renewable resource","an indispensable role","human survival","societal development","accurate forecasting","water quality","early identification","future pollutant concentrations","water quality indices","evidence-based decision-making and targeted environmental interventions","the emergence","advanced computational technologies","particularly deep learning","considerable interest","researchers","applications","water quality prediction","its robust data analytics capabilities","this article","the deployment","deep learning methodologies","water quality forecasting","single-model and mixed-model approaches","we","optimization strategies","data fusion techniques","other factors","the efficacy","deep learning-based water quality prediction models","understanding","these factors","accurate water quality prediction","challenges","data scarcity","long-term prediction accuracy","limited deployments","large-scale models","future research","these limitations","refining prediction algorithms","high-dimensional datasets","model performance","large-scale model application","these efforts","precise water resource management","environmental conservation"]},{"title":"Implementing a Hierarchical Deep Learning Approach for Simulating Multilevel Auction Data","authors":["Igor Sadoune","Marcelin Joanis","Andrea Lodi"],"abstract":"We present a deep learning solution to address the challenges of simulating realistic synthetic first-price sealed-bid auction data. The complexities encountered in this type of auction data include high-cardinality discrete feature spaces and a multilevel structure arising from multiple bids associated with a single auction instance. Our methodology combines deep generative modeling (DGM) with an artificial learner that predicts the conditional bid distribution based on auction characteristics, contributing to advancements in simulation-based research. This approach lays the groundwork for creating realistic auction environments suitable for agent-based learning and modeling applications. Our contribution is twofold: we introduce a comprehensive methodology for simulating multilevel discrete auction data, and we underscore the potential of DGM as a powerful instrument for refining simulation techniques and fostering the development of economic models grounded in generative AI.","doi":"10.1007\/s10614-024-10622-4","cleaned_title":"implementing a hierarchical deep learning approach for simulating multilevel auction data","cleaned_abstract":"we present a deep learning solution to address the challenges of simulating realistic synthetic first-price sealed-bid auction data. the complexities encountered in this type of auction data include high-cardinality discrete feature spaces and a multilevel structure arising from multiple bids associated with a single auction instance. our methodology combines deep generative modeling (dgm) with an artificial learner that predicts the conditional bid distribution based on auction characteristics, contributing to advancements in simulation-based research. this approach lays the groundwork for creating realistic auction environments suitable for agent-based learning and modeling applications. our contribution is twofold: we introduce a comprehensive methodology for simulating multilevel discrete auction data, and we underscore the potential of dgm as a powerful instrument for refining simulation techniques and fostering the development of economic models grounded in generative ai.","key_phrases":["we","a deep learning solution","the challenges","realistic synthetic first-price sealed-bid auction data","the complexities","this type","auction data","high-cardinality discrete feature spaces","a multilevel structure","multiple bids","a single auction instance","our methodology","deep generative modeling","(dgm","an artificial learner","that","the conditional bid distribution","auction characteristics","advancements","simulation-based research","this approach","the groundwork","realistic auction environments","agent-based learning and modeling applications","our contribution","we","a comprehensive methodology","multilevel discrete auction data","we","the potential","dgm","a powerful instrument","simulation techniques","the development","economic models","generative ai","first"]},{"title":"A deep learning based approach for image retrieval extraction in mobile edge computing","authors":["Jamal Alasadi","Ghassan F. Bati","Ahmed Al Hilli"],"abstract":"Deep learning has been widely explored in 5G applications, including computer vision, the Internet of Things (IoT), and intermedia classification. However, applying the deep learning approach in limited-resource mobile devices is one of the most challenging issues. At the same time, users\u2019 experience in terms of Quality of Service (QoS) (e.g., service latency, outcome accuracy, and achievable data rate) performs poorly while interacting with machine learning applications. Mobile edge computing (MEC) has been introduced as a cooperative approach to bring computation resources in proximity to end-user devices to overcome these limitations. This article aims to design a novel image reiterative extraction algorithm based on convolution neural network (CNN) learning and computational task offloading to support machine learning-based mobile applications in resource-limited and uncertain environments. Accordingly, we leverage the framework of image retrieval extraction and introduce three approaches. First, privacy preservation is strict and aims to protect personal data. Second, network traffic reduction. Third, minimizing feature matching time. Our simulation results associated with real-time experiments on a small-scale MEC server have shown the effectiveness of the proposed deep learning-based approach over existing schemes. The source code is available here: https:\/\/github.com\/jamalalasadi\/CNN_Image_retrieval.","doi":"10.1007\/s43995-024-00060-6","cleaned_title":"a deep learning based approach for image retrieval extraction in mobile edge computing","cleaned_abstract":"deep learning has been widely explored in 5g applications, including computer vision, the internet of things (iot), and intermedia classification. however, applying the deep learning approach in limited-resource mobile devices is one of the most challenging issues. at the same time, users\u2019 experience in terms of quality of service (qos) (e.g., service latency, outcome accuracy, and achievable data rate) performs poorly while interacting with machine learning applications. mobile edge computing (mec) has been introduced as a cooperative approach to bring computation resources in proximity to end-user devices to overcome these limitations. this article aims to design a novel image reiterative extraction algorithm based on convolution neural network (cnn) learning and computational task offloading to support machine learning-based mobile applications in resource-limited and uncertain environments. accordingly, we leverage the framework of image retrieval extraction and introduce three approaches. first, privacy preservation is strict and aims to protect personal data. second, network traffic reduction. third, minimizing feature matching time. our simulation results associated with real-time experiments on a small-scale mec server have shown the effectiveness of the proposed deep learning-based approach over existing schemes. the source code is available here: https:\/\/github.com\/jamalalasadi\/cnn_image_retrieval.","key_phrases":["deep learning","5g applications","computer vision","the internet","things","iot","intermedia classification","the deep learning approach","limited-resource mobile devices","the most challenging issues","the same time","users\u2019 experience","terms","quality","service","qos","machine learning applications","mobile edge computing","(mec","a cooperative approach","computation resources","proximity","end-user devices","these limitations","this article","a novel image reiterative extraction algorithm","convolution neural network","(cnn) learning","computational task","machine learning-based mobile applications","resource-limited and uncertain environments","we","the framework","image retrieval extraction","three approaches","privacy preservation","personal data","feature matching time","our simulation results","real-time experiments","a small-scale mec server","the effectiveness","the proposed deep learning-based approach","existing schemes","the source code","https:\/\/github.com\/jamalalasadi\/cnn_image_retrieval","5","cnn","three","first","second","third"]},{"title":"Deep learning-based automated angle measurement for flatfoot diagnosis in weight-bearing lateral radiographs","authors":["Won-Jun Noh","Mu Sook Lee","Byoung-Dai Lee"],"abstract":"This study aimed to develop and evaluate a deep learning-based system for the automatic measurement of angles (specifically, Meary\u2019s angle and calcaneal pitch) in weight-bearing lateral radiographs of the foot for flatfoot diagnosis. We utilized 3960 lateral radiographs, either from the left or right foot, sourced from a pool of 4000 patients to construct and evaluate a deep learning-based model. These radiographs were captured between June and November 2021, and patients who had undergone total ankle replacement surgery or ankle arthrodesis surgery were excluded. Various methods, including correlation analysis, Bland\u2013Altman plots, and paired T-tests, were employed to assess the concordance between the angles automatically measured using the system and those assessed by clinical experts. The evaluation dataset comprised 150 weight-bearing radiographs from 150 patients. In all test cases, the angles automatically computed using the deep learning-based system were in good agreement with the reference standards (Meary\u2019s angle: Pearson correlation coefficient (PCC)\u2009=\u20090.964, intraclass correlation coefficient (ICC)\u2009=\u20090.963, concordance correlation coefficient (CCC)\u2009=\u20090.963, p-value\u2009=\u20090.632, mean absolute error (MAE)\u2009=\u20091.59\u00b0; calcaneal pitch: PCC\u2009=\u20090.988, ICC\u2009=\u20090.987, CCC\u2009=\u20090.987, p-value\u2009=\u20090.055, MAE\u2009=\u20090.63\u00b0). The average time required for angle measurement using only the CPU to execute the deep learning-based system was 11\u2009\u00b1\u20091\u00a0s. The deep learning-based automatic angle measurement system, a tool for diagnosing flatfoot, demonstrated comparable accuracy and reliability with the results obtained by medical professionals for patients without internal fixation devices.","doi":"10.1038\/s41598-024-69549-3","cleaned_title":"deep learning-based automated angle measurement for flatfoot diagnosis in weight-bearing lateral radiographs","cleaned_abstract":"this study aimed to develop and evaluate a deep learning-based system for the automatic measurement of angles (specifically, meary\u2019s angle and calcaneal pitch) in weight-bearing lateral radiographs of the foot for flatfoot diagnosis. we utilized 3960 lateral radiographs, either from the left or right foot, sourced from a pool of 4000 patients to construct and evaluate a deep learning-based model. these radiographs were captured between june and november 2021, and patients who had undergone total ankle replacement surgery or ankle arthrodesis surgery were excluded. various methods, including correlation analysis, bland\u2013altman plots, and paired t-tests, were employed to assess the concordance between the angles automatically measured using the system and those assessed by clinical experts. the evaluation dataset comprised 150 weight-bearing radiographs from 150 patients. in all test cases, the angles automatically computed using the deep learning-based system were in good agreement with the reference standards (meary\u2019s angle: pearson correlation coefficient (pcc) = 0.964, intraclass correlation coefficient (icc) = 0.963, concordance correlation coefficient (ccc) = 0.963, p-value = 0.632, mean absolute error (mae) = 1.59\u00b0; calcaneal pitch: pcc = 0.988, icc = 0.987, ccc = 0.987, p-value = 0.055, mae = 0.63\u00b0). the average time required for angle measurement using only the cpu to execute the deep learning-based system was 11 \u00b1 1 s. the deep learning-based automatic angle measurement system, a tool for diagnosing flatfoot, demonstrated comparable accuracy and reliability with the results obtained by medical professionals for patients without internal fixation devices.","key_phrases":["this study","a deep learning-based system","the automatic measurement","angles","(specifically, meary\u2019s angle","calcaneal pitch","weight-bearing lateral radiographs","the foot","flatfoot diagnosis","we","3960 lateral radiographs","the left or right foot","a pool","4000 patients","a deep learning-based model","these radiographs","june","november","patients","who","total ankle replacement surgery","ankle arthrodesis surgery","various methods","correlation analysis","bland\u2013altman plots","t-tests","the concordance","the angles","the system","those","clinical experts","the evaluation dataset","150 weight-bearing","150 patients","all test cases","the angles","the deep learning-based system","good agreement","the reference standards","meary\u2019s angle","pcc","intraclass correlation coefficient","icc","concordance correlation coefficient","ccc","absolute error","mae","1.59\u00b0","calcaneal pitch","the average time","angle measurement","only the cpu","the deep learning-based system","11 \u00b1","the deep learning-based automatic angle measurement system","a tool","comparable accuracy","reliability","the results","medical professionals","patients","internal fixation devices","3960","4000","between june and november 2021","150","150","0.964","0.963","0.963","0.632","1.59","0.988","0.987","0.987","0.055","0.63","11","1"]},{"title":"Presenting a three layer stacking ensemble classifier of deep learning and machine learning for skin cancer classification","authors":["Bahman Jafari Tabaghsar","Reza Tavoli","Mohammad Mahdi Alizadeh Toosi"],"abstract":"One of the most common types of cancer in the world is skin cancer. Despite the different types of skin diseases with different shapes, the classification of skin diseases is a very difficult task. As a result, considering such a problem, a combination model of deep learning algorithms and machine has been proposed for skin disease classification.In this paper, a three-layer architecture based on ensemble learning is presented. In the first layer, the training input is given to convolutional neural network and EfficientNET. The output of the first layer is given to the classifiers of the second layer including machine learning classifiers. The output of the best decision of these classifiers is sent to the third layer classifier and the final prediction is made.The reason for using the three-layer architecture based on group learning was the lack of correct recognition of some classes by simple classifications. On the other hand, some diseases with different classes are classified in the same class. This model helps to correctly identify input samples with the correct combination of classifications in different layers.HAM10000 data set has been used to test and validate the proposed method. The mentioned dataset includes 10,015 images of skin lesions in seven different classes and includes different types of skin diseases. The accuracy is 99.97 on the testing set, which was much better than the previous heavy models.","doi":"10.1007\/s11042-024-19195-8","cleaned_title":"presenting a three layer stacking ensemble classifier of deep learning and machine learning for skin cancer classification","cleaned_abstract":"one of the most common types of cancer in the world is skin cancer. despite the different types of skin diseases with different shapes, the classification of skin diseases is a very difficult task. as a result, considering such a problem, a combination model of deep learning algorithms and machine has been proposed for skin disease classification.in this paper, a three-layer architecture based on ensemble learning is presented. in the first layer, the training input is given to convolutional neural network and efficientnet. the output of the first layer is given to the classifiers of the second layer including machine learning classifiers. the output of the best decision of these classifiers is sent to the third layer classifier and the final prediction is made.the reason for using the three-layer architecture based on group learning was the lack of correct recognition of some classes by simple classifications. on the other hand, some diseases with different classes are classified in the same class. this model helps to correctly identify input samples with the correct combination of classifications in different layers.ham10000 data set has been used to test and validate the proposed method. the mentioned dataset includes 10,015 images of skin lesions in seven different classes and includes different types of skin diseases. the accuracy is 99.97 on the testing set, which was much better than the previous heavy models.","key_phrases":["the most common types","cancer","the world","skin cancer","the different types","skin diseases","different shapes","the classification","skin diseases","a very difficult task","a result","such a problem","a combination model","deep learning algorithms","machine","a three-layer architecture","ensemble learning","the first layer","the training input","convolutional neural network","efficientnet","the output","the first layer","the classifiers","the second layer","machine learning classifiers","the output","the best decision","these classifiers","the third layer classifier","the final prediction","made.the reason","the three-layer architecture","group learning","the lack","correct recognition","some classes","simple classifications","the other hand","some diseases","different classes","the same class","this model","input samples","the correct combination","classifications","different layers.ham10000 data set","the proposed method","the mentioned dataset","10,015 images","skin lesions","seven different classes","different types","skin diseases","the accuracy","the testing set","which","the previous heavy models","one","three","first","first","second","third","three","10,015","seven","99.97"]},{"title":"Beschleunigte muskuloskeletale Magnetresonanztomographie mit Deep-Learning-gest\u00fctzter Bildrekonstruktion bei 0,55\u202fT\u20133\u202fT","authors":["Jan Vosshenrich","Jan Fritz"],"abstract":"Klinisches\/methodisches ProblemDie Magnetresonanztomographie (MRT) ist ein zentraler Bestandteil der muskuloskeletalen Diagnostik. Lange Akquisitionszeiten k\u00f6nnen jedoch zu Einschr\u00e4nkungen in der klinischen Praxis f\u00fchren.Radiologische StandardverfahrenDie MRT hat sich aufgrund des hohen Aufl\u00f6sungsverm\u00f6gens und Signal-zu-Rausch-Verh\u00e4ltnisses (SNR) sowie des exzellenten Weichteilkontrastes als Modalit\u00e4t der Wahl in der Diagnostik von Verletzungen und Erkrankungen des muskuloskeletalen Systems etabliert.Methodische InnovationenKontinuierliche Weiterentwicklungen in der Hard- und Softwaretechnologie haben eine bildqualit\u00e4ts- und genauigkeitsneutrale Beschleunigung von 2D-Turbo-Spin-Echo(TSE)-Sequenzen um den Faktor\u00a04 erm\u00f6glicht. K\u00fcrzlich vorgestellte, auf Deep Learning (DL) basierende Bildrekonstruktionsalgorithmen helfen, die Abh\u00e4ngigkeit zwischen SNR, r\u00e4umlicher Aufl\u00f6sung und Akquisitionszeit weiter zu minimieren und erlauben die Anwendung h\u00f6herer Beschleunigungsfaktoren.Leistungsf\u00e4higkeitDie kombinierte Anwendung fortschrittlicher Beschleunigungstechniken und DL-basierter Bildrekonstruktion birgt enormes Potenzial, um die Effizienz, den Patientenkomfort und die Zug\u00e4nglichkeit der muskuloskeletalen MRT bei gleichbleibend hoher diagnostischer Genauigkeit zu maximieren.BewertungDL-rekonstruierte beschleunigte MRT-Untersuchungen haben ihre Praxisreife und ihren Mehrwert innerhalb k\u00fcrzester Zeit unter Beweis gestellt. Aktuelle wissenschaftliche Erkenntnisse legen nahe, dass das Potenzial dieser Technologie noch nicht ausgesch\u00f6pft ist.Empfehlung f\u00fcr die PraxisBeschleunigte MRT-Untersuchungen mit DL-gest\u00fctzter Bildrekonstruktion k\u00f6nnen zuverl\u00e4ssig in der Prim\u00e4rdiagnostik und Verlaufskontrolle muskuloskeletaler Fragestellungen eingesetzt werden.","doi":"10.1007\/s00117-024-01325-w","cleaned_title":"beschleunigte muskuloskeletale magnetresonanztomographie mit deep-learning-gest\u00fctzter bildrekonstruktion bei 0,55 t\u20133 t","cleaned_abstract":"klinisches\/methodisches problemdie magnetresonanztomographie (mrt) ist ein zentraler bestandteil der muskuloskeletalen diagnostik. lange akquisitionszeiten k\u00f6nnen jedoch zu einschr\u00e4nkungen in der klinischen praxis f\u00fchren.radiologische standardverfahrendie mrt hat sich aufgrund des hohen aufl\u00f6sungsverm\u00f6gens und signal-zu-rausch-verh\u00e4ltnisses (snr) sowie des exzellenten weichteilkontrastes als modalit\u00e4t der wahl in der diagnostik von verletzungen und erkrankungen des muskuloskeletalen systems etabliert.methodische innovationenkontinuierliche weiterentwicklungen in der hard- und softwaretechnologie haben eine bildqualit\u00e4ts- und genauigkeitsneutrale beschleunigung von 2d-turbo-spin-echo(tse)-sequenzen um den faktor 4 erm\u00f6glicht. k\u00fcrzlich vorgestellte, auf deep learning (dl) basierende bildrekonstruktionsalgorithmen helfen, die abh\u00e4ngigkeit zwischen snr, r\u00e4umlicher aufl\u00f6sung und akquisitionszeit weiter zu minimieren und erlauben die anwendung h\u00f6herer beschleunigungsfaktoren.leistungsf\u00e4higkeitdie kombinierte anwendung fortschrittlicher beschleunigungstechniken und dl-basierter bildrekonstruktion birgt enormes potenzial, um die effizienz, den patientenkomfort und die zug\u00e4nglichkeit der muskuloskeletalen mrt bei gleichbleibend hoher diagnostischer genauigkeit zu maximieren.bewertungdl-rekonstruierte beschleunigte mrt-untersuchungen haben ihre praxisreife und ihren mehrwert innerhalb k\u00fcrzester zeit unter beweis gestellt. aktuelle wissenschaftliche erkenntnisse legen nahe, dass das potenzial dieser technologie noch nicht ausgesch\u00f6pft ist.empfehlung f\u00fcr die praxisbeschleunigte mrt-untersuchungen mit dl-gest\u00fctzter bildrekonstruktion k\u00f6nnen zuverl\u00e4ssig in der prim\u00e4rdiagnostik und verlaufskontrolle muskuloskeletaler fragestellungen eingesetzt werden.","key_phrases":["klinisches\/methodisches problemdie magnetresonanztomographie","(mrt","lange akquisitionszeiten k\u00f6nnen jedoch zu einschr\u00e4nkungen","der","standardverfahrendie mrt hat","des hohen aufl\u00f6sungsverm\u00f6gens","signal-zu-rausch-verh\u00e4ltnisses","snr","als modalit\u00e4t der","der hard- und softwaretechnologie","haben eine","und genauigkeitsneutrale beschleunigung von 2d-turbo-spin-echo(tse)-sequenzen um den faktor 4 erm\u00f6glicht","k\u00fcrzlich vorgestellte","auf","deep learning","dl","bildrekonstruktionsalgorithmen","abh\u00e4ngigkeit zwischen snr","r\u00e4umlicher aufl\u00f6sung und akquisitionszeit weiter zu minimieren und erlauben die anwendung h\u00f6herer beschleunigungsfaktoren.leistungsf\u00e4higkeitdie kombinierte anwendung fortschrittlicher","und dl-basierter bildrekonstruktion birgt","die effizienz, den patientenkomfort und die zug\u00e4nglichkeit der","mrt bei","hoher diagnostischer genauigkeit","zu maximieren.bewertungdl-rekonstruierte beschleunigte mrt-untersuchungen","ihre praxisreife","ihren mehrwert","k\u00fcrzester zeit unter beweis gestellt","dass das potenzial dieser technologie noch nicht","praxisbeschleunigte mrt-untersuchungen mit dl-gest\u00fctzter bildrekonstruktion k\u00f6nnen zuverl\u00e4ssig","der prim\u00e4rdiagnostik und verlaufskontrolle muskuloskeletaler fragestellungen eingesetzt werden","klinisches","problemdie magnetresonanztomographie","ein zentraler","bestandteil der muskuloskeletalen diagnostik","f\u00fchren.radiologische standardverfahrendie mrt","von verletzungen und erkrankungen des muskuloskeletalen","beschleunigung von","2d","4","k\u00fcrzlich vorgestellte","bildrekonstruktionsalgorithmen helfen","zwischen snr","akquisitionszeit","zu minimieren","anwendung h\u00f6herer","anwendung fortschrittlicher beschleunigungstechniken","zug\u00e4nglichkeit der muskuloskeletalen mrt","ihren mehrwert","k\u00fcrzester zeit","aktuelle wissenschaftliche erkenntnisse legen","das","verlaufskontrolle","fragestellungen"]},{"title":"Deep-learning based supervisory monitoring of robotized DE-GMAW process through learning from human welders","authors":["Rui Yu","Yue Cao","Jennifer Martin","Otto Chiang","YuMing Zhang"],"abstract":"Double-electrode gas metal arc welding (DE-GMAW) modifies GMAW by adding a second electrode to bypass a portion of the current flowing from the wire. This reduces the current to, and the heat input on, the workpiece. Successful bypassing depends on the relative position of the bypass electrode to the continuously varying wire tip. To ensure proper operation, we propose robotizing the system using a follower robot to carry and adaptively adjust the bypass electrode. The primary information for monitoring this process is the arc image, which directly shows desired and undesired modes. However, developing a robust algorithm for processing the complex arc image is time-consuming and challenging. Employing a deep learning approach requires labeling numerous arc images for the corresponding DE-GMAW modes, which is not practically feasible. To introduce alternative labels, we analyze arc phenomena in various DE-GMAW modes and correlate them with distinct arc systems having varying voltages. These voltages serve as automatically derived labels to train the deep-learning network. The results demonstrated reliable process monitoring.","doi":"10.1007\/s40194-023-01635-y","cleaned_title":"deep-learning based supervisory monitoring of robotized de-gmaw process through learning from human welders","cleaned_abstract":"double-electrode gas metal arc welding (de-gmaw) modifies gmaw by adding a second electrode to bypass a portion of the current flowing from the wire. this reduces the current to, and the heat input on, the workpiece. successful bypassing depends on the relative position of the bypass electrode to the continuously varying wire tip. to ensure proper operation, we propose robotizing the system using a follower robot to carry and adaptively adjust the bypass electrode. the primary information for monitoring this process is the arc image, which directly shows desired and undesired modes. however, developing a robust algorithm for processing the complex arc image is time-consuming and challenging. employing a deep learning approach requires labeling numerous arc images for the corresponding de-gmaw modes, which is not practically feasible. to introduce alternative labels, we analyze arc phenomena in various de-gmaw modes and correlate them with distinct arc systems having varying voltages. these voltages serve as automatically derived labels to train the deep-learning network. the results demonstrated reliable process monitoring.","key_phrases":["double-electrode gas metal arc welding","de","-","gmaw","gmaw","a second electrode","a portion","the current","the wire","this","the heat input","the workpiece","successful bypassing","the relative position","the bypass","the continuously varying wire tip","proper operation","we","the system","a follower robot","the bypass","the primary information","this process","the arc image","which","desired and undesired modes","a robust algorithm","the complex arc image","a deep learning approach","numerous arc images","the corresponding de-gmaw modes","which","alternative labels","we","arc phenomena","various de-gmaw modes","them","distinct arc systems","varying voltages","these voltages","automatically derived labels","the deep-learning network","the results","reliable process monitoring","second"]},{"title":"Fluid dynamic control and optimization using deep reinforcement learning","authors":["Innyoung Kim","Donghyun You"],"abstract":"This paper presents a review of recent research on applying deep reinforcement learning in fluid dynamics. Reinforcement learning is a technique in which the agent autonomously learns optimal action strategies while interacting with the environment, mimicking human learning mechanisms. Combined with artificial intelligence technology, it is providing a new direction in fluid dynamic control and optimization, which were challenging due to the nonlinear and high-dimensional characteristics of the fluid. In the section on fluid dynamic control, control strategies for drag reduction and research on controlling biological motion are reviewed. The optimization section focuses on shape optimization and automation of computational fluid dynamics. Current challenges and possible future developments are also described.Graphical Abstract","doi":"10.1007\/s42791-024-00067-z","cleaned_title":"fluid dynamic control and optimization using deep reinforcement learning","cleaned_abstract":"this paper presents a review of recent research on applying deep reinforcement learning in fluid dynamics. reinforcement learning is a technique in which the agent autonomously learns optimal action strategies while interacting with the environment, mimicking human learning mechanisms. combined with artificial intelligence technology, it is providing a new direction in fluid dynamic control and optimization, which were challenging due to the nonlinear and high-dimensional characteristics of the fluid. in the section on fluid dynamic control, control strategies for drag reduction and research on controlling biological motion are reviewed. the optimization section focuses on shape optimization and automation of computational fluid dynamics. current challenges and possible future developments are also described.graphical abstract","key_phrases":["this paper","a review","recent research","deep reinforcement learning","fluid dynamics","reinforcement learning","a technique","which","the agent","optimal action strategies","the environment","human learning mechanisms","artificial intelligence technology","it","a new direction","fluid dynamic control","optimization","which","the nonlinear and high-dimensional characteristics","the fluid","the section","fluid dynamic control","control strategies","drag reduction","research","biological motion","the optimization section","shape optimization","automation","computational fluid dynamics","current challenges","possible future developments"]},{"title":"Deep-learning based artificial intelligence tool for melt pools and defect segmentation","authors":["Amra Peles","Vincent C. Paquit","Ryan R. Dehoff"],"abstract":"Accelerating fabrication of additively manufactured components with precise microstructures is important for quality and qualification of built parts, as well as for a fundamental understanding of process improvement. Accomplishing this requires fast and robust characterization of melt pool geometries and structural defects in images. This paper proposes a pragmatic approach based on implementation of deep learning models and self-consistent workflow that enable systematic segmentation of defects and melt pools in optical images. Deep learning is based on an image-to-image translation\u2013conditional generative adversarial neural network architecture. An artificial intelligence (AI) tool based on this deep learning model enables fast and incrementally more accurate predictions of the prevalent geometric features, including melt pool boundaries and printing-induced structural defects. We present statistical analysis of geometric features that is enabled by the AI tool, showing strong spatial correlation of defects and the melt pool boundaries. The correlations of widths and heights of melt pools with dataset processing parameters show the highest sensitivity to thermal influences resulting from laser passes in adjacent and subsequent layer passes. The presented models and tools are demonstrated on the aluminum alloy and datasets produced with different sets of processing parameters. However, they have universal quality and could easily be adapted to different material compositions. The method can be easily generalized to microstructural characterizations other than optical microscopy.","doi":"10.1007\/s10845-024-02457-5","cleaned_title":"deep-learning based artificial intelligence tool for melt pools and defect segmentation","cleaned_abstract":"accelerating fabrication of additively manufactured components with precise microstructures is important for quality and qualification of built parts, as well as for a fundamental understanding of process improvement. accomplishing this requires fast and robust characterization of melt pool geometries and structural defects in images. this paper proposes a pragmatic approach based on implementation of deep learning models and self-consistent workflow that enable systematic segmentation of defects and melt pools in optical images. deep learning is based on an image-to-image translation\u2013conditional generative adversarial neural network architecture. an artificial intelligence (ai) tool based on this deep learning model enables fast and incrementally more accurate predictions of the prevalent geometric features, including melt pool boundaries and printing-induced structural defects. we present statistical analysis of geometric features that is enabled by the ai tool, showing strong spatial correlation of defects and the melt pool boundaries. the correlations of widths and heights of melt pools with dataset processing parameters show the highest sensitivity to thermal influences resulting from laser passes in adjacent and subsequent layer passes. the presented models and tools are demonstrated on the aluminum alloy and datasets produced with different sets of processing parameters. however, they have universal quality and could easily be adapted to different material compositions. the method can be easily generalized to microstructural characterizations other than optical microscopy.","key_phrases":["accelerating fabrication","additively manufactured components","precise microstructures","quality","qualification","built parts","a fundamental understanding","process improvement","this","fast and robust characterization","melt pool geometries","structural defects","images","this paper","a pragmatic approach","implementation","deep learning models","self-consistent workflow","that","systematic segmentation","defects","pools","optical images","deep learning","image","conditional generative adversarial neural network architecture","an artificial intelligence","(ai) tool","this deep learning model","fast and incrementally more accurate predictions","the prevalent geometric features","pool boundaries","printing-induced structural defects","we","statistical analysis","geometric features","that","the ai tool","strong spatial correlation","defects","the melt pool boundaries","the correlations","widths","heights","melt pools","dataset processing parameters","the highest sensitivity","thermal influences","laser passes","adjacent and subsequent layer passes","the presented models","tools","the aluminum alloy","datasets","different sets","processing parameters","they","universal quality","different material compositions","the method","microstructural characterizations","optical microscopy"]},{"title":"Multi-echelon inventory optimization using deep reinforcement learning","authors":["Kevin Geevers","Lotte van Hezewijk","Martijn R. K. Mes"],"abstract":"This paper studies the applicability of a deep reinforcement learning approach to three different multi-echelon inventory systems, with the objective of minimizing the holding and backorder costs. First, we conduct an extensive literature review to map the current applications of reinforcement learning in multi-echelon inventory systems. Next, we apply our deep reinforcement learning method to three cases with different network structures (linear, divergent, and general structures). The linear and divergent cases are derived from literature, whereas the general case is based on a real-life manufacturer. We apply the proximal policy optimization (PPO) algorithm, with a continuous action space, and show that it consistently outperforms the benchmark solution. It achieves an average improvement of 16.4% for the linear case, 11.3% for the divergent case, and 6.6% for the general case. We explain the limitations of our approach and propose avenues for future research.","doi":"10.1007\/s10100-023-00872-2","cleaned_title":"multi-echelon inventory optimization using deep reinforcement learning","cleaned_abstract":"this paper studies the applicability of a deep reinforcement learning approach to three different multi-echelon inventory systems, with the objective of minimizing the holding and backorder costs. first, we conduct an extensive literature review to map the current applications of reinforcement learning in multi-echelon inventory systems. next, we apply our deep reinforcement learning method to three cases with different network structures (linear, divergent, and general structures). the linear and divergent cases are derived from literature, whereas the general case is based on a real-life manufacturer. we apply the proximal policy optimization (ppo) algorithm, with a continuous action space, and show that it consistently outperforms the benchmark solution. it achieves an average improvement of 16.4% for the linear case, 11.3% for the divergent case, and 6.6% for the general case. we explain the limitations of our approach and propose avenues for future research.","key_phrases":["this paper","the applicability","a deep reinforcement learning approach","three different multi-echelon inventory systems","the objective","the holding and backorder costs","we","an extensive literature review","the current applications","reinforcement learning","multi-echelon inventory systems","we","our deep reinforcement learning method","three cases","different network structures","general structures","the linear and divergent cases","literature","the general case","a real-life manufacturer","we","the proximal policy optimization","ppo","algorithm","a continuous action space","it","the benchmark solution","it","an average improvement","16.4%","the linear case","11.3%","the divergent case","6.6%","the general case","we","the limitations","our approach","avenues","future research","three","first","three","linear","linear","16.4%","11.3%","6.6%"]},{"title":"An Analysis of Plant Diseases on Detection and Classification: From Machine Learning to Deep Learning Techniques","authors":["P. K. Midhunraj","K. S. Thivya","M. Anand"],"abstract":"Plants are acknowledged as being crucial because they are the main source of human energy generation due to their nutritional, therapeutic, and other benefits. Therefore, it is necessary to increase crop productivity. One of these significant factors contributing to reduced agricultural yields is the prevalence of bacterial, fungal, and viral illnesses. Applying techniques for plant disease identification can stop and treat these diseases. So, numerous machine learning (ML) and deep learning (DL) methods were created and tested by researchers to identify plant diseases. Therefore, this study gives a detailed discussion of the various research studies conducted in plant disease detection utilizing\u00a0ML and DL-based techniques. This review offers research advancements in plant disease recognition from ML to DL techniques. Additionally, many datasets about plant diseases are thoroughly examined. It also addresses the difficulties and issues with the current systems.","doi":"10.1007\/s11042-023-17600-2","cleaned_title":"an analysis of plant diseases on detection and classification: from machine learning to deep learning techniques","cleaned_abstract":"plants are acknowledged as being crucial because they are the main source of human energy generation due to their nutritional, therapeutic, and other benefits. therefore, it is necessary to increase crop productivity. one of these significant factors contributing to reduced agricultural yields is the prevalence of bacterial, fungal, and viral illnesses. applying techniques for plant disease identification can stop and treat these diseases. so, numerous machine learning (ml) and deep learning (dl) methods were created and tested by researchers to identify plant diseases. therefore, this study gives a detailed discussion of the various research studies conducted in plant disease detection utilizing ml and dl-based techniques. this review offers research advancements in plant disease recognition from ml to dl techniques. additionally, many datasets about plant diseases are thoroughly examined. it also addresses the difficulties and issues with the current systems.","key_phrases":["plants","they","the main source","human energy generation","their nutritional, therapeutic, and other benefits","it","crop productivity","these significant factors","reduced agricultural yields","the prevalence","bacterial, fungal, and viral illnesses","techniques","plant disease identification","these diseases","numerous machine learning","ml","deep learning (dl) methods","researchers","plant diseases","this study","a detailed discussion","the various research studies","plant disease detection","ml","dl-based techniques","this review","research advancements","plant disease recognition","ml","dl techniques","many datasets","plant diseases","it","the difficulties","issues","the current systems","one"]},{"title":"The Mori\u2013Zwanzig formulation of deep learning","authors":["Daniele Venturi","Xiantao Li"],"abstract":"We develop a new formulation of deep learning based on the Mori\u2013Zwanzig (MZ) formalism of irreversible statistical mechanics. The new formulation is built upon the well-known duality between deep neural networks and discrete dynamical systems, and it allows us to directly propagate quantities of interest (conditional expectations and probability density functions) forward and backward through the network by means of exact linear operator equations. Such new equations can be used as a starting point to develop new effective parameterizations of deep neural networks and provide a new framework to study deep learning via operator-theoretic methods. The proposed MZ formulation of deep learning naturally introduces a new concept, i.e., the memory of the neural network, which plays a fundamental role in low-dimensional modeling and parameterization. By using the theory of contraction mappings, we develop sufficient conditions for the memory of the neural network to decay with the number of layers. This allows us to rigorously transform deep networks into shallow ones, e.g., by reducing the number of neurons per layer (using projection operators), or by reducing the total number of layers (using the decay property of the memory operator).","doi":"10.1007\/s40687-023-00390-2","cleaned_title":"the mori\u2013zwanzig formulation of deep learning","cleaned_abstract":"we develop a new formulation of deep learning based on the mori\u2013zwanzig (mz) formalism of irreversible statistical mechanics. the new formulation is built upon the well-known duality between deep neural networks and discrete dynamical systems, and it allows us to directly propagate quantities of interest (conditional expectations and probability density functions) forward and backward through the network by means of exact linear operator equations. such new equations can be used as a starting point to develop new effective parameterizations of deep neural networks and provide a new framework to study deep learning via operator-theoretic methods. the proposed mz formulation of deep learning naturally introduces a new concept, i.e., the memory of the neural network, which plays a fundamental role in low-dimensional modeling and parameterization. by using the theory of contraction mappings, we develop sufficient conditions for the memory of the neural network to decay with the number of layers. this allows us to rigorously transform deep networks into shallow ones, e.g., by reducing the number of neurons per layer (using projection operators), or by reducing the total number of layers (using the decay property of the memory operator).","key_phrases":["we","a new formulation","deep learning","the mori","zwanzig (mz) formalism","irreversible statistical mechanics","the new formulation","the well-known duality","deep neural networks","discrete dynamical systems","it","us","quantities","interest","conditional expectations","probability density functions","the network","means","exact linear operator equations","such new equations","a starting point","new effective parameterizations","deep neural networks","a new framework","deep learning","operator-theoretic methods","the proposed mz formulation","deep learning","a new concept","the neural network","which","a fundamental role","low-dimensional modeling","parameterization","the theory","contraction mappings","we","sufficient conditions","the memory","the neural network","the number","layers","this","us","deep networks","shallow ones","the number","neurons","layer","projection operators","the total number","layers","the decay property","the memory operator","linear"]},{"title":"XDeMo: a novel deep learning framework for DNA motif mining using transformer models","authors":["Rajashree Chaurasia","Udayan Ghose"],"abstract":"Motivation: Recognizing and studying DNA patterns is crucial for improving knowledge of illnesses, cell function, and gene control. Motifs determine which transcription factor a protein may bind to, leading to a better unraveling of gene expression. Advancements in the fields of deep learning and high-throughput sequencing have made possible the exploration of motif discovery anew, with greater accuracy and performance. Methodology: In this paper, a novel deep learning framework (XDeMo \u2013 Transformer-based Deep Motifs) for DNA motif mining using Transformer models is proposed. Furthermore, a hybrid encoding scheme is also introduced, called \u2018blended\u2019 encoding specifically designed for use with deep learning transformer models that are trained using DNA sequences. Results: Our proposed transformer-based framework for DNA motif discovery augmented by blended encoding outperforms many state-of-the-art deep learning models on many baseline performance metrics when trained on the standard datasets. Our models demonstrated robust performance in predicting motifs with high discriminative power, precision, recall, and F1 score. Conclusion: The model\u2019s ability to capture intricate sequence patterns and long-range dependencies led to the discovery of biologically meaningful motifs that were verified from known transcription factor binding motif databases. This shows that our novel framework can be effectively used to find DNA motifs and therefore, aid in further downstream analyses for biomedical and biotechnological applications.","doi":"10.1007\/s13721-024-00463-4","cleaned_title":"xdemo: a novel deep learning framework for dna motif mining using transformer models","cleaned_abstract":"motivation: recognizing and studying dna patterns is crucial for improving knowledge of illnesses, cell function, and gene control. motifs determine which transcription factor a protein may bind to, leading to a better unraveling of gene expression. advancements in the fields of deep learning and high-throughput sequencing have made possible the exploration of motif discovery anew, with greater accuracy and performance. methodology: in this paper, a novel deep learning framework (xdemo \u2013 transformer-based deep motifs) for dna motif mining using transformer models is proposed. furthermore, a hybrid encoding scheme is also introduced, called \u2018blended\u2019 encoding specifically designed for use with deep learning transformer models that are trained using dna sequences. results: our proposed transformer-based framework for dna motif discovery augmented by blended encoding outperforms many state-of-the-art deep learning models on many baseline performance metrics when trained on the standard datasets. our models demonstrated robust performance in predicting motifs with high discriminative power, precision, recall, and f1 score. conclusion: the model\u2019s ability to capture intricate sequence patterns and long-range dependencies led to the discovery of biologically meaningful motifs that were verified from known transcription factor binding motif databases. this shows that our novel framework can be effectively used to find dna motifs and therefore, aid in further downstream analyses for biomedical and biotechnological applications.","key_phrases":["dna patterns","knowledge","illnesses","cell function","gene control","motifs","which transcription factor","a protein","a better unraveling","gene expression","advancements","the fields","deep learning","high-throughput sequencing","the exploration","motif discovery","greater accuracy","performance","methodology","this paper","a novel deep learning framework","xdemo \u2013 transformer-based deep motifs","dna motif mining","transformer models","a hybrid encoding scheme","\u2018blended\u2019 encoding","use","deep learning transformer models","that","dna sequences","results","our proposed transformer-based framework","dna motif discovery","blended encoding outperforms","the-art","many baseline performance metrics","the standard datasets","our models","robust performance","motifs","high discriminative power","precision","recall","f1 score","conclusion","the model\u2019s ability","intricate sequence patterns","long-range dependencies","the discovery","biologically meaningful motifs","that","known transcription factor binding motif databases","this","our novel framework","dna motifs",", aid","further downstream analyses","biomedical and biotechnological applications"]},{"title":"DeepDOF-SE: affordable deep-learning microscopy platform for slide-free histology","authors":["Lingbo Jin","Yubo Tang","Jackson B. Coole","Melody T. Tan","Xuan Zhao","Hawraa Badaoui","Jacob T. Robinson","Michelle D. Williams","Nadarajah Vigneswaran","Ann M. Gillenwater","Rebecca R. Richards-Kortum","Ashok Veeraraghavan"],"abstract":"Histopathology plays a critical role in the diagnosis and surgical management of cancer. However, access to histopathology services, especially frozen section pathology during surgery, is limited in resource-constrained settings because preparing slides from resected tissue is time-consuming, labor-intensive, and requires expensive infrastructure. Here, we report a deep-learning-enabled microscope, named DeepDOF-SE, to rapidly scan intact tissue at cellular resolution without the need for physical sectioning. Three key features jointly make DeepDOF-SE practical. First, tissue specimens are stained directly with inexpensive vital fluorescent dyes and optically sectioned with ultra-violet excitation that localizes fluorescent emission to a thin surface layer. Second, a deep-learning algorithm extends the depth-of-field, allowing rapid acquisition of in-focus images from large areas of tissue even when the tissue surface is highly irregular. Finally, a semi-supervised generative adversarial network virtually stains DeepDOF-SE fluorescence images with hematoxylin-and-eosin appearance, facilitating image interpretation by pathologists without significant additional training. We developed the DeepDOF-SE platform using a data-driven approach and validated its performance by imaging surgical resections of suspected oral tumors. Our results show that DeepDOF-SE provides histological information of diagnostic importance, offering a rapid and affordable slide-free histology platform for intraoperative tumor margin assessment and in low-resource settings.","doi":"10.1038\/s41467-024-47065-2","cleaned_title":"deepdof-se: affordable deep-learning microscopy platform for slide-free histology","cleaned_abstract":"histopathology plays a critical role in the diagnosis and surgical management of cancer. however, access to histopathology services, especially frozen section pathology during surgery, is limited in resource-constrained settings because preparing slides from resected tissue is time-consuming, labor-intensive, and requires expensive infrastructure. here, we report a deep-learning-enabled microscope, named deepdof-se, to rapidly scan intact tissue at cellular resolution without the need for physical sectioning. three key features jointly make deepdof-se practical. first, tissue specimens are stained directly with inexpensive vital fluorescent dyes and optically sectioned with ultra-violet excitation that localizes fluorescent emission to a thin surface layer. second, a deep-learning algorithm extends the depth-of-field, allowing rapid acquisition of in-focus images from large areas of tissue even when the tissue surface is highly irregular. finally, a semi-supervised generative adversarial network virtually stains deepdof-se fluorescence images with hematoxylin-and-eosin appearance, facilitating image interpretation by pathologists without significant additional training. we developed the deepdof-se platform using a data-driven approach and validated its performance by imaging surgical resections of suspected oral tumors. our results show that deepdof-se provides histological information of diagnostic importance, offering a rapid and affordable slide-free histology platform for intraoperative tumor margin assessment and in low-resource settings.","key_phrases":["histopathology","a critical role","the diagnosis","surgical management","cancer","access","histopathology services","especially frozen section pathology","surgery","resource-constrained settings","slides","resected tissue","expensive infrastructure","we","a deep-learning-enabled microscope","deepdof-se","intact tissue","cellular resolution","the need","physical sectioning","three key features","deepdof-se","tissue specimens","inexpensive vital fluorescent dyes","ultra-violet excitation","that","fluorescent emission","a thin surface layer","a deep-learning algorithm","the depth","field","rapid acquisition","focus","large areas","tissue","the tissue surface","a semi-supervised generative adversarial network","deepdof-se fluorescence images","hematoxylin-and-eosin appearance","image interpretation","pathologists","significant additional training","we","the deepdof-se platform","a data-driven approach","its performance","surgical resections","suspected oral tumors","our results","deepdof-se","histological information","diagnostic importance","a rapid and affordable slide-free histology platform","intraoperative tumor margin assessment","low-resource settings","three","first","second","hematoxylin"]},{"title":"Identification of rural courtyards\u2019 utilization status using deep learning and machine learning methods on unmanned aerial vehicle images in north China","authors":["Maojun Wang","Wenyu Xu","Guangzhong Cao","Tao Liu"],"abstract":"The issue of unoccupied or abandoned homesteads (courtyards) in China emerges given the increasing aging population, rapid urbanization and massive rural-urban migration. From the aspect of rural vitalization, land-use planning, and policy making, determining the number of unoccupied courtyards is important. Field and questionnaire-based surveys were currently the main approaches, but these traditional methods were often expensive and laborious. A new workflow is explored using deep learning and machine learning algorithms on unmanned aerial vehicle (UAV) images. Initially, features of the built environment were extracted using deep learning to evaluate the courtyard management, including extracting complete or collapsed farmhouses by Alexnet, detecting solar water heaters by YOLOv5s, calculating green looking ratio (GLR) by FCN. Their precisions exceeded 98%. Then, seven machine learning algorithms (Adaboost, binomial logistic regression, neural network, random forest, support vector machine, decision trees, and XGBoost algorithms) were applied to identify the rural courtyards\u2019 utilization status. The Adaboost algorithm showed the best performance with the comprehensive consideration of most metrics (Accuracy: 0.933, Precision: 0.932, Recall: 0.984, F1-score: 0.957). Results showed that identifying the courtyards\u2019 utilization statuses based on the courtyard built environment is feasible. It is transferable and cost-effective for large-scale village surveys, and may contribute to the intensive and sustainable approach to rural land use.","doi":"10.1007\/s12273-023-1099-9","cleaned_title":"identification of rural courtyards\u2019 utilization status using deep learning and machine learning methods on unmanned aerial vehicle images in north china","cleaned_abstract":"the issue of unoccupied or abandoned homesteads (courtyards) in china emerges given the increasing aging population, rapid urbanization and massive rural-urban migration. from the aspect of rural vitalization, land-use planning, and policy making, determining the number of unoccupied courtyards is important. field and questionnaire-based surveys were currently the main approaches, but these traditional methods were often expensive and laborious. a new workflow is explored using deep learning and machine learning algorithms on unmanned aerial vehicle (uav) images. initially, features of the built environment were extracted using deep learning to evaluate the courtyard management, including extracting complete or collapsed farmhouses by alexnet, detecting solar water heaters by yolov5s, calculating green looking ratio (glr) by fcn. their precisions exceeded 98%. then, seven machine learning algorithms (adaboost, binomial logistic regression, neural network, random forest, support vector machine, decision trees, and xgboost algorithms) were applied to identify the rural courtyards\u2019 utilization status. the adaboost algorithm showed the best performance with the comprehensive consideration of most metrics (accuracy: 0.933, precision: 0.932, recall: 0.984, f1-score: 0.957). results showed that identifying the courtyards\u2019 utilization statuses based on the courtyard built environment is feasible. it is transferable and cost-effective for large-scale village surveys, and may contribute to the intensive and sustainable approach to rural land use.","key_phrases":["the issue","unoccupied or abandoned homesteads","courtyards","china","the increasing aging population","rapid urbanization","massive rural-urban migration","the aspect","rural vitalization","land-use planning","policy making","the number","unoccupied courtyards","field","questionnaire-based surveys","the main approaches","these traditional methods","a new workflow","deep learning and machine learning algorithms","unmanned aerial vehicle","(uav) images","features","the built environment","deep learning","the courtyard management","complete or collapsed farmhouses","alexnet","solar water heaters","yolov5s","green looking ratio","glr","fcn","their precisions","seven machine learning algorithms","adaboost, binomial logistic regression","neural network","random forest","support vector machine","decision trees","xgboost algorithms","the rural courtyards\u2019 utilization status","the adaboost algorithm","the best performance","the comprehensive consideration","most metrics","accuracy","precision","recall","f1-score","results","the courtyards\u2019 utilization statuses","the courtyard built environment","it","large-scale village surveys","the intensive and sustainable approach","rural land use","china","98%","seven","0.933","0.932","0.984","0.957"]},{"title":"A Transfer Learning-Based CNN Deep Learning Model for Unfavorable Driving State Recognition","authors":["Jichi Chen","Hong Wang","Enqiu He"],"abstract":"The detection of unfavorable driving states (UDS) of drivers based on electroencephalogram (EEG) measures has received continuous attention from extensive scholars on account of directly reflecting brain neural activity with high temporal resolution and low risk of being deceived. However, the existing EEG-based driver UDS detection methods involve limited exploration of the functional connectivity patterns and interaction relationships within the brain network. Therefore, there is still room for improvement in the accuracy of detection. In this project, we propose three pretrained convolutional neural network (CNN)-based automatic detection frameworks for UDS of drivers with 30-channel EEG signals. The frameworks are investigated by adjusting the learning rate and choosing the optimization solver, etc. Two different conditions of driving experiments are performed, collecting EEG signals from sixteen subjects. The acquired 1-dimensional 30-channel EEG signals are converted into 2-dimensional matrices by the Granger causality (GC) method to form the functional connectivity graphs of the brain (FCGB). Then, the FCGB are fed into pretrained deep learning models that employed transfer learning strategy for feature extraction and judgment of different EEG signal types. Furthermore, we adopt two visualization interpretability techniques, named, activation visualization and gradient-weighted class activation mapping (Grad-CAM) for better visualizing and understanding the predictions of the pretrained models after fine-tuning. The experimental outcomes show that Resnet 18 model yields the highest average recognition accuracy of 90% using the rmsprop optimizer with a learning rate of 1e\u2009\u2212\u20093. The overall outcomes suggest that cooperating of biologically inspired functional connectivity graphs of the brain and pretrained transfer learning algorithms is a prospective approach in reducing the rate of major traffic accidents caused by driver unfavorable driving states.","doi":"10.1007\/s12559-023-10196-7","cleaned_title":"a transfer learning-based cnn deep learning model for unfavorable driving state recognition","cleaned_abstract":"the detection of unfavorable driving states (uds) of drivers based on electroencephalogram (eeg) measures has received continuous attention from extensive scholars on account of directly reflecting brain neural activity with high temporal resolution and low risk of being deceived. however, the existing eeg-based driver uds detection methods involve limited exploration of the functional connectivity patterns and interaction relationships within the brain network. therefore, there is still room for improvement in the accuracy of detection. in this project, we propose three pretrained convolutional neural network (cnn)-based automatic detection frameworks for uds of drivers with 30-channel eeg signals. the frameworks are investigated by adjusting the learning rate and choosing the optimization solver, etc. two different conditions of driving experiments are performed, collecting eeg signals from sixteen subjects. the acquired 1-dimensional 30-channel eeg signals are converted into 2-dimensional matrices by the granger causality (gc) method to form the functional connectivity graphs of the brain (fcgb). then, the fcgb are fed into pretrained deep learning models that employed transfer learning strategy for feature extraction and judgment of different eeg signal types. furthermore, we adopt two visualization interpretability techniques, named, activation visualization and gradient-weighted class activation mapping (grad-cam) for better visualizing and understanding the predictions of the pretrained models after fine-tuning. the experimental outcomes show that resnet 18 model yields the highest average recognition accuracy of 90% using the rmsprop optimizer with a learning rate of 1e \u2212 3. the overall outcomes suggest that cooperating of biologically inspired functional connectivity graphs of the brain and pretrained transfer learning algorithms is a prospective approach in reducing the rate of major traffic accidents caused by driver unfavorable driving states.","key_phrases":["the detection","unfavorable driving states","uds","drivers","electroencephalogram (eeg) measures","continuous attention","extensive scholars","account","brain neural activity","high temporal resolution","low risk","the existing eeg-based driver uds detection methods","limited exploration","the functional connectivity patterns","interaction relationships","the brain network","room","improvement","the accuracy","detection","this project","we","three pretrained convolutional neural network","cnn)-based automatic detection frameworks","uds","drivers","30-channel eeg signals","the frameworks","the learning rate","the optimization","two different conditions","driving experiments","eeg signals","sixteen subjects","the acquired 1-dimensional 30-channel eeg signals","2-dimensional matrices","the granger causality","(gc) method","the functional connectivity graphs","the brain","fcgb","the fcgb","pretrained deep learning models","that","transfer learning strategy","feature extraction","judgment","different eeg signal types","we","two visualization interpretability techniques","grad-cam","better visualizing","the predictions","the pretrained models","fine-tuning","the experimental outcomes","resnet","18 model","the highest average recognition accuracy","90%","the rmsprop optimizer","a learning rate","the overall outcomes","biologically inspired functional connectivity graphs","the brain","pretrained transfer learning algorithms","a prospective approach","the rate","major traffic accidents","driver unfavorable driving states","three","30","two","1","30","2","fed","two","18","90%","1e \u2212","3"]},{"title":"A hybrid deep model with cumulative learning for few-shot learning","authors":["Jiehao Liu","Zhao Yang","Liufei Luo","Mingkai Luo","Luyu Hu","Jiahao Li"],"abstract":"Few-shot learning (FSL) aims to recognize unseen classes with only a few samples for each class. This challenging research endeavors to narrow the gap between the computer vision technology and the human visual system. Recently, mainstream approaches for FSL can be grouped into meta-learning and classification learning. These two methods train the FSL model from local and global classification viewpoints respectively. In our work, we find the former method can effectively learn transferable knowledge (generalization capacity) with an episodic training paradigm but encounters the problem of slow convergence. The latter method can build an essential classification ability quickly (classification capacity) with a mini-batch training paradigm but easily causes an over-fitting problem. In light of this issue, we propose a hybrid deep model with cumulative learning to tackle the FSL problem by absorbing the advantages of the both methods. The proposed hybrid deep model innovatively integrates meta-learning and classification learning (IMC) in a unified two-branch network framework in which a meta-learning branch and a classification learning branch can work simultaneously. Besides, by considering the different characteristics of the two branches, we propose a cumulative learning strategy to take care of both generalization capacity learning and classification capacity learning in our IMC model training. With the proposed method, the model can quickly build the basic classification capability at the initial stage and continually mine discriminative class information during the remaining training for better generalization. Extensive experiments on CIFAR-FS, FC100, mini-ImageNet and tiered-ImageNet datasets are implemented to demonstrate the promising performance of our method.","doi":"10.1007\/s11042-022-14218-8","cleaned_title":"a hybrid deep model with cumulative learning for few-shot learning","cleaned_abstract":"few-shot learning (fsl) aims to recognize unseen classes with only a few samples for each class. this challenging research endeavors to narrow the gap between the computer vision technology and the human visual system. recently, mainstream approaches for fsl can be grouped into meta-learning and classification learning. these two methods train the fsl model from local and global classification viewpoints respectively. in our work, we find the former method can effectively learn transferable knowledge (generalization capacity) with an episodic training paradigm but encounters the problem of slow convergence. the latter method can build an essential classification ability quickly (classification capacity) with a mini-batch training paradigm but easily causes an over-fitting problem. in light of this issue, we propose a hybrid deep model with cumulative learning to tackle the fsl problem by absorbing the advantages of the both methods. the proposed hybrid deep model innovatively integrates meta-learning and classification learning (imc) in a unified two-branch network framework in which a meta-learning branch and a classification learning branch can work simultaneously. besides, by considering the different characteristics of the two branches, we propose a cumulative learning strategy to take care of both generalization capacity learning and classification capacity learning in our imc model training. with the proposed method, the model can quickly build the basic classification capability at the initial stage and continually mine discriminative class information during the remaining training for better generalization. extensive experiments on cifar-fs, fc100, mini-imagenet and tiered-imagenet datasets are implemented to demonstrate the promising performance of our method.","key_phrases":["few-shot learning","unseen classes","only a few samples","each class","this challenging research endeavors","the gap","the computer vision technology","the human visual system","mainstream approaches","fsl","meta-learning and classification learning","these two methods","the fsl model","local and global classification viewpoints","our work","we","the former method","transferable knowledge","generalization capacity","an episodic training paradigm","the problem","slow convergence","the latter method","an essential classification ability","classification capacity","a mini-batch training paradigm","an over-fitting problem","light","this issue","we","a hybrid deep model","cumulative learning","the fsl problem","the advantages","the both methods","the proposed hybrid deep model","meta-learning and classification learning","imc","a unified two-branch network framework","which","a meta-learning branch","a classification learning branch","the different characteristics","the two branches","we","a cumulative learning strategy","care","both generalization capacity learning and classification capacity","our imc model training","the proposed method","the model","the basic classification capability","the initial stage","continually mine discriminative class information","the remaining training","better generalization","extensive experiments","cifar-fs, fc100, mini","-","imagenet","tiered-imagenet datasets","the promising performance","our method","two","two","two"]},{"title":"Distributed source DOA estimation based on deep learning networks","authors":["Quan Tian","Ruiyan Cai","Gongrun Qiu","Yang Luo"],"abstract":"With space electromagnetic environments becoming increasingly complex, the direction of arrival (DOA) estimation based on the point source model can no longer meet the requirements of spatial target location. Based on the characteristics of the distributed source, a new DOA estimation algorithm based on deep learning is proposed. The algorithm first maps the distributed source model into the point source model via a generative adversarial network (GAN) and further combines the subspace-based method to achieve central DOA estimation. Second, by constructing a deep neural network (DNN), the covariance matrix of the received signals is used as the input to estimate the angular spread of the distributed source. The experimental results show that the proposed algorithm can achieve better performance than the existing methods for a distributed source.","doi":"10.1007\/s11760-024-03402-y","cleaned_title":"distributed source doa estimation based on deep learning networks","cleaned_abstract":"with space electromagnetic environments becoming increasingly complex, the direction of arrival (doa) estimation based on the point source model can no longer meet the requirements of spatial target location. based on the characteristics of the distributed source, a new doa estimation algorithm based on deep learning is proposed. the algorithm first maps the distributed source model into the point source model via a generative adversarial network (gan) and further combines the subspace-based method to achieve central doa estimation. second, by constructing a deep neural network (dnn), the covariance matrix of the received signals is used as the input to estimate the angular spread of the distributed source. the experimental results show that the proposed algorithm can achieve better performance than the existing methods for a distributed source.","key_phrases":["space electromagnetic environments","the direction","arrival","(doa) estimation","the point source model","the requirements","spatial target location","the characteristics","the distributed source","a new doa estimation algorithm","deep learning","algorithm","the distributed source model","the point source model","a generative adversarial network","gan","the subspace-based method","central doa estimation","a deep neural network","dnn","the covariance matrix","the received signals","the input","the angular spread","the distributed source","the experimental results","the proposed algorithm","better performance","the existing methods","a distributed source","first","second"]},{"title":"Deep learning reconstruction for lumbar spine MRI acceleration: a prospective study","authors":["Hui Tang","Ming Hong","Lu Yu","Yang Song","Mengqiu Cao","Lei Xiang","Yan Zhou","Shiteng Suo"],"abstract":"BackgroundWe compared magnetic resonance imaging (MRI) turbo spin-echo images reconstructed using a deep learning technique (TSE-DL) with standard turbo spin-echo (TSE-SD) images of the lumbar spine regarding image quality and detection performance of common degenerative pathologies.MethodsThis prospective, single-center study included 31 patients (15 males and 16 females; aged 51\u2009\u00b1\u200916\u00a0years (mean\u2009\u00b1\u2009standard deviation)) who underwent lumbar spine exams with both TSE-SD and TSE-DL acquisitions for degenerative spine diseases. Images were analyzed by two radiologists and assessed for qualitative image quality using a 4-point Likert scale, quantitative signal-to-noise ratio (SNR) of anatomic landmarks, and detection of common pathologies. Paired-sample t, Wilcoxon, and McNemar tests, unweighted\/linearly weighted Cohen \u03ba statistics, and intraclass correlation coefficients were used.ResultsScan time for TSE-DL and TSE-SD protocols was 2:55 and 5:17\u00a0min:s, respectively. The overall image quality was either significantly higher for TSE-DL or not significantly different between TSE-SD and TSE-DL. TSE-DL demonstrated higher SNR and subject noise scores than TSE-SD. For pathology detection, the interreader agreement was substantial to almost perfect for TSE-DL, with \u03ba values ranging from 0.61 to 1.00; the interprotocol agreement was almost perfect for both readers, with \u03ba values ranging from 0.84 to 1.00. There was no significant difference in the diagnostic confidence or detection rate of common pathologies between the two sequences (p\u2009\u2265\u20090.081).ConclusionsTSE-DL allowed for a 45% reduction in scan time over TSE-SD in lumbar spine MRI without compromising the overall image quality and showed comparable detection performance of common pathologies in the evaluation of degenerative lumbar spine changes.Relevance statementDeep learning-reconstructed lumbar spine MRI protocol enabled a 45% reduction in scan time compared with conventional reconstruction, with comparable image quality and detection performance of common degenerative pathologies.Key points\u2022 Lumbar spine MRI with deep learning reconstruction has broad application prospects.\u2022 Deep learning reconstruction of lumbar spine MRI saved 45% scan time without compromising overall image quality.\u2022 When compared with standard sequences, deep learning reconstruction showed similar detection performance of common degenerative lumbar spine pathologies.Graphical Abstract","doi":"10.1186\/s41747-024-00470-0","cleaned_title":"deep learning reconstruction for lumbar spine mri acceleration: a prospective study","cleaned_abstract":"backgroundwe compared magnetic resonance imaging (mri) turbo spin-echo images reconstructed using a deep learning technique (tse-dl) with standard turbo spin-echo (tse-sd) images of the lumbar spine regarding image quality and detection performance of common degenerative pathologies.methodsthis prospective, single-center study included 31 patients (15 males and 16 females; aged 51 \u00b1 16 years (mean \u00b1 standard deviation)) who underwent lumbar spine exams with both tse-sd and tse-dl acquisitions for degenerative spine diseases. images were analyzed by two radiologists and assessed for qualitative image quality using a 4-point likert scale, quantitative signal-to-noise ratio (snr) of anatomic landmarks, and detection of common pathologies. paired-sample t, wilcoxon, and mcnemar tests, unweighted\/linearly weighted cohen \u03ba statistics, and intraclass correlation coefficients were used.resultsscan time for tse-dl and tse-sd protocols was 2:55 and 5:17 min:s, respectively. the overall image quality was either significantly higher for tse-dl or not significantly different between tse-sd and tse-dl. tse-dl demonstrated higher snr and subject noise scores than tse-sd. for pathology detection, the interreader agreement was substantial to almost perfect for tse-dl, with \u03ba values ranging from 0.61 to 1.00; the interprotocol agreement was almost perfect for both readers, with \u03ba values ranging from 0.84 to 1.00. there was no significant difference in the diagnostic confidence or detection rate of common pathologies between the two sequences (p \u2265 0.081).conclusionstse-dl allowed for a 45% reduction in scan time over tse-sd in lumbar spine mri without compromising the overall image quality and showed comparable detection performance of common pathologies in the evaluation of degenerative lumbar spine changes.relevance statementdeep learning-reconstructed lumbar spine mri protocol enabled a 45% reduction in scan time compared with conventional reconstruction, with comparable image quality and detection performance of common degenerative pathologies.key points\u2022 lumbar spine mri with deep learning reconstruction has broad application prospects.\u2022 deep learning reconstruction of lumbar spine mri saved 45% scan time without compromising overall image quality.\u2022 when compared with standard sequences, deep learning reconstruction showed similar detection performance of common degenerative lumbar spine pathologies.graphical abstract","key_phrases":["backgroundwe","mri","a deep learning technique","tse-dl","standard turbo spin-echo (tse-sd) images","the lumbar spine","image quality and detection performance","common degenerative pathologies.methodsthis prospective, single-center study","31 patients","15 males","16 females","51 \u00b1","\u00b1 standard deviation","who","lumbar spine exams","both tse-sd and tse-dl acquisitions","degenerative spine diseases","images","two radiologists","qualitative image quality","a 4-point likert scale","noise","snr","anatomic landmarks","detection","common pathologies","paired-sample t","wilcoxon","mcnemar tests","cohen \u03ba statistics","correlation coefficients","used.resultsscan time","tse-dl and tse-sd protocols","min",":","s","the overall image quality","tse-dl","tse-sd and tse-dl. tse-dl","higher snr and subject noise scores","tse-sd","pathology detection","the interreader agreement","tse-dl","\u03ba values","the interprotocol agreement","both readers","\u03ba values","no significant difference","the diagnostic confidence or detection rate","common pathologies","the two sequences","p \u2265","0.081).conclusionstse-dl","a 45% reduction","scan time","tse-sd","lumbar spine mri","the overall image quality","comparable detection performance","common pathologies","the evaluation","degenerative lumbar spine changes.relevance","learning-reconstructed lumbar spine mri protocol","a 45% reduction","scan time","conventional reconstruction","comparable image quality and detection performance","lumbar spine mri","deep learning reconstruction","broad application prospects.\u2022","reconstruction","lumbar spine mri","45% scan time","overall image","standard sequences","deep learning reconstruction","similar detection performance","common degenerative lumbar spine","31","15","16","51 \u00b1 16 years","two","4","mcnemar","used.resultsscan","2:55","5:17","0.61","1.00","0.84","1.00","two","45%","45%","45%"]},{"title":"Machine learning vs deep learning in stock market investment: an international evidence","authors":["Jing Hao","Feng He","Feng Ma","Shibo Zhang","Xiaotao Zhang"],"abstract":"Machine learning and deep learning are powerful tools for quantitative investment. To examine the effectiveness of the models in different markets, this paper applies random forest and DNN models to forecast stock prices and construct statistical arbitrage strategies in five stock markets, including mainland China, the United States, the United Kingdom, Canada and Japan. Each model is applied to the price of major stock indices constituting stocks in these markets from 2005 to 2020 to construct a long-short portfolio with 20 selected stocks by the model. The results show that the a particular model obtains significantly different profits in different markets, among which DNN has the best performance, especially in the Chinese stock market. We find that DNN models generally perform better than other machine learning models in all markets.","doi":"10.1007\/s10479-023-05286-6","cleaned_title":"machine learning vs deep learning in stock market investment: an international evidence","cleaned_abstract":"machine learning and deep learning are powerful tools for quantitative investment. to examine the effectiveness of the models in different markets, this paper applies random forest and dnn models to forecast stock prices and construct statistical arbitrage strategies in five stock markets, including mainland china, the united states, the united kingdom, canada and japan. each model is applied to the price of major stock indices constituting stocks in these markets from 2005 to 2020 to construct a long-short portfolio with 20 selected stocks by the model. the results show that the a particular model obtains significantly different profits in different markets, among which dnn has the best performance, especially in the chinese stock market. we find that dnn models generally perform better than other machine learning models in all markets.","key_phrases":["machine learning","deep learning","powerful tools","quantitative investment","the effectiveness","the models","different markets","this paper","random forest","dnn models","stock prices","statistical arbitrage strategies","five stock markets","mainland china","the united states","the united kingdom","canada","japan","each model","the price","major stock indices","stocks","these markets","a long-short portfolio","20 selected stocks","the model","the results","the a particular model","significantly different profits","different markets","which","dnn","the best performance","the chinese stock market","we","dnn models","other machine learning models","all markets","five","china","the united states","the united kingdom","canada","japan","2005","2020","20","chinese"]},{"title":"Deep learning implementations in mining applications: a compact critical review","authors":["Faris Azhari","Charlotte C. Sennersten","Craig A. Lindley","Ewan Sellers"],"abstract":"Deep learning is a sub-field of artificial intelligence that combines feature engineering and classification in one method. It is a data-driven technique that optimises a predictive model via learning from a large dataset. Digitisation in industry has included acquisition and storage of a variety of large datasets for interpretation and decision making. This has led to the adoption of deep learning in different industries, such as transportation, manufacturing, medicine and agriculture. However, in the mining industry, the adoption and development of new technologies, including deep learning methods, has not progressed at the same rate as in other industries. Nevertheless, in the past 5 years, applications of deep learning have been increasing in the mining research space. Deep learning has been implemented to solve a variety of problems related to mine exploration, ore and metal extraction and reclamation processes. The increased automation adoption in mining provides an avenue for wider application of deep learning as an element within a mine automation framework. This\u00a0work provides a compact, comprehensive review of deep learning implementations in mining-related applications. The trends of these implementations in terms of years, venues, deep learning network types, tasks and general implementation, categorised by the value chain operations of exploration, extraction and reclamation are outlined. The review enables shortcomings regarding progress within the research context to be highlighted such as the proprietary nature of data, small datasets (tens to thousands of data points) limited to single operations with unique geology, mine design and equipment, lack of large scale publicly available mining related datasets and limited sensor types leading to the majority of applications being image-based analysis. Gaps identified for future research and application includes the usage of a wider range of sensor data, improved understanding of the outputs by mining practitioners, adversarial testing of the deep learning models, development of public datasets covering the extensive range of conditions experienced in mines.","doi":"10.1007\/s10462-023-10500-9","cleaned_title":"deep learning implementations in mining applications: a compact critical review","cleaned_abstract":"deep learning is a sub-field of artificial intelligence that combines feature engineering and classification in one method. it is a data-driven technique that optimises a predictive model via learning from a large dataset. digitisation in industry has included acquisition and storage of a variety of large datasets for interpretation and decision making. this has led to the adoption of deep learning in different industries, such as transportation, manufacturing, medicine and agriculture. however, in the mining industry, the adoption and development of new technologies, including deep learning methods, has not progressed at the same rate as in other industries. nevertheless, in the past 5 years, applications of deep learning have been increasing in the mining research space. deep learning has been implemented to solve a variety of problems related to mine exploration, ore and metal extraction and reclamation processes. the increased automation adoption in mining provides an avenue for wider application of deep learning as an element within a mine automation framework. this work provides a compact, comprehensive review of deep learning implementations in mining-related applications. the trends of these implementations in terms of years, venues, deep learning network types, tasks and general implementation, categorised by the value chain operations of exploration, extraction and reclamation are outlined. the review enables shortcomings regarding progress within the research context to be highlighted such as the proprietary nature of data, small datasets (tens to thousands of data points) limited to single operations with unique geology, mine design and equipment, lack of large scale publicly available mining related datasets and limited sensor types leading to the majority of applications being image-based analysis. gaps identified for future research and application includes the usage of a wider range of sensor data, improved understanding of the outputs by mining practitioners, adversarial testing of the deep learning models, development of public datasets covering the extensive range of conditions experienced in mines.","key_phrases":["deep learning","a sub","-","field","artificial intelligence","that","feature engineering","classification","one method","it","a data-driven technique","that","a predictive model","a large dataset","digitisation","industry","acquisition","storage","a variety","large datasets","interpretation","decision making","this","the adoption","deep learning","different industries","transportation","manufacturing","medicine","agriculture","the mining industry","new technologies","deep learning methods","the same rate","other industries","the past 5 years","applications","deep learning","the mining research space","deep learning","a variety","problems","mine exploration, ore and metal extraction and reclamation processes","the increased automation adoption","mining","an avenue","wider application","deep learning","an element","a mine automation framework","this work","a compact, comprehensive review","deep learning implementations","mining-related applications","the trends","these implementations","terms","years","venues","deep learning network types","tasks","general implementation","the value chain operations","exploration","extraction","reclamation","the review","shortcomings","progress","the research context","the proprietary nature","data","small datasets","tens to thousands","data points","single operations","unique geology","mine design","equipment","large scale publicly available mining related datasets","limited sensor types","the majority","applications","image-based analysis","gaps","future research","application","the usage","a wider range","sensor data","understanding","the outputs","mining practitioners","adversarial testing","the deep learning models","development","public datasets","the extensive range","conditions","mines","one","the past 5 years","tens to thousands"]},{"title":"Surface wave inversion with unknown number of soil layers based on a hybrid learning procedure of deep learning and genetic algorithm","authors":["Zan Zhou","Thomas Man-Hoi Lok","Wan-Huan Zhou"],"abstract":"Surface wave inversion is a key step in the application of surface waves to soil velocity profiling. Currently, a common practice for the process of inversion is that the number of soil layers is assumed to be known before using heuristic search algorithms to compute the shear wave velocity profile or the number of soil layers is considered as an optimization variable. However, an improper selection of the number of layers may lead to an incorrect shear wave velocity profile. In this study, a deep learning and genetic algorithm hybrid learning procedure is proposed to perform the surface wave inversion without the need to assume the number of soil layers. First, a deep neural network is adapted to learn from a large number of synthetic dispersion curves for inferring the layer number. Then, the shear-wave velocity profile is determined by a genetic algorithm with the known layer number. By applying this procedure to both simulated and real-world cases, the results indicate that the proposed method is reliable and efficient for surface wave inversion.","doi":"10.1007\/s11803-024-2240-1","cleaned_title":"surface wave inversion with unknown number of soil layers based on a hybrid learning procedure of deep learning and genetic algorithm","cleaned_abstract":"surface wave inversion is a key step in the application of surface waves to soil velocity profiling. currently, a common practice for the process of inversion is that the number of soil layers is assumed to be known before using heuristic search algorithms to compute the shear wave velocity profile or the number of soil layers is considered as an optimization variable. however, an improper selection of the number of layers may lead to an incorrect shear wave velocity profile. in this study, a deep learning and genetic algorithm hybrid learning procedure is proposed to perform the surface wave inversion without the need to assume the number of soil layers. first, a deep neural network is adapted to learn from a large number of synthetic dispersion curves for inferring the layer number. then, the shear-wave velocity profile is determined by a genetic algorithm with the known layer number. by applying this procedure to both simulated and real-world cases, the results indicate that the proposed method is reliable and efficient for surface wave inversion.","key_phrases":["surface wave inversion","a key step","the application","surface waves","velocity profiling","a common practice","the process","inversion","the number","soil layers","heuristic search algorithms","the shear wave velocity profile","the number","soil layers","an optimization variable","an improper selection","the number","layers","an incorrect shear wave velocity profile","this study","a deep learning and genetic algorithm hybrid learning procedure","the surface wave inversion","the need","the number","soil layers","a deep neural network","a large number","synthetic dispersion curves","the layer number","the shear-wave velocity profile","a genetic algorithm","the known layer number","this procedure","both simulated and real-world cases","the results","the proposed method","surface wave inversion","first"]},{"title":"An Extensive Review on Deep Learning and Machine Learning Intervention in Prediction and Classification of Types of Aneurysms","authors":["Renugadevi Ammapalayam Sinnaswamy","Natesan Palanisamy","Kavitha Subramaniam","Suresh Muthusamy","Ravita Lamba","Sreejith Sekaran"],"abstract":"Aneurysm (Rupture of blood vessels) may happen in the cerebrum, abdominal aorta and thoracic aorta of humans, which has a high fatal rate. The advancement of the artificial technologies specifically machine learning algorithms and deep learning models have attempted to predict the aneurysm, which may reduce the death rate. The main objective of this paper is to provide the review of various algorithms and models for the early prediction of the various types of aneurysms. The focused literature review was conducted from the preferred journals from 2007 to 2022 on various parameters such as way of collecting images, the techniques used, number of images used in data set, performance metrics and future work. The summarized overview of advances in prediction of aneurysms using the machine learning algorithms from non linear kernel support regression algorithm to 3D Unet architecture of deep learning models starting from CT scan images to final performance analysis in prediction. The range of sensitivity, specificity and area under receiving operating characteristic was from 0. 7 to 1 for the abdominal aortic aneurysm detection, intracranial aneurysm detection. The thoracic aortic aneurysm was not concentrated much in the literature review, so the prediction of thoracic aortic aneurysm using machine learning as well as deep learning model is recommended.","doi":"10.1007\/s11277-023-10532-y","cleaned_title":"an extensive review on deep learning and machine learning intervention in prediction and classification of types of aneurysms","cleaned_abstract":"aneurysm (rupture of blood vessels) may happen in the cerebrum, abdominal aorta and thoracic aorta of humans, which has a high fatal rate. the advancement of the artificial technologies specifically machine learning algorithms and deep learning models have attempted to predict the aneurysm, which may reduce the death rate. the main objective of this paper is to provide the review of various algorithms and models for the early prediction of the various types of aneurysms. the focused literature review was conducted from the preferred journals from 2007 to 2022 on various parameters such as way of collecting images, the techniques used, number of images used in data set, performance metrics and future work. the summarized overview of advances in prediction of aneurysms using the machine learning algorithms from non linear kernel support regression algorithm to 3d unet architecture of deep learning models starting from ct scan images to final performance analysis in prediction. the range of sensitivity, specificity and area under receiving operating characteristic was from 0. 7 to 1 for the abdominal aortic aneurysm detection, intracranial aneurysm detection. the thoracic aortic aneurysm was not concentrated much in the literature review, so the prediction of thoracic aortic aneurysm using machine learning as well as deep learning model is recommended.","key_phrases":["aneurysm","rupture","blood vessels","the cerebrum","abdominal aorta","aorta","humans","which","a high fatal rate","the advancement","the artificial technologies","specifically machine learning algorithms","deep learning models","the aneurysm","which","the death rate","the main objective","this paper","the review","various algorithms","models","the early prediction","the various types","aneurysms","the focused literature review","the preferred journals","various parameters","way","images","the techniques","images","data set","performance metrics","future work","the summarized overview","advances","prediction","aneurysms","the machine learning algorithms","non linear kernel support regression algorithm","3d unet architecture","deep learning models","ct scan images","final performance analysis","prediction","the range","sensitivity","specificity","area","operating characteristic","the abdominal aortic aneurysm detection","intracranial aneurysm detection","the thoracic aortic aneurysm","the literature review","the prediction","thoracic aortic aneurysm","machine learning","deep learning model","2007","2022","3d","scan","0","7 to 1","intracranial aneurysm detection","the thoracic aortic aneurysm"]},{"title":"Deep learning based active image steganalysis: a review","authors":["Punam Bedi","Anuradha Singhal","Veenu Bhasin"],"abstract":"Steganalysis plays a vital role in cybersecurity in today\u2019s digital era where exchange of malicious information can be done easily across web pages. Steganography techniques are used to hide data in an object where the existence of hidden information is also obscured. Steganalysis is the process for detection of steganography within an object and can be categorized as active and passive steganalysis. Passive steganalysis tries to classify a given object as a clean or modified object. Active steganalysis aims to extract more details about hidden contents such as length of embedded message, region of inserted message, key used for embedding, required by cybersecurity experts for comprehensive analysis. Images being a viable source of exchange of information in the era of internet, social media are the most susceptible source for such transmission. Many researchers have worked and developed techniques required to detect and alert about such counterfeit exchanges over the internet. Literature present in passive and active image steganalysis techniques, addresses these issues by detecting and unveiling details of such obscured communication respectively. This paper provides a systematic and comprehensive review of work done on active image steganalysis techniques using deep learning techniques. This review will be helpful to the new researchers to become aware and build a strong foundation of literature present in active image steganalysis using deep learning techniques. The paper also includes various steganographic algorithms, dataset and performance evaluation metrics used in literature. Open research challenges and possible future research directions are also discussed in the paper.","doi":"10.1007\/s13198-023-02203-9","cleaned_title":"deep learning based active image steganalysis: a review","cleaned_abstract":"steganalysis plays a vital role in cybersecurity in today\u2019s digital era where exchange of malicious information can be done easily across web pages. steganography techniques are used to hide data in an object where the existence of hidden information is also obscured. steganalysis is the process for detection of steganography within an object and can be categorized as active and passive steganalysis. passive steganalysis tries to classify a given object as a clean or modified object. active steganalysis aims to extract more details about hidden contents such as length of embedded message, region of inserted message, key used for embedding, required by cybersecurity experts for comprehensive analysis. images being a viable source of exchange of information in the era of internet, social media are the most susceptible source for such transmission. many researchers have worked and developed techniques required to detect and alert about such counterfeit exchanges over the internet. literature present in passive and active image steganalysis techniques, addresses these issues by detecting and unveiling details of such obscured communication respectively. this paper provides a systematic and comprehensive review of work done on active image steganalysis techniques using deep learning techniques. this review will be helpful to the new researchers to become aware and build a strong foundation of literature present in active image steganalysis using deep learning techniques. the paper also includes various steganographic algorithms, dataset and performance evaluation metrics used in literature. open research challenges and possible future research directions are also discussed in the paper.","key_phrases":["steganalysis","a vital role","cybersecurity","today\u2019s digital era","exchange","malicious information","web pages","steganography techniques","data","an object","the existence","hidden information","steganalysis","the process","detection","steganography","an object","active and passive steganalysis","passive steganalysis","a given object","a clean or modified object","active steganalysis","more details","hidden contents","length","embedded message","region","inserted message","cybersecurity experts","comprehensive analysis","images","a viable source","exchange","information","the era","internet","social media","the most susceptible source","such transmission","many researchers","techniques","such counterfeit exchanges","the internet","literature","passive and active image steganalysis techniques","these issues","details","such obscured communication","this paper","a systematic and comprehensive review","work","active image steganalysis techniques","deep learning techniques","this review","the new researchers","a strong foundation","literature","active image steganalysis","deep learning techniques","the paper","various steganographic algorithms","dataset","performance evaluation metrics","literature","open research challenges","possible future research directions","the paper","today"]},{"title":"Early detection and prediction of Heart Disease using Wearable devices and Deep Learning algorithms","authors":["S. Sivasubramaniam","S. P. Balamurugan"],"abstract":"In this paper, we propose a multimodal deep learning algorithm that combines convolutional neural networks (CNNs) and long short-term memory (LSTM) networks for early detection and prediction of heart disease using data collected from wearable devices. This combined multi-model deep learning algorithm is used to detect the accurate precision and accuracy value. At first, we consider, ECG and PPG signals, which are collected from the dataset. Then, the features from ECG and PPG are extracted using CNN and the accelerometer features are extracted using the LSTM model. The combined features are then classified using hybrid CNN-LSTM network architecture. The algorithm is evaluated using a publicly available benchmark dataset. The model achieved an accuracy of 99.33% in detecting heart disease, outperforming several state-of-the-art deep learning models. In addition, the model can predict the likelihood of developing heart disease with a precision of 99.33%, providing an early warning system for at-risk patients. The results demonstrate the potential of a multimodal approach for early detection and prediction of heart disease using wearable devices and deep learning algorithms.","doi":"10.1007\/s11042-024-19127-6","cleaned_title":"early detection and prediction of heart disease using wearable devices and deep learning algorithms","cleaned_abstract":"in this paper, we propose a multimodal deep learning algorithm that combines convolutional neural networks (cnns) and long short-term memory (lstm) networks for early detection and prediction of heart disease using data collected from wearable devices. this combined multi-model deep learning algorithm is used to detect the accurate precision and accuracy value. at first, we consider, ecg and ppg signals, which are collected from the dataset. then, the features from ecg and ppg are extracted using cnn and the accelerometer features are extracted using the lstm model. the combined features are then classified using hybrid cnn-lstm network architecture. the algorithm is evaluated using a publicly available benchmark dataset. the model achieved an accuracy of 99.33% in detecting heart disease, outperforming several state-of-the-art deep learning models. in addition, the model can predict the likelihood of developing heart disease with a precision of 99.33%, providing an early warning system for at-risk patients. the results demonstrate the potential of a multimodal approach for early detection and prediction of heart disease using wearable devices and deep learning algorithms.","key_phrases":["this paper","we","a multimodal deep learning algorithm","that","cnns","lstm","early detection","prediction","heart disease","data","wearable devices","this combined multi-model deep learning algorithm","the accurate precision","accuracy value","we","ecg and ppg signals","which","the dataset","the features","ecg","ppg","cnn","the accelerometer features","the lstm model","the combined features","hybrid cnn-lstm network architecture","the algorithm","a publicly available benchmark dataset","the model","an accuracy","99.33%","heart disease","the-art","addition","the model","the likelihood","heart disease","a precision","99.33%","an early warning system","risk","the results","the potential","a multimodal approach","early detection","prediction","heart disease","wearable devices","deep learning algorithms","first","cnn","cnn","99.33%","99.33%"]},{"title":"Deep cross-domain transfer for emotion recognition via joint learning","authors":["Dung Nguyen","Duc Thanh Nguyen","Sridha Sridharan","Mohamed Abdelrazek","Simon Denman","Son N. Tran","Rui Zeng","Clinton Fookes"],"abstract":"Deep learning has been applied to achieve significant progress in emotion recognition from multimedia data. Despite such substantial progress, existing approaches are hindered by insufficient training data, leading to weak generalisation under mismatched conditions. To address these challenges, we propose a learning strategy which jointly transfers emotional knowledge learnt from rich datasets to source-poor datasets. Our method is also able to learn cross-domain features, leading to improved recognition performance. To demonstrate the robustness of the proposed learning strategy, we conducted extensive experiments on several benchmark datasets including eNTERFACE, SAVEE, EMODB, and RAVDESS. Experimental results show that the proposed method surpassed existing transfer learning schemes by a significant margin.","doi":"10.1007\/s11042-023-15441-7","cleaned_title":"deep cross-domain transfer for emotion recognition via joint learning","cleaned_abstract":"deep learning has been applied to achieve significant progress in emotion recognition from multimedia data. despite such substantial progress, existing approaches are hindered by insufficient training data, leading to weak generalisation under mismatched conditions. to address these challenges, we propose a learning strategy which jointly transfers emotional knowledge learnt from rich datasets to source-poor datasets. our method is also able to learn cross-domain features, leading to improved recognition performance. to demonstrate the robustness of the proposed learning strategy, we conducted extensive experiments on several benchmark datasets including enterface, savee, emodb, and ravdess. experimental results show that the proposed method surpassed existing transfer learning schemes by a significant margin.","key_phrases":["deep learning","significant progress","emotion recognition","multimedia data","such substantial progress","existing approaches","insufficient training data","weak generalisation","mismatched conditions","these challenges","we","a learning strategy","which","emotional knowledge","rich datasets","source-poor datasets","our method","cross-domain features","improved recognition performance","the robustness","the proposed learning strategy","we","extensive experiments","several benchmark datasets","enterface","emodb","ravdess","experimental results","the proposed method","existing transfer learning schemes","a significant margin"]},{"title":"Scoring method of English composition integrating deep learning in higher vocational colleges","authors":["Shuo Feng","Lixia Yu","Fen Liu"],"abstract":"Along with the progress of natural language processing technology and deep learning, the subjectivity, slow feedback, and long grading time of traditional English essay grading have been addressed. Intelligent English automatic scoring has been widely concerned by scholars. Given the limitations of topic relevance feature extraction methods and traditional automatic grading methods for English compositions, a topic decision model is proposed to calculate the topic relevance score of the topic richness in English composition. Then, based on the Score of Relevance Based on Topic Richness (TRSR) calculation method, an intelligent English composition scoring method combining artificial feature extraction and deep learning is designed. From the findings, the Topic Decision (TD) model achieved the best effect only when it was iterated 80 times. The corresponding accuracy, recall and F1 value were 0.97, 0.93 and 0.95 respectively. The model training loss finally stabilized at 0.03. The Intelligent English Composition Grading Method Integrating Deep Learning (DLIECG) method has the best overall performance and the best performance on dataset P. To sum up, the intelligent English composition scoring method has better effectiveness and reliability.","doi":"10.1038\/s41598-024-57419-x","cleaned_title":"scoring method of english composition integrating deep learning in higher vocational colleges","cleaned_abstract":"along with the progress of natural language processing technology and deep learning, the subjectivity, slow feedback, and long grading time of traditional english essay grading have been addressed. intelligent english automatic scoring has been widely concerned by scholars. given the limitations of topic relevance feature extraction methods and traditional automatic grading methods for english compositions, a topic decision model is proposed to calculate the topic relevance score of the topic richness in english composition. then, based on the score of relevance based on topic richness (trsr) calculation method, an intelligent english composition scoring method combining artificial feature extraction and deep learning is designed. from the findings, the topic decision (td) model achieved the best effect only when it was iterated 80 times. the corresponding accuracy, recall and f1 value were 0.97, 0.93 and 0.95 respectively. the model training loss finally stabilized at 0.03. the intelligent english composition grading method integrating deep learning (dliecg) method has the best overall performance and the best performance on dataset p. to sum up, the intelligent english composition scoring method has better effectiveness and reliability.","key_phrases":["the progress","natural language processing technology","deep learning","the subjectivity","slow feedback","long grading time","intelligent english automatic scoring","scholars","the limitations","topic relevance feature extraction methods","traditional automatic grading methods","english compositions","a topic decision model","the topic relevance score","the topic richness","english composition","the score","relevance","topic richness (trsr) calculation method","an intelligent english composition scoring method","artificial feature extraction","deep learning","the findings","the topic decision","(td) model","the best effect","it","the corresponding accuracy","recall","f1 value","the model training loss","the intelligent english composition","method","deep learning (dliecg) method","the best overall performance","the best performance","dataset p.","the intelligent english composition scoring method","better effectiveness","reliability","english","english","english","english","english","80","0.97","0.93","0.95","0.03","english","english"]},{"title":"Machine learning and deep learning techniques for breast cancer diagnosis and classification: a comprehensive review of medical imaging studies","authors":["Mehran Radak","Haider Yabr Lafta","Hossein Fallahi"],"abstract":"BackgroundBreast cancer is a major public health concern, and early diagnosis and classification are critical for effective treatment. Machine learning and deep learning techniques have shown great promise in the classification and diagnosis of breast cancer.PurposeIn this review, we examine studies that have used these techniques for breast cancer classification and diagnosis, focusing on five groups of medical images: mammography, ultrasound, MRI, histology, and thermography. We discuss the use of five popular machine learning techniques, including Nearest Neighbor, SVM, Naive Bayesian Network, DT, and ANN, as well as deep learning architectures and convolutional neural networks.ConclusionOur review finds that machine learning and deep learning techniques have achieved high accuracy rates in breast cancer classification and diagnosis across various medical imaging modalities. Furthermore, these techniques have the potential to improve clinical decision-making and ultimately lead to better patient outcomes.","doi":"10.1007\/s00432-023-04956-z","cleaned_title":"machine learning and deep learning techniques for breast cancer diagnosis and classification: a comprehensive review of medical imaging studies","cleaned_abstract":"backgroundbreast cancer is a major public health concern, and early diagnosis and classification are critical for effective treatment. machine learning and deep learning techniques have shown great promise in the classification and diagnosis of breast cancer.purposein this review, we examine studies that have used these techniques for breast cancer classification and diagnosis, focusing on five groups of medical images: mammography, ultrasound, mri, histology, and thermography. we discuss the use of five popular machine learning techniques, including nearest neighbor, svm, naive bayesian network, dt, and ann, as well as deep learning architectures and convolutional neural networks.conclusionour review finds that machine learning and deep learning techniques have achieved high accuracy rates in breast cancer classification and diagnosis across various medical imaging modalities. furthermore, these techniques have the potential to improve clinical decision-making and ultimately lead to better patient outcomes.","key_phrases":["backgroundbreast cancer","a major public health concern","early diagnosis","classification","effective treatment","machine learning","deep learning techniques","great promise","the classification","diagnosis","breast cancer.purposein","this review","we","studies","that","these techniques","breast cancer classification","diagnosis","five groups","medical images","mammography","ultrasound","mri","histology","thermography","we","the use","five popular machine learning techniques","neighbor","svm","naive bayesian network","dt","ann","deep learning architectures","convolutional neural networks.conclusionour review","machine learning","deep learning techniques","high accuracy rates","breast cancer classification","diagnosis","various medical imaging modalities","these techniques","the potential","clinical decision-making","better patient outcomes","five","five"]},{"title":"Deep learning based vessel arrivals monitoring via autoregressive statistical control charts","authors":["Sara El Mekkaoui","Ghait Boukachab","Loubna Benabbou","Abdelaziz Berrado"],"abstract":"This paper introduces a methodology for monitoring the vessel arrival process, a critical factor in enhancing maritime operational efficiency. This approach uses deep learning sequence models and Statistical Process Control Charts to track the variability in a vessel arrival process. The proposed solution uses the predictive deep learning model to get a vessel\u2019s estimated time of arrival, produces quality characteristics, and applies statistical control charts to monitor their variability. The paper presents the results of applying the proposed methodology for vessel arrivals at a coal terminal, which demonstrates the effectiveness of the method. By enabling precise monitoring of arrival times, this methodology not only supports efficient ship and port operations planning but also aids in the timely adoption of operational adjustments. This can significantly contribute to operational measures aimed at reducing shipping emissions and optimizing resource utilization.","doi":"10.1007\/s13437-024-00342-9","cleaned_title":"deep learning based vessel arrivals monitoring via autoregressive statistical control charts","cleaned_abstract":"this paper introduces a methodology for monitoring the vessel arrival process, a critical factor in enhancing maritime operational efficiency. this approach uses deep learning sequence models and statistical process control charts to track the variability in a vessel arrival process. the proposed solution uses the predictive deep learning model to get a vessel\u2019s estimated time of arrival, produces quality characteristics, and applies statistical control charts to monitor their variability. the paper presents the results of applying the proposed methodology for vessel arrivals at a coal terminal, which demonstrates the effectiveness of the method. by enabling precise monitoring of arrival times, this methodology not only supports efficient ship and port operations planning but also aids in the timely adoption of operational adjustments. this can significantly contribute to operational measures aimed at reducing shipping emissions and optimizing resource utilization.","key_phrases":["this paper","a methodology","the vessel arrival process","a critical factor","maritime operational efficiency","this approach","deep learning sequence models","statistical process control charts","the variability","a vessel arrival process","the proposed solution","the predictive deep learning model","a vessel\u2019s estimated time","arrival","quality characteristics","statistical control charts","their variability","the paper","the results","the proposed methodology","vessel arrivals","a coal terminal","which","the effectiveness","the method","precise monitoring","arrival times","this methodology","efficient ship and port operations planning","the timely adoption","operational adjustments","this","operational measures","shipping emissions","resource utilization"]},{"title":"SAKMR: Industrial control anomaly detection based on semi-supervised hybrid deep learning","authors":["Shijie Tang","Yong Ding","Meng Zhao","Huiyong Wang"],"abstract":"With the advent of Industry 4.0, industrial control systems (ICS) are more and more closely connected with the Internet, leading to a rapid increase in the types and quantities of security threats that arise from ICS. Anomaly detection is an effective defense measure against attacks. At present, it is the main trend to use hybrid deep learning methods to realize ICS anomaly detection. However, we found that many ICS anomaly detection methods based on hybrid deep learning adopt phased learning, in which each phase is optimized separately with optimization goals deviating from the overall goal. In view of this issue, we propose an end-to-end anomaly detection method SAKMR based on hybrid deep learning. Our method uses radial basis function network (RBFN) to realize K-means clustering, and combines it with stacked auto-encoder (SAE), which is conducive to defining reconstruction error and clustering error into an objective function to ensure joint optimization of feature extraction and classification. Experiments were conducted on the commonly used KDDCUP99 and SWAT datasets. The results show that SAKMR is effective in detecting abnormal industrial control data and outperforms the baseline methods on multiple performance indicators such as F1-Measure.","doi":"10.1007\/s12083-023-01586-7","cleaned_title":"sakmr: industrial control anomaly detection based on semi-supervised hybrid deep learning","cleaned_abstract":"with the advent of industry 4.0, industrial control systems (ics) are more and more closely connected with the internet, leading to a rapid increase in the types and quantities of security threats that arise from ics. anomaly detection is an effective defense measure against attacks. at present, it is the main trend to use hybrid deep learning methods to realize ics anomaly detection. however, we found that many ics anomaly detection methods based on hybrid deep learning adopt phased learning, in which each phase is optimized separately with optimization goals deviating from the overall goal. in view of this issue, we propose an end-to-end anomaly detection method sakmr based on hybrid deep learning. our method uses radial basis function network (rbfn) to realize k-means clustering, and combines it with stacked auto-encoder (sae), which is conducive to defining reconstruction error and clustering error into an objective function to ensure joint optimization of feature extraction and classification. experiments were conducted on the commonly used kddcup99 and swat datasets. the results show that sakmr is effective in detecting abnormal industrial control data and outperforms the baseline methods on multiple performance indicators such as f1-measure.","key_phrases":["the advent","industry","industrial control systems","ics","the internet","a rapid increase","the types","quantities","security threats","that","ics","anomaly detection","an effective defense measure","attacks","present","it","the main trend","hybrid deep learning methods","ics anomaly detection","we","that many ics anomaly detection methods","hybrid deep learning adopt phased learning","which","each phase","optimization goals","the overall goal","view","this issue","we","end","hybrid deep learning","our method","radial basis function network","rbfn","k","it","stacked auto-encoder","sae","which","reconstruction error","error","an objective function","joint optimization","feature extraction","classification","experiments","the commonly used kddcup99","swat datasets","the results","sakmr","abnormal industrial control data","the baseline methods","multiple performance indicators","f1-measure","4.0","anomaly detection","anomaly detection method","kddcup99"]},{"title":"Single sample face recognition using deep learning: a survey","authors":["Vivek Tomar","Nitin Kumar","Ayush Raj Srivastava"],"abstract":"Face recognition has become popular in the last few decades among researchers across the globe due to its applicability in several domains. This problem becomes more challenging when only a single training image is available and is popularly known as single sample face recognition (SSFR) problem. SSFR becomes even more complex when images are captured under varying illumination conditions, different poses, occlusion, and expression. Further, deep learning methods have shown performance at par with humans recently. Due to the emergence of deep learning methods in the last decade, it has been made possible to recognize faces with excellent accuracy even in a single sample scenario. In this paper, we present a comprehensive survey of SSFR using deep learning. We also propose a novel taxonomy and broadly divide these methods into three categories viz. virtual sample generation, feature-based, and hybrid methods. Performance comparison of these methods as reported in the literature has also been performed. Finally, we review publicly available databases used by the researchers and give some important future research directions which will help aspiring researchers in this fascinating area.","doi":"10.1007\/s10462-023-10551-y","cleaned_title":"single sample face recognition using deep learning: a survey","cleaned_abstract":"face recognition has become popular in the last few decades among researchers across the globe due to its applicability in several domains. this problem becomes more challenging when only a single training image is available and is popularly known as single sample face recognition (ssfr) problem. ssfr becomes even more complex when images are captured under varying illumination conditions, different poses, occlusion, and expression. further, deep learning methods have shown performance at par with humans recently. due to the emergence of deep learning methods in the last decade, it has been made possible to recognize faces with excellent accuracy even in a single sample scenario. in this paper, we present a comprehensive survey of ssfr using deep learning. we also propose a novel taxonomy and broadly divide these methods into three categories viz. virtual sample generation, feature-based, and hybrid methods. performance comparison of these methods as reported in the literature has also been performed. finally, we review publicly available databases used by the researchers and give some important future research directions which will help aspiring researchers in this fascinating area.","key_phrases":["face recognition","the last few decades","researchers","the globe","its applicability","several domains","this problem","only a single training image","single sample face recognition (ssfr) problem","images","illumination conditions","different poses","occlusion","expression","deep learning methods","performance","par","humans","the emergence","deep learning methods","the last decade","it","faces","excellent accuracy","a single sample scenario","this paper","we","a comprehensive survey","deep learning","we","a novel taxonomy","these methods","three categories","virtual sample generation","feature-based, and hybrid methods","performance comparison","these methods","the literature","we","publicly available databases","the researchers","some important future research directions","which","researchers","this fascinating area","the last few decades","the last decade","three"]},{"title":"RNA contact prediction by data efficient deep learning","authors":["Oskar Taubert","Fabrice von der Lehr","Alina Bazarova","Christian Faber","Philipp Knechtges","Marie Weiel","Charlotte Debus","Daniel Coquelin","Achim Basermann","Achim Streit","Stefan Kesselheim","Markus G\u00f6tz","Alexander Schug"],"abstract":"On the path to full understanding of the structure-function relationship or even design of RNA, structure prediction would offer an intriguing complement to experimental efforts. Any deep learning on RNA structure, however, is hampered by the sparsity of labeled training data. Utilizing the limited data available, we here focus on predicting spatial adjacencies (\"contact maps\u201d) as a proxy for 3D structure. Our model, BARNACLE, combines the utilization of unlabeled data through self-supervised pre-training and efficient use of the sparse labeled data through an XGBoost classifier. BARNACLE shows a considerable improvement over both the established classical baseline and a deep neural network. In order to demonstrate that our approach can be applied to tasks with similar data constraints, we show that our findings generalize to the related setting of accessible surface area prediction.","doi":"10.1038\/s42003-023-05244-9","cleaned_title":"rna contact prediction by data efficient deep learning","cleaned_abstract":"on the path to full understanding of the structure-function relationship or even design of rna, structure prediction would offer an intriguing complement to experimental efforts. any deep learning on rna structure, however, is hampered by the sparsity of labeled training data. utilizing the limited data available, we here focus on predicting spatial adjacencies (\"contact maps\u201d) as a proxy for 3d structure. our model, barnacle, combines the utilization of unlabeled data through self-supervised pre-training and efficient use of the sparse labeled data through an xgboost classifier. barnacle shows a considerable improvement over both the established classical baseline and a deep neural network. in order to demonstrate that our approach can be applied to tasks with similar data constraints, we show that our findings generalize to the related setting of accessible surface area prediction.","key_phrases":["the path","full understanding","the structure-function relationship","even design","rna","structure prediction","an intriguing complement","experimental efforts","any deep learning","rna structure","the sparsity","labeled training data","the limited data","we","spatial adjacencies","(\"contact maps","a proxy","3d structure","our model","barnacle","the utilization","unlabeled data","self-supervised pre","-","training and efficient use","the sparse","data","an xgboost classifier","barnacle","a considerable improvement","both the established classical baseline","a deep neural network","order","our approach","tasks","similar data constraints","we","our findings","the related setting","accessible surface area prediction","3d"]},{"title":"Deep learning for the harmonization of structural MRI scans: a survey","authors":["Soolmaz Abbasi","Haoyu Lan","Jeiran Choupan","Nasim Sheikh-Bahaei","Gaurav Pandey","Bino Varghese"],"abstract":"Medical imaging datasets for research are frequently collected from multiple imaging centers using different scanners, protocols, and settings. These variations affect data consistency and compatibility across different sources. Image harmonization is a critical step to mitigate the effects of factors like inherent differences between various vendors, hardware upgrades, protocol changes, and scanner calibration drift, as well as to ensure consistent data for medical image processing techniques. Given the critical importance and widespread relevance of this issue, a vast array of image harmonization methodologies have emerged, with deep learning-based approaches driving substantial advancements in recent times. The goal of this review paper is to examine the latest deep learning techniques employed for image harmonization by analyzing cutting-edge architectural approaches in the field of medical image harmonization, evaluating both their strengths and limitations. This paper begins by providing a comprehensive fundamental overview of image harmonization strategies, covering three critical aspects: established imaging datasets, commonly used evaluation metrics, and characteristics of different scanners. Subsequently, this paper analyzes recent structural MRI (Magnetic Resonance Imaging) harmonization techniques based on network architecture, network learning algorithm, network supervision strategy, and network output. The underlying architectures include U-Net, Generative Adversarial Networks (GANs), Variational Autoencoders (VAEs), flow-based generative models, transformer-based approaches, as well as custom-designed network architectures. This paper investigates the effectiveness of Disentangled Representation Learning (DRL) as a pivotal learning algorithm in harmonization. Lastly, the review highlights the primary limitations in harmonization techniques, specifically the lack of comprehensive quantitative comparisons across different methods. The overall aim of this review is to serve as a guide for researchers and practitioners to select appropriate architectures based on their specific conditions and requirements. It also aims to foster discussions around ongoing challenges in the field and shed light on promising future research directions with the potential for significant advancements.","doi":"10.1186\/s12938-024-01280-6","cleaned_title":"deep learning for the harmonization of structural mri scans: a survey","cleaned_abstract":"medical imaging datasets for research are frequently collected from multiple imaging centers using different scanners, protocols, and settings. these variations affect data consistency and compatibility across different sources. image harmonization is a critical step to mitigate the effects of factors like inherent differences between various vendors, hardware upgrades, protocol changes, and scanner calibration drift, as well as to ensure consistent data for medical image processing techniques. given the critical importance and widespread relevance of this issue, a vast array of image harmonization methodologies have emerged, with deep learning-based approaches driving substantial advancements in recent times. the goal of this review paper is to examine the latest deep learning techniques employed for image harmonization by analyzing cutting-edge architectural approaches in the field of medical image harmonization, evaluating both their strengths and limitations. this paper begins by providing a comprehensive fundamental overview of image harmonization strategies, covering three critical aspects: established imaging datasets, commonly used evaluation metrics, and characteristics of different scanners. subsequently, this paper analyzes recent structural mri (magnetic resonance imaging) harmonization techniques based on network architecture, network learning algorithm, network supervision strategy, and network output. the underlying architectures include u-net, generative adversarial networks (gans), variational autoencoders (vaes), flow-based generative models, transformer-based approaches, as well as custom-designed network architectures. this paper investigates the effectiveness of disentangled representation learning (drl) as a pivotal learning algorithm in harmonization. lastly, the review highlights the primary limitations in harmonization techniques, specifically the lack of comprehensive quantitative comparisons across different methods. the overall aim of this review is to serve as a guide for researchers and practitioners to select appropriate architectures based on their specific conditions and requirements. it also aims to foster discussions around ongoing challenges in the field and shed light on promising future research directions with the potential for significant advancements.","key_phrases":["medical imaging datasets","research","multiple imaging centers","different scanners","protocols","settings","these variations","data consistency","compatibility","different sources","image harmonization","a critical step","the effects","factors","inherent differences","various vendors","hardware upgrades","protocol changes","scanner calibration drift","consistent data","medical image processing techniques","the critical importance","widespread relevance","this issue","a vast array","image harmonization methodologies","deep learning-based approaches","substantial advancements","recent times","the goal","this review paper","the latest deep learning techniques","image harmonization","cutting-edge architectural approaches","the field","medical image harmonization","both their strengths","limitations","this paper","a comprehensive fundamental overview","image harmonization strategies","three critical aspects","imaging datasets","commonly used evaluation metrics","characteristics","different scanners","this paper","recent structural mri","magnetic resonance imaging","harmonization techniques","network architecture","network learning algorithm","network supervision strategy","network output","the underlying architectures","u","-","net","generative adversarial networks","gans","variational autoencoders","flow-based generative models","transformer-based approaches","custom-designed network architectures","this paper","the effectiveness","drl","a pivotal learning algorithm","harmonization","the review","the primary limitations","harmonization techniques","specifically the lack","comprehensive quantitative comparisons","different methods","the overall aim","this review","a guide","researchers","practitioners","appropriate architectures","their specific conditions","requirements","it","discussions","ongoing challenges","the field","light","future research directions","the potential","significant advancements","three"]},{"title":"A review on vision-based deep learning techniques for damage detection in bolted joints","authors":["Zahir Malik","Ansh Mirani","Tanneru Gopi","Mallika Alapati"],"abstract":"Bolted connections are widely used in steel structures. Detection of bolt loosening is the prime concern in the bolted joints to avoid sudden failure leading to catastrophe. Loosening of the bolts causes interfacial movement by reducing the pre-torque when subjected to vibrations due to dynamic loads. With the advent of computing capabilities, sensor technologies, and machine learning model accuracy in bolt loosening detection, damage recognition efficiency in bolted joints has increased. Integrating deep learning with machine vision, effective models can be proposed without human interventions. The present paper summarizes the research review on bolt loosening detection using machine vision and deep learning techniques from the past decade.","doi":"10.1007\/s42107-024-01139-0","cleaned_title":"a review on vision-based deep learning techniques for damage detection in bolted joints","cleaned_abstract":"bolted connections are widely used in steel structures. detection of bolt loosening is the prime concern in the bolted joints to avoid sudden failure leading to catastrophe. loosening of the bolts causes interfacial movement by reducing the pre-torque when subjected to vibrations due to dynamic loads. with the advent of computing capabilities, sensor technologies, and machine learning model accuracy in bolt loosening detection, damage recognition efficiency in bolted joints has increased. integrating deep learning with machine vision, effective models can be proposed without human interventions. the present paper summarizes the research review on bolt loosening detection using machine vision and deep learning techniques from the past decade.","key_phrases":["bolted connections","steel structures","detection","bolt loosening","the prime concern","the bolted joints","sudden failure","catastrophe","loosening","the bolts","interfacial movement","the pre","torque","vibrations","dynamic loads","the advent","computing capabilities","sensor technologies","machine learning model accuracy","bolt loosening detection","damage recognition efficiency","bolted joints","deep learning","machine vision","effective models","human interventions","the present paper","the research review","bolt loosening detection","machine vision","deep learning techniques","the past decade","the past decade"]},{"title":"Real-time thermography for breast cancer detection with deep learning","authors":["Mohammed Abdulla Salim Al Husaini","Mohamed Hadi Habaebi","Md Rafiqul Islam"],"abstract":"In this study, we propose a framework that enhances breast cancer classification accuracy by preserving spatial features and leveraging in situ cooling support. The framework utilizes real-time thermography video streaming for early breast cancer detection using Deep Learning models. Inception v3, Inception v4, and a modified Inception Mv4 were developed using MATLAB 2019. However, the thermal camera was connected to a mobile phone to capture images of the breast area for classification of normal and abnormal breast. This study\u2019s training dataset included 1000 thermal images, where a FLIR One Pro thermal camera connected to a mobile device was used for the imaging process. Of the 1000 images obtained, 700 images were considered for the normal breast thermography class while the 300 images were suitable for the abnormal class. We evaluate Deep Convolutional Neural Network models, such as Inception v3, Inception v4, and a modified Inception Mv4. Our results demonstrate that Inception Mv4, with real-time video streaming, efficiently detects even the slightest temperature contrast in breast tissue sequences achieving a 99.748% accuracy in comparison to a 99.712% and 96.8% for Inception v4 and v3, respectively. The use of in situ cooling gel further enhances image acquisition efficiency and detection accuracy. Interestingly, increasing the tumor surface temperature by 0.1% leads to an average 7% improvement in detection and classification accuracy. Our findings support the effectiveness of Inception Mv4 for real-time breast cancer detection, especially when combined with in situ cooling gel and varying tumor temperatures. In conclusion, future research directions should focus on incorporating thermal video clips into the thermal images database, utilizing high-quality thermal cameras, and exploring alternative Deep Learning models for improved breast cancer detection.","doi":"10.1007\/s44163-024-00157-w","cleaned_title":"real-time thermography for breast cancer detection with deep learning","cleaned_abstract":"in this study, we propose a framework that enhances breast cancer classification accuracy by preserving spatial features and leveraging in situ cooling support. the framework utilizes real-time thermography video streaming for early breast cancer detection using deep learning models. inception v3, inception v4, and a modified inception mv4 were developed using matlab 2019. however, the thermal camera was connected to a mobile phone to capture images of the breast area for classification of normal and abnormal breast. this study\u2019s training dataset included 1000 thermal images, where a flir one pro thermal camera connected to a mobile device was used for the imaging process. of the 1000 images obtained, 700 images were considered for the normal breast thermography class while the 300 images were suitable for the abnormal class. we evaluate deep convolutional neural network models, such as inception v3, inception v4, and a modified inception mv4. our results demonstrate that inception mv4, with real-time video streaming, efficiently detects even the slightest temperature contrast in breast tissue sequences achieving a 99.748% accuracy in comparison to a 99.712% and 96.8% for inception v4 and v3, respectively. the use of in situ cooling gel further enhances image acquisition efficiency and detection accuracy. interestingly, increasing the tumor surface temperature by 0.1% leads to an average 7% improvement in detection and classification accuracy. our findings support the effectiveness of inception mv4 for real-time breast cancer detection, especially when combined with in situ cooling gel and varying tumor temperatures. in conclusion, future research directions should focus on incorporating thermal video clips into the thermal images database, utilizing high-quality thermal cameras, and exploring alternative deep learning models for improved breast cancer detection.","key_phrases":["this study","we","a framework","that","breast cancer classification accuracy","spatial features","situ cooling support","the framework","real-time thermography video streaming","early breast cancer detection","deep learning models","inception v3","inception v4","a modified inception mv4","matlab","the thermal camera","a mobile phone","images","the breast area","classification","normal and abnormal breast","this study\u2019s training dataset","1000 thermal images","a flir one pro thermal camera","a mobile device","the imaging process","the 1000 images","700 images","the normal breast thermography class","the 300 images","the abnormal class","we","deep convolutional neural network models","inception v3","inception v4","a modified inception mv4","our results","inception mv4","real-time video streaming","even the slightest temperature contrast","breast tissue sequences","a 99.748% accuracy","comparison","a 99.712%","96.8%","inception v4","v3","the use","situ","gel","image acquisition efficiency","detection accuracy","the tumor surface temperature","0.1%","an average 7% improvement","detection and classification accuracy","our findings","the effectiveness","inception mv4","real-time breast cancer detection","gel and varying tumor temperatures","conclusion","future research directions","thermal video clips","the thermal images database","high-quality thermal cameras","alternative deep learning models","improved breast cancer detection","v3","2019","1000","one","1000","700","300","99.748%","99.712%","96.8%","v3","0.1%","an average","7%"]},{"title":"How deep learning is complementing deep thinking in ATLAS","authors":["Deepak Kar"],"abstract":"ATLAS collaboration uses machine learning (ML) algorithms in many different ways in its physics programme, starting from object reconstruction, simulation of calorimeter showers, signal to background discrimination in searches and measurements, tagging jets based on their origin and so on. Anomaly detection (AD) techniques are also gaining popularity where they are used to find hidden patterns in the data, with lesser dependence on simulated samples as in the case of supervised learning-based methods. ML methods used in detector simulation and in jet tagging in ATLAS will be discussed, along with four searches using ML\/AD techniques.","doi":"10.1140\/epjs\/s11734-024-01238-8","cleaned_title":"how deep learning is complementing deep thinking in atlas","cleaned_abstract":"atlas collaboration uses machine learning (ml) algorithms in many different ways in its physics programme, starting from object reconstruction, simulation of calorimeter showers, signal to background discrimination in searches and measurements, tagging jets based on their origin and so on. anomaly detection (ad) techniques are also gaining popularity where they are used to find hidden patterns in the data, with lesser dependence on simulated samples as in the case of supervised learning-based methods. ml methods used in detector simulation and in jet tagging in atlas will be discussed, along with four searches using ml\/ad techniques.","key_phrases":["atlas collaboration","machine learning","ml","many different ways","its physics programme","object reconstruction","simulation","calorimeter showers","background discrimination","searches","measurements","tagging jets","their origin","anomaly detection","(ad) techniques","popularity","they","hidden patterns","the data","lesser dependence","simulated samples","the case","supervised learning-based methods","ml methods","detector simulation","jet tagging","atlas","four searches","ml\/ad techniques","four"]},{"title":"Multiclass skin lesion classification using deep learning networks optimal information fusion","authors":["Muhammad Attique Khan","Ameer Hamza","Mohammad Shabaz","Seifeine Kadry","Saddaf Rubab","Muhammad Abdullah Bilal","Muhammad Naeem Akbar","Suresh Manic Kesavan"],"abstract":"A serious, all-encompassing, and deadly cancer that affects every part of the body is skin cancer. The most prevalent causes of skin lesions are UV radiation, which can damage human skin, and moles. If skin cancer is discovered early, it may be adequately treated. In order to diagnose skin lesions with less effort, dermatologists are increasingly turning to machine learning (ML) techniques and computer-aided diagnostic (CAD) systems. This paper proposes a computerized method for multiclass lesion classification using a fusion of optimal deep-learning model features. The dataset used in this work, ISIC2018, is imbalanced; therefore, augmentation is performed based on a few mathematical operations. After that, two pre-trained deep learning models (DarkNet-19 and MobileNet-V2) have been fine-tuned and trained on the selected dataset. After training, features are extracted from the average pool layer and optimized using a hybrid firefly optimization technique. The selected features are fused in two ways: (i) original serial approach and (ii) proposed threshold approach. Machine learning classifiers are used to classify the chosen features at the end. Using the ISIC2018 dataset, the experimental procedure produced an accuracy of 89.0%. Whereas, 87.34, 87.57, and 87.45 are sensitivity, precision, and F1 score respectively. At the end, comparison is also conducted with recent techniques, and it shows the proposed method shows improved accuracy along with other performance measures.","doi":"10.1007\/s42452-024-05998-9","cleaned_title":"multiclass skin lesion classification using deep learning networks optimal information fusion","cleaned_abstract":"a serious, all-encompassing, and deadly cancer that affects every part of the body is skin cancer. the most prevalent causes of skin lesions are uv radiation, which can damage human skin, and moles. if skin cancer is discovered early, it may be adequately treated. in order to diagnose skin lesions with less effort, dermatologists are increasingly turning to machine learning (ml) techniques and computer-aided diagnostic (cad) systems. this paper proposes a computerized method for multiclass lesion classification using a fusion of optimal deep-learning model features. the dataset used in this work, isic2018, is imbalanced; therefore, augmentation is performed based on a few mathematical operations. after that, two pre-trained deep learning models (darknet-19 and mobilenet-v2) have been fine-tuned and trained on the selected dataset. after training, features are extracted from the average pool layer and optimized using a hybrid firefly optimization technique. the selected features are fused in two ways: (i) original serial approach and (ii) proposed threshold approach. machine learning classifiers are used to classify the chosen features at the end. using the isic2018 dataset, the experimental procedure produced an accuracy of 89.0%. whereas, 87.34, 87.57, and 87.45 are sensitivity, precision, and f1 score respectively. at the end, comparison is also conducted with recent techniques, and it shows the proposed method shows improved accuracy along with other performance measures.","key_phrases":["a serious, all-encompassing, and deadly cancer","that","every part","the body","skin cancer","the most prevalent causes","skin lesions","uv radiation","which","human skin","skin cancer","it","order","skin lesions","less effort","dermatologists","machine learning","(ml) techniques","computer-aided diagnostic (cad) systems","this paper","a computerized method","multiclass lesion classification","a fusion","optimal deep-learning model features","the dataset","this work","isic2018","augmentation","a few mathematical operations","that","two pre-trained deep learning models","darknet-19","mobilenet-v2","the selected dataset","training","features","the average pool layer","a hybrid firefly optimization technique","the selected features","two ways","(i) original serial approach","(ii) proposed threshold approach","machine learning classifiers","the chosen features","the end","the isic2018 dataset","the experimental procedure","an accuracy","89.0%","sensitivity","precision","f1 score","the end","comparison","recent techniques","it","the proposed method","improved accuracy","other performance measures","isic2018","two","darknet-19","two","isic2018","89.0%","87.34","87.57","87.45"]},{"title":"REDQT: a method for automated mobile application GUI testing based on deep reinforcement learning algorithms","authors":["Fengyu Wang","Chuanqi Tao","Jerry Gao"],"abstract":"As mobile applications become increasingly prevalent in daily life, the demand for their functionality and reliability continues to grow. Traditional mobile application testing methods, particularly graphical user interface (GUI) testing, face challenges of limited automation and adaptability.Despite the application of various machine learning approaches to GUI testing, enhancing the utilization of limited component samples in complex mobile application environments remains an overlooked issue in many automated testing methods. This study introduces a mobile application testing method based on deep reinforcement learning, aimed at improving performance and adaptability during the testing process. By integrating the feature recognition capabilities of deep learning with the decision-making mechanisms of reinforcement learning, our method can effectively simulate user operations and identify potential application pitfalls. Initially, the study analyzes the limitations of traditional mobile application testing methods and explores the advantages of deep reinforcement learning in handling complex tasks. Subsequently, we present RedqT: an automated mobile application GUI testing method based on a deep reinforcement learning algorithm (REDQ), aimed at enhancing the utilization of application information through the characteristics of the REDQ algorithm. A study testing 18 open-source Android applications on GitHub demonstrated that our method shows promising performance in terms of code coverage and testing speed.","doi":"10.1007\/s11761-024-00413-y","cleaned_title":"redqt: a method for automated mobile application gui testing based on deep reinforcement learning algorithms","cleaned_abstract":"as mobile applications become increasingly prevalent in daily life, the demand for their functionality and reliability continues to grow. traditional mobile application testing methods, particularly graphical user interface (gui) testing, face challenges of limited automation and adaptability.despite the application of various machine learning approaches to gui testing, enhancing the utilization of limited component samples in complex mobile application environments remains an overlooked issue in many automated testing methods. this study introduces a mobile application testing method based on deep reinforcement learning, aimed at improving performance and adaptability during the testing process. by integrating the feature recognition capabilities of deep learning with the decision-making mechanisms of reinforcement learning, our method can effectively simulate user operations and identify potential application pitfalls. initially, the study analyzes the limitations of traditional mobile application testing methods and explores the advantages of deep reinforcement learning in handling complex tasks. subsequently, we present redqt: an automated mobile application gui testing method based on a deep reinforcement learning algorithm (redq), aimed at enhancing the utilization of application information through the characteristics of the redq algorithm. a study testing 18 open-source android applications on github demonstrated that our method shows promising performance in terms of code coverage and testing speed.","key_phrases":["mobile applications","daily life","the demand","their functionality","reliability","traditional mobile application testing methods","particularly graphical user interface","gui","testing","challenges","limited automation","the application","various machine learning approaches","gui testing","the utilization","limited component samples","complex mobile application environments","an overlooked issue","many automated testing methods","this study","a mobile application testing method","deep reinforcement learning","performance","adaptability","the testing process","the feature recognition capabilities","deep learning","the decision-making mechanisms","reinforcement learning","our method","user operations","potential application pitfalls","the study","the limitations","traditional mobile application testing methods","the advantages","deep reinforcement learning","complex tasks","we","an automated mobile application gui testing method","a deep reinforcement learning algorithm","redq","the utilization","application information","the characteristics","the redq","a study","18 open-source android applications","github","our method","promising performance","terms","code coverage","testing","speed","daily","18"]},{"title":"Streamflow Prediction Utilizing Deep Learning and Machine Learning Algorithms for Sustainable Water Supply Management","authors":["Sarmad Dashti Latif","Ali Najah Ahmed"],"abstract":"As a result of global climate change, sustainable water supply management is becoming increasingly difficult. Dams and reservoirs are key tools for controlling and managing water resources; they have benefited human cultures in a variety of ways, including enhanced human health, increased food production, water supply for domestic and industrial use, economic growth, irrigation, hydro-power generation, and flood control. This study aims to compare the application of deep learning and conventional machine learning algorithms for predicting daily reservoir inflow. Long short-term memory (LSTM) has been applied as a deep learning algorithm and boosted regression tree (BRT) has been implemented as a machine learning algorithm. Five statistical indices have been selected to evaluate the performance of the proposed models. The selected statistical measurements are mean absolute error (MAE), root mean square error (RMSE), correlation coefficient (R), coefficient of determination (R2), mean square error (MSE), Nash Sutcliffe Model Efficiency Coefficient (NSE), and the RMSE-observations standard deviation ratio (RSR). The findings showed that LSTM outperformed BRT with a significant difference in terms of accuracy.","doi":"10.1007\/s11269-023-03499-9","cleaned_title":"streamflow prediction utilizing deep learning and machine learning algorithms for sustainable water supply management","cleaned_abstract":"as a result of global climate change, sustainable water supply management is becoming increasingly difficult. dams and reservoirs are key tools for controlling and managing water resources; they have benefited human cultures in a variety of ways, including enhanced human health, increased food production, water supply for domestic and industrial use, economic growth, irrigation, hydro-power generation, and flood control. this study aims to compare the application of deep learning and conventional machine learning algorithms for predicting daily reservoir inflow. long short-term memory (lstm) has been applied as a deep learning algorithm and boosted regression tree (brt) has been implemented as a machine learning algorithm. five statistical indices have been selected to evaluate the performance of the proposed models. the selected statistical measurements are mean absolute error (mae), root mean square error (rmse), correlation coefficient (r), coefficient of determination (r2), mean square error (mse), nash sutcliffe model efficiency coefficient (nse), and the rmse-observations standard deviation ratio (rsr). the findings showed that lstm outperformed brt with a significant difference in terms of accuracy.","key_phrases":["a result","global climate change","sustainable water supply management","dams","reservoirs","key tools","water resources","they","human cultures","a variety","ways","enhanced human health","increased food production","water supply","domestic and industrial use","economic growth","irrigation","hydro-power generation","flood control","this study","the application","deep learning","conventional machine","algorithms","daily reservoir inflow","long short-term memory","lstm","a deep learning algorithm","regression tree","brt","a machine learning algorithm","five statistical indices","the performance","the proposed models","the selected statistical measurements","mean absolute error","mae","root mean square error","rmse","correlation coefficient","r","coefficient","determination","r2","square error","mse","nash sutcliffe model efficiency coefficient","nse","the rmse-observations standard deviation ratio","rsr","the findings","lstm","brt","a significant difference","terms","accuracy","daily","five","nash sutcliffe"]},{"title":"Deep learning enables fast, gentle STED microscopy","authors":["Vahid Ebrahimi","Till Stephan","Jiah Kim","Pablo Carravilla","Christian Eggeling","Stefan Jakobs","Kyu Young Han"],"abstract":"STED microscopy is widely used to image subcellular structures with super-resolution. Here, we report that restoring STED images with deep learning can mitigate photobleaching and photodamage by reducing the pixel dwell time by one or two orders of magnitude. Our method allows for efficient and robust restoration of noisy 2D and 3D STED images with multiple targets and facilitates long-term imaging of mitochondrial dynamics.","doi":"10.1038\/s42003-023-05054-z","cleaned_title":"deep learning enables fast, gentle sted microscopy","cleaned_abstract":"sted microscopy is widely used to image subcellular structures with super-resolution. here, we report that restoring sted images with deep learning can mitigate photobleaching and photodamage by reducing the pixel dwell time by one or two orders of magnitude. our method allows for efficient and robust restoration of noisy 2d and 3d sted images with multiple targets and facilitates long-term imaging of mitochondrial dynamics.","key_phrases":["sted microscopy","subcellular structures","super","-","resolution","we","sted images","deep learning","photobleaching","photodamage","the pixel","dwell time","one or two orders","magnitude","our method","efficient and robust restoration","noisy 2d","3d sted images","multiple targets","long-term imaging","mitochondrial dynamics","one","two","2d","3d"]},{"title":"Biological gender identification in Turkish news text using deep learning models","authors":["P\u0131nar T\u00fcfekci","Melike Bekta\u015f K\u00f6sesoy"],"abstract":"Identifying the biological gender of authors based on the content of their written work is a crucial task in Natural Language Processing (NLP). Accurate biological gender identification finds numerous applications in fields such as linguistics, sociology, and marketing. However, achieving high accuracy in identifying the biological gender of the author is heavily dependent on the quality of the collected data and its proper splitting. Therefore, determining the best-performing model necessitates experimental evaluation. This study aimed to develop and evaluate four learning algorithms for biological gender identification in news texts. To this end, a comprehensive dataset, IAG-TNKU, was created from a Turkish newspaper, comprising 43,292 news articles. Four models utilizing popular machine learning algorithms, including Naive Bayes and Random Forest, and two deep learning algorithms, Long Short Term Memory and Convolutional Neural Networks, were developed and evaluated rigorously. The results indicated that the Long Short Term Memory (LSTM) algorithm outperformed the other three models, exhibiting an exceptional accuracy of 88.51%. This model's outstanding performance underpins the importance of utilizing innovative deep learning algorithms for biological gender identification tasks in NLP. The present study contributes to extant literature by developing a new dataset for biological gender identification in news texts and evaluating four machine learning algorithms. Our findings highlight the significance of utilizing innovative techniques for biological gender identification tasks. The dataset and deep learning algorithm can be applied in many areas such as sociolinguistics, marketing research, and journalism, where the identification of biological gender in written content plays a pivotal role.","doi":"10.1007\/s11042-023-17622-w","cleaned_title":"biological gender identification in turkish news text using deep learning models","cleaned_abstract":"identifying the biological gender of authors based on the content of their written work is a crucial task in natural language processing (nlp). accurate biological gender identification finds numerous applications in fields such as linguistics, sociology, and marketing. however, achieving high accuracy in identifying the biological gender of the author is heavily dependent on the quality of the collected data and its proper splitting. therefore, determining the best-performing model necessitates experimental evaluation. this study aimed to develop and evaluate four learning algorithms for biological gender identification in news texts. to this end, a comprehensive dataset, iag-tnku, was created from a turkish newspaper, comprising 43,292 news articles. four models utilizing popular machine learning algorithms, including naive bayes and random forest, and two deep learning algorithms, long short term memory and convolutional neural networks, were developed and evaluated rigorously. the results indicated that the long short term memory (lstm) algorithm outperformed the other three models, exhibiting an exceptional accuracy of 88.51%. this model's outstanding performance underpins the importance of utilizing innovative deep learning algorithms for biological gender identification tasks in nlp. the present study contributes to extant literature by developing a new dataset for biological gender identification in news texts and evaluating four machine learning algorithms. our findings highlight the significance of utilizing innovative techniques for biological gender identification tasks. the dataset and deep learning algorithm can be applied in many areas such as sociolinguistics, marketing research, and journalism, where the identification of biological gender in written content plays a pivotal role.","key_phrases":["the biological gender","authors","the content","their written work","a crucial task","natural language processing","nlp","accurate biological gender identification","numerous applications","fields","linguistics","sociology","marketing","high accuracy","the biological gender","the author","the quality","the collected data","its proper splitting","the best-performing model necessitates experimental evaluation","this study","four learning algorithms","biological gender identification","news texts","this end","a comprehensive dataset","iag","tnku","a turkish newspaper","43,292 news articles","four models","popular machine learning algorithms","naive bayes","random forest","two deep learning algorithms","long short term memory","convolutional neural networks","the results","the long short term memory","lstm","algorithm","the other three models","an exceptional accuracy","88.51%","this model's outstanding performance","the importance","innovative deep learning algorithms","biological gender identification tasks","nlp","the present study","extant literature","a new dataset","biological gender identification","news texts","four machine learning algorithms","our findings","the significance","innovative techniques","biological gender identification tasks","the dataset and deep learning algorithm","many areas","sociolinguistics","marketing research","journalism","the identification","biological gender","written content","a pivotal role","four","turkish","43,292","four","two","three","88.51%","four"]},{"title":"Analysis, characterization, prediction, and attribution of extreme atmospheric events with machine learning and deep learning techniques: a review","authors":["Sancho Salcedo-Sanz","Jorge P\u00e9rez-Aracil","Guido Ascenso","Javier Del Ser","David Casillas-P\u00e9rez","Christopher Kadow","Du\u0161an Fister","David Barriopedro","Ricardo Garc\u00eda-Herrera","Matteo Giuliani","Andrea Castelletti"],"abstract":"Atmospheric extreme events cause severe damage to human societies and ecosystems. The frequency and intensity of extremes and other associated events are continuously increasing due to climate change and global warming. The accurate prediction, characterization, and attribution of atmospheric extreme events is, therefore, a key research field in which many groups are currently working by applying different methodologies and computational tools. Machine learning and deep learning methods have arisen in the last years as powerful techniques to tackle many of the problems related to atmospheric extreme events. This paper reviews machine learning and deep learning approaches applied to the analysis, characterization, prediction, and attribution of the most important atmospheric extremes. A summary of the most used machine learning and deep learning techniques in this area, and a comprehensive critical review of literature related to ML in EEs, are provided. The critical literature review has been extended to extreme events related to rainfall and floods, heatwaves and extreme temperatures, droughts, severe weather events and fog, and low-visibility episodes. A case study focused on the analysis of extreme atmospheric temperature prediction with ML and DL techniques is also presented in the paper. Conclusions, perspectives, and outlooks on the field are finally drawn.","doi":"10.1007\/s00704-023-04571-5","cleaned_title":"analysis, characterization, prediction, and attribution of extreme atmospheric events with machine learning and deep learning techniques: a review","cleaned_abstract":"atmospheric extreme events cause severe damage to human societies and ecosystems. the frequency and intensity of extremes and other associated events are continuously increasing due to climate change and global warming. the accurate prediction, characterization, and attribution of atmospheric extreme events is, therefore, a key research field in which many groups are currently working by applying different methodologies and computational tools. machine learning and deep learning methods have arisen in the last years as powerful techniques to tackle many of the problems related to atmospheric extreme events. this paper reviews machine learning and deep learning approaches applied to the analysis, characterization, prediction, and attribution of the most important atmospheric extremes. a summary of the most used machine learning and deep learning techniques in this area, and a comprehensive critical review of literature related to ml in ees, are provided. the critical literature review has been extended to extreme events related to rainfall and floods, heatwaves and extreme temperatures, droughts, severe weather events and fog, and low-visibility episodes. a case study focused on the analysis of extreme atmospheric temperature prediction with ml and dl techniques is also presented in the paper. conclusions, perspectives, and outlooks on the field are finally drawn.","key_phrases":["atmospheric extreme events","severe damage","human societies","ecosystems","the frequency","intensity","extremes","other associated events","climate change","global warming","the accurate prediction","characterization","attribution","atmospheric extreme events","a key research field","which","many groups","different methodologies","computational tools","machine learning","deep learning methods","the last years","powerful techniques","the problems","atmospheric extreme events","this paper","machine learning","deep learning approaches","the analysis","characterization","prediction","attribution","the most important atmospheric extremes","a summary","the most used machine learning","deep learning techniques","this area","a comprehensive critical review","literature","ml","ees","the critical literature review","extreme events","rainfall","floods","heatwaves","extreme temperatures","droughts","severe weather events","fog","low-visibility episodes","a case study","the analysis","extreme atmospheric temperature prediction","ml","dl techniques","the paper","conclusions","perspectives","outlooks","the field","the last years"]},{"title":"Deep-learning based sleep apnea detection using sleep sound, SpO2, and pulse\u00a0rate","authors":["Chutinan Singtothong","Thitirat Siriborvornratanakul"],"abstract":"Sleep apnea, a common sleep disorder where breathing is repeatedly interrupted during sleep, poses significant health risks. Traditional diagnostic methods like overnight polysomnography (PSG) are complex and costly, limiting widespread screening. This study suggests a deep learning model to identify sleep apnea using sleep sounds, oxygen saturation (SpO2), and pulse rate. Mel-spectrogram, computed from PSG data of 24 patients, serves as input. The study shows the effectiveness of deep learning, with the combined model achieving 96% accuracy in inferring apnea severity, outperforming individual models using SpO2 and pulse rate (79%) and sleep sound (83%).","doi":"10.1007\/s41870-024-01906-x","cleaned_title":"deep-learning based sleep apnea detection using sleep sound, spo2, and pulse rate","cleaned_abstract":"sleep apnea, a common sleep disorder where breathing is repeatedly interrupted during sleep, poses significant health risks. traditional diagnostic methods like overnight polysomnography (psg) are complex and costly, limiting widespread screening. this study suggests a deep learning model to identify sleep apnea using sleep sounds, oxygen saturation (spo2), and pulse rate. mel-spectrogram, computed from psg data of 24 patients, serves as input. the study shows the effectiveness of deep learning, with the combined model achieving 96% accuracy in inferring apnea severity, outperforming individual models using spo2 and pulse rate (79%) and sleep sound (83%).","key_phrases":["sleep apnea","a common sleep disorder","breathing","sleep","significant health risks","traditional diagnostic methods","overnight polysomnography","psg","widespread screening","this study","a deep learning model","sleep apnea","sleep sounds","oxygen saturation","spo2","pulse rate","mel-spectrogram","psg data","24 patients","input","the study","the effectiveness","deep learning","the combined model","96% accuracy","apnea severity","outperforming individual models","spo2","pulse rate","79%","sleep sound","83%","overnight","mel-spectrogram","24","96%","spo2","79%","83%"]},{"title":"Deep learning automatically assesses 2-\u00b5m laser-induced skin damage OCT images","authors":["Changke Wang","Qiong Ma","Yu Wei","Qi Liu","Yuqing Wang","Chenliang Xu","Caihui Li","Qingyu Cai","Haiyang Sun","Xiaoan Tang","Hongxiang Kang"],"abstract":"The present study proposed a noninvasive, automated, in vivo assessment method based on optical coherence tomography (OCT) and deep learning techniques to qualitatively and quantitatively analyze the biological effects of 2-\u00b5m laser-induced skin damage at different irradiation doses. Different doses of 2-\u00b5m laser irradiation established a mouse skin damage model, after which the skin-damaged tissues were imaged non-invasively in vivo using OCT. The acquired images were preprocessed to construct the dataset required for deep learning. The deep learning models used were U-Net, DeepLabV3+, PSP-Net, and HR-Net, and the trained models were used to segment the damage images and further quantify the damage volume of mouse skin under different irradiation doses. The comparison of the qualitative and quantitative results of the four network models showed that HR-Net had the best performance, the highest agreement between the segmentation results and real values, and the smallest error in the quantitative assessment of the damage volume. Based on HR-Net to segment the damage image and quantify the damage volume, the irradiation doses 5.41, 9.55, 13.05, 20.85, 32.71, 52.92, 76.71, and 97.24\u00a0J\/cm\u00b2 corresponded to a damage volume of 4.58, 12.56, 16.74, 20.88, 24.52, 30.75, 34.13, and 37.32\u00a0mm\u00b3. The damage volume increased in a radiation dose-dependent manner.","doi":"10.1007\/s10103-024-04053-8","cleaned_title":"deep learning automatically assesses 2-\u00b5m laser-induced skin damage oct images","cleaned_abstract":"the present study proposed a noninvasive, automated, in vivo assessment method based on optical coherence tomography (oct) and deep learning techniques to qualitatively and quantitatively analyze the biological effects of 2-\u00b5m laser-induced skin damage at different irradiation doses. different doses of 2-\u00b5m laser irradiation established a mouse skin damage model, after which the skin-damaged tissues were imaged non-invasively in vivo using oct. the acquired images were preprocessed to construct the dataset required for deep learning. the deep learning models used were u-net, deeplabv3+, psp-net, and hr-net, and the trained models were used to segment the damage images and further quantify the damage volume of mouse skin under different irradiation doses. the comparison of the qualitative and quantitative results of the four network models showed that hr-net had the best performance, the highest agreement between the segmentation results and real values, and the smallest error in the quantitative assessment of the damage volume. based on hr-net to segment the damage image and quantify the damage volume, the irradiation doses 5.41, 9.55, 13.05, 20.85, 32.71, 52.92, 76.71, and 97.24 j\/cm\u00b2 corresponded to a damage volume of 4.58, 12.56, 16.74, 20.88, 24.52, 30.75, 34.13, and 37.32 mm\u00b3. the damage volume increased in a radiation dose-dependent manner.","key_phrases":["the present study","vivo assessment method","optical coherence tomography","oct","techniques","the biological effects","2-\u00b5m laser-induced skin damage","different irradiation doses","different doses","2-\u00b5m laser irradiation","a mouse skin damage model","which","the skin-damaged tissues","vivo","oct","the acquired images","the dataset","deep learning","the deep learning models","u","-","net","psp-net","hr-net","the trained models","the damage images","the damage volume","mouse skin","different irradiation doses","the comparison","the qualitative and quantitative results","the four network models","hr-net","the best performance","the highest agreement","the segmentation results","real values","the smallest error","the quantitative assessment","the damage volume","hr-net","the damage image","the damage volume","the irradiation","97.24 j\/cm\u00b2","a damage volume","37.32 mm\u00b3.","the damage volume","a radiation dose-dependent manner","2-\u00b5m","2-\u00b5m","oct","four","5.41","9.55","13.05","20.85","32.71","52.92","76.71","97.24","4.58","12.56","16.74","20.88","24.52","30.75","34.13","37.32"]},{"title":"A hybrid deep model with cumulative learning for few-shot learning","authors":["Jiehao Liu","Zhao Yang","Liufei Luo","Mingkai Luo","Luyu Hu","Jiahao Li"],"abstract":"Few-shot learning (FSL) aims to recognize unseen classes with only a few samples for each class. This challenging research endeavors to narrow the gap between the computer vision technology and the human visual system. Recently, mainstream approaches for FSL can be grouped into meta-learning and classification learning. These two methods train the FSL model from local and global classification viewpoints respectively. In our work, we find the former method can effectively learn transferable knowledge (generalization capacity) with an episodic training paradigm but encounters the problem of slow convergence. The latter method can build an essential classification ability quickly (classification capacity) with a mini-batch training paradigm but easily causes an over-fitting problem. In light of this issue, we propose a hybrid deep model with cumulative learning to tackle the FSL problem by absorbing the advantages of the both methods. The proposed hybrid deep model innovatively integrates meta-learning and classification learning (IMC) in a unified two-branch network framework in which a meta-learning branch and a classification learning branch can work simultaneously. Besides, by considering the different characteristics of the two branches, we propose a cumulative learning strategy to take care of both generalization capacity learning and classification capacity learning in our IMC model training. With the proposed method, the model can quickly build the basic classification capability at the initial stage and continually mine discriminative class information during the remaining training for better generalization. Extensive experiments on CIFAR-FS, FC100, mini-ImageNet and tiered-ImageNet datasets are implemented to demonstrate the promising performance of our method.","doi":"10.1007\/s11042-022-14218-8","cleaned_title":"a hybrid deep model with cumulative learning for few-shot learning","cleaned_abstract":"few-shot learning (fsl) aims to recognize unseen classes with only a few samples for each class. this challenging research endeavors to narrow the gap between the computer vision technology and the human visual system. recently, mainstream approaches for fsl can be grouped into meta-learning and classification learning. these two methods train the fsl model from local and global classification viewpoints respectively. in our work, we find the former method can effectively learn transferable knowledge (generalization capacity) with an episodic training paradigm but encounters the problem of slow convergence. the latter method can build an essential classification ability quickly (classification capacity) with a mini-batch training paradigm but easily causes an over-fitting problem. in light of this issue, we propose a hybrid deep model with cumulative learning to tackle the fsl problem by absorbing the advantages of the both methods. the proposed hybrid deep model innovatively integrates meta-learning and classification learning (imc) in a unified two-branch network framework in which a meta-learning branch and a classification learning branch can work simultaneously. besides, by considering the different characteristics of the two branches, we propose a cumulative learning strategy to take care of both generalization capacity learning and classification capacity learning in our imc model training. with the proposed method, the model can quickly build the basic classification capability at the initial stage and continually mine discriminative class information during the remaining training for better generalization. extensive experiments on cifar-fs, fc100, mini-imagenet and tiered-imagenet datasets are implemented to demonstrate the promising performance of our method.","key_phrases":["few-shot learning","unseen classes","only a few samples","each class","this challenging research endeavors","the gap","the computer vision technology","the human visual system","mainstream approaches","fsl","meta-learning and classification learning","these two methods","the fsl model","local and global classification viewpoints","our work","we","the former method","transferable knowledge","generalization capacity","an episodic training paradigm","the problem","slow convergence","the latter method","an essential classification ability","classification capacity","a mini-batch training paradigm","an over-fitting problem","light","this issue","we","a hybrid deep model","cumulative learning","the fsl problem","the advantages","the both methods","the proposed hybrid deep model","meta-learning and classification learning","imc","a unified two-branch network framework","which","a meta-learning branch","a classification learning branch","the different characteristics","the two branches","we","a cumulative learning strategy","care","both generalization capacity learning and classification capacity","our imc model training","the proposed method","the model","the basic classification capability","the initial stage","continually mine discriminative class information","the remaining training","better generalization","extensive experiments","cifar-fs, fc100, mini","-","imagenet","tiered-imagenet datasets","the promising performance","our method","two","two","two"]},{"title":"Cytopathic Effect Detection and Clonal Selection using Deep Learning","authors":["Yu Yuan","Tony Wang","Jordan Sims","Kim Le","Cenk Undey","Erdal Oruklu"],"abstract":"Purpose:In biotechnology, microscopic cell imaging is often used to identify and analyze cell morphology and cell state for a variety of applications. For example, microscopy can be used to detect the presence of cytopathic effects (CPE) in cell culture samples to determine virus contamination. Another application of microscopy is to verify clonality during cell line development. Conventionally, inspection of these microscopy images is performed manually by human analysts. This is both tedious and time consuming. In this paper, we propose using supervised deep learning algorithms to automate the cell detection processes mentioned above.Methods:The proposed algorithms utilize image processing techniques and convolutional neural networks (CNN) to detect the presence of CPE and to verify the clonality in cell line development.Results:We train and test the algorithms on image data which have been collected and labeled by domain experts. Our experiments have shown promising results in terms of both accuracy and speed.Conclusion:Deep learning algorithms achieve high accuracy (more than 95%) on both CPE detection and clonal selection applications, resulting in a highly efficient and cost-effective automation process.","doi":"10.1007\/s11095-024-03749-4","cleaned_title":"cytopathic effect detection and clonal selection using deep learning","cleaned_abstract":"purpose:in biotechnology, microscopic cell imaging is often used to identify and analyze cell morphology and cell state for a variety of applications. for example, microscopy can be used to detect the presence of cytopathic effects (cpe) in cell culture samples to determine virus contamination. another application of microscopy is to verify clonality during cell line development. conventionally, inspection of these microscopy images is performed manually by human analysts. this is both tedious and time consuming. in this paper, we propose using supervised deep learning algorithms to automate the cell detection processes mentioned above.methods:the proposed algorithms utilize image processing techniques and convolutional neural networks (cnn) to detect the presence of cpe and to verify the clonality in cell line development.results:we train and test the algorithms on image data which have been collected and labeled by domain experts. our experiments have shown promising results in terms of both accuracy and speed.conclusion:deep learning algorithms achieve high accuracy (more than 95%) on both cpe detection and clonal selection applications, resulting in a highly efficient and cost-effective automation process.","key_phrases":["purpose","biotechnology","microscopic cell imaging","cell morphology","cell state","a variety","applications","example","the presence","cytopathic effects","cell culture samples","virus contamination","another application","microscopy","clonality","cell line development","inspection","these microscopy images","human analysts","this","this paper","we","supervised deep learning algorithms","the cell detection processes","above.methods","the proposed algorithms","image processing techniques","convolutional neural networks","cnn","the presence","cpe","the clonality","cell line development.results","we","the algorithms","image data","which","domain experts","our experiments","promising results","terms","both accuracy","speed.conclusion","deep learning algorithms","high accuracy","(more than 95%","both cpe detection and clonal selection applications","a highly efficient and cost-effective automation process","cnn","more than 95%"]},{"title":"A secured deep learning based smart home automation system","authors":["Chitukula Sanjay","Konda Jahnavi","Shyam Karanth"],"abstract":"With the expansion of modern technologies and the Internet of Things (IoT), the concept of smart homes has gained tremendous popularity with a view to making people\u2019s lives easier by ensuring a secured environment. Several home automation systems have been developed to report suspicious activities by capturing the movements of residents. However, these systems are associated with challenges such as weak security, lack of interoperability and integration with IoT devices, timely reporting of suspicious movements, etc. Therefore, the given paper proposes a novel smart home automation framework for controlling home appliances by integrating with sensors, IoT devices, and microcontrollers, which would in turn monitor the movements and send notifications about suspicious movements on the resident\u2019s smartphone. The proposed framework makes use of convolutional neural networks (CNNs) for motion detection and classification based on pre-processing of images. The images related to the movements of residents are captured by a spy camera installed in the system. It helps in identification of outsiders based on differentiation of motion patterns. The performance of the framework is compared with existing deep learning models used in recent studies based on evaluation metrics such as accuracy (%), precision (%), recall (%), and f-1 measure (%). The results show that the proposed framework attains the highest accuracy (98.67%), thereby surpassing the existing deep learning models used in smart home automation systems.","doi":"10.1007\/s41870-024-02097-1","cleaned_title":"a secured deep learning based smart home automation system","cleaned_abstract":"with the expansion of modern technologies and the internet of things (iot), the concept of smart homes has gained tremendous popularity with a view to making people\u2019s lives easier by ensuring a secured environment. several home automation systems have been developed to report suspicious activities by capturing the movements of residents. however, these systems are associated with challenges such as weak security, lack of interoperability and integration with iot devices, timely reporting of suspicious movements, etc. therefore, the given paper proposes a novel smart home automation framework for controlling home appliances by integrating with sensors, iot devices, and microcontrollers, which would in turn monitor the movements and send notifications about suspicious movements on the resident\u2019s smartphone. the proposed framework makes use of convolutional neural networks (cnns) for motion detection and classification based on pre-processing of images. the images related to the movements of residents are captured by a spy camera installed in the system. it helps in identification of outsiders based on differentiation of motion patterns. the performance of the framework is compared with existing deep learning models used in recent studies based on evaluation metrics such as accuracy (%), precision (%), recall (%), and f-1 measure (%). the results show that the proposed framework attains the highest accuracy (98.67%), thereby surpassing the existing deep learning models used in smart home automation systems.","key_phrases":["the expansion","modern technologies","the internet","things","iot","the concept","smart homes","tremendous popularity","a view","people","a secured environment","several home automation systems","suspicious activities","the movements","residents","these systems","challenges","weak security","lack","interoperability","integration","iot devices","timely reporting","suspicious movements","the given paper","a novel smart home automation framework","home appliances","sensors","iot devices","microcontrollers","which","turn","the movements","notifications","suspicious movements","the resident\u2019s smartphone","the proposed framework","use","convolutional neural networks","cnns","motion detection","classification","pre","processing","images","the images","the movements","residents","a spy camera","the system","it","identification","outsiders","differentiation","motion patterns","the performance","the framework","existing deep learning models","recent studies","evaluation metrics","accuracy","precision","recall","f-1 measure","the results","the proposed framework","the highest accuracy","98.67%","the existing deep learning models","smart home automation systems","98.67%"]},{"title":"Minimization of occurrence of retained surgical items using machine learning and deep learning techniques: a review","authors":["Mohammed Abo-Zahhad","Ahmed H. Abd El-Malek","Mohammed S. Sayed","Susan Njeri Gitau"],"abstract":"Retained surgical items (RSIs) pose significant risks to patients and healthcare professionals, prompting extensive efforts to reduce their incidence. RSIs are objects inadvertently left within patients\u2019 bodies after surgery, which can lead to severe consequences such as infections and death. The repercussions highlight the critical need to address this issue. Machine learning (ML) and deep learning (DL) have displayed considerable potential for enhancing the prevention of RSIs through heightened precision and decreased reliance on human involvement. ML techniques are finding an expanding number of applications in medicine, ranging from automated imaging analysis to diagnosis. DL has enabled substantial advances in the prediction capabilities of computers by combining the availability of massive volumes of data with extremely effective learning algorithms. This paper reviews and evaluates recently published articles on the application of ML and DL in RSIs prevention and diagnosis, stressing the need for a multi-layered approach that leverages each method\u2019s strengths to mitigate RSI risks. It highlights the key findings, advantages, and limitations of the different techniques used. Extensive datasets for training ML and DL models could enhance RSI detection systems. This paper also discusses the various datasets used by researchers for training the models. In addition, future directions for improving these technologies for RSI diagnosis and prevention are considered. By merging ML and DL with current procedures, it is conceivable to substantially minimize RSIs, enhance patient safety, and elevate surgical care standards.","doi":"10.1186\/s13040-024-00367-z","cleaned_title":"minimization of occurrence of retained surgical items using machine learning and deep learning techniques: a review","cleaned_abstract":"retained surgical items (rsis) pose significant risks to patients and healthcare professionals, prompting extensive efforts to reduce their incidence. rsis are objects inadvertently left within patients\u2019 bodies after surgery, which can lead to severe consequences such as infections and death. the repercussions highlight the critical need to address this issue. machine learning (ml) and deep learning (dl) have displayed considerable potential for enhancing the prevention of rsis through heightened precision and decreased reliance on human involvement. ml techniques are finding an expanding number of applications in medicine, ranging from automated imaging analysis to diagnosis. dl has enabled substantial advances in the prediction capabilities of computers by combining the availability of massive volumes of data with extremely effective learning algorithms. this paper reviews and evaluates recently published articles on the application of ml and dl in rsis prevention and diagnosis, stressing the need for a multi-layered approach that leverages each method\u2019s strengths to mitigate rsi risks. it highlights the key findings, advantages, and limitations of the different techniques used. extensive datasets for training ml and dl models could enhance rsi detection systems. this paper also discusses the various datasets used by researchers for training the models. in addition, future directions for improving these technologies for rsi diagnosis and prevention are considered. by merging ml and dl with current procedures, it is conceivable to substantially minimize rsis, enhance patient safety, and elevate surgical care standards.","key_phrases":["surgical items","rsis","significant risks","patients","healthcare professionals","extensive efforts","their incidence","rsis","objects","patients\u2019 bodies","surgery","which","severe consequences","infections","death","the repercussions","the critical need","this issue","machine learning","ml","deep learning","dl","considerable potential","the prevention","rsis","heightened precision","reliance","human involvement","techniques","an expanding number","applications","medicine","automated imaging analysis","diagnosis","dl","substantial advances","the prediction capabilities","computers","the availability","massive volumes","data","extremely effective learning algorithms","this paper reviews","evaluates","articles","the application","ml","dl","rsis prevention","diagnosis","the need","a multi-layered approach","that","each method\u2019s strengths","rsi risks","it","the key findings","advantages","limitations","the different techniques","extensive datasets","training ml","dl models","rsi detection systems","this paper","the various datasets","researchers","the models","addition","future directions","these technologies","rsi diagnosis","prevention","ml","dl","current procedures","it","rsis","patient safety","surgical care standards"]},{"title":"FPGA implementation of deep learning architecture for kidney cancer detection from histopathological images","authors":["Shyam Lal","Amit Kumar Chanchal","Jyoti Kini","Gopal Krishna Upadhyay"],"abstract":"Kidney cancer is the most common type of cancer, and designing an automated system to accurately classify the cancer grade is of paramount importance for a better prognosis of the disease from histopathological kidney cancer images. Application of deep learning neural networks (DLNNs) for histopathological image classification is thriving and implementation of these networks on edge devices has been gaining the ground correspondingly due to high computational power and low latency requirements. This paper designs an automated system that classifies histopathological kidney cancer images. For experimentation, we have collected Kidney histopathological images of Non-cancerous, cancerous, and their respective grade of Renal Cell Carcinoma (RCC) from Kasturba Medical College (KMC), Mangalore, Karnataka, India. We have implemented and analyzed performances of deep learning architectures on a Field Programmable Gate Array (FPGA) board. Results yield that the Inception-V3 network provides better accuracy for kidney cancer detection as compared to other deep learning models on Kidney histopathological images. Further, the DenseNet-169 network provides better accuracy for kidney cancer grading as compared to other existing deep learning architecture on the FPGA board.","doi":"10.1007\/s11042-023-17895-1","cleaned_title":"fpga implementation of deep learning architecture for kidney cancer detection from histopathological images","cleaned_abstract":"kidney cancer is the most common type of cancer, and designing an automated system to accurately classify the cancer grade is of paramount importance for a better prognosis of the disease from histopathological kidney cancer images. application of deep learning neural networks (dlnns) for histopathological image classification is thriving and implementation of these networks on edge devices has been gaining the ground correspondingly due to high computational power and low latency requirements. this paper designs an automated system that classifies histopathological kidney cancer images. for experimentation, we have collected kidney histopathological images of non-cancerous, cancerous, and their respective grade of renal cell carcinoma (rcc) from kasturba medical college (kmc), mangalore, karnataka, india. we have implemented and analyzed performances of deep learning architectures on a field programmable gate array (fpga) board. results yield that the inception-v3 network provides better accuracy for kidney cancer detection as compared to other deep learning models on kidney histopathological images. further, the densenet-169 network provides better accuracy for kidney cancer grading as compared to other existing deep learning architecture on the fpga board.","key_phrases":["kidney cancer","the most common type","cancer","an automated system","the cancer grade","paramount importance","a better prognosis","the disease","histopathological kidney cancer images","application","neural networks","dlnns","histopathological image classification","implementation","these networks","edge devices","the ground","high computational power","low latency requirements","this paper","an automated system","that","histopathological kidney cancer images","experimentation","we","kidney histopathological images","their respective grade","renal cell carcinoma","rcc","kasturba medical college","mangalore","karnataka","india","we","analyzed performances","deep learning architectures","a field programmable gate array","(fpga) board","results","the inception-v3 network","better accuracy","kidney cancer detection","other deep learning models","kidney histopathological images","the densenet-169 network","better accuracy","kidney cancer","other existing deep learning architecture","the fpga board","rcc","kasturba medical college (kmc","karnataka","india"]},{"title":"Sapotech Reveal CAST\u2014Deep Learning Based Surface Quality Analysis of Hot Steel Slabs in Continuous Casting","authors":["Saku Kaukonen","Eemil Kiviahde","Hannu Suopaj\u00e4rvi"],"abstract":"In this contribution, the Sapotech Reveal CAST machine vision-based surface inspection solution is presented in the context of assessing the surface quality of hot steel slabs during the process of continuous casting.Further, a\u00a0Deep Learning (DL) based approach for defect detection and classification is presented. Deep Learning enables the assessment of surface quality and the detection of defects, which is extremely challenging to be performed using classical machine vision methods. In Reveal CAST, the tools and methods for training, testing, and applying Deep Learning are integrated into a\u00a0single product.Finally, a\u00a0cloud enabled Reveal CAST case for Deep Learning is presented with example results from a\u00a0hybrid cloud-edge reference production system.","doi":"10.1007\/s00501-023-01404-w","cleaned_title":"sapotech reveal cast\u2014deep learning based surface quality analysis of hot steel slabs in continuous casting","cleaned_abstract":"in this contribution, the sapotech reveal cast machine vision-based surface inspection solution is presented in the context of assessing the surface quality of hot steel slabs during the process of continuous casting.further, a deep learning (dl) based approach for defect detection and classification is presented. deep learning enables the assessment of surface quality and the detection of defects, which is extremely challenging to be performed using classical machine vision methods. in reveal cast, the tools and methods for training, testing, and applying deep learning are integrated into a single product.finally, a cloud enabled reveal cast case for deep learning is presented with example results from a hybrid cloud-edge reference production system.","key_phrases":["this contribution","the sapotech","cast machine vision-based surface inspection solution","the context","the surface quality","hot steel slabs","the process","continuous casting.further","a deep learning","(dl) based approach","defect detection","classification","deep learning","the assessment","surface quality","the detection","defects","which","classical machine vision methods","reveal cast","the tools","methods","training","testing","deep learning","a cloud","reveal cast case","deep learning","example results","a hybrid cloud-edge reference production system"]},{"title":"Classification of Plant Leaf Disease Using Deep Learning","authors":["K. Indira","H. Mallika"],"abstract":"Agriculture is critical in human existence. Practically, 60% of the population is occupied with some sort of farming, either straight forwardly or in a roundabout way. The important factor which determines the quality of crops grown in fields is early detection, classification and severity of plant leaf diseases. Traditional plant disease detection was based on manual design of features and classifiers. The traditional methods faced many challenges in real complex environment such as low contrast, noise in the image being captured. In recent years, deep learning model widely used in image classification task is convolutional neural network, and hence, plant disease detection based on deep learning has important academic research value. In this paper, a CNN model with different convolution layers, AlexNet and MobileNet, was designed to classify the plant leaf diseases of pepper bell, tomato, potato, rice, apple and sorghum with a total of 26 classes. The dataset for pepper bell, tomato, potato, rice and apple is taken from kaggle, whereas sorghum dataset is generated by taking pictures from the field. Along with identification of disease, severity detection is performed on two selected diseases of tomato crop using the best model. Using an available dataset of 24,156 images of diseased and healthy plant leaves, the performance of the model was evaluated. Simulation results revealed an accuracy of 84.24% with CNN model consisting of five layers and an accuracy of 91.19% with pretrained AlexNet model and an accuracy of 97.33% with MobileNet with a learning rate of 0.001. Severity detection was performed on tomato early blight and tomato bacterial spot with levels 1\u20135 using MobileNet. Simulation result revealed an accuracy of 87.08% and 88.75% for tomato early blight and tomato bacterial spot, respectively.","doi":"10.1007\/s40031-024-00993-5","cleaned_title":"classification of plant leaf disease using deep learning","cleaned_abstract":"agriculture is critical in human existence. practically, 60% of the population is occupied with some sort of farming, either straight forwardly or in a roundabout way. the important factor which determines the quality of crops grown in fields is early detection, classification and severity of plant leaf diseases. traditional plant disease detection was based on manual design of features and classifiers. the traditional methods faced many challenges in real complex environment such as low contrast, noise in the image being captured. in recent years, deep learning model widely used in image classification task is convolutional neural network, and hence, plant disease detection based on deep learning has important academic research value. in this paper, a cnn model with different convolution layers, alexnet and mobilenet, was designed to classify the plant leaf diseases of pepper bell, tomato, potato, rice, apple and sorghum with a total of 26 classes. the dataset for pepper bell, tomato, potato, rice and apple is taken from kaggle, whereas sorghum dataset is generated by taking pictures from the field. along with identification of disease, severity detection is performed on two selected diseases of tomato crop using the best model. using an available dataset of 24,156 images of diseased and healthy plant leaves, the performance of the model was evaluated. simulation results revealed an accuracy of 84.24% with cnn model consisting of five layers and an accuracy of 91.19% with pretrained alexnet model and an accuracy of 97.33% with mobilenet with a learning rate of 0.001. severity detection was performed on tomato early blight and tomato bacterial spot with levels 1\u20135 using mobilenet. simulation result revealed an accuracy of 87.08% and 88.75% for tomato early blight and tomato bacterial spot, respectively.","key_phrases":["agriculture","human existence","60%","the population","some sort","farming","a roundabout way","the important factor","which","the quality","crops","fields","early detection","classification","severity","plant leaf diseases","traditional plant disease detection","manual design","features","classifiers","the traditional methods","many challenges","real complex environment","low contrast","the image","recent years","deep learning model","image classification task","convolutional neural network","plant disease detection","deep learning","important academic research value","this paper","a cnn model","different convolution layers","alexnet","mobilenet","the plant leaf diseases","pepper bell","tomato","potato","rice","apple","sorghum","a total","26 classes","the dataset","pepper bell","kaggle","sorghum dataset","pictures","the field","identification","disease","severity detection","two selected diseases","tomato crop","the best model","an available dataset","24,156 images","diseased and healthy plant leaves","the performance","the model","simulation results","an accuracy","84.24%","cnn model","five layers","an accuracy","91.19%","pretrained alexnet model","an accuracy","97.33%","mobilenet","a learning rate","severity detection","tomato early blight","tomato bacterial spot","levels","mobilenet","simulation result","an accuracy","87.08%","88.75%","tomato early blight","tomato bacterial spot","60%","recent years","cnn","tomato, potato","26","tomato, potato","kaggle","two","24,156","84.24%","cnn","five","91.19%","97.33%","0.001","1\u20135","87.08% and","88.75%"]},{"title":"DeepDOF-SE: affordable deep-learning microscopy platform for slide-free histology","authors":["Lingbo Jin","Yubo Tang","Jackson B. Coole","Melody T. Tan","Xuan Zhao","Hawraa Badaoui","Jacob T. Robinson","Michelle D. Williams","Nadarajah Vigneswaran","Ann M. Gillenwater","Rebecca R. Richards-Kortum","Ashok Veeraraghavan"],"abstract":"Histopathology plays a critical role in the diagnosis and surgical management of cancer. However, access to histopathology services, especially frozen section pathology during surgery, is limited in resource-constrained settings because preparing slides from resected tissue is time-consuming, labor-intensive, and requires expensive infrastructure. Here, we report a deep-learning-enabled microscope, named DeepDOF-SE, to rapidly scan intact tissue at cellular resolution without the need for physical sectioning. Three key features jointly make DeepDOF-SE practical. First, tissue specimens are stained directly with inexpensive vital fluorescent dyes and optically sectioned with ultra-violet excitation that localizes fluorescent emission to a thin surface layer. Second, a deep-learning algorithm extends the depth-of-field, allowing rapid acquisition of in-focus images from large areas of tissue even when the tissue surface is highly irregular. Finally, a semi-supervised generative adversarial network virtually stains DeepDOF-SE fluorescence images with hematoxylin-and-eosin appearance, facilitating image interpretation by pathologists without significant additional training. We developed the DeepDOF-SE platform using a data-driven approach and validated its performance by imaging surgical resections of suspected oral tumors. Our results show that DeepDOF-SE provides histological information of diagnostic importance, offering a rapid and affordable slide-free histology platform for intraoperative tumor margin assessment and in low-resource settings.","doi":"10.1038\/s41467-024-47065-2","cleaned_title":"deepdof-se: affordable deep-learning microscopy platform for slide-free histology","cleaned_abstract":"histopathology plays a critical role in the diagnosis and surgical management of cancer. however, access to histopathology services, especially frozen section pathology during surgery, is limited in resource-constrained settings because preparing slides from resected tissue is time-consuming, labor-intensive, and requires expensive infrastructure. here, we report a deep-learning-enabled microscope, named deepdof-se, to rapidly scan intact tissue at cellular resolution without the need for physical sectioning. three key features jointly make deepdof-se practical. first, tissue specimens are stained directly with inexpensive vital fluorescent dyes and optically sectioned with ultra-violet excitation that localizes fluorescent emission to a thin surface layer. second, a deep-learning algorithm extends the depth-of-field, allowing rapid acquisition of in-focus images from large areas of tissue even when the tissue surface is highly irregular. finally, a semi-supervised generative adversarial network virtually stains deepdof-se fluorescence images with hematoxylin-and-eosin appearance, facilitating image interpretation by pathologists without significant additional training. we developed the deepdof-se platform using a data-driven approach and validated its performance by imaging surgical resections of suspected oral tumors. our results show that deepdof-se provides histological information of diagnostic importance, offering a rapid and affordable slide-free histology platform for intraoperative tumor margin assessment and in low-resource settings.","key_phrases":["histopathology","a critical role","the diagnosis","surgical management","cancer","access","histopathology services","especially frozen section pathology","surgery","resource-constrained settings","slides","resected tissue","expensive infrastructure","we","a deep-learning-enabled microscope","deepdof-se","intact tissue","cellular resolution","the need","physical sectioning","three key features","deepdof-se","tissue specimens","inexpensive vital fluorescent dyes","ultra-violet excitation","that","fluorescent emission","a thin surface layer","a deep-learning algorithm","the depth","field","rapid acquisition","focus","large areas","tissue","the tissue surface","a semi-supervised generative adversarial network","deepdof-se fluorescence images","hematoxylin-and-eosin appearance","image interpretation","pathologists","significant additional training","we","the deepdof-se platform","a data-driven approach","its performance","surgical resections","suspected oral tumors","our results","deepdof-se","histological information","diagnostic importance","a rapid and affordable slide-free histology platform","intraoperative tumor margin assessment","low-resource settings","three","first","second","hematoxylin"]},{"title":"Enhancing deep learning classification performance of tongue lesions in imbalanced data: mosaic-based soft labeling with curriculum learning","authors":["Sung-Jae Lee","Hyun Jun Oh","Young-Don Son","Jong-Hoon Kim","Ik-Jae Kwon","Bongju Kim","Jong-Ho Lee","Hang-Keun Kim"],"abstract":"BackgroundOral potentially malignant disorders (OPMDs) are associated with an increased risk of cancer of the oral cavity including the tongue. The early detection of oral cavity cancers and OPMDs is critical for reducing cancer-specific morbidity and mortality. Recently, there have been studies to apply the rapidly advancing technology of deep learning for diagnosing oral cavity cancer and OPMDs. However, several challenging issues such as class imbalance must be resolved to effectively train a deep learning model for medical imaging classification tasks. The aim of this study is to evaluate a new technique of artificial intelligence to improve the classification performance in an imbalanced tongue lesion dataset.MethodsA total of 1,810 tongue images were used for the classification. The class-imbalanced dataset consisted of 372 instances of cancer, 141 instances of OPMDs, and 1,297 instances of noncancerous lesions. The EfficientNet model was used as the feature extraction model for classification. Mosaic data augmentation, soft labeling, and curriculum learning (CL) were employed to improve the classification performance of the convolutional neural network.ResultsUtilizing a mosaic-augmented dataset in conjunction with CL, the final model achieved an accuracy rate of 0.9444, surpassing conventional oversampling and weight balancing methods. The relative precision improvement rate for the minority class OPMD was 21.2%, while the relative \\({F}_{1}\\) score improvement rate of OPMD was 4.9%.ConclusionsThe present study demonstrates that the integration of mosaic-based soft labeling and curriculum learning improves the classification performance of tongue lesions compared to previous methods, establishing a foundation for future research on effectively learning from imbalanced data.","doi":"10.1186\/s12903-024-03898-3","cleaned_title":"enhancing deep learning classification performance of tongue lesions in imbalanced data: mosaic-based soft labeling with curriculum learning","cleaned_abstract":"backgroundoral potentially malignant disorders (opmds) are associated with an increased risk of cancer of the oral cavity including the tongue. the early detection of oral cavity cancers and opmds is critical for reducing cancer-specific morbidity and mortality. recently, there have been studies to apply the rapidly advancing technology of deep learning for diagnosing oral cavity cancer and opmds. however, several challenging issues such as class imbalance must be resolved to effectively train a deep learning model for medical imaging classification tasks. the aim of this study is to evaluate a new technique of artificial intelligence to improve the classification performance in an imbalanced tongue lesion dataset.methodsa total of 1,810 tongue images were used for the classification. the class-imbalanced dataset consisted of 372 instances of cancer, 141 instances of opmds, and 1,297 instances of noncancerous lesions. the efficientnet model was used as the feature extraction model for classification. mosaic data augmentation, soft labeling, and curriculum learning (cl) were employed to improve the classification performance of the convolutional neural network.resultsutilizing a mosaic-augmented dataset in conjunction with cl, the final model achieved an accuracy rate of 0.9444, surpassing conventional oversampling and weight balancing methods. the relative precision improvement rate for the minority class opmd was 21.2%, while the relative \\({f}_{1}\\) score improvement rate of opmd was 4.9%.conclusionsthe present study demonstrates that the integration of mosaic-based soft labeling and curriculum learning improves the classification performance of tongue lesions compared to previous methods, establishing a foundation for future research on effectively learning from imbalanced data.","key_phrases":["backgroundoral potentially malignant disorders","an increased risk","cancer","the oral cavity","the tongue","the early detection","oral cavity cancers","opmds","cancer-specific morbidity","mortality","studies","the rapidly advancing technology","deep learning","oral cavity cancer","opmds","several challenging issues","class imbalance","a deep learning model","medical imaging classification tasks","the aim","this study","a new technique","artificial intelligence","the classification performance","an imbalanced tongue lesion","dataset.methodsa total","1,810 tongue images","the classification","the class-imbalanced dataset","372 instances","cancer","141 instances","opmds","1,297 instances","noncancerous lesions","the efficientnet model","the feature extraction model","classification","mosaic data augmentation","soft labeling","curriculum learning","cl","the classification performance","the convolutional neural","a mosaic-augmented dataset","conjunction","cl","the final model","an accuracy rate","conventional oversampling","weight balancing methods","the relative precision improvement rate","the minority class","21.2%","the relative \\({f}_{1}\\) score improvement rate","4.9%.conclusionsthe present study","the integration","mosaic-based soft labeling","curriculum learning","the classification performance","tongue lesions","previous methods","a foundation","future research","imbalanced data","dataset.methodsa","1,810","372","141","1,297","mosaic data","mosaic","0.9444","21.2%"]},{"title":"A deep learning-based car accident detection approach in video-based traffic surveillance","authors":["Xinyu Wu","Tingting Li"],"abstract":"Car accident detection plays a crucial role in video-based traffic surveillance systems, contributing to prompt response and improved road safety. In the literature, various methods have been investigated for accident detection, among which deep learning approaches have shown superior accuracy compared to other methods. The popularity of deep learning stems from its ability to automatically learn complex features from data. However, the current research challenge in deep learning-based accident detection lies in achieving high accuracy rates while meeting real-time requirements. To address this challenge, this study introduces a deep learning approach using convolutional neural networks (CNNs) to enhance car accident detection, prioritizing accuracy and real-time performance. It includes a tailored dataset for evaluation, and the F1-scores reveal reasonably accurate detection for \u201cdamaged-rear-window\u201d (62%) and \u201cdamaged-window\u201d (63%), while \u201cdamaged-windscreen\u201d exhibits exceptional performance at 83%. These results demonstrate the potential of CNNs in improving car accident detection, particularly for certain classes. Following extensive experiments and performance analysis, the proposed method demonstrates accurate results, significantly enhancing car accident detection in video-based traffic surveillance scenarios.","doi":"10.1007\/s12596-023-01581-4","cleaned_title":"a deep learning-based car accident detection approach in video-based traffic surveillance","cleaned_abstract":"car accident detection plays a crucial role in video-based traffic surveillance systems, contributing to prompt response and improved road safety. in the literature, various methods have been investigated for accident detection, among which deep learning approaches have shown superior accuracy compared to other methods. the popularity of deep learning stems from its ability to automatically learn complex features from data. however, the current research challenge in deep learning-based accident detection lies in achieving high accuracy rates while meeting real-time requirements. to address this challenge, this study introduces a deep learning approach using convolutional neural networks (cnns) to enhance car accident detection, prioritizing accuracy and real-time performance. it includes a tailored dataset for evaluation, and the f1-scores reveal reasonably accurate detection for \u201cdamaged-rear-window\u201d (62%) and \u201cdamaged-window\u201d (63%), while \u201cdamaged-windscreen\u201d exhibits exceptional performance at 83%. these results demonstrate the potential of cnns in improving car accident detection, particularly for certain classes. following extensive experiments and performance analysis, the proposed method demonstrates accurate results, significantly enhancing car accident detection in video-based traffic surveillance scenarios.","key_phrases":["car accident detection","a crucial role","video-based traffic surveillance systems","response","the literature","various methods","accident detection","which","deep learning approaches","superior accuracy","other methods","the popularity","deep learning","its ability","complex features","data","the current research challenge","deep learning-based accident detection","high accuracy rates","real-time requirements","this challenge","this study","a deep learning approach","convolutional neural networks","cnns","car accident detection","accuracy","real-time performance","it","a tailored dataset","evaluation","the f1-scores","reasonably accurate detection","\u201cdamaged-rear-window","62%","\u201cdamaged-window","63%","exceptional performance","83%","these results","the potential","cnns","car accident detection","certain classes","extensive experiments","performance analysis","the proposed method","accurate results","car accident detection","video-based traffic surveillance scenarios","62%","63%","83%"]},{"title":"Deep learning model for pleural effusion detection via active learning and pseudo-labeling: a multisite study","authors":["Joseph Chang","Bo-Ru Lin","Ti-Hao Wang","Chung-Ming Chen"],"abstract":"BackgroundThe study aimed to develop and validate a deep learning-based Computer Aided Triage (CADt) algorithm for detecting pleural effusion in chest radiographs using an active learning (AL) framework. This is aimed at addressing the critical need for a clinical grade algorithm that can timely diagnose pleural effusion, which affects approximately 1.5\u00a0million people annually in the United States.MethodsIn this multisite study, 10,599 chest radiographs from 2006 to 2018 were retrospectively collected from an institution in Taiwan to train the deep learning algorithm. The AL framework utilized significantly reduced the need for expert annotations. For external validation, the algorithm was tested on a multisite dataset of 600 chest radiographs from 22 clinical sites in the United States and Taiwan, which were annotated by three U.S. board-certified radiologists.ResultsThe CADt algorithm demonstrated high effectiveness in identifying pleural effusion, achieving a sensitivity of 0.95 (95% CI: [0.92, 0.97]) and a specificity of 0.97 (95% CI: [0.95, 0.99]). The area under the receiver operating characteristic curve (AUC) was 0.97 (95% DeLong\u2019s CI: [0.95, 0.99]). Subgroup analyses showed that the algorithm maintained robust performance across various demographics and clinical settings.ConclusionThis study presents a novel approach in developing clinical grade CADt solutions for the diagnosis of pleural effusion. The AL-based CADt algorithm not only achieved high accuracy in detecting pleural effusion but also significantly reduced the workload required for clinical experts in annotating medical data. This method enhances the feasibility of employing advanced technological solutions for prompt and accurate diagnosis in medical settings.","doi":"10.1186\/s12880-024-01260-1","cleaned_title":"deep learning model for pleural effusion detection via active learning and pseudo-labeling: a multisite study","cleaned_abstract":"backgroundthe study aimed to develop and validate a deep learning-based computer aided triage (cadt) algorithm for detecting pleural effusion in chest radiographs using an active learning (al) framework. this is aimed at addressing the critical need for a clinical grade algorithm that can timely diagnose pleural effusion, which affects approximately 1.5 million people annually in the united states.methodsin this multisite study, 10,599 chest radiographs from 2006 to 2018 were retrospectively collected from an institution in taiwan to train the deep learning algorithm. the al framework utilized significantly reduced the need for expert annotations. for external validation, the algorithm was tested on a multisite dataset of 600 chest radiographs from 22 clinical sites in the united states and taiwan, which were annotated by three u.s. board-certified radiologists.resultsthe cadt algorithm demonstrated high effectiveness in identifying pleural effusion, achieving a sensitivity of 0.95 (95% ci: [0.92, 0.97]) and a specificity of 0.97 (95% ci: [0.95, 0.99]). the area under the receiver operating characteristic curve (auc) was 0.97 (95% delong\u2019s ci: [0.95, 0.99]). subgroup analyses showed that the algorithm maintained robust performance across various demographics and clinical settings.conclusionthis study presents a novel approach in developing clinical grade cadt solutions for the diagnosis of pleural effusion. the al-based cadt algorithm not only achieved high accuracy in detecting pleural effusion but also significantly reduced the workload required for clinical experts in annotating medical data. this method enhances the feasibility of employing advanced technological solutions for prompt and accurate diagnosis in medical settings.","key_phrases":["backgroundthe study","a deep learning-based computer aided triage","cadt","algorithm","pleural effusion","chest","an active learning","(al) framework","this","the critical need","a clinical grade","algorithm","that","pleural effusion","which","approximately 1.5 million people","the united states.methodsin","this multisite study","10,599 chest","an institution","taiwan","the deep learning algorithm","the al framework","the need","expert annotations","external validation","the algorithm","a multisite dataset","600 chest","22 clinical sites","the united states","taiwan","which","radiologists.resultsthe cadt algorithm","high effectiveness","pleural effusion","a sensitivity","95% ci","a specificity","(95% ci","the area","the receiver operating characteristic curve","auc","(95% delong","subgroup analyses","the algorithm","robust performance","various demographics","clinical settings.conclusionthis study","a novel approach","clinical grade cadt solutions","the diagnosis","pleural effusion","the al-based cadt","algorithm","high accuracy","pleural effusion","the workload","clinical experts","medical data","this method","the feasibility","advanced technological solutions","prompt and accurate diagnosis","medical settings","approximately 1.5 million","annually","the united states.methodsin","10,599","2006","2018","taiwan","al","600","22","the united states","taiwan","three","u.s.","0.95","95%","0.92","0.97","0.97","95%","0.95","0.99","0.97","95%","0.95","0.99","al"]},{"title":"Deep Learning-Based Empirical and Sub-Space Decomposition for Speech Enhancement","authors":["Khaoula Mraihi","Mohamed Anouar Ben Messaoud"],"abstract":"This research presents a single-channel speech enhancement approach based on the combination of the adaptive empirical wavelet transform and the improved sub-space decomposition method followed by a deep learning network. The adaptive empirical wavelet transform is used to determine the boundaries of the segments, then we decompose the obtained spectrogram of the noisy speech into three sub-spaces to determine the low-rank matrix and the sparse matrix of the spectrogram under the perturbation of the residual matrix. The residual noise affecting the speech quality is avoided by the low-rank decomposition using the nonnegative factorization. Then, a cross-domain learning framework is developed to specify the correlations along the frequency and time axes and avoid the disadvantages of the time\u2013frequency domain. Experimental results show that the proposed approach outperforms several competing speech enhancement methods and achieves the highest PESQ, Cov and STOI under different types of noise and at low SNR values in the two datasets. The proposed model is tested on a hardware-level manual design to accelerate the execution of the developed deep learning model on an FPGA.","doi":"10.1007\/s00034-024-02606-4","cleaned_title":"deep learning-based empirical and sub-space decomposition for speech enhancement","cleaned_abstract":"this research presents a single-channel speech enhancement approach based on the combination of the adaptive empirical wavelet transform and the improved sub-space decomposition method followed by a deep learning network. the adaptive empirical wavelet transform is used to determine the boundaries of the segments, then we decompose the obtained spectrogram of the noisy speech into three sub-spaces to determine the low-rank matrix and the sparse matrix of the spectrogram under the perturbation of the residual matrix. the residual noise affecting the speech quality is avoided by the low-rank decomposition using the nonnegative factorization. then, a cross-domain learning framework is developed to specify the correlations along the frequency and time axes and avoid the disadvantages of the time\u2013frequency domain. experimental results show that the proposed approach outperforms several competing speech enhancement methods and achieves the highest pesq, cov and stoi under different types of noise and at low snr values in the two datasets. the proposed model is tested on a hardware-level manual design to accelerate the execution of the developed deep learning model on an fpga.","key_phrases":["this research","a single-channel speech enhancement approach","the combination","the adaptive empirical wavelet transform","the improved sub-space decomposition method","a deep learning network","the adaptive empirical wavelet transform","the boundaries","the segments","we","the obtained spectrogram","the noisy speech","three sub","-","spaces","the low-rank matrix","the sparse matrix","the spectrogram","the perturbation","the residual matrix","the residual noise","the speech quality","the low-rank decomposition","the nonnegative factorization","a cross-domain learning framework","the correlations","the frequency and time axes","the disadvantages","the time","\u2013frequency domain","experimental results","the proposed approach","several competing speech enhancement methods","the highest pesq","cov","stoi","different types","noise","low snr values","the two datasets","the proposed model","a hardware-level manual design","the execution","the developed deep learning model","an fpga","three","two"]},{"title":"Distributed source DOA estimation based on deep learning networks","authors":["Quan Tian","Ruiyan Cai","Gongrun Qiu","Yang Luo"],"abstract":"With space electromagnetic environments becoming increasingly complex, the direction of arrival (DOA) estimation based on the point source model can no longer meet the requirements of spatial target location. Based on the characteristics of the distributed source, a new DOA estimation algorithm based on deep learning is proposed. The algorithm first maps the distributed source model into the point source model via a generative adversarial network (GAN) and further combines the subspace-based method to achieve central DOA estimation. Second, by constructing a deep neural network (DNN), the covariance matrix of the received signals is used as the input to estimate the angular spread of the distributed source. The experimental results show that the proposed algorithm can achieve better performance than the existing methods for a distributed source.","doi":"10.1007\/s11760-024-03402-y","cleaned_title":"distributed source doa estimation based on deep learning networks","cleaned_abstract":"with space electromagnetic environments becoming increasingly complex, the direction of arrival (doa) estimation based on the point source model can no longer meet the requirements of spatial target location. based on the characteristics of the distributed source, a new doa estimation algorithm based on deep learning is proposed. the algorithm first maps the distributed source model into the point source model via a generative adversarial network (gan) and further combines the subspace-based method to achieve central doa estimation. second, by constructing a deep neural network (dnn), the covariance matrix of the received signals is used as the input to estimate the angular spread of the distributed source. the experimental results show that the proposed algorithm can achieve better performance than the existing methods for a distributed source.","key_phrases":["space electromagnetic environments","the direction","arrival","(doa) estimation","the point source model","the requirements","spatial target location","the characteristics","the distributed source","a new doa estimation algorithm","deep learning","algorithm","the distributed source model","the point source model","a generative adversarial network","gan","the subspace-based method","central doa estimation","a deep neural network","dnn","the covariance matrix","the received signals","the input","the angular spread","the distributed source","the experimental results","the proposed algorithm","better performance","the existing methods","a distributed source","first","second"]},{"title":"Car crash detection using ensemble deep learning","authors":["Vani Suthamathi Saravanarajan","Rung-Ching Chen","Christine Dewi","Long-Sheng Chen","Lata Ganesan"],"abstract":"With the recent advancements in Autonomous Vehicles (AVs), two important factors that play a vital role to avoid accidents and collisions are obstacles and track detection. AVs must implement an accident detection model to detect accident vehicles and avoid running into rollover vehicles. At present many trajectories-based and sensor-based multiple-vehicle accident prediction models exist. In Taiwan, the AV Tesla sedan's failure to detect overturned vehicles shows that an efficient deep learning model is still required to detect a single-car crash by taking appropriate actions like slowing down, tracking changes, and informing the concerned authorities. This paper proposes a novel car crash detection system for various car crashes using three deep learning models, namely VGG16(feature extractor using transfer learning), RPN (region proposal network), and CNN8L (region-based detector). The CNN8L is a novel lightweight sequential convolutional neural network for region-based classification and detection. The model is trained using a customized dataset, evaluated using different metrics and compared with various state-of-the-art models. The experimental results show that the VGG16 combined with the CNN8L model performed much better when compared to other models. The proposed system accurately recognizes car accidents with an Accident Detection Rate (ADR) of 86.25% and False Alarm Rate (FAR) of 33.00%.","doi":"10.1007\/s11042-023-15906-9","cleaned_title":"car crash detection using ensemble deep learning","cleaned_abstract":"with the recent advancements in autonomous vehicles (avs), two important factors that play a vital role to avoid accidents and collisions are obstacles and track detection. avs must implement an accident detection model to detect accident vehicles and avoid running into rollover vehicles. at present many trajectories-based and sensor-based multiple-vehicle accident prediction models exist. in taiwan, the av tesla sedan's failure to detect overturned vehicles shows that an efficient deep learning model is still required to detect a single-car crash by taking appropriate actions like slowing down, tracking changes, and informing the concerned authorities. this paper proposes a novel car crash detection system for various car crashes using three deep learning models, namely vgg16(feature extractor using transfer learning), rpn (region proposal network), and cnn8l (region-based detector). the cnn8l is a novel lightweight sequential convolutional neural network for region-based classification and detection. the model is trained using a customized dataset, evaluated using different metrics and compared with various state-of-the-art models. the experimental results show that the vgg16 combined with the cnn8l model performed much better when compared to other models. the proposed system accurately recognizes car accidents with an accident detection rate (adr) of 86.25% and false alarm rate (far) of 33.00%.","key_phrases":["the recent advancements","autonomous vehicles","two important factors","that","a vital role","accidents","collisions","obstacles","track detection","avs","an accident detection model","accident vehicles","rollover vehicles","present many trajectories-based and sensor-based multiple-vehicle accident prediction models","taiwan","the av tesla sedan's failure","overturned vehicles","an efficient deep learning model","a single-car crash","appropriate actions","changes","the concerned authorities","this paper","a novel car crash detection system","various car crashes","three deep learning models","namely vgg16(feature extractor","transfer learning","rpn","region proposal network","cnn8l","(region-based detector","the cnn8l","a novel lightweight sequential convolutional neural network","region-based classification","detection","the model","a customized dataset","different metrics","the-art","the experimental results","the vgg16","the cnn8l model","other models","the proposed system","car accidents","an accident detection rate","adr","86.25%","false alarm rate","33.00%","two","taiwan","three","86.25%","33.00%"]},{"title":"Meta-heuristic-based hybrid deep learning model for vulnerability detection and prevention in software system","authors":["Lijin Shaji","R. Suji Pramila"],"abstract":"Software vulnerabilities are flaws that may be exploited to cause loss or harm. Various automated machine-learning techniques have been developed in preceding studies to detect software vulnerabilities. This work tries to develop a technique for securing the software on the basis of their vulnerabilities that are already known, by developing a hybrid deep learning model to detect those vulnerabilities. Moreover, certain countermeasures are suggested based on the types of vulnerability to prevent the attack further. For different software projects taken as the dataset, feature fusion is done by utilizing canonical correlation analysis together with Deep Residual Network (DRN). A hybrid deep learning technique trained using AdamW-Rat Swarm Optimizer (AdamW-RSO) is designed to detect software vulnerability. Hybrid deep learning makes use of the Deep Belief Network (DBN) and Generative Adversarial Network (GAN). For every vulnerability, its location of occurrence within the software development procedures and techniques of alleviation via implementation level or design level activities are described. Thus, it helps in understanding the appearance of vulnerabilities, suggesting the use of various countermeasures during the initial phases of software design, and therefore, assures software security. Evaluating the performance of vulnerability detection by the proposed technique regarding recall, precision, and f-measure, it is found to be more effective than the existing methods.","doi":"10.1007\/s10878-024-01185-z","cleaned_title":"meta-heuristic-based hybrid deep learning model for vulnerability detection and prevention in software system","cleaned_abstract":"software vulnerabilities are flaws that may be exploited to cause loss or harm. various automated machine-learning techniques have been developed in preceding studies to detect software vulnerabilities. this work tries to develop a technique for securing the software on the basis of their vulnerabilities that are already known, by developing a hybrid deep learning model to detect those vulnerabilities. moreover, certain countermeasures are suggested based on the types of vulnerability to prevent the attack further. for different software projects taken as the dataset, feature fusion is done by utilizing canonical correlation analysis together with deep residual network (drn). a hybrid deep learning technique trained using adamw-rat swarm optimizer (adamw-rso) is designed to detect software vulnerability. hybrid deep learning makes use of the deep belief network (dbn) and generative adversarial network (gan). for every vulnerability, its location of occurrence within the software development procedures and techniques of alleviation via implementation level or design level activities are described. thus, it helps in understanding the appearance of vulnerabilities, suggesting the use of various countermeasures during the initial phases of software design, and therefore, assures software security. evaluating the performance of vulnerability detection by the proposed technique regarding recall, precision, and f-measure, it is found to be more effective than the existing methods.","key_phrases":["software vulnerabilities","flaws","that","loss","harm","various automated machine-learning techniques","studies","software vulnerabilities","this work","a technique","the software","the basis","their vulnerabilities","that","a hybrid deep learning model","those vulnerabilities","certain countermeasures","the types","vulnerability","the attack","different software projects","the dataset","feature fusion","canonical correlation analysis","deep residual network","drn","a hybrid deep learning technique","adamw-rat swarm optimizer","adamw-rso","software vulnerability","hybrid deep learning","use","the deep belief network","dbn","adversarial network","gan","every vulnerability","its location","occurrence","the software development procedures","techniques","alleviation","implementation level or design level activities","it","the appearance","vulnerabilities","the use","various countermeasures","the initial phases","software design","therefore, assures software security","the performance","vulnerability detection","the proposed technique","recall","precision","f-measure","it","the existing methods"]},{"title":"Deep learning implementations in mining applications: a compact critical review","authors":["Faris Azhari","Charlotte C. Sennersten","Craig A. Lindley","Ewan Sellers"],"abstract":"Deep learning is a sub-field of artificial intelligence that combines feature engineering and classification in one method. It is a data-driven technique that optimises a predictive model via learning from a large dataset. Digitisation in industry has included acquisition and storage of a variety of large datasets for interpretation and decision making. This has led to the adoption of deep learning in different industries, such as transportation, manufacturing, medicine and agriculture. However, in the mining industry, the adoption and development of new technologies, including deep learning methods, has not progressed at the same rate as in other industries. Nevertheless, in the past 5 years, applications of deep learning have been increasing in the mining research space. Deep learning has been implemented to solve a variety of problems related to mine exploration, ore and metal extraction and reclamation processes. The increased automation adoption in mining provides an avenue for wider application of deep learning as an element within a mine automation framework. This\u00a0work provides a compact, comprehensive review of deep learning implementations in mining-related applications. The trends of these implementations in terms of years, venues, deep learning network types, tasks and general implementation, categorised by the value chain operations of exploration, extraction and reclamation are outlined. The review enables shortcomings regarding progress within the research context to be highlighted such as the proprietary nature of data, small datasets (tens to thousands of data points) limited to single operations with unique geology, mine design and equipment, lack of large scale publicly available mining related datasets and limited sensor types leading to the majority of applications being image-based analysis. Gaps identified for future research and application includes the usage of a wider range of sensor data, improved understanding of the outputs by mining practitioners, adversarial testing of the deep learning models, development of public datasets covering the extensive range of conditions experienced in mines.","doi":"10.1007\/s10462-023-10500-9","cleaned_title":"deep learning implementations in mining applications: a compact critical review","cleaned_abstract":"deep learning is a sub-field of artificial intelligence that combines feature engineering and classification in one method. it is a data-driven technique that optimises a predictive model via learning from a large dataset. digitisation in industry has included acquisition and storage of a variety of large datasets for interpretation and decision making. this has led to the adoption of deep learning in different industries, such as transportation, manufacturing, medicine and agriculture. however, in the mining industry, the adoption and development of new technologies, including deep learning methods, has not progressed at the same rate as in other industries. nevertheless, in the past 5 years, applications of deep learning have been increasing in the mining research space. deep learning has been implemented to solve a variety of problems related to mine exploration, ore and metal extraction and reclamation processes. the increased automation adoption in mining provides an avenue for wider application of deep learning as an element within a mine automation framework. this work provides a compact, comprehensive review of deep learning implementations in mining-related applications. the trends of these implementations in terms of years, venues, deep learning network types, tasks and general implementation, categorised by the value chain operations of exploration, extraction and reclamation are outlined. the review enables shortcomings regarding progress within the research context to be highlighted such as the proprietary nature of data, small datasets (tens to thousands of data points) limited to single operations with unique geology, mine design and equipment, lack of large scale publicly available mining related datasets and limited sensor types leading to the majority of applications being image-based analysis. gaps identified for future research and application includes the usage of a wider range of sensor data, improved understanding of the outputs by mining practitioners, adversarial testing of the deep learning models, development of public datasets covering the extensive range of conditions experienced in mines.","key_phrases":["deep learning","a sub","-","field","artificial intelligence","that","feature engineering","classification","one method","it","a data-driven technique","that","a predictive model","a large dataset","digitisation","industry","acquisition","storage","a variety","large datasets","interpretation","decision making","this","the adoption","deep learning","different industries","transportation","manufacturing","medicine","agriculture","the mining industry","new technologies","deep learning methods","the same rate","other industries","the past 5 years","applications","deep learning","the mining research space","deep learning","a variety","problems","mine exploration, ore and metal extraction and reclamation processes","the increased automation adoption","mining","an avenue","wider application","deep learning","an element","a mine automation framework","this work","a compact, comprehensive review","deep learning implementations","mining-related applications","the trends","these implementations","terms","years","venues","deep learning network types","tasks","general implementation","the value chain operations","exploration","extraction","reclamation","the review","shortcomings","progress","the research context","the proprietary nature","data","small datasets","tens to thousands","data points","single operations","unique geology","mine design","equipment","large scale publicly available mining related datasets","limited sensor types","the majority","applications","image-based analysis","gaps","future research","application","the usage","a wider range","sensor data","understanding","the outputs","mining practitioners","adversarial testing","the deep learning models","development","public datasets","the extensive range","conditions","mines","one","the past 5 years","tens to thousands"]},{"title":"Sentiment analysis of Canadian maritime case law: a sentiment case law and deep learning approach","authors":["Bola Abimbola","Qing Tan","Enrique A. De La Cal Mar\u00edn"],"abstract":"Historical information in the Canadian Maritime Judiciary increases with time because of the need to archive data to be utilized in case references and for later application when determining verdicts for similar cases. However, such data are typically stored in multiple systems, making its reachability technical. Utilizing technologies like deep learning and sentiment analysis provides chances to facilitate faster access to court records. Such practice enhances impartial verdicts, minimizing workloads for court employees, and decreases the time used in legal proceedings for claims during maritime contracts such as shipping disputes between parties. This paper seeks to develop a sentiment analysis framework that uses deep learning, distributed learning, and machine learning to improve access to statutes, laws, and cases used by maritime judges in making judgments to back their claims. The suggested approach uses deep learning models, including convolutional neural networks (CNNs), deep neural networks, long short-term memory (LSTM), and recurrent neural networks. It extracts court records having crucial sentiments or statements for maritime court verdicts. The suggested approach has been used successfully during sentiment analysis by emphasizing feature selection from a legal repository. The LSTM\u2009+\u2009CNN model has shown promising results in obtaining sentiments and records from multiple devices and sufficiently proposing practical guidance to judicial personnel regarding the regulations applicable to various situations.","doi":"10.1007\/s41870-024-01820-2","cleaned_title":"sentiment analysis of canadian maritime case law: a sentiment case law and deep learning approach","cleaned_abstract":"historical information in the canadian maritime judiciary increases with time because of the need to archive data to be utilized in case references and for later application when determining verdicts for similar cases. however, such data are typically stored in multiple systems, making its reachability technical. utilizing technologies like deep learning and sentiment analysis provides chances to facilitate faster access to court records. such practice enhances impartial verdicts, minimizing workloads for court employees, and decreases the time used in legal proceedings for claims during maritime contracts such as shipping disputes between parties. this paper seeks to develop a sentiment analysis framework that uses deep learning, distributed learning, and machine learning to improve access to statutes, laws, and cases used by maritime judges in making judgments to back their claims. the suggested approach uses deep learning models, including convolutional neural networks (cnns), deep neural networks, long short-term memory (lstm), and recurrent neural networks. it extracts court records having crucial sentiments or statements for maritime court verdicts. the suggested approach has been used successfully during sentiment analysis by emphasizing feature selection from a legal repository. the lstm + cnn model has shown promising results in obtaining sentiments and records from multiple devices and sufficiently proposing practical guidance to judicial personnel regarding the regulations applicable to various situations.","key_phrases":["historical information","the canadian maritime judiciary","time","the need","data","case references","later application","verdicts","similar cases","such data","multiple systems","its reachability","technologies","deep learning","sentiment analysis","chances","faster access","court records","such practice","impartial verdicts","workloads","court employees","the time","legal proceedings","claims","maritime contracts","shipping disputes","parties","this paper","a sentiment analysis framework","that","deep learning","learning","access","statutes","laws","cases","maritime judges","judgments","their claims","the suggested approach","deep learning models","convolutional neural networks","cnns","deep neural networks","long short-term memory","lstm","neural networks","it","court records","crucial sentiments","statements","maritime court verdicts","the suggested approach","sentiment analysis","feature selection","a legal repository","the lstm + cnn model","promising results","sentiments","records","multiple devices","practical guidance","judicial personnel","the regulations","various situations","canadian"]},{"title":"Multi-step framework for glaucoma diagnosis in retinal fundus images using deep learning","authors":["Sanli Yi","Lingxiang Zhou"],"abstract":"Glaucoma is one of the most common causes of blindness in the world. Screening glaucoma from retinal fundus images based on deep learning is a common method at present. In the diagnosis of glaucoma based on deep learning, the blood vessels within the optic disc interfere with the diagnosis, and there is also some pathological information outside the optic disc in fundus images. Therefore, integrating the original fundus image with the vessel-removed optic disc image can improve diagnostic efficiency. In this paper, we propose a novel multi-step framework named MSGC-CNN that can better diagnose glaucoma. In the framework, (1) we combine glaucoma pathological knowledge with deep learning model, fuse the features of original fundus image and optic disc region in which the interference of blood vessel is specifically removed by U-Net, and make glaucoma diagnosis based on the fused features. (2) Aiming at the characteristics of glaucoma fundus images, such as small amount of data, high resolution, and rich feature information, we design a new feature extraction network RA-ResNet and combined it with transfer learning. In order to verify our method, we conduct binary classification experiments on three public datasets, Drishti-GS, RIM-ONE-R3, and ACRIMA, with accuracy of 92.01%, 93.75%, and 97.87%. The results demonstrate a significant improvement over earlier results.Graphical abstract","doi":"10.1007\/s11517-024-03172-2","cleaned_title":"multi-step framework for glaucoma diagnosis in retinal fundus images using deep learning","cleaned_abstract":"glaucoma is one of the most common causes of blindness in the world. screening glaucoma from retinal fundus images based on deep learning is a common method at present. in the diagnosis of glaucoma based on deep learning, the blood vessels within the optic disc interfere with the diagnosis, and there is also some pathological information outside the optic disc in fundus images. therefore, integrating the original fundus image with the vessel-removed optic disc image can improve diagnostic efficiency. in this paper, we propose a novel multi-step framework named msgc-cnn that can better diagnose glaucoma. in the framework, (1) we combine glaucoma pathological knowledge with deep learning model, fuse the features of original fundus image and optic disc region in which the interference of blood vessel is specifically removed by u-net, and make glaucoma diagnosis based on the fused features. (2) aiming at the characteristics of glaucoma fundus images, such as small amount of data, high resolution, and rich feature information, we design a new feature extraction network ra-resnet and combined it with transfer learning. in order to verify our method, we conduct binary classification experiments on three public datasets, drishti-gs, rim-one-r3, and acrima, with accuracy of 92.01%, 93.75%, and 97.87%. the results demonstrate a significant improvement over earlier results.graphical abstract","key_phrases":["glaucoma","the most common causes","blindness","the world","glaucoma","retinal fundus images","deep learning","a common method","present","the diagnosis","glaucoma","deep learning","the blood vessels","the optic disc","the diagnosis","some pathological information","the optic disc","fundus images","the original fundus image","the vessel-removed optic disc image","diagnostic efficiency","this paper","we","a novel multi-step framework","msgc-cnn","that","glaucoma","the framework","we","glaucoma pathological knowledge","deep learning model","the features","original fundus image","optic disc region","which","the interference","blood vessel","u","-","net","glaucoma diagnosis","the fused features","the characteristics","glaucoma fundus images","small amount","data","high resolution","rich feature information","we","a new feature extraction network","ra-resnet","it","transfer learning","order","our method","we","binary classification experiments","three public datasets","drishti-gs","rim-one-r3","acrima","accuracy","92.01%","93.75%","97.87%","the results","a significant improvement","glaucoma","msgc-cnn","1","2","three","one-r3","92.01%","93.75%","97.87%"]},{"title":"Early detection and prediction of Heart Disease using Wearable devices and Deep Learning algorithms","authors":["S. Sivasubramaniam","S. P. Balamurugan"],"abstract":"In this paper, we propose a multimodal deep learning algorithm that combines convolutional neural networks (CNNs) and long short-term memory (LSTM) networks for early detection and prediction of heart disease using data collected from wearable devices. This combined multi-model deep learning algorithm is used to detect the accurate precision and accuracy value. At first, we consider, ECG and PPG signals, which are collected from the dataset. Then, the features from ECG and PPG are extracted using CNN and the accelerometer features are extracted using the LSTM model. The combined features are then classified using hybrid CNN-LSTM network architecture. The algorithm is evaluated using a publicly available benchmark dataset. The model achieved an accuracy of 99.33% in detecting heart disease, outperforming several state-of-the-art deep learning models. In addition, the model can predict the likelihood of developing heart disease with a precision of 99.33%, providing an early warning system for at-risk patients. The results demonstrate the potential of a multimodal approach for early detection and prediction of heart disease using wearable devices and deep learning algorithms.","doi":"10.1007\/s11042-024-19127-6","cleaned_title":"early detection and prediction of heart disease using wearable devices and deep learning algorithms","cleaned_abstract":"in this paper, we propose a multimodal deep learning algorithm that combines convolutional neural networks (cnns) and long short-term memory (lstm) networks for early detection and prediction of heart disease using data collected from wearable devices. this combined multi-model deep learning algorithm is used to detect the accurate precision and accuracy value. at first, we consider, ecg and ppg signals, which are collected from the dataset. then, the features from ecg and ppg are extracted using cnn and the accelerometer features are extracted using the lstm model. the combined features are then classified using hybrid cnn-lstm network architecture. the algorithm is evaluated using a publicly available benchmark dataset. the model achieved an accuracy of 99.33% in detecting heart disease, outperforming several state-of-the-art deep learning models. in addition, the model can predict the likelihood of developing heart disease with a precision of 99.33%, providing an early warning system for at-risk patients. the results demonstrate the potential of a multimodal approach for early detection and prediction of heart disease using wearable devices and deep learning algorithms.","key_phrases":["this paper","we","a multimodal deep learning algorithm","that","cnns","lstm","early detection","prediction","heart disease","data","wearable devices","this combined multi-model deep learning algorithm","the accurate precision","accuracy value","we","ecg and ppg signals","which","the dataset","the features","ecg","ppg","cnn","the accelerometer features","the lstm model","the combined features","hybrid cnn-lstm network architecture","the algorithm","a publicly available benchmark dataset","the model","an accuracy","99.33%","heart disease","the-art","addition","the model","the likelihood","heart disease","a precision","99.33%","an early warning system","risk","the results","the potential","a multimodal approach","early detection","prediction","heart disease","wearable devices","deep learning algorithms","first","cnn","cnn","99.33%","99.33%"]},{"title":"A new content-aware image resizing based on R\u00e9nyi entropy and deep learning","authors":["Jila Ayubi","Mehdi Chehel Amirani","Morteza Valizadeh"],"abstract":"One of the most popular techniques for changing the purpose of an image or resizing a digital image with content awareness is the seam-carving method. The performance of image resizing algorithms based on seam machining shows that these algorithms are highly dependent on the extraction of importance map techniques and the detection of salient objects. So far, various algorithms have been proposed to extract the importance map. In this paper, a new method based on R\u00e9nyi entropy is proposed to extract the importance map. Also, a deep learning network has been used to detect salient objects. The simulator results showed that combining R\u00e9nyi\u2019s importance map with a deep network of salient object detection performed better than classical seam-carving and other extended seam-carving algorithms based on deep learning.","doi":"10.1007\/s00521-024-09517-0","cleaned_title":"a new content-aware image resizing based on r\u00e9nyi entropy and deep learning","cleaned_abstract":"one of the most popular techniques for changing the purpose of an image or resizing a digital image with content awareness is the seam-carving method. the performance of image resizing algorithms based on seam machining shows that these algorithms are highly dependent on the extraction of importance map techniques and the detection of salient objects. so far, various algorithms have been proposed to extract the importance map. in this paper, a new method based on r\u00e9nyi entropy is proposed to extract the importance map. also, a deep learning network has been used to detect salient objects. the simulator results showed that combining r\u00e9nyi\u2019s importance map with a deep network of salient object detection performed better than classical seam-carving and other extended seam-carving algorithms based on deep learning.","key_phrases":["the most popular techniques","the purpose","an image","a digital image","content awareness","the seam-carving method","the performance","image resizing algorithms","seam machining","these algorithms","the extraction","importance map techniques","the detection","salient objects","various algorithms","the importance map","this paper","a new method","r\u00e9nyi entropy","the importance map","a deep learning network","salient objects","the simulator results","r\u00e9nyi\u2019s importance map","a deep network","salient object detection","classical seam-carving","other extended seam-carving algorithms","deep learning","one"]},{"title":"Deep learning based active image steganalysis: a review","authors":["Punam Bedi","Anuradha Singhal","Veenu Bhasin"],"abstract":"Steganalysis plays a vital role in cybersecurity in today\u2019s digital era where exchange of malicious information can be done easily across web pages. Steganography techniques are used to hide data in an object where the existence of hidden information is also obscured. Steganalysis is the process for detection of steganography within an object and can be categorized as active and passive steganalysis. Passive steganalysis tries to classify a given object as a clean or modified object. Active steganalysis aims to extract more details about hidden contents such as length of embedded message, region of inserted message, key used for embedding, required by cybersecurity experts for comprehensive analysis. Images being a viable source of exchange of information in the era of internet, social media are the most susceptible source for such transmission. Many researchers have worked and developed techniques required to detect and alert about such counterfeit exchanges over the internet. Literature present in passive and active image steganalysis techniques, addresses these issues by detecting and unveiling details of such obscured communication respectively. This paper provides a systematic and comprehensive review of work done on active image steganalysis techniques using deep learning techniques. This review will be helpful to the new researchers to become aware and build a strong foundation of literature present in active image steganalysis using deep learning techniques. The paper also includes various steganographic algorithms, dataset and performance evaluation metrics used in literature. Open research challenges and possible future research directions are also discussed in the paper.","doi":"10.1007\/s13198-023-02203-9","cleaned_title":"deep learning based active image steganalysis: a review","cleaned_abstract":"steganalysis plays a vital role in cybersecurity in today\u2019s digital era where exchange of malicious information can be done easily across web pages. steganography techniques are used to hide data in an object where the existence of hidden information is also obscured. steganalysis is the process for detection of steganography within an object and can be categorized as active and passive steganalysis. passive steganalysis tries to classify a given object as a clean or modified object. active steganalysis aims to extract more details about hidden contents such as length of embedded message, region of inserted message, key used for embedding, required by cybersecurity experts for comprehensive analysis. images being a viable source of exchange of information in the era of internet, social media are the most susceptible source for such transmission. many researchers have worked and developed techniques required to detect and alert about such counterfeit exchanges over the internet. literature present in passive and active image steganalysis techniques, addresses these issues by detecting and unveiling details of such obscured communication respectively. this paper provides a systematic and comprehensive review of work done on active image steganalysis techniques using deep learning techniques. this review will be helpful to the new researchers to become aware and build a strong foundation of literature present in active image steganalysis using deep learning techniques. the paper also includes various steganographic algorithms, dataset and performance evaluation metrics used in literature. open research challenges and possible future research directions are also discussed in the paper.","key_phrases":["steganalysis","a vital role","cybersecurity","today\u2019s digital era","exchange","malicious information","web pages","steganography techniques","data","an object","the existence","hidden information","steganalysis","the process","detection","steganography","an object","active and passive steganalysis","passive steganalysis","a given object","a clean or modified object","active steganalysis","more details","hidden contents","length","embedded message","region","inserted message","cybersecurity experts","comprehensive analysis","images","a viable source","exchange","information","the era","internet","social media","the most susceptible source","such transmission","many researchers","techniques","such counterfeit exchanges","the internet","literature","passive and active image steganalysis techniques","these issues","details","such obscured communication","this paper","a systematic and comprehensive review","work","active image steganalysis techniques","deep learning techniques","this review","the new researchers","a strong foundation","literature","active image steganalysis","deep learning techniques","the paper","various steganographic algorithms","dataset","performance evaluation metrics","literature","open research challenges","possible future research directions","the paper","today"]},{"title":"A deep learning-based car accident detection approach in video-based traffic surveillance","authors":["Xinyu Wu","Tingting Li"],"abstract":"Car accident detection plays a crucial role in video-based traffic surveillance systems, contributing to prompt response and improved road safety. In the literature, various methods have been investigated for accident detection, among which deep learning approaches have shown superior accuracy compared to other methods. The popularity of deep learning stems from its ability to automatically learn complex features from data. However, the current research challenge in deep learning-based accident detection lies in achieving high accuracy rates while meeting real-time requirements. To address this challenge, this study introduces a deep learning approach using convolutional neural networks (CNNs) to enhance car accident detection, prioritizing accuracy and real-time performance. It includes a tailored dataset for evaluation, and the F1-scores reveal reasonably accurate detection for \u201cdamaged-rear-window\u201d (62%) and \u201cdamaged-window\u201d (63%), while \u201cdamaged-windscreen\u201d exhibits exceptional performance at 83%. These results demonstrate the potential of CNNs in improving car accident detection, particularly for certain classes. Following extensive experiments and performance analysis, the proposed method demonstrates accurate results, significantly enhancing car accident detection in video-based traffic surveillance scenarios.","doi":"10.1007\/s12596-023-01581-4","cleaned_title":"a deep learning-based car accident detection approach in video-based traffic surveillance","cleaned_abstract":"car accident detection plays a crucial role in video-based traffic surveillance systems, contributing to prompt response and improved road safety. in the literature, various methods have been investigated for accident detection, among which deep learning approaches have shown superior accuracy compared to other methods. the popularity of deep learning stems from its ability to automatically learn complex features from data. however, the current research challenge in deep learning-based accident detection lies in achieving high accuracy rates while meeting real-time requirements. to address this challenge, this study introduces a deep learning approach using convolutional neural networks (cnns) to enhance car accident detection, prioritizing accuracy and real-time performance. it includes a tailored dataset for evaluation, and the f1-scores reveal reasonably accurate detection for \u201cdamaged-rear-window\u201d (62%) and \u201cdamaged-window\u201d (63%), while \u201cdamaged-windscreen\u201d exhibits exceptional performance at 83%. these results demonstrate the potential of cnns in improving car accident detection, particularly for certain classes. following extensive experiments and performance analysis, the proposed method demonstrates accurate results, significantly enhancing car accident detection in video-based traffic surveillance scenarios.","key_phrases":["car accident detection","a crucial role","video-based traffic surveillance systems","response","the literature","various methods","accident detection","which","deep learning approaches","superior accuracy","other methods","the popularity","deep learning","its ability","complex features","data","the current research challenge","deep learning-based accident detection","high accuracy rates","real-time requirements","this challenge","this study","a deep learning approach","convolutional neural networks","cnns","car accident detection","accuracy","real-time performance","it","a tailored dataset","evaluation","the f1-scores","reasonably accurate detection","\u201cdamaged-rear-window","62%","\u201cdamaged-window","63%","exceptional performance","83%","these results","the potential","cnns","car accident detection","certain classes","extensive experiments","performance analysis","the proposed method","accurate results","car accident detection","video-based traffic surveillance scenarios","62%","63%","83%"]},{"title":"Scoring method of English composition integrating deep learning in higher vocational colleges","authors":["Shuo Feng","Lixia Yu","Fen Liu"],"abstract":"Along with the progress of natural language processing technology and deep learning, the subjectivity, slow feedback, and long grading time of traditional English essay grading have been addressed. Intelligent English automatic scoring has been widely concerned by scholars. Given the limitations of topic relevance feature extraction methods and traditional automatic grading methods for English compositions, a topic decision model is proposed to calculate the topic relevance score of the topic richness in English composition. Then, based on the Score of Relevance Based on Topic Richness (TRSR) calculation method, an intelligent English composition scoring method combining artificial feature extraction and deep learning is designed. From the findings, the Topic Decision (TD) model achieved the best effect only when it was iterated 80 times. The corresponding accuracy, recall and F1 value were 0.97, 0.93 and 0.95 respectively. The model training loss finally stabilized at 0.03. The Intelligent English Composition Grading Method Integrating Deep Learning (DLIECG) method has the best overall performance and the best performance on dataset P. To sum up, the intelligent English composition scoring method has better effectiveness and reliability.","doi":"10.1038\/s41598-024-57419-x","cleaned_title":"scoring method of english composition integrating deep learning in higher vocational colleges","cleaned_abstract":"along with the progress of natural language processing technology and deep learning, the subjectivity, slow feedback, and long grading time of traditional english essay grading have been addressed. intelligent english automatic scoring has been widely concerned by scholars. given the limitations of topic relevance feature extraction methods and traditional automatic grading methods for english compositions, a topic decision model is proposed to calculate the topic relevance score of the topic richness in english composition. then, based on the score of relevance based on topic richness (trsr) calculation method, an intelligent english composition scoring method combining artificial feature extraction and deep learning is designed. from the findings, the topic decision (td) model achieved the best effect only when it was iterated 80 times. the corresponding accuracy, recall and f1 value were 0.97, 0.93 and 0.95 respectively. the model training loss finally stabilized at 0.03. the intelligent english composition grading method integrating deep learning (dliecg) method has the best overall performance and the best performance on dataset p. to sum up, the intelligent english composition scoring method has better effectiveness and reliability.","key_phrases":["the progress","natural language processing technology","deep learning","the subjectivity","slow feedback","long grading time","intelligent english automatic scoring","scholars","the limitations","topic relevance feature extraction methods","traditional automatic grading methods","english compositions","a topic decision model","the topic relevance score","the topic richness","english composition","the score","relevance","topic richness (trsr) calculation method","an intelligent english composition scoring method","artificial feature extraction","deep learning","the findings","the topic decision","(td) model","the best effect","it","the corresponding accuracy","recall","f1 value","the model training loss","the intelligent english composition","method","deep learning (dliecg) method","the best overall performance","the best performance","dataset p.","the intelligent english composition scoring method","better effectiveness","reliability","english","english","english","english","english","80","0.97","0.93","0.95","0.03","english","english"]},{"title":"A deep learning framework for students' academic performance analysis","authors":["Sumati Pathak","Hiral Raja","Sumit Srivastava","Neelam Sahu","Rohit Raja","Amit Kumar Dewangan"],"abstract":"Students Performance (SP) analysis is regarded as one of the most important steps in the educational system for supporting students' academic success and the institutions' overall outcomes. Nevertheless, it is tremendously challenging due to the numerous details that many students have. Data Mining (DM) is the most widely used approach for SP prediction that extracts imperative information from a bigger raw data set. Even though there are various DM-centered performance prediction approaches, they all have low accuracy and high training time and don't produce the desired output. This paper proposes a hybrid deep learning framework using Deer Hunting Optimization based Deep Learning Neural Networks (DH-DLNN). A self-structured questionnaire covers all aspects of using information and communication technology, including increased access, knowledge building, learning, performance, motivation, classroom management and interaction, collaborative learning, and satisfaction. Data Cleaning and data conversion preprocess the dataset. The prediction of the student's level is then performed by extracting imperative features from the preprocessed data, followed by feature ranking using entropy calculations. The obtained entropy values are inputted into the DH-DLNN, which predicts the students' academic performance. Finally, the accuracy of the proposed system is evaluated using K-fold cross-validation. The experiment results revealed that DH-DLNN outperforms the other classification approaches with an accuracy of 96.33%.","doi":"10.1007\/s40012-023-00388-9","cleaned_title":"a deep learning framework for students' academic performance analysis","cleaned_abstract":"students performance (sp) analysis is regarded as one of the most important steps in the educational system for supporting students' academic success and the institutions' overall outcomes. nevertheless, it is tremendously challenging due to the numerous details that many students have. data mining (dm) is the most widely used approach for sp prediction that extracts imperative information from a bigger raw data set. even though there are various dm-centered performance prediction approaches, they all have low accuracy and high training time and don't produce the desired output. this paper proposes a hybrid deep learning framework using deer hunting optimization based deep learning neural networks (dh-dlnn). a self-structured questionnaire covers all aspects of using information and communication technology, including increased access, knowledge building, learning, performance, motivation, classroom management and interaction, collaborative learning, and satisfaction. data cleaning and data conversion preprocess the dataset. the prediction of the student's level is then performed by extracting imperative features from the preprocessed data, followed by feature ranking using entropy calculations. the obtained entropy values are inputted into the dh-dlnn, which predicts the students' academic performance. finally, the accuracy of the proposed system is evaluated using k-fold cross-validation. the experiment results revealed that dh-dlnn outperforms the other classification approaches with an accuracy of 96.33%.","key_phrases":["students","performance (sp) analysis","the most important steps","the educational system","students' academic success","the institutions' overall outcomes","it","the numerous details","many students","data mining","(dm","the most widely used approach","sp prediction","that","imperative information","a bigger raw data set","various dm-centered performance prediction approaches","they","all","low accuracy","high training time","the desired output","this paper","a hybrid deep learning framework","deer hunting optimization","neural networks","dh-dlnn","a self-structured questionnaire","all aspects","information and communication technology","increased access","knowledge building","learning","performance","motivation","classroom management","interaction","collaborative learning","satisfaction","data cleaning","data conversion","the dataset","the prediction","the student's level","imperative features","the preprocessed data","feature","entropy calculations","the obtained entropy values","the dh-dlnn","which","the students' academic performance","the accuracy","the proposed system","k","fold cross","-","validation","the experiment results","dh-dlnn","the other classification approaches","an accuracy","96.33%","one","96.33%"]},{"title":"Car crash detection using ensemble deep learning","authors":["Vani Suthamathi Saravanarajan","Rung-Ching Chen","Christine Dewi","Long-Sheng Chen","Lata Ganesan"],"abstract":"With the recent advancements in Autonomous Vehicles (AVs), two important factors that play a vital role to avoid accidents and collisions are obstacles and track detection. AVs must implement an accident detection model to detect accident vehicles and avoid running into rollover vehicles. At present many trajectories-based and sensor-based multiple-vehicle accident prediction models exist. In Taiwan, the AV Tesla sedan's failure to detect overturned vehicles shows that an efficient deep learning model is still required to detect a single-car crash by taking appropriate actions like slowing down, tracking changes, and informing the concerned authorities. This paper proposes a novel car crash detection system for various car crashes using three deep learning models, namely VGG16(feature extractor using transfer learning), RPN (region proposal network), and CNN8L (region-based detector). The CNN8L is a novel lightweight sequential convolutional neural network for region-based classification and detection. The model is trained using a customized dataset, evaluated using different metrics and compared with various state-of-the-art models. The experimental results show that the VGG16 combined with the CNN8L model performed much better when compared to other models. The proposed system accurately recognizes car accidents with an Accident Detection Rate (ADR) of 86.25% and False Alarm Rate (FAR) of 33.00%.","doi":"10.1007\/s11042-023-15906-9","cleaned_title":"car crash detection using ensemble deep learning","cleaned_abstract":"with the recent advancements in autonomous vehicles (avs), two important factors that play a vital role to avoid accidents and collisions are obstacles and track detection. avs must implement an accident detection model to detect accident vehicles and avoid running into rollover vehicles. at present many trajectories-based and sensor-based multiple-vehicle accident prediction models exist. in taiwan, the av tesla sedan's failure to detect overturned vehicles shows that an efficient deep learning model is still required to detect a single-car crash by taking appropriate actions like slowing down, tracking changes, and informing the concerned authorities. this paper proposes a novel car crash detection system for various car crashes using three deep learning models, namely vgg16(feature extractor using transfer learning), rpn (region proposal network), and cnn8l (region-based detector). the cnn8l is a novel lightweight sequential convolutional neural network for region-based classification and detection. the model is trained using a customized dataset, evaluated using different metrics and compared with various state-of-the-art models. the experimental results show that the vgg16 combined with the cnn8l model performed much better when compared to other models. the proposed system accurately recognizes car accidents with an accident detection rate (adr) of 86.25% and false alarm rate (far) of 33.00%.","key_phrases":["the recent advancements","autonomous vehicles","two important factors","that","a vital role","accidents","collisions","obstacles","track detection","avs","an accident detection model","accident vehicles","rollover vehicles","present many trajectories-based and sensor-based multiple-vehicle accident prediction models","taiwan","the av tesla sedan's failure","overturned vehicles","an efficient deep learning model","a single-car crash","appropriate actions","changes","the concerned authorities","this paper","a novel car crash detection system","various car crashes","three deep learning models","namely vgg16(feature extractor","transfer learning","rpn","region proposal network","cnn8l","(region-based detector","the cnn8l","a novel lightweight sequential convolutional neural network","region-based classification","detection","the model","a customized dataset","different metrics","the-art","the experimental results","the vgg16","the cnn8l model","other models","the proposed system","car accidents","an accident detection rate","adr","86.25%","false alarm rate","33.00%","two","taiwan","three","86.25%","33.00%"]},{"title":"Deep learning based vessel arrivals monitoring via autoregressive statistical control charts","authors":["Sara El Mekkaoui","Ghait Boukachab","Loubna Benabbou","Abdelaziz Berrado"],"abstract":"This paper introduces a methodology for monitoring the vessel arrival process, a critical factor in enhancing maritime operational efficiency. This approach uses deep learning sequence models and Statistical Process Control Charts to track the variability in a vessel arrival process. The proposed solution uses the predictive deep learning model to get a vessel\u2019s estimated time of arrival, produces quality characteristics, and applies statistical control charts to monitor their variability. The paper presents the results of applying the proposed methodology for vessel arrivals at a coal terminal, which demonstrates the effectiveness of the method. By enabling precise monitoring of arrival times, this methodology not only supports efficient ship and port operations planning but also aids in the timely adoption of operational adjustments. This can significantly contribute to operational measures aimed at reducing shipping emissions and optimizing resource utilization.","doi":"10.1007\/s13437-024-00342-9","cleaned_title":"deep learning based vessel arrivals monitoring via autoregressive statistical control charts","cleaned_abstract":"this paper introduces a methodology for monitoring the vessel arrival process, a critical factor in enhancing maritime operational efficiency. this approach uses deep learning sequence models and statistical process control charts to track the variability in a vessel arrival process. the proposed solution uses the predictive deep learning model to get a vessel\u2019s estimated time of arrival, produces quality characteristics, and applies statistical control charts to monitor their variability. the paper presents the results of applying the proposed methodology for vessel arrivals at a coal terminal, which demonstrates the effectiveness of the method. by enabling precise monitoring of arrival times, this methodology not only supports efficient ship and port operations planning but also aids in the timely adoption of operational adjustments. this can significantly contribute to operational measures aimed at reducing shipping emissions and optimizing resource utilization.","key_phrases":["this paper","a methodology","the vessel arrival process","a critical factor","maritime operational efficiency","this approach","deep learning sequence models","statistical process control charts","the variability","a vessel arrival process","the proposed solution","the predictive deep learning model","a vessel\u2019s estimated time","arrival","quality characteristics","statistical control charts","their variability","the paper","the results","the proposed methodology","vessel arrivals","a coal terminal","which","the effectiveness","the method","precise monitoring","arrival times","this methodology","efficient ship and port operations planning","the timely adoption","operational adjustments","this","operational measures","shipping emissions","resource utilization"]},{"title":"RNA contact prediction by data efficient deep learning","authors":["Oskar Taubert","Fabrice von der Lehr","Alina Bazarova","Christian Faber","Philipp Knechtges","Marie Weiel","Charlotte Debus","Daniel Coquelin","Achim Basermann","Achim Streit","Stefan Kesselheim","Markus G\u00f6tz","Alexander Schug"],"abstract":"On the path to full understanding of the structure-function relationship or even design of RNA, structure prediction would offer an intriguing complement to experimental efforts. Any deep learning on RNA structure, however, is hampered by the sparsity of labeled training data. Utilizing the limited data available, we here focus on predicting spatial adjacencies (\"contact maps\u201d) as a proxy for 3D structure. Our model, BARNACLE, combines the utilization of unlabeled data through self-supervised pre-training and efficient use of the sparse labeled data through an XGBoost classifier. BARNACLE shows a considerable improvement over both the established classical baseline and a deep neural network. In order to demonstrate that our approach can be applied to tasks with similar data constraints, we show that our findings generalize to the related setting of accessible surface area prediction.","doi":"10.1038\/s42003-023-05244-9","cleaned_title":"rna contact prediction by data efficient deep learning","cleaned_abstract":"on the path to full understanding of the structure-function relationship or even design of rna, structure prediction would offer an intriguing complement to experimental efforts. any deep learning on rna structure, however, is hampered by the sparsity of labeled training data. utilizing the limited data available, we here focus on predicting spatial adjacencies (\"contact maps\u201d) as a proxy for 3d structure. our model, barnacle, combines the utilization of unlabeled data through self-supervised pre-training and efficient use of the sparse labeled data through an xgboost classifier. barnacle shows a considerable improvement over both the established classical baseline and a deep neural network. in order to demonstrate that our approach can be applied to tasks with similar data constraints, we show that our findings generalize to the related setting of accessible surface area prediction.","key_phrases":["the path","full understanding","the structure-function relationship","even design","rna","structure prediction","an intriguing complement","experimental efforts","any deep learning","rna structure","the sparsity","labeled training data","the limited data","we","spatial adjacencies","(\"contact maps","a proxy","3d structure","our model","barnacle","the utilization","unlabeled data","self-supervised pre","-","training and efficient use","the sparse","data","an xgboost classifier","barnacle","a considerable improvement","both the established classical baseline","a deep neural network","order","our approach","tasks","similar data constraints","we","our findings","the related setting","accessible surface area prediction","3d"]},{"title":"Analysis, characterization, prediction, and attribution of extreme atmospheric events with machine learning and deep learning techniques: a review","authors":["Sancho Salcedo-Sanz","Jorge P\u00e9rez-Aracil","Guido Ascenso","Javier Del Ser","David Casillas-P\u00e9rez","Christopher Kadow","Du\u0161an Fister","David Barriopedro","Ricardo Garc\u00eda-Herrera","Matteo Giuliani","Andrea Castelletti"],"abstract":"Atmospheric extreme events cause severe damage to human societies and ecosystems. The frequency and intensity of extremes and other associated events are continuously increasing due to climate change and global warming. The accurate prediction, characterization, and attribution of atmospheric extreme events is, therefore, a key research field in which many groups are currently working by applying different methodologies and computational tools. Machine learning and deep learning methods have arisen in the last years as powerful techniques to tackle many of the problems related to atmospheric extreme events. This paper reviews machine learning and deep learning approaches applied to the analysis, characterization, prediction, and attribution of the most important atmospheric extremes. A summary of the most used machine learning and deep learning techniques in this area, and a comprehensive critical review of literature related to ML in EEs, are provided. The critical literature review has been extended to extreme events related to rainfall and floods, heatwaves and extreme temperatures, droughts, severe weather events and fog, and low-visibility episodes. A case study focused on the analysis of extreme atmospheric temperature prediction with ML and DL techniques is also presented in the paper. Conclusions, perspectives, and outlooks on the field are finally drawn.","doi":"10.1007\/s00704-023-04571-5","cleaned_title":"analysis, characterization, prediction, and attribution of extreme atmospheric events with machine learning and deep learning techniques: a review","cleaned_abstract":"atmospheric extreme events cause severe damage to human societies and ecosystems. the frequency and intensity of extremes and other associated events are continuously increasing due to climate change and global warming. the accurate prediction, characterization, and attribution of atmospheric extreme events is, therefore, a key research field in which many groups are currently working by applying different methodologies and computational tools. machine learning and deep learning methods have arisen in the last years as powerful techniques to tackle many of the problems related to atmospheric extreme events. this paper reviews machine learning and deep learning approaches applied to the analysis, characterization, prediction, and attribution of the most important atmospheric extremes. a summary of the most used machine learning and deep learning techniques in this area, and a comprehensive critical review of literature related to ml in ees, are provided. the critical literature review has been extended to extreme events related to rainfall and floods, heatwaves and extreme temperatures, droughts, severe weather events and fog, and low-visibility episodes. a case study focused on the analysis of extreme atmospheric temperature prediction with ml and dl techniques is also presented in the paper. conclusions, perspectives, and outlooks on the field are finally drawn.","key_phrases":["atmospheric extreme events","severe damage","human societies","ecosystems","the frequency","intensity","extremes","other associated events","climate change","global warming","the accurate prediction","characterization","attribution","atmospheric extreme events","a key research field","which","many groups","different methodologies","computational tools","machine learning","deep learning methods","the last years","powerful techniques","the problems","atmospheric extreme events","this paper","machine learning","deep learning approaches","the analysis","characterization","prediction","attribution","the most important atmospheric extremes","a summary","the most used machine learning","deep learning techniques","this area","a comprehensive critical review","literature","ml","ees","the critical literature review","extreme events","rainfall","floods","heatwaves","extreme temperatures","droughts","severe weather events","fog","low-visibility episodes","a case study","the analysis","extreme atmospheric temperature prediction","ml","dl techniques","the paper","conclusions","perspectives","outlooks","the field","the last years"]},{"title":"Application of deep learning and XGBoost in predicting pathological staging of breast cancer MR images","authors":["Yue Miao","Siyuan Tang","Zhuqiang Zhang","Jukun Song","Zhi Liu","Qiang Chen","Miao Zhang"],"abstract":"The methods of deep learning and traditional radiomics feature extraction were preliminarily discussed, and a multimodal data prediction model for breast cancer clinical stage was established. The MR images and clinical staging data of breast cancer were obtained from the official websites of the American Cancer Center TCGA and TCIA, respectively, with a total of 139 patient samples. The region of interest was delineated on the enhanced image of breast cancer MR, and then the feature extraction of radiomics and deep learning was performed, and 108 radiomics features and 1024 deep-learning features were extracted for each case. After feature screening and processing, clinical data were integrated, and a machine-learning model was used to predict clinical stage I and non-stage I. Results 26 radiomic features and 12 deep features related to staging were screened out by LASSO algorithm, and a classification model was constructed based on XGBoost machine learning. The patients were predicted with an accuracy rate of 80.00%, and the area under the curve of the receiver operating characteristic curve was 0.833. It is feasible to predict the clinical stage of breast cancer through radiomics and deep-learning feature extraction and machine-learning technology. The classification model based on multimodal data established by using machine-learning classifier can distinguish clinical stage I and non-stage I in breast cancer and have higher accuracy. This study confirms the feasibility and accuracy of combining data from different modalities to contribute to clinical staging prediction. The research contributions include demonstrating the superiority of deep-learning models for feature extraction and classification, as well as highlighting the potential of combining deep learning and traditional machine-learning algorithms for improved classification performance.","doi":"10.1007\/s11227-023-05797-w","cleaned_title":"application of deep learning and xgboost in predicting pathological staging of breast cancer mr images","cleaned_abstract":"the methods of deep learning and traditional radiomics feature extraction were preliminarily discussed, and a multimodal data prediction model for breast cancer clinical stage was established. the mr images and clinical staging data of breast cancer were obtained from the official websites of the american cancer center tcga and tcia, respectively, with a total of 139 patient samples. the region of interest was delineated on the enhanced image of breast cancer mr, and then the feature extraction of radiomics and deep learning was performed, and 108 radiomics features and 1024 deep-learning features were extracted for each case. after feature screening and processing, clinical data were integrated, and a machine-learning model was used to predict clinical stage i and non-stage i. results 26 radiomic features and 12 deep features related to staging were screened out by lasso algorithm, and a classification model was constructed based on xgboost machine learning. the patients were predicted with an accuracy rate of 80.00%, and the area under the curve of the receiver operating characteristic curve was 0.833. it is feasible to predict the clinical stage of breast cancer through radiomics and deep-learning feature extraction and machine-learning technology. the classification model based on multimodal data established by using machine-learning classifier can distinguish clinical stage i and non-stage i in breast cancer and have higher accuracy. this study confirms the feasibility and accuracy of combining data from different modalities to contribute to clinical staging prediction. the research contributions include demonstrating the superiority of deep-learning models for feature extraction and classification, as well as highlighting the potential of combining deep learning and traditional machine-learning algorithms for improved classification performance.","key_phrases":["the methods","deep learning","traditional radiomics","feature extraction","a multimodal data prediction model","breast cancer clinical stage","the mr images","clinical staging data","breast cancer","the official websites","the american cancer center","tcia","a total","139 patient samples","the region","interest","the enhanced image","breast cancer","then the feature extraction","radiomics","deep learning","108 radiomics features","1024 deep-learning features","each case","feature screening","processing","clinical data","a machine-learning model","clinical stage","i","non-stage i.","26 radiomic features","12 deep features","staging","lasso algorithm","a classification model","xgboost machine learning","the patients","an accuracy rate","80.00%","the area","the curve","the receiver operating characteristic curve","it","the clinical stage","breast cancer","radiomics","deep-learning feature extraction and machine-learning technology","the classification model","multimodal data","machine-learning classifier","clinical stage","i","i","breast cancer","higher accuracy","this study","the feasibility","accuracy","data","different modalities","clinical staging prediction","the research contributions","the superiority","deep-learning models","feature extraction","classification","the potential","deep learning","traditional machine-learning algorithms","improved classification performance","american","139","108","1024","26","12","lasso","80.00%","0.833","distinguish"]},{"title":"Enhancing deep learning classification performance of tongue lesions in imbalanced data: mosaic-based soft labeling with curriculum learning","authors":["Sung-Jae Lee","Hyun Jun Oh","Young-Don Son","Jong-Hoon Kim","Ik-Jae Kwon","Bongju Kim","Jong-Ho Lee","Hang-Keun Kim"],"abstract":"BackgroundOral potentially malignant disorders (OPMDs) are associated with an increased risk of cancer of the oral cavity including the tongue. The early detection of oral cavity cancers and OPMDs is critical for reducing cancer-specific morbidity and mortality. Recently, there have been studies to apply the rapidly advancing technology of deep learning for diagnosing oral cavity cancer and OPMDs. However, several challenging issues such as class imbalance must be resolved to effectively train a deep learning model for medical imaging classification tasks. The aim of this study is to evaluate a new technique of artificial intelligence to improve the classification performance in an imbalanced tongue lesion dataset.MethodsA total of 1,810 tongue images were used for the classification. The class-imbalanced dataset consisted of 372 instances of cancer, 141 instances of OPMDs, and 1,297 instances of noncancerous lesions. The EfficientNet model was used as the feature extraction model for classification. Mosaic data augmentation, soft labeling, and curriculum learning (CL) were employed to improve the classification performance of the convolutional neural network.ResultsUtilizing a mosaic-augmented dataset in conjunction with CL, the final model achieved an accuracy rate of 0.9444, surpassing conventional oversampling and weight balancing methods. The relative precision improvement rate for the minority class OPMD was 21.2%, while the relative \\({F}_{1}\\) score improvement rate of OPMD was 4.9%.ConclusionsThe present study demonstrates that the integration of mosaic-based soft labeling and curriculum learning improves the classification performance of tongue lesions compared to previous methods, establishing a foundation for future research on effectively learning from imbalanced data.","doi":"10.1186\/s12903-024-03898-3","cleaned_title":"enhancing deep learning classification performance of tongue lesions in imbalanced data: mosaic-based soft labeling with curriculum learning","cleaned_abstract":"backgroundoral potentially malignant disorders (opmds) are associated with an increased risk of cancer of the oral cavity including the tongue. the early detection of oral cavity cancers and opmds is critical for reducing cancer-specific morbidity and mortality. recently, there have been studies to apply the rapidly advancing technology of deep learning for diagnosing oral cavity cancer and opmds. however, several challenging issues such as class imbalance must be resolved to effectively train a deep learning model for medical imaging classification tasks. the aim of this study is to evaluate a new technique of artificial intelligence to improve the classification performance in an imbalanced tongue lesion dataset.methodsa total of 1,810 tongue images were used for the classification. the class-imbalanced dataset consisted of 372 instances of cancer, 141 instances of opmds, and 1,297 instances of noncancerous lesions. the efficientnet model was used as the feature extraction model for classification. mosaic data augmentation, soft labeling, and curriculum learning (cl) were employed to improve the classification performance of the convolutional neural network.resultsutilizing a mosaic-augmented dataset in conjunction with cl, the final model achieved an accuracy rate of 0.9444, surpassing conventional oversampling and weight balancing methods. the relative precision improvement rate for the minority class opmd was 21.2%, while the relative \\({f}_{1}\\) score improvement rate of opmd was 4.9%.conclusionsthe present study demonstrates that the integration of mosaic-based soft labeling and curriculum learning improves the classification performance of tongue lesions compared to previous methods, establishing a foundation for future research on effectively learning from imbalanced data.","key_phrases":["backgroundoral potentially malignant disorders","an increased risk","cancer","the oral cavity","the tongue","the early detection","oral cavity cancers","opmds","cancer-specific morbidity","mortality","studies","the rapidly advancing technology","deep learning","oral cavity cancer","opmds","several challenging issues","class imbalance","a deep learning model","medical imaging classification tasks","the aim","this study","a new technique","artificial intelligence","the classification performance","an imbalanced tongue lesion","dataset.methodsa total","1,810 tongue images","the classification","the class-imbalanced dataset","372 instances","cancer","141 instances","opmds","1,297 instances","noncancerous lesions","the efficientnet model","the feature extraction model","classification","mosaic data augmentation","soft labeling","curriculum learning","cl","the classification performance","the convolutional neural","a mosaic-augmented dataset","conjunction","cl","the final model","an accuracy rate","conventional oversampling","weight balancing methods","the relative precision improvement rate","the minority class","21.2%","the relative \\({f}_{1}\\) score improvement rate","4.9%.conclusionsthe present study","the integration","mosaic-based soft labeling","curriculum learning","the classification performance","tongue lesions","previous methods","a foundation","future research","imbalanced data","dataset.methodsa","1,810","372","141","1,297","mosaic data","mosaic","0.9444","21.2%"]},{"title":"Deep learning enables fast, gentle STED microscopy","authors":["Vahid Ebrahimi","Till Stephan","Jiah Kim","Pablo Carravilla","Christian Eggeling","Stefan Jakobs","Kyu Young Han"],"abstract":"STED microscopy is widely used to image subcellular structures with super-resolution. Here, we report that restoring STED images with deep learning can mitigate photobleaching and photodamage by reducing the pixel dwell time by one or two orders of magnitude. Our method allows for efficient and robust restoration of noisy 2D and 3D STED images with multiple targets and facilitates long-term imaging of mitochondrial dynamics.","doi":"10.1038\/s42003-023-05054-z","cleaned_title":"deep learning enables fast, gentle sted microscopy","cleaned_abstract":"sted microscopy is widely used to image subcellular structures with super-resolution. here, we report that restoring sted images with deep learning can mitigate photobleaching and photodamage by reducing the pixel dwell time by one or two orders of magnitude. our method allows for efficient and robust restoration of noisy 2d and 3d sted images with multiple targets and facilitates long-term imaging of mitochondrial dynamics.","key_phrases":["sted microscopy","subcellular structures","super","-","resolution","we","sted images","deep learning","photobleaching","photodamage","the pixel","dwell time","one or two orders","magnitude","our method","efficient and robust restoration","noisy 2d","3d sted images","multiple targets","long-term imaging","mitochondrial dynamics","one","two","2d","3d"]},{"title":"Biological gender identification in Turkish news text using deep learning models","authors":["P\u0131nar T\u00fcfekci","Melike Bekta\u015f K\u00f6sesoy"],"abstract":"Identifying the biological gender of authors based on the content of their written work is a crucial task in Natural Language Processing (NLP). Accurate biological gender identification finds numerous applications in fields such as linguistics, sociology, and marketing. However, achieving high accuracy in identifying the biological gender of the author is heavily dependent on the quality of the collected data and its proper splitting. Therefore, determining the best-performing model necessitates experimental evaluation. This study aimed to develop and evaluate four learning algorithms for biological gender identification in news texts. To this end, a comprehensive dataset, IAG-TNKU, was created from a Turkish newspaper, comprising 43,292 news articles. Four models utilizing popular machine learning algorithms, including Naive Bayes and Random Forest, and two deep learning algorithms, Long Short Term Memory and Convolutional Neural Networks, were developed and evaluated rigorously. The results indicated that the Long Short Term Memory (LSTM) algorithm outperformed the other three models, exhibiting an exceptional accuracy of 88.51%. This model's outstanding performance underpins the importance of utilizing innovative deep learning algorithms for biological gender identification tasks in NLP. The present study contributes to extant literature by developing a new dataset for biological gender identification in news texts and evaluating four machine learning algorithms. Our findings highlight the significance of utilizing innovative techniques for biological gender identification tasks. The dataset and deep learning algorithm can be applied in many areas such as sociolinguistics, marketing research, and journalism, where the identification of biological gender in written content plays a pivotal role.","doi":"10.1007\/s11042-023-17622-w","cleaned_title":"biological gender identification in turkish news text using deep learning models","cleaned_abstract":"identifying the biological gender of authors based on the content of their written work is a crucial task in natural language processing (nlp). accurate biological gender identification finds numerous applications in fields such as linguistics, sociology, and marketing. however, achieving high accuracy in identifying the biological gender of the author is heavily dependent on the quality of the collected data and its proper splitting. therefore, determining the best-performing model necessitates experimental evaluation. this study aimed to develop and evaluate four learning algorithms for biological gender identification in news texts. to this end, a comprehensive dataset, iag-tnku, was created from a turkish newspaper, comprising 43,292 news articles. four models utilizing popular machine learning algorithms, including naive bayes and random forest, and two deep learning algorithms, long short term memory and convolutional neural networks, were developed and evaluated rigorously. the results indicated that the long short term memory (lstm) algorithm outperformed the other three models, exhibiting an exceptional accuracy of 88.51%. this model's outstanding performance underpins the importance of utilizing innovative deep learning algorithms for biological gender identification tasks in nlp. the present study contributes to extant literature by developing a new dataset for biological gender identification in news texts and evaluating four machine learning algorithms. our findings highlight the significance of utilizing innovative techniques for biological gender identification tasks. the dataset and deep learning algorithm can be applied in many areas such as sociolinguistics, marketing research, and journalism, where the identification of biological gender in written content plays a pivotal role.","key_phrases":["the biological gender","authors","the content","their written work","a crucial task","natural language processing","nlp","accurate biological gender identification","numerous applications","fields","linguistics","sociology","marketing","high accuracy","the biological gender","the author","the quality","the collected data","its proper splitting","the best-performing model necessitates experimental evaluation","this study","four learning algorithms","biological gender identification","news texts","this end","a comprehensive dataset","iag","tnku","a turkish newspaper","43,292 news articles","four models","popular machine learning algorithms","naive bayes","random forest","two deep learning algorithms","long short term memory","convolutional neural networks","the results","the long short term memory","lstm","algorithm","the other three models","an exceptional accuracy","88.51%","this model's outstanding performance","the importance","innovative deep learning algorithms","biological gender identification tasks","nlp","the present study","extant literature","a new dataset","biological gender identification","news texts","four machine learning algorithms","our findings","the significance","innovative techniques","biological gender identification tasks","the dataset and deep learning algorithm","many areas","sociolinguistics","marketing research","journalism","the identification","biological gender","written content","a pivotal role","four","turkish","43,292","four","two","three","88.51%","four"]},{"title":"Deep learning model for pleural effusion detection via active learning and pseudo-labeling: a multisite study","authors":["Joseph Chang","Bo-Ru Lin","Ti-Hao Wang","Chung-Ming Chen"],"abstract":"BackgroundThe study aimed to develop and validate a deep learning-based Computer Aided Triage (CADt) algorithm for detecting pleural effusion in chest radiographs using an active learning (AL) framework. This is aimed at addressing the critical need for a clinical grade algorithm that can timely diagnose pleural effusion, which affects approximately 1.5\u00a0million people annually in the United States.MethodsIn this multisite study, 10,599 chest radiographs from 2006 to 2018 were retrospectively collected from an institution in Taiwan to train the deep learning algorithm. The AL framework utilized significantly reduced the need for expert annotations. For external validation, the algorithm was tested on a multisite dataset of 600 chest radiographs from 22 clinical sites in the United States and Taiwan, which were annotated by three U.S. board-certified radiologists.ResultsThe CADt algorithm demonstrated high effectiveness in identifying pleural effusion, achieving a sensitivity of 0.95 (95% CI: [0.92, 0.97]) and a specificity of 0.97 (95% CI: [0.95, 0.99]). The area under the receiver operating characteristic curve (AUC) was 0.97 (95% DeLong\u2019s CI: [0.95, 0.99]). Subgroup analyses showed that the algorithm maintained robust performance across various demographics and clinical settings.ConclusionThis study presents a novel approach in developing clinical grade CADt solutions for the diagnosis of pleural effusion. The AL-based CADt algorithm not only achieved high accuracy in detecting pleural effusion but also significantly reduced the workload required for clinical experts in annotating medical data. This method enhances the feasibility of employing advanced technological solutions for prompt and accurate diagnosis in medical settings.","doi":"10.1186\/s12880-024-01260-1","cleaned_title":"deep learning model for pleural effusion detection via active learning and pseudo-labeling: a multisite study","cleaned_abstract":"backgroundthe study aimed to develop and validate a deep learning-based computer aided triage (cadt) algorithm for detecting pleural effusion in chest radiographs using an active learning (al) framework. this is aimed at addressing the critical need for a clinical grade algorithm that can timely diagnose pleural effusion, which affects approximately 1.5 million people annually in the united states.methodsin this multisite study, 10,599 chest radiographs from 2006 to 2018 were retrospectively collected from an institution in taiwan to train the deep learning algorithm. the al framework utilized significantly reduced the need for expert annotations. for external validation, the algorithm was tested on a multisite dataset of 600 chest radiographs from 22 clinical sites in the united states and taiwan, which were annotated by three u.s. board-certified radiologists.resultsthe cadt algorithm demonstrated high effectiveness in identifying pleural effusion, achieving a sensitivity of 0.95 (95% ci: [0.92, 0.97]) and a specificity of 0.97 (95% ci: [0.95, 0.99]). the area under the receiver operating characteristic curve (auc) was 0.97 (95% delong\u2019s ci: [0.95, 0.99]). subgroup analyses showed that the algorithm maintained robust performance across various demographics and clinical settings.conclusionthis study presents a novel approach in developing clinical grade cadt solutions for the diagnosis of pleural effusion. the al-based cadt algorithm not only achieved high accuracy in detecting pleural effusion but also significantly reduced the workload required for clinical experts in annotating medical data. this method enhances the feasibility of employing advanced technological solutions for prompt and accurate diagnosis in medical settings.","key_phrases":["backgroundthe study","a deep learning-based computer aided triage","cadt","algorithm","pleural effusion","chest","an active learning","(al) framework","this","the critical need","a clinical grade","algorithm","that","pleural effusion","which","approximately 1.5 million people","the united states.methodsin","this multisite study","10,599 chest","an institution","taiwan","the deep learning algorithm","the al framework","the need","expert annotations","external validation","the algorithm","a multisite dataset","600 chest","22 clinical sites","the united states","taiwan","which","radiologists.resultsthe cadt algorithm","high effectiveness","pleural effusion","a sensitivity","95% ci","a specificity","(95% ci","the area","the receiver operating characteristic curve","auc","(95% delong","subgroup analyses","the algorithm","robust performance","various demographics","clinical settings.conclusionthis study","a novel approach","clinical grade cadt solutions","the diagnosis","pleural effusion","the al-based cadt","algorithm","high accuracy","pleural effusion","the workload","clinical experts","medical data","this method","the feasibility","advanced technological solutions","prompt and accurate diagnosis","medical settings","approximately 1.5 million","annually","the united states.methodsin","10,599","2006","2018","taiwan","al","600","22","the united states","taiwan","three","u.s.","0.95","95%","0.92","0.97","0.97","95%","0.95","0.99","0.97","95%","0.95","0.99","al"]},{"title":"Learning key steps to attack deep reinforcement learning agents","authors":["Chien-Min Yu","Ming-Hsin Chen","Hsuan-Tien Lin"],"abstract":"Deep reinforcement learning agents are vulnerable to adversarial attacks. In particular, recent studies have shown that attacking a few key steps can effectively decrease the agent\u2019s cumulative reward. However, all existing attacking methods define those key steps with human-designed heuristics, and it is not clear how more effective key steps can be identified. This paper introduces a novel reinforcement learning framework that learns key steps through interacting with the agent. The proposed framework does not require any human heuristics nor knowledge, and can be flexibly coupled with any white-box or black-box adversarial attack scenarios. Experiments on benchmark Atari games across different scenarios demonstrate that the proposed framework is superior to existing methods for identifying effective key steps. The results highlight the weakness of RL agents even under budgeted attacks.","doi":"10.1007\/s10994-023-06318-9","cleaned_title":"learning key steps to attack deep reinforcement learning agents","cleaned_abstract":"deep reinforcement learning agents are vulnerable to adversarial attacks. in particular, recent studies have shown that attacking a few key steps can effectively decrease the agent\u2019s cumulative reward. however, all existing attacking methods define those key steps with human-designed heuristics, and it is not clear how more effective key steps can be identified. this paper introduces a novel reinforcement learning framework that learns key steps through interacting with the agent. the proposed framework does not require any human heuristics nor knowledge, and can be flexibly coupled with any white-box or black-box adversarial attack scenarios. experiments on benchmark atari games across different scenarios demonstrate that the proposed framework is superior to existing methods for identifying effective key steps. the results highlight the weakness of rl agents even under budgeted attacks.","key_phrases":["deep reinforcement learning agents","adversarial attacks","recent studies","a few key steps","the agent\u2019s cumulative reward","all existing attacking methods","those key steps","human-designed heuristics","it","how more effective key steps","this paper","a novel reinforcement learning framework","that","key steps","the agent","the proposed framework","any human heuristics","knowledge","any white-box or black-box adversarial attack scenarios","experiments","benchmark atari games","different scenarios","the proposed framework","existing methods","effective key steps","the results","the weakness","rl agents","budgeted attacks"]},{"title":"An active learning approach to train a deep learning algorithm for tumor segmentation from brain MR images","authors":["Andrew S. Boehringer","Amirhossein Sanaat","Hossein Arabi","Habib Zaidi"],"abstract":"PurposeThis study focuses on assessing the performance of active learning techniques to train a brain MRI glioma segmentation model.MethodsThe publicly available training dataset provided for the 2021 RSNA-ASNR-MICCAI Brain Tumor Segmentation (BraTS) Challenge was used in this study, consisting of 1251 multi-institutional, multi-parametric MR images. Post-contrast T1, T2, and T2 FLAIR images as well as ground truth manual segmentation were used as input for the model. The data were split into a training set of 1151 cases and testing set of 100 cases, with the testing set remaining constant throughout. Deep convolutional neural network segmentation models were trained using the NiftyNet platform. To test the viability of active learning in training a segmentation model, an initial reference model was trained using all 1151 training cases followed by two additional models using only 575 cases and 100 cases. The resulting predicted segmentations of these two additional models on the remaining training cases were then addended to the training dataset for additional training.ResultsIt was demonstrated that an active learning approach for manual segmentation can lead to comparable model performance for segmentation of brain gliomas (0.906 reference Dice score vs 0.868 active learning Dice score) while only requiring manual annotation for 28.6% of the data.ConclusionThe active learning approach when applied to model training can drastically reduce the time and labor spent on preparation of ground truth training data.Critical relevance statementActive learning concepts were applied to a deep learning-assisted segmentation of brain gliomas from MR images to assess their viability in reducing the required amount of manually annotated ground truth data in model training.Key points\u2022 This study focuses on assessing the performance of active learning techniques to train a brain MRI glioma segmentation model.\u2022 The active learning approach for manual segmentation can lead to comparable model performance for segmentation of brain gliomas.\u2022 Active learning when applied to model training can drastically reduce the time and labor spent on preparation of ground truth training data.Graphical Abstract","doi":"10.1186\/s13244-023-01487-6","cleaned_title":"an active learning approach to train a deep learning algorithm for tumor segmentation from brain mr images","cleaned_abstract":"purposethis study focuses on assessing the performance of active learning techniques to train a brain mri glioma segmentation model.methodsthe publicly available training dataset provided for the 2021 rsna-asnr-miccai brain tumor segmentation (brats) challenge was used in this study, consisting of 1251 multi-institutional, multi-parametric mr images. post-contrast t1, t2, and t2 flair images as well as ground truth manual segmentation were used as input for the model. the data were split into a training set of 1151 cases and testing set of 100 cases, with the testing set remaining constant throughout. deep convolutional neural network segmentation models were trained using the niftynet platform. to test the viability of active learning in training a segmentation model, an initial reference model was trained using all 1151 training cases followed by two additional models using only 575 cases and 100 cases. the resulting predicted segmentations of these two additional models on the remaining training cases were then addended to the training dataset for additional training.resultsit was demonstrated that an active learning approach for manual segmentation can lead to comparable model performance for segmentation of brain gliomas (0.906 reference dice score vs 0.868 active learning dice score) while only requiring manual annotation for 28.6% of the data.conclusionthe active learning approach when applied to model training can drastically reduce the time and labor spent on preparation of ground truth training data.critical relevance statementactive learning concepts were applied to a deep learning-assisted segmentation of brain gliomas from mr images to assess their viability in reducing the required amount of manually annotated ground truth data in model training.key points\u2022 this study focuses on assessing the performance of active learning techniques to train a brain mri glioma segmentation model.\u2022 the active learning approach for manual segmentation can lead to comparable model performance for segmentation of brain gliomas.\u2022 active learning when applied to model training can drastically reduce the time and labor spent on preparation of ground truth training data.graphical abstract","key_phrases":["purposethis study","the performance","active learning techniques","a brain mri glioma segmentation","model.methodsthe publicly available training dataset","the 2021 rsna-asnr-miccai brain tumor segmentation (brats) challenge","this study","1251 multi-institutional, multi-parametric mr images","post-contrast t1","t2","flair images","ground truth manual segmentation","input","the model","the data","a training set","1151 cases","testing","100 cases","the testing","deep convolutional neural network segmentation models","the niftynet platform","the viability","active learning","a segmentation model","an initial reference model","all 1151 training cases","two additional models","only 575 cases","100 cases","the resulting predicted segmentations","these two additional models","the remaining training cases","the training dataset","additional training.resultsit","an active learning approach","manual segmentation","comparable model performance","segmentation","brain gliomas","0.906 reference dice score","0.868 active learning dice score","manual annotation","28.6%","the data.conclusionthe active learning approach","model training","the time","labor","preparation","ground truth","data.critical relevance statementactive learning concepts","a deep learning-assisted segmentation","brain gliomas","mr images","their viability","the required amount","manually annotated ground truth data","this study","the performance","active learning techniques","a brain mri glioma segmentation","the active learning approach","manual segmentation","comparable model performance","segmentation","brain gliomas.\u2022 active learning","model training","the time","labor","preparation","ground truth","data.graphical abstract","2021","1251","1151","100","1151","two","only 575","100","two","0.906 reference","0.868","28.6%"]},{"title":"Efficient shallow learning as an alternative to deep learning","authors":["Yuval Meir","Ofek Tevet","Yarden Tzach","Shiri Hodassman","Ronit D. Gross","Ido Kanter"],"abstract":"The realization of complex classification tasks requires training of deep learning (DL) architectures consisting of tens or even hundreds of convolutional and fully connected hidden layers, which is far from the reality of the human brain. According to the DL rationale, the first convolutional layer reveals localized patterns in the input and large-scale patterns in the following layers, until it reliably characterizes a class of inputs. Here, we demonstrate that with a fixed ratio between the depths of the first and second convolutional layers, the error rates of the generalized shallow LeNet architecture, consisting of only five layers, decay as a power law with the number of filters in the first convolutional layer. The extrapolation of this power law indicates that the generalized LeNet can achieve small error rates that were previously obtained for the CIFAR-10 database using DL architectures. A power law with a similar exponent also characterizes the generalized VGG-16 architecture. However, this results in a significantly increased number of operations required to achieve a given error rate with respect to LeNet. This power law phenomenon governs various generalized LeNet and VGG-16 architectures, hinting at its universal behavior and suggesting a quantitative hierarchical time\u2013space complexity among machine learning architectures. Additionally, the conservation law along the convolutional layers, which is the square-root of their size times their depth, is found to asymptotically minimize error rates. The efficient shallow learning that is demonstrated in this study calls for further quantitative examination using various databases and architectures and its accelerated implementation using future dedicated hardware developments.","doi":"10.1038\/s41598-023-32559-8","cleaned_title":"efficient shallow learning as an alternative to deep learning","cleaned_abstract":"the realization of complex classification tasks requires training of deep learning (dl) architectures consisting of tens or even hundreds of convolutional and fully connected hidden layers, which is far from the reality of the human brain. according to the dl rationale, the first convolutional layer reveals localized patterns in the input and large-scale patterns in the following layers, until it reliably characterizes a class of inputs. here, we demonstrate that with a fixed ratio between the depths of the first and second convolutional layers, the error rates of the generalized shallow lenet architecture, consisting of only five layers, decay as a power law with the number of filters in the first convolutional layer. the extrapolation of this power law indicates that the generalized lenet can achieve small error rates that were previously obtained for the cifar-10 database using dl architectures. a power law with a similar exponent also characterizes the generalized vgg-16 architecture. however, this results in a significantly increased number of operations required to achieve a given error rate with respect to lenet. this power law phenomenon governs various generalized lenet and vgg-16 architectures, hinting at its universal behavior and suggesting a quantitative hierarchical time\u2013space complexity among machine learning architectures. additionally, the conservation law along the convolutional layers, which is the square-root of their size times their depth, is found to asymptotically minimize error rates. the efficient shallow learning that is demonstrated in this study calls for further quantitative examination using various databases and architectures and its accelerated implementation using future dedicated hardware developments.","key_phrases":["the realization","complex classification tasks","training","deep learning","dl","tens","even hundreds","convolutional and fully connected hidden layers","which","the reality","the human brain","the dl rationale","the first convolutional layer","localized patterns","the input","large-scale patterns","the following layers","it","a class","inputs","we","a fixed ratio","the depths","the first and second convolutional layers","the error rates","the generalized shallow lenet architecture","only five layers","a power law","the number","filters","the first convolutional layer","the extrapolation","this power law","the generalized lenet","small error rates","that","the cifar-10 database","dl architectures","a power law","a similar exponent","the generalized vgg-16 architecture","this","a significantly increased number","operations","a given error rate","respect","lenet","this power law phenomenon","various generalized lenet","vgg-16","architectures","its universal behavior","a quantitative hierarchical time\u2013space complexity","machine learning architectures","the conservation law","the convolutional layers","which","the square-root","their size","their depth","error rates","the efficient shallow learning","that","this study","further quantitative examination","various databases","architectures","its accelerated implementation","future dedicated hardware developments","tens or even hundreds","first","first","second","only five","first","cifar-10","vgg-16"]},{"title":"Crop type classification with hyperspectral images using deep learning : a transfer learning approach","authors":["Usha Patel","Mohib Pathan","Preeti Kathiria","Vibha Patel"],"abstract":"Crop classification plays a vital role in felicitating agriculture statistics to the state and national government in decision-making. In recent years, due to advancements in remote sensing, high-resolution hyperspectral images (HSIs) are available for land cover classification. HSIs can classify the different crop categories precisely due to their narrow and continuous spectral band reflection. With improvements in computing power and evolution in deep learning technology, Deep learning is rapidly being used for HSIs classification. However, to train deep neural networks, many labeled samples are needed. The labeling of HSIs is time-consuming and costly. A transfer learning approach is used in many applications where a labeled dataset is challenging. This paper opts for the heterogeneous transfer learning models on benchmark HSIs datasets to discuss the performance accuracy of well-defined deep learning models\u2014VGG16, VGG19, ResNet, and DenseNet for crop classification. Also, it discusses the performance accuracy of customized 2-dimensional Convolutional neural network (2DCNN) and 3-dimensional Convolutional neural network (3DCNN) deep learning models using homogeneous transfer learning models on benchmark HSIs datasets for crop classification. The results show that although HSIs datasets contain few samples, the transfer learning models perform better with limited labeled samples. The results achieved 99% of accuracy for the Indian Pines and Pavia University dataset with 15% of labeled training samples with heterogeneous transfer learning. As per the overall accuracy, homogeneous transfer learning with 2DCNN and 3DCNN models pre-trained on the Indian Pines dataset and adjusted on the Salinas scene dataset performs far better than heterogeneous transfer learning.","doi":"10.1007\/s40808-022-01608-y","cleaned_title":"crop type classification with hyperspectral images using deep learning : a transfer learning approach","cleaned_abstract":"crop classification plays a vital role in felicitating agriculture statistics to the state and national government in decision-making. in recent years, due to advancements in remote sensing, high-resolution hyperspectral images (hsis) are available for land cover classification. hsis can classify the different crop categories precisely due to their narrow and continuous spectral band reflection. with improvements in computing power and evolution in deep learning technology, deep learning is rapidly being used for hsis classification. however, to train deep neural networks, many labeled samples are needed. the labeling of hsis is time-consuming and costly. a transfer learning approach is used in many applications where a labeled dataset is challenging. this paper opts for the heterogeneous transfer learning models on benchmark hsis datasets to discuss the performance accuracy of well-defined deep learning models\u2014vgg16, vgg19, resnet, and densenet for crop classification. also, it discusses the performance accuracy of customized 2-dimensional convolutional neural network (2dcnn) and 3-dimensional convolutional neural network (3dcnn) deep learning models using homogeneous transfer learning models on benchmark hsis datasets for crop classification. the results show that although hsis datasets contain few samples, the transfer learning models perform better with limited labeled samples. the results achieved 99% of accuracy for the indian pines and pavia university dataset with 15% of labeled training samples with heterogeneous transfer learning. as per the overall accuracy, homogeneous transfer learning with 2dcnn and 3dcnn models pre-trained on the indian pines dataset and adjusted on the salinas scene dataset performs far better than heterogeneous transfer learning.","key_phrases":["crop classification","a vital role","agriculture statistics","the state and national government","decision-making","recent years","advancements","remote sensing","high-resolution hyperspectral images","hsis","land cover classification","hsis","the different crop categories","their narrow and continuous spectral band reflection","improvements","computing power","evolution","deep learning technology","deep learning","hsis classification","deep neural networks","many labeled samples","the labeling","hsis","a transfer learning approach","many applications","a labeled dataset","the heterogeneous transfer learning models","benchmark hsis datasets","the performance accuracy","well-defined deep learning models","vgg16","vgg19","resnet","densenet","crop classification","it","the performance accuracy","customized 2-dimensional convolutional neural network","3-dimensional convolutional neural network (3dcnn) deep learning models","homogeneous transfer learning models","benchmark hsis datasets","crop classification","the results","hsis datasets","few samples","the transfer learning models","limited labeled samples","the results","99%","accuracy","the indian pines","pavia university","15%","labeled training samples","heterogeneous transfer learning","the overall accuracy","3dcnn models","the indian pines","the salinas scene dataset","heterogeneous transfer learning","recent years","2","2dcnn","3","3dcnn","99%","indian","15%","2dcnn","3dcnn","indian"]},{"title":"Novel deep learning models for yoga pose estimator","authors":["Amira Samy Talaat"],"abstract":"Yoga pose recognition and correction are artificial intelligent techniques to provide standardized and appropriate yoga poses. Incorrect yoga poses can cause serious injuries and long-term complications. Analyzing human posture can identify and rectify abnormal positions, improving well-being at home. A posture estimator extracts yoga asana attributes from properly represented images. These extracted features are then utilized directly as inputs for various neural networks and machine learning models. These models serve the purpose of evaluating and predicting the accuracy of specific yoga poses. The objective of this research is to explore multiple methods for classifying yoga poses. The LGDeep model is introduced, which combines a novel residual convolutional neural network with three deep learning approaches: Xception, VGGNet, and SqueezeNet. Additionally, the LGDeep model incorporates feature extraction methods such as LDA and GDA. Experimental results demonstrate that the LGDeep classifier outperforms other approaches and achieves the highest classification accuracy ratio.","doi":"10.1007\/s42452-023-05581-8","cleaned_title":"novel deep learning models for yoga pose estimator","cleaned_abstract":"yoga pose recognition and correction are artificial intelligent techniques to provide standardized and appropriate yoga poses. incorrect yoga poses can cause serious injuries and long-term complications. analyzing human posture can identify and rectify abnormal positions, improving well-being at home. a posture estimator extracts yoga asana attributes from properly represented images. these extracted features are then utilized directly as inputs for various neural networks and machine learning models. these models serve the purpose of evaluating and predicting the accuracy of specific yoga poses. the objective of this research is to explore multiple methods for classifying yoga poses. the lgdeep model is introduced, which combines a novel residual convolutional neural network with three deep learning approaches: xception, vggnet, and squeezenet. additionally, the lgdeep model incorporates feature extraction methods such as lda and gda. experimental results demonstrate that the lgdeep classifier outperforms other approaches and achieves the highest classification accuracy ratio.","key_phrases":["yoga","recognition","correction","artificial intelligent techniques","standardized and appropriate yoga poses","incorrect yoga poses","serious injuries","long-term complications","human posture","abnormal positions","well-being","home","a posture estimator","yoga asana attributes","properly represented images","these extracted features","various neural networks","machine learning models","these models","the purpose","the accuracy","specific yoga poses","the objective","this research","multiple methods","yoga poses","the lgdeep model","which","a novel residual convolutional neural network","three deep learning approaches","xception","vggnet","squeezenet","the lgdeep model","feature extraction methods","lda","gda","experimental results","the lgdeep classifier","other approaches","the highest classification accuracy ratio","three"]},{"title":"GyroFlow+: Gyroscope-Guided Unsupervised Deep Homography and Optical Flow Learning","authors":["Haipeng Li","Kunming Luo","Bing Zeng","Shuaicheng Liu"],"abstract":"Existing homography and optical flow methods are erroneous in challenging scenes, such as fog, rain, night, and snow because the basic assumptions such as brightness and gradient constancy are broken. To address this issue, we present an unsupervised learning approach that fuses gyroscope into homography and optical flow learning. Specifically, we first convert gyroscope readings into motion fields named gyro field. Second, we design a self-guided fusion module (SGF) to fuse the background motion extracted from the gyro field with the optical flow and guide the network to focus on motion details. Meanwhile, we propose a homography decoder module (HD) to combine gyro field and intermediate results of SGF to produce the homography. To the best of our knowledge, this is the first deep learning framework that fuses gyroscope data and image content for both deep homography and optical flow learning. To validate our method, we propose a new dataset that covers regular and challenging scenes. Experiments show that our method outperforms the state-of-the-art methods in both regular and challenging scenes. The code and dataset are available at https:\/\/github.com\/lhaippp\/GyroFlowPlus.","doi":"10.1007\/s11263-023-01978-5","cleaned_title":"gyroflow+: gyroscope-guided unsupervised deep homography and optical flow learning","cleaned_abstract":"existing homography and optical flow methods are erroneous in challenging scenes, such as fog, rain, night, and snow because the basic assumptions such as brightness and gradient constancy are broken. to address this issue, we present an unsupervised learning approach that fuses gyroscope into homography and optical flow learning. specifically, we first convert gyroscope readings into motion fields named gyro field. second, we design a self-guided fusion module (sgf) to fuse the background motion extracted from the gyro field with the optical flow and guide the network to focus on motion details. meanwhile, we propose a homography decoder module (hd) to combine gyro field and intermediate results of sgf to produce the homography. to the best of our knowledge, this is the first deep learning framework that fuses gyroscope data and image content for both deep homography and optical flow learning. to validate our method, we propose a new dataset that covers regular and challenging scenes. experiments show that our method outperforms the state-of-the-art methods in both regular and challenging scenes. the code and dataset are available at https:\/\/github.com\/lhaippp\/gyroflowplus.","key_phrases":["existing homography","optical flow methods","challenging scenes","fog","rain","night","snow","the basic assumptions","brightness and gradient constancy","this issue","we","an unsupervised learning approach","that","fuses","homography","optical flow learning","we","gyroscope readings","motion fields","gyro field","we","a self-guided fusion module","sgf","the background motion","the gyro field","the optical flow","the network","motion details","we","a homography decoder module","hd","gyro field","intermediate results","sgf","the homography","our knowledge","this","the first deep learning framework","that","data","image content","both deep homography","optical flow learning","our method","we","a new dataset","that","regular and challenging scenes","experiments","our method","the-art","both regular and challenging scenes","the code","dataset","https:\/\/github.com\/lhaippp\/gyroflowplus","night","first","second","first","deep homography"]},{"title":"An effective facial spoofing detection approach based on weighted deep ensemble learning","authors":["My Abdelouahed Sabri","Assia Ennouni","Abdellah Aarab"],"abstract":"Deep learning has seen successful implementation in various domains, such as natural language processing, image classification, and object detection in recent times. In the field of biometrics, deep learning has also been used to develop effective anti-spoofing systems. Facial spoofing, the act of presenting fake facial information to deceive a biometric system, poses a significant threat to the security of face recognition systems. To address this challenge, we propose, in this paper, an effective and robust facial spoofing detection approach based on weighted deep ensemble learning. Our method combines the strengths of two powerful deep learning architectures, DenseNet201 and MiniVGG. The choice of these two architectures is based on a comparative study between DenseNet201, DenseNet169, VGG16, MiniVGG, and ResNet50, where DenseNet201 and MiniVGG obtained the best recall and precision scores, respectively. Our proposed weighted voting ensemble leverages each architecture-specific capabilities to make the final prediction. We assign weights to each classification model based on its performance, which are determined by a mathematical formulation considering the trade-off between recall and precision. To validate the effectiveness of our proposed approach, we evaluate it on the challenging ROSE-Youtu face liveness detection dataset. Our experimental results demonstrate that our proposed method achieves an impressive accuracy rate of 99% in accurately detecting facial spoofing attacks.","doi":"10.1007\/s11760-023-02818-2","cleaned_title":"an effective facial spoofing detection approach based on weighted deep ensemble learning","cleaned_abstract":"deep learning has seen successful implementation in various domains, such as natural language processing, image classification, and object detection in recent times. in the field of biometrics, deep learning has also been used to develop effective anti-spoofing systems. facial spoofing, the act of presenting fake facial information to deceive a biometric system, poses a significant threat to the security of face recognition systems. to address this challenge, we propose, in this paper, an effective and robust facial spoofing detection approach based on weighted deep ensemble learning. our method combines the strengths of two powerful deep learning architectures, densenet201 and minivgg. the choice of these two architectures is based on a comparative study between densenet201, densenet169, vgg16, minivgg, and resnet50, where densenet201 and minivgg obtained the best recall and precision scores, respectively. our proposed weighted voting ensemble leverages each architecture-specific capabilities to make the final prediction. we assign weights to each classification model based on its performance, which are determined by a mathematical formulation considering the trade-off between recall and precision. to validate the effectiveness of our proposed approach, we evaluate it on the challenging rose-youtu face liveness detection dataset. our experimental results demonstrate that our proposed method achieves an impressive accuracy rate of 99% in accurately detecting facial spoofing attacks.","key_phrases":["deep learning","successful implementation","various domains","natural language processing","image classification","object","detection","recent times","the field","biometrics","deep learning","effective anti-spoofing systems","facial spoofing","the act","fake facial information","a biometric system","a significant threat","the security","face recognition systems","this challenge","we","this paper","an effective and robust facial spoofing detection approach","weighted deep ensemble learning","our method","the strengths","two powerful deep learning architectures","densenet201","the choice","these two architectures","a comparative study","densenet201","densenet169","vgg16","minivgg","resnet50","densenet201","minivgg","the best recall","precision scores","voting ensemble leverages","each architecture-specific capabilities","the final prediction","we","weights","each classification model","its performance","which","a mathematical formulation","the trade-off","recall","precision","the effectiveness","our proposed approach","we","it","the challenging rose-youtu","liveness detection dataset","our experimental results","our proposed method","an impressive accuracy rate","99%","facial spoofing attacks","two","two","resnet50","99%"]},{"title":"Deep active learning with high structural discriminability for molecular mutagenicity prediction","authors":["Huiyan Xu","Yanpeng Zhao","Yixin Zhang","Junshan Han","Peng Zan","Song He","Xiaochen Bo"],"abstract":"The assessment of mutagenicity is essential in drug discovery, as it may lead to cancer and germ cells damage. Although in silico methods have been proposed for mutagenicity prediction, their performance is hindered by the scarcity of labeled molecules. However, experimental mutagenicity testing can be time-consuming and costly. One solution to reduce the annotation cost is active learning, where the algorithm actively selects the most valuable molecules from a vast chemical space and presents them to the oracle (e.g., a human expert) for annotation, thereby rapidly improving the model\u2019s predictive performance with a smaller annotation cost. In this paper, we propose muTOX-AL, a deep active learning framework, which can actively explore the chemical space and identify the most valuable molecules, resulting in competitive performance with a small number of labeled samples. The experimental results show that, compared to the random sampling strategy, muTOX-AL can reduce the number of training molecules by about 57%. Additionally, muTOX-AL exhibits outstanding molecular structural discriminability, allowing it to pick molecules with high structural similarity but opposite properties.","doi":"10.1038\/s42003-024-06758-6","cleaned_title":"deep active learning with high structural discriminability for molecular mutagenicity prediction","cleaned_abstract":"the assessment of mutagenicity is essential in drug discovery, as it may lead to cancer and germ cells damage. although in silico methods have been proposed for mutagenicity prediction, their performance is hindered by the scarcity of labeled molecules. however, experimental mutagenicity testing can be time-consuming and costly. one solution to reduce the annotation cost is active learning, where the algorithm actively selects the most valuable molecules from a vast chemical space and presents them to the oracle (e.g., a human expert) for annotation, thereby rapidly improving the model\u2019s predictive performance with a smaller annotation cost. in this paper, we propose mutox-al, a deep active learning framework, which can actively explore the chemical space and identify the most valuable molecules, resulting in competitive performance with a small number of labeled samples. the experimental results show that, compared to the random sampling strategy, mutox-al can reduce the number of training molecules by about 57%. additionally, mutox-al exhibits outstanding molecular structural discriminability, allowing it to pick molecules with high structural similarity but opposite properties.","key_phrases":["the assessment","mutagenicity","drug discovery","it","silico methods","mutagenicity prediction","their performance","the scarcity","labeled molecules","experimental mutagenicity testing","one solution","the annotation cost","active learning","the algorithm","the most valuable molecules","a vast chemical space","them","the oracle","e.g., a human expert","annotation","the model\u2019s predictive performance","a smaller annotation cost","this paper","we","mutox-al","a deep active learning framework","which","the chemical space","the most valuable molecules","competitive performance","a small number","labeled samples","the experimental results","the random sampling strategy","mutox-al","the number","training molecules","about 57%","mutox-al","outstanding molecular structural discriminability","it","molecules","high structural similarity","opposite properties","one","mutox-al","about 57%"]},{"title":"A privacy-preserving approach for detecting smishing attacks using federated deep learning","authors":["Mohamed Abdelkarim Remmide","Fatima Boumahdi","Bousmaha Ilhem","Narhimene Boustia"],"abstract":"Smishing is a type of social engineering attack that involves sending fraudulent SMS messages to trick recipients into revealing sensitive information. In recent years, it has become a significant threat to mobile communications. In this study, we introduce a novel smishing detection method based on federated learning, which is a decentralized approach ensuring data privacy. We develop a robust detection model within a federated learning framework based on deep learning methods such as Long Short-Term Memory (LSTM) and Bidirectional LSTM (Bi-LSTM). Our experiments show that the federated learning method using Bi-LSTM achieves an accuracy of 88.78%, highlighting its effectiveness in tackling smishing detection while preserving user privacy. This approach not only offers a promising solution to smishing attacks but also lays the groundwork for future research in mobile security and privacy-preserving machine learning.","doi":"10.1007\/s41870-024-02144-x","cleaned_title":"a privacy-preserving approach for detecting smishing attacks using federated deep learning","cleaned_abstract":"smishing is a type of social engineering attack that involves sending fraudulent sms messages to trick recipients into revealing sensitive information. in recent years, it has become a significant threat to mobile communications. in this study, we introduce a novel smishing detection method based on federated learning, which is a decentralized approach ensuring data privacy. we develop a robust detection model within a federated learning framework based on deep learning methods such as long short-term memory (lstm) and bidirectional lstm (bi-lstm). our experiments show that the federated learning method using bi-lstm achieves an accuracy of 88.78%, highlighting its effectiveness in tackling smishing detection while preserving user privacy. this approach not only offers a promising solution to smishing attacks but also lays the groundwork for future research in mobile security and privacy-preserving machine learning.","key_phrases":["a type","social engineering attack","that","fraudulent sms messages","recipients","sensitive information","recent years","it","a significant threat","mobile communications","this study","we","a novel","detection method","federated learning","which","a decentralized approach","data privacy","we","a robust detection model","a federated learning framework","deep learning methods","long short-term memory","lstm","bidirectional lstm","bi","-","lstm","our experiments","the federated learning method","bi","-","lstm","an accuracy","88.78%","its effectiveness","detection","user privacy","this approach","a promising solution","attacks","the groundwork","future research","mobile security","privacy-preserving machine learning","recent years","88.78%"]},{"title":"SlumberNet: deep learning classification of sleep stages using residual neural networks","authors":["Pawan K. Jha","Utham K. Valekunja","Akhilesh B. Reddy"],"abstract":"Sleep research is fundamental to understanding health and well-being, as proper sleep is essential for maintaining optimal physiological function. Here we present SlumberNet, a novel deep learning model based on residual network (ResNet) architecture, designed to classify sleep states in mice using electroencephalogram (EEG) and electromyogram (EMG) signals. Our model was trained and tested on data from mice undergoing baseline sleep, sleep deprivation, and recovery sleep, enabling it to handle a wide range of sleep conditions. Employing k-fold cross-validation and data augmentation techniques, SlumberNet achieved high levels of overall performance (accuracy\u2009=\u200997%; F1 score\u2009=\u200996%) in predicting sleep stages and showed robust performance even with a small and diverse training dataset. Comparison of SlumberNet's performance to manual sleep stage classification revealed a significant reduction in analysis time (~\u200950\u2009\u00d7\u2009faster), without sacrificing accuracy. Our study showcases the potential of deep learning to facilitate sleep research by providing a more efficient, accurate, and scalable method for sleep stage classification. Our work with SlumberNet further demonstrates the power of deep learning in mouse sleep research.","doi":"10.1038\/s41598-024-54727-0","cleaned_title":"slumbernet: deep learning classification of sleep stages using residual neural networks","cleaned_abstract":"sleep research is fundamental to understanding health and well-being, as proper sleep is essential for maintaining optimal physiological function. here we present slumbernet, a novel deep learning model based on residual network (resnet) architecture, designed to classify sleep states in mice using electroencephalogram (eeg) and electromyogram (emg) signals. our model was trained and tested on data from mice undergoing baseline sleep, sleep deprivation, and recovery sleep, enabling it to handle a wide range of sleep conditions. employing k-fold cross-validation and data augmentation techniques, slumbernet achieved high levels of overall performance (accuracy = 97%; f1 score = 96%) in predicting sleep stages and showed robust performance even with a small and diverse training dataset. comparison of slumbernet's performance to manual sleep stage classification revealed a significant reduction in analysis time (~ 50 \u00d7 faster), without sacrificing accuracy. our study showcases the potential of deep learning to facilitate sleep research by providing a more efficient, accurate, and scalable method for sleep stage classification. our work with slumbernet further demonstrates the power of deep learning in mouse sleep research.","key_phrases":["sleep research","health","well-being","proper sleep","optimal physiological function","we","slumbernet","a novel deep learning model","residual network","resnet) architecture","sleep states","mice","eeg","emg","our model","data","mice","baseline sleep","sleep deprivation","recovery sleep","it","a wide range","sleep conditions","k-fold cross-validation and data augmentation techniques","slumbernet","high levels","overall performance","f1 score","sleep stages","robust performance","a small and diverse training dataset","comparison","slumbernet's performance","manual sleep stage classification","a significant reduction","analysis time","accuracy","our study","the potential","deep learning","sleep research","a more efficient, accurate, and scalable method","sleep stage classification","our work","slumbernet","the power","deep learning","mouse sleep research","97%","96%","50"]},{"title":"A comparative analysis and classification of cancerous brain tumors detection based on classical machine learning and deep transfer learning models","authors":["Yajuvendra Pratap Singh","D.K Lobiyal"],"abstract":"Brain tumor can be fatal for human life. Therefore, proper and timely diagnosis and treatment is important to save human lives. The similarity and variety between the normal and tumor tissues make it difficult for diagnosis through human assisted techniques. In the recent past, machine learning and deep learning techniques have been applied for the classification and segmentation of brain tumor. These techniques have shown promising results by improving the accuracy of classification and segmentation. In this paper, we proposed to implement various classical machine-learning techniques support vector machine, Naive Bayes classifier, K- Nearest Neighbor, random forest, and deep-learning CNN-based models Xception, Inceptionv3, VGG19, and DenseNet201 techniques to classify gliomas, meningiomas, and pituitary tumors and compare their performance. Further, we proposed to modify Xception model to improve the performance of classification and segmentation. In this research, we used the standard Figshare dataset consisting of 3064 images of size \\(112\\times 112\\) each. The performance of these models is measured and compared in terms of precision, recall, F1-score, and accuracy. The classical Machine learning model gives scores varying from 88% to 93%. For the above four metrics. However, all the deep learning models give their scores of more than 96% for the above metrics. Our proposed modified Xception model gives scores more than 98% in all the above four metrics, which is comparable to the best scores reported in the literature.","doi":"10.1007\/s11042-023-16637-7","cleaned_title":"a comparative analysis and classification of cancerous brain tumors detection based on classical machine learning and deep transfer learning models","cleaned_abstract":"brain tumor can be fatal for human life. therefore, proper and timely diagnosis and treatment is important to save human lives. the similarity and variety between the normal and tumor tissues make it difficult for diagnosis through human assisted techniques. in the recent past, machine learning and deep learning techniques have been applied for the classification and segmentation of brain tumor. these techniques have shown promising results by improving the accuracy of classification and segmentation. in this paper, we proposed to implement various classical machine-learning techniques support vector machine, naive bayes classifier, k- nearest neighbor, random forest, and deep-learning cnn-based models xception, inceptionv3, vgg19, and densenet201 techniques to classify gliomas, meningiomas, and pituitary tumors and compare their performance. further, we proposed to modify xception model to improve the performance of classification and segmentation. in this research, we used the standard figshare dataset consisting of 3064 images of size \\(112\\times 112\\) each. the performance of these models is measured and compared in terms of precision, recall, f1-score, and accuracy. the classical machine learning model gives scores varying from 88% to 93%. for the above four metrics. however, all the deep learning models give their scores of more than 96% for the above metrics. our proposed modified xception model gives scores more than 98% in all the above four metrics, which is comparable to the best scores reported in the literature.","key_phrases":["brain tumor","human life","proper and timely diagnosis","treatment","human lives","the similarity","variety","the normal and tumor tissues","it","diagnosis","human assisted techniques","the recent past","machine learning","deep learning techniques","the classification","segmentation","brain tumor","these techniques","promising results","the accuracy","classification","segmentation","this paper","we","various classical machine-learning techniques support vector machine","naive bayes classifier","neighbor","random forest","deep-learning cnn-based models xception","inceptionv3","vgg19","densenet201 techniques","gliomas","meningiomas","pituitary tumors","their performance","we","xception model","the performance","classification","segmentation","this research","we","the standard figshare dataset","3064 images","size","\\(112\\times 112\\","each","the performance","these models","terms","precision","f1-score","accuracy","the classical machine learning model","scores","88%","to 93%","the above four metrics","all the deep learning models","their scores","more than 96%","the above metrics","our proposed modified xception model","scores","more than 98%","all the above four metrics","which","the best scores","the literature","k-","cnn","inceptionv3","meningiomas","3064","112\\","88% to 93%","four","more than 96%","more than 98%","four"]},{"title":"Implementing Deep Learning-Based Intelligent Inspection for Investment Castings","authors":["Nabhan Yousef","Amit Sata"],"abstract":"In this study, a user-friendly intelligent inspection device was developed employing deep learning techniques to effectively detect and characterize surface defects in investment castings. The inspection techniques encompassed a range from basic visual inspection to more specialized methods like liquid penetration and magnetic particle tests. The developed device leveraged convolutional neural networks (CNN), residual neural networks (ResNet), and recurrent neural networks (R-CNN) models, with a dataset of 3600 images depicting industrial castings, both defective and defect-free. A majority of the dataset was allocated for training (approximately 80%), while the remainder was used for testing. Among the deep learning models considered, ResNet exhibited superior accuracy in defect inspection, trailed by CNN and R-CNN. The resultant intelligent inspection device was successfully implemented in an industrial setting, enabling efficient defect identification in investment castings through the trained ResNet model.","doi":"10.1007\/s13369-023-08240-7","cleaned_title":"implementing deep learning-based intelligent inspection for investment castings","cleaned_abstract":"in this study, a user-friendly intelligent inspection device was developed employing deep learning techniques to effectively detect and characterize surface defects in investment castings. the inspection techniques encompassed a range from basic visual inspection to more specialized methods like liquid penetration and magnetic particle tests. the developed device leveraged convolutional neural networks (cnn), residual neural networks (resnet), and recurrent neural networks (r-cnn) models, with a dataset of 3600 images depicting industrial castings, both defective and defect-free. a majority of the dataset was allocated for training (approximately 80%), while the remainder was used for testing. among the deep learning models considered, resnet exhibited superior accuracy in defect inspection, trailed by cnn and r-cnn. the resultant intelligent inspection device was successfully implemented in an industrial setting, enabling efficient defect identification in investment castings through the trained resnet model.","key_phrases":["this study","a user-friendly intelligent inspection device","deep learning techniques","surface defects","investment castings","the inspection techniques","a range","basic visual inspection","more specialized methods","liquid penetration","magnetic particle tests","the developed device leveraged convolutional neural networks","cnn","residual neural networks","resnet","recurrent neural networks","r-cnn) models","a dataset","3600 images","industrial castings","a majority","the dataset","training","approximately 80%","the remainder","testing","the deep learning models","resnet","superior accuracy","defect inspection","cnn","r-cnn","the resultant intelligent inspection device","an industrial setting","efficient defect identification","investment castings","the trained resnet model","cnn","3600","approximately 80%","cnn"]},{"title":"Revealing the mechanisms of semantic satiation with deep learning models","authors":["Xinyu Zhang","Jing Lian","Zhaofei Yu","Huajin Tang","Dong Liang","Jizhao Liu","Jian K. Liu"],"abstract":"The phenomenon of semantic satiation, which refers to the loss of meaning of a word or phrase after being repeated many times, is a well-known psychological phenomenon. However, the microscopic neural computational principles responsible for these mechanisms remain unknown. In this study, we use a deep learning model of continuous coupled neural networks to investigate the mechanism underlying semantic satiation and precisely describe this process with neuronal components. Our results suggest that, from a mesoscopic perspective, semantic satiation may be a bottom-up process. Unlike existing macroscopic psychological studies that suggest that semantic satiation is a top-down process, our simulations use a similar experimental paradigm as classical psychology experiments and observe similar results. Satiation of semantic objectives, similar to the learning process of our network model used for object recognition, relies on continuous learning and switching between objects. The underlying neural coupling strengthens or weakens satiation. Taken together, both neural and network mechanisms play a role in controlling semantic satiation.","doi":"10.1038\/s42003-024-06162-0","cleaned_title":"revealing the mechanisms of semantic satiation with deep learning models","cleaned_abstract":"the phenomenon of semantic satiation, which refers to the loss of meaning of a word or phrase after being repeated many times, is a well-known psychological phenomenon. however, the microscopic neural computational principles responsible for these mechanisms remain unknown. in this study, we use a deep learning model of continuous coupled neural networks to investigate the mechanism underlying semantic satiation and precisely describe this process with neuronal components. our results suggest that, from a mesoscopic perspective, semantic satiation may be a bottom-up process. unlike existing macroscopic psychological studies that suggest that semantic satiation is a top-down process, our simulations use a similar experimental paradigm as classical psychology experiments and observe similar results. satiation of semantic objectives, similar to the learning process of our network model used for object recognition, relies on continuous learning and switching between objects. the underlying neural coupling strengthens or weakens satiation. taken together, both neural and network mechanisms play a role in controlling semantic satiation.","key_phrases":["the phenomenon","semantic satiation","which","the loss","meaning","a word","phrase","a well-known psychological phenomenon","the microscopic neural computational principles","these mechanisms","this study","we","a deep learning model","continuous coupled neural networks","the mechanism","semantic satiation","this process","neuronal components","our results","a mesoscopic perspective","semantic satiation","a bottom-up process","existing macroscopic psychological studies","that","semantic satiation","a top-down process","our simulations","a similar experimental paradigm","classical psychology experiments","similar results","satiation","semantic objectives","the learning process","our network model","object recognition","continuous learning","switching","objects","the underlying neural coupling strengthens","satiation","both neural and network mechanisms","a role","semantic satiation"]},{"title":"Incremental\u2013decremental data transformation based ensemble deep learning model (IDT-eDL) for temperature prediction","authors":["Vipin Kumar","Rana Kumar"],"abstract":"Human life heavily depends on weather conditions, which affect the necessary operations like agriculture, aviation, tourism, industries, etc., where the temperature plays a vital role in deciding the weather conditions along with other meteorological variables. Therefore, temperature forecasting has drawn considerable attention from researchers because of its significant effect on daily life activities and the ever-challenging forecasting task. These research objectives are to investigate the transformation of data based on incremental and decremental approaches and to find the practical ensemble approach over proposed models for effective temperature prediction, where the proposed model is called the Incremental\u2013Decremental Data Transformation-Based Ensemble Deep Learning Model (IDT-eDL). The temperature dataset from Delhi, India, has been utilized to compare proposed and traditional deep learning models over various performance measures. The proposed IDT-eDL with BiLSTM deep learning model (i.e., IDT-eDL_BiLSTM ) has performed the best among the proposed models and traditional deep learning model and achieved Performance over measures MSE: 1.36, RMSE: 1.16, MAE: 0.89, MAPE: 4.13 and \\(R^2\\):0.999. Additionally, non-parametric statistical analysis of Friedman ranking is also performed to validate the effectiveness of the proposed IDT-eDL model, which also shows a higher ranking of the proposed model than the traditional deep learning models.","doi":"10.1007\/s40808-024-01953-0","cleaned_title":"incremental\u2013decremental data transformation based ensemble deep learning model (idt-edl) for temperature prediction","cleaned_abstract":"human life heavily depends on weather conditions, which affect the necessary operations like agriculture, aviation, tourism, industries, etc., where the temperature plays a vital role in deciding the weather conditions along with other meteorological variables. therefore, temperature forecasting has drawn considerable attention from researchers because of its significant effect on daily life activities and the ever-challenging forecasting task. these research objectives are to investigate the transformation of data based on incremental and decremental approaches and to find the practical ensemble approach over proposed models for effective temperature prediction, where the proposed model is called the incremental\u2013decremental data transformation-based ensemble deep learning model (idt-edl). the temperature dataset from delhi, india, has been utilized to compare proposed and traditional deep learning models over various performance measures. the proposed idt-edl with bilstm deep learning model (i.e., idt-edl_bilstm ) has performed the best among the proposed models and traditional deep learning model and achieved performance over measures mse: 1.36, rmse: 1.16, mae: 0.89, mape: 4.13 and \\(r^2\\):0.999. additionally, non-parametric statistical analysis of friedman ranking is also performed to validate the effectiveness of the proposed idt-edl model, which also shows a higher ranking of the proposed model than the traditional deep learning models.","key_phrases":["human life","weather conditions","which","the necessary operations","agriculture","aviation","tourism","industries","the temperature","a vital role","the weather conditions","other meteorological variables","temperature forecasting","considerable attention","researchers","its significant effect","daily life activities","the ever-challenging forecasting task","these research objectives","the transformation","data","incremental and decremental approaches","the practical ensemble approach","proposed models","effective temperature prediction","the proposed model","the incremental\u2013decremental data transformation-based ensemble deep learning model","idt-edl","the temperature","delhi","india","proposed and traditional deep learning models","various performance measures","the proposed idt-edl","bilstm deep learning model","i.e., idt-edl_bilstm","the proposed models","traditional deep learning model","performance","measures mse","rmse","mape","\\(r^2\\):0.999","non-parametric statistical analysis","friedman ranking","the effectiveness","the proposed idt-edl model","which","a higher ranking","the proposed model","the traditional deep learning models","delhi","india","1.36","rmse","1.16","0.89","4.13"]},{"title":"Deep Learning Framework for Predicting Essential Proteins with Temporal Convolutional Networks","authors":["Pengli Lu\u00a0\n            (\u5362\u9e4f\u4e3d)","Peishi Yang\u00a0\n            (\u6768\u57f9\u5b9e)","Yonggang Liao\u00a0\n            (\u5ed6\u6c38\u521a)"],"abstract":"Essential proteins are an indispensable part of cells and play an extremely significant role in genetic disease diagnosis and drug development. Therefore, the prediction of essential proteins has received extensive attention from researchers. Many centrality methods and machine learning algorithms have been proposed to predict essential proteins. Nevertheless, the topological characteristics learned by the centrality method are not comprehensive enough, resulting in low accuracy. In addition, machine learning algorithms need sufficient prior knowledge to select features, and the ability to solve imbalanced classification problems needs to be further strengthened. These two factors greatly affect the performance of predicting essential proteins. In this paper, we propose a deep learning framework based on temporal convolutional networks to predict essential proteins by integrating gene expression data and protein-protein interaction (PPI) network. We make use of the method of network embedding to automatically learn more abundant features of proteins in the PPI network. For gene expression data, we treat it as sequence data, and use temporal convolutional networks to extract sequence features. Finally, the two types of features are integrated and put into the multi-layer neural network to complete the final classification task. The performance of our method is evaluated by comparing with seven centrality methods, six machine learning algorithms, and two deep learning models. The results of the experiment show that our method is more effective than the comparison methods for predicting essential proteins.","doi":"10.1007\/s12204-023-2632-9","cleaned_title":"deep learning framework for predicting essential proteins with temporal convolutional networks","cleaned_abstract":"essential proteins are an indispensable part of cells and play an extremely significant role in genetic disease diagnosis and drug development. therefore, the prediction of essential proteins has received extensive attention from researchers. many centrality methods and machine learning algorithms have been proposed to predict essential proteins. nevertheless, the topological characteristics learned by the centrality method are not comprehensive enough, resulting in low accuracy. in addition, machine learning algorithms need sufficient prior knowledge to select features, and the ability to solve imbalanced classification problems needs to be further strengthened. these two factors greatly affect the performance of predicting essential proteins. in this paper, we propose a deep learning framework based on temporal convolutional networks to predict essential proteins by integrating gene expression data and protein-protein interaction (ppi) network. we make use of the method of network embedding to automatically learn more abundant features of proteins in the ppi network. for gene expression data, we treat it as sequence data, and use temporal convolutional networks to extract sequence features. finally, the two types of features are integrated and put into the multi-layer neural network to complete the final classification task. the performance of our method is evaluated by comparing with seven centrality methods, six machine learning algorithms, and two deep learning models. the results of the experiment show that our method is more effective than the comparison methods for predicting essential proteins.","key_phrases":["essential proteins","an indispensable part","cells","an extremely significant role","genetic disease diagnosis and drug development","the prediction","essential proteins","extensive attention","researchers","many centrality methods","machine learning algorithms","essential proteins","the topological characteristics","the centrality method","low accuracy","addition","machine learning algorithms","sufficient prior knowledge","features","the ability","imbalanced classification problems","these two factors","the performance","essential proteins","this paper","we","a deep learning framework","temporal convolutional networks","essential proteins","gene expression data","protein-protein interaction","(ppi) network","we","use","the method","network","more abundant features","proteins","the ppi network","gene expression data","we","it","sequence data","temporal convolutional networks","sequence features","the two types","features","the multi-layer neural network","the final classification task","the performance","our method","seven centrality methods","six machine learning algorithms","two deep learning models","the results","the experiment","our method","the comparison methods","essential proteins","two","seven","six","two"]},{"title":"Unsupervised deep representation learning enables phenotype discovery for genetic association studies of brain imaging","authors":["Khush Patel","Ziqian Xie","Hao Yuan","Sheikh Muhammad Saiful Islam","Yaochen Xie","Wei He","Wanheng Zhang","Assaf Gottlieb","Han Chen","Luca Giancardo","Alexander Knaack","Evan Fletcher","Myriam Fornage","Shuiwang Ji","Degui Zhi"],"abstract":"Understanding the genetic architecture of brain structure is challenging, partly due to difficulties in designing robust, non-biased descriptors of brain morphology. Until recently, brain measures for genome-wide association studies (GWAS) consisted of traditionally expert-defined or software-derived image-derived phenotypes (IDPs) that are often based on theoretical preconceptions or computed from limited amounts of data. Here, we present an approach to derive brain imaging phenotypes using unsupervised deep representation learning. We train a 3-D convolutional autoencoder model with reconstruction loss on 6130 UK Biobank (UKBB) participants\u2019 T1 or T2-FLAIR (T2) brain MRIs to create a 128-dimensional representation known as Unsupervised Deep learning derived Imaging Phenotypes (UDIPs). GWAS of these UDIPs in held-out UKBB subjects (n\u2009=\u200922,880 discovery and n\u2009=\u200912,359\/11,265 replication cohorts for T1\/T2) identified 9457 significant SNPs organized into 97 independent genetic loci of which 60 loci were replicated. Twenty-six loci were not reported in earlier T1 and T2 IDP-based UK Biobank GWAS. We developed a perturbation-based decoder interpretation approach to show that these loci are associated with UDIPs mapped to multiple relevant brain regions. Our results established unsupervised deep learning can derive robust, unbiased, heritable, and interpretable brain imaging phenotypes.","doi":"10.1038\/s42003-024-06096-7","cleaned_title":"unsupervised deep representation learning enables phenotype discovery for genetic association studies of brain imaging","cleaned_abstract":"understanding the genetic architecture of brain structure is challenging, partly due to difficulties in designing robust, non-biased descriptors of brain morphology. until recently, brain measures for genome-wide association studies (gwas) consisted of traditionally expert-defined or software-derived image-derived phenotypes (idps) that are often based on theoretical preconceptions or computed from limited amounts of data. here, we present an approach to derive brain imaging phenotypes using unsupervised deep representation learning. we train a 3-d convolutional autoencoder model with reconstruction loss on 6130 uk biobank (ukbb) participants\u2019 t1 or t2-flair (t2) brain mris to create a 128-dimensional representation known as unsupervised deep learning derived imaging phenotypes (udips). gwas of these udips in held-out ukbb subjects (n = 22,880 discovery and n = 12,359\/11,265 replication cohorts for t1\/t2) identified 9457 significant snps organized into 97 independent genetic loci of which 60 loci were replicated. twenty-six loci were not reported in earlier t1 and t2 idp-based uk biobank gwas. we developed a perturbation-based decoder interpretation approach to show that these loci are associated with udips mapped to multiple relevant brain regions. our results established unsupervised deep learning can derive robust, unbiased, heritable, and interpretable brain imaging phenotypes.","key_phrases":["the genetic architecture","brain structure","difficulties","robust, non-biased descriptors","brain morphology","brain measures","genome-wide association studies","gwas","traditionally expert-defined or software-derived image-derived phenotypes","idps","that","theoretical preconceptions","limited amounts","data","we","an approach","brain imaging phenotypes","unsupervised deep representation learning","we","a 3-d convolutional autoencoder model","reconstruction loss","6130 uk biobank (ukbb) participants\u2019 t1 or t2-flair (t2) brain mris","a 128-dimensional representation","unsupervised deep learning derived imaging phenotypes","udips","gwas","these udips","held-out ukbb subjects","n = 22,880 discovery","n = 12,359\/11,265 replication cohorts","t1\/t2","9457 significant snps","97 independent genetic loci","which","60 loci","twenty-six loci","earlier t1 and t2 idp-based uk biobank gwas","we","a perturbation-based decoder interpretation approach","these loci","udips","multiple relevant brain regions","our results","unsupervised deep learning","robust, unbiased, heritable, and interpretable brain imaging phenotypes","3","6130","mris","128","22,880","12,359\/11,265","9457","97","60","twenty-six"]},{"title":"Deep learning-based weld defect classification using VGG16 transfer learning adaptive fine-tuning","authors":["Samuel Kumaresan","K. S. Jai Aultrin","S. S. Kumar","M. Dev Anand"],"abstract":"Welding is a vital joining process; however, occurrences of weld defects often degrade the quality of the welded part. The risk of occurrence of a variety of defects has led to the development of advanced weld defects detection systems such as automated weld defects detection and classification. The present work is a novel approach that proposes and investigates a unique image-centered method based on a deep learning model trained by a small X-ray image dataset. A data augmentation method able to process images on the go was used to offset the limitation of the small X-ray dataset. Fine-tuned transfer learning techniques were used to train two convolutional neural network based architectures with VGG16 and ResNet50 as the base models for the augmented sets. Out of the networks we fine-tuned, VGG16 based model performed well with a relatively higher average accuracy of 90%. Even though the small dataset was spread across 15 different classes in an unbalanced way, the learning curves showed acceptable model generalization characteristics.","doi":"10.1007\/s12008-023-01327-3","cleaned_title":"deep learning-based weld defect classification using vgg16 transfer learning adaptive fine-tuning","cleaned_abstract":"welding is a vital joining process; however, occurrences of weld defects often degrade the quality of the welded part. the risk of occurrence of a variety of defects has led to the development of advanced weld defects detection systems such as automated weld defects detection and classification. the present work is a novel approach that proposes and investigates a unique image-centered method based on a deep learning model trained by a small x-ray image dataset. a data augmentation method able to process images on the go was used to offset the limitation of the small x-ray dataset. fine-tuned transfer learning techniques were used to train two convolutional neural network based architectures with vgg16 and resnet50 as the base models for the augmented sets. out of the networks we fine-tuned, vgg16 based model performed well with a relatively higher average accuracy of 90%. even though the small dataset was spread across 15 different classes in an unbalanced way, the learning curves showed acceptable model generalization characteristics.","key_phrases":["welding","a vital joining process","occurrences","weld defects","the quality","the welded part","the risk","occurrence","a variety","defects","the development","advanced weld defects detection systems","automated weld defects detection","classification","the present work","a novel approach","that","a unique image-centered method","a deep learning model","a small x-ray image dataset","a data augmentation method","images","the go","the limitation","the small x-ray dataset","fine-tuned transfer learning techniques","two convolutional neural network based architectures","vgg16","resnet50","the base models","the augmented sets","the networks","we","vgg16 based model","a relatively higher average accuracy","90%","the small dataset","15 different classes","an unbalanced way","the learning curves","acceptable model generalization characteristics","two","resnet50","90%","15"]},{"title":"Deep-learning-assisted online surface roughness monitoring in ultraprecision fly cutting","authors":["Adeel Shehzad","XiaoTing Rui","YuanYuan Ding","JianShu Zhang","Yu Chang","HanJing Lu","YiHeng Chen"],"abstract":"Surface roughness is one of the most critical attributes of machined components, especially those used in high-performance systems. Online surface roughness monitoring offers advancements comparable to post-process inspection methods, reducing inspection time and costs and concurrently reducing the likelihood of defects. Currently, online monitoring approaches for surface roughness are constrained by several limitations, including the reliance on handcrafted feature extraction, which necessitates the involvement of human experts and entails time-consuming processes. Moreover, the prediction models trained under one set of cutting conditions exhibit poor performance when applied to different experimental settings. To address these challenges, this work presents a novel deep-learning-assisted online surface roughness monitoring method for ultraprecision fly cutting of copper workpieces under different cutting conditions. Tooltip acceleration signals were acquired during each cutting experiment to develop two datasets, and no handcrafted features were extracted. Five deep learning models were developed and evaluated using standard performance metrics. A convolutional neural network stacked on a long short-term memory network outperformed all other network models, yielding exceptional results, including a mean absolute percentage error as low as 1.51% and anR2 value of 96.6%. Furthermore, the robustness of the proposed model was assessed via a validation cohort analysis using experimental data obtained using cutting parameters different from those previously employed. The performance of the model remained consistent and commendable under varied conditions, asserting its applicability in real-world scenarios.","doi":"10.1007\/s11431-023-2615-4","cleaned_title":"deep-learning-assisted online surface roughness monitoring in ultraprecision fly cutting","cleaned_abstract":"surface roughness is one of the most critical attributes of machined components, especially those used in high-performance systems. online surface roughness monitoring offers advancements comparable to post-process inspection methods, reducing inspection time and costs and concurrently reducing the likelihood of defects. currently, online monitoring approaches for surface roughness are constrained by several limitations, including the reliance on handcrafted feature extraction, which necessitates the involvement of human experts and entails time-consuming processes. moreover, the prediction models trained under one set of cutting conditions exhibit poor performance when applied to different experimental settings. to address these challenges, this work presents a novel deep-learning-assisted online surface roughness monitoring method for ultraprecision fly cutting of copper workpieces under different cutting conditions. tooltip acceleration signals were acquired during each cutting experiment to develop two datasets, and no handcrafted features were extracted. five deep learning models were developed and evaluated using standard performance metrics. a convolutional neural network stacked on a long short-term memory network outperformed all other network models, yielding exceptional results, including a mean absolute percentage error as low as 1.51% and anr2 value of 96.6%. furthermore, the robustness of the proposed model was assessed via a validation cohort analysis using experimental data obtained using cutting parameters different from those previously employed. the performance of the model remained consistent and commendable under varied conditions, asserting its applicability in real-world scenarios.","key_phrases":["surface roughness","the most critical attributes","machined components","especially those","high-performance systems","online surface roughness monitoring","advancements","post-process inspection methods","inspection time","costs","the likelihood","defects","online monitoring approaches","surface roughness","several limitations","the reliance","handcrafted feature extraction","which","the involvement","human experts","time-consuming processes","the prediction models","one set","conditions","poor performance","different experimental settings","these challenges","this work","a novel deep-learning-assisted online surface roughness","method","ultraprecision fly cutting","copper workpieces","different cutting conditions","tooltip acceleration signals","each cutting experiment","two datasets","no handcrafted features","five deep learning models","standard performance metrics","a convolutional neural network","a long short-term memory network","all other network models","exceptional results","a mean absolute percentage error","1.51%","anr2 value","96.6%","the robustness","the proposed model","a validation cohort analysis","experimental data","cutting parameters","those","the performance","the model","varied conditions","its applicability","real-world scenarios","two","five","as low as 1.51%","96.6%"]},{"title":"Robust Deep Learning for Accurate Landslide Identification and Prediction","authors":["T. Bhuvaneswari","R. Chandra Guru Sekar","M. Chengathir Selvi","J. Jemima Rubavathi","V. Kaviyaa"],"abstract":"AbstractLandslide is the most common natural risk in mountainous regions on all five continents and they can pose a serious threat in these areas. Strong earthquakes, unusual weather events such as storms and eruptions of volcanoes, and human-caused events such as creating roadways that crossed the slopes are the main causes of landslides and they cause significant dangers to residential properties and society as a whole. The Landslide4sense dataset is used for identifying landslides, which contains 3799 training samples and 245 testing samples. These image patches are taken from the Sentinel-2 sensor, while the slope and Digital Elevation Model (DEM) are from the ALOS PALSAR sensor. Data was gathered from four distinct geographical areas namely\u00a0Kodagu, Iburi, Taiwan, and Gorkha. We use Deep Learning (DL) models such as ResNet18, U-Net, and VGG16 to predict the landslide. By comparing the above models with the evaluation metrics like loss, precision, recall, F1 score and accuracy, ResNet18 model is selected as the best model for landslide identification.","doi":"10.1134\/S1028334X23602961","cleaned_title":"robust deep learning for accurate landslide identification and prediction","cleaned_abstract":"abstractlandslide is the most common natural risk in mountainous regions on all five continents and they can pose a serious threat in these areas. strong earthquakes, unusual weather events such as storms and eruptions of volcanoes, and human-caused events such as creating roadways that crossed the slopes are the main causes of landslides and they cause significant dangers to residential properties and society as a whole. the landslide4sense dataset is used for identifying landslides, which contains 3799 training samples and 245 testing samples. these image patches are taken from the sentinel-2 sensor, while the slope and digital elevation model (dem) are from the alos palsar sensor. data was gathered from four distinct geographical areas namely kodagu, iburi, taiwan, and gorkha. we use deep learning (dl) models such as resnet18, u-net, and vgg16 to predict the landslide. by comparing the above models with the evaluation metrics like loss, precision, recall, f1 score and accuracy, resnet18 model is selected as the best model for landslide identification.","key_phrases":["abstractlandslide","the most common natural risk","mountainous regions","all five continents","they","a serious threat","these areas","strong earthquakes","unusual weather events","storms","eruptions","volcanoes","human-caused events","roadways","that","the slopes","the main causes","landslides","they","significant dangers","residential properties","society","a whole","the landslide4sense dataset","landslides","which","3799 training samples","245 testing samples","these image patches","the sentinel-2 sensor","the slope","digital elevation model","dem","the alos palsar sensor","data","four distinct geographical areas","we","deep learning (dl) models","resnet18","u","-","net","vgg16","the landslide","the above models","the evaluation metrics","loss","precision","recall","f1 score","accuracy","resnet18 model","the best model","landslide identification","five","3799","245","dem","four","kodagu","iburi","taiwan","resnet18","resnet18"]},{"title":"DeepEPhishNet: a deep learning framework for email phishing detection using word embedding algorithms","authors":["M Somesha","Alwyn Roshan Pais"],"abstract":"Email phishing is a social engineering scheme that uses spoofed emails intended to trick the user into disclosing legitimate business and personal credentials. Many phishing email detection techniques exist based on machine learning, deep learning, and word embedding. In this paper, we propose a new technique for the detection of phishing emails using word embedding (Word2Vec, FastText, and TF-IDF) and deep learning techniques (DNN and BiLSTM network). Our proposed technique makes use of only four header based (From, Returnpath, Subject, Message-ID) features of the emails for the email classification. We applied several word embeddings for the evaluation of our models. From the experimental evaluation, we observed that the DNN model with FastText-SkipGram achieved an accuracy of 99.52% and BiLSTM model with FastText-SkipGram achieved an accuracy of 99.42%. Among these two techniques, DNN outperformed BiLSTM using the same word embedding (FastText-SkipGram) techniques with an accuracy of 99.52%.","doi":"10.1007\/s12046-024-02538-4","cleaned_title":"deepephishnet: a deep learning framework for email phishing detection using word embedding algorithms","cleaned_abstract":"email phishing is a social engineering scheme that uses spoofed emails intended to trick the user into disclosing legitimate business and personal credentials. many phishing email detection techniques exist based on machine learning, deep learning, and word embedding. in this paper, we propose a new technique for the detection of phishing emails using word embedding (word2vec, fasttext, and tf-idf) and deep learning techniques (dnn and bilstm network). our proposed technique makes use of only four header based (from, returnpath, subject, message-id) features of the emails for the email classification. we applied several word embeddings for the evaluation of our models. from the experimental evaluation, we observed that the dnn model with fasttext-skipgram achieved an accuracy of 99.52% and bilstm model with fasttext-skipgram achieved an accuracy of 99.42%. among these two techniques, dnn outperformed bilstm using the same word embedding (fasttext-skipgram) techniques with an accuracy of 99.52%.","key_phrases":["email phishing","a social engineering scheme","that","spoofed emails","the user","legitimate business","personal credentials","many phishing email detection techniques","machine learning","deep learning","word","this paper","we","a new technique","the detection","emails","word","tf-idf","deep learning techniques","dnn and bilstm network","our proposed technique","use","only four header based (from, returnpath, subject, message-id) features","the emails","the email classification","we","several word embeddings","the evaluation","our models","the experimental evaluation","we","the dnn model","fasttext-skipgram","an accuracy","99.52%","bilstm model","fasttext-skipgram","an accuracy","99.42%","these two techniques","dnn","the same word","fasttext-skipgram","an accuracy","99.52%","only four","99.52%","99.42%","two","99.52%"]},{"title":"Quantum mechanics-based deep learning framework considering near-zero variance data","authors":["Eunseo Oh","Hyunsoo Lee"],"abstract":"AbstractWith the development of automation technology, big data is collected during operation processes, and among various machine learning analysis techniques using such data, deep neural network (DNN) has high analysis performance. However, most industrial data has low-variance or near-zero variance data from the refined processes in the collected data itself. This reduces deep learning analysis performance, which is affected by data quality. To overcome this, in this study, the weight learning pattern of an applied DNN is modeled as a stochastic differential equation (SDE) based on quantum mechanics. Through the drift and diffuse terms of quantum mechanics, the patterns of the DNN and data are quickly acquired, and the data with near-zero variance is effectively analyzed simultaneously. To demonstrate the superiority of the proposed framework, DNN analysis was performed using data with near-zero variance issues, and it was proved that the proposed framework is effective in processing near-zero variance data compared with other existing algorithms.Graphical abstract","doi":"10.1007\/s10489-024-05465-3","cleaned_title":"quantum mechanics-based deep learning framework considering near-zero variance data","cleaned_abstract":"abstractwith the development of automation technology, big data is collected during operation processes, and among various machine learning analysis techniques using such data, deep neural network (dnn) has high analysis performance. however, most industrial data has low-variance or near-zero variance data from the refined processes in the collected data itself. this reduces deep learning analysis performance, which is affected by data quality. to overcome this, in this study, the weight learning pattern of an applied dnn is modeled as a stochastic differential equation (sde) based on quantum mechanics. through the drift and diffuse terms of quantum mechanics, the patterns of the dnn and data are quickly acquired, and the data with near-zero variance is effectively analyzed simultaneously. to demonstrate the superiority of the proposed framework, dnn analysis was performed using data with near-zero variance issues, and it was proved that the proposed framework is effective in processing near-zero variance data compared with other existing algorithms.graphical abstract","key_phrases":["the development","automation technology","big data","operation processes","various machine","analysis techniques","such data","deep neural network","dnn","high analysis performance","most industrial data","low-variance or near-zero variance data","the refined processes","the collected data","itself","this","deep learning analysis performance","which","data quality","this","this study","the weight learning pattern","an applied dnn","a stochastic differential equation","sde","quantum mechanics","the drift","diffuse","terms","quantum mechanics","the patterns","the dnn","data","the data","near-zero variance","the superiority","the proposed framework","dnn analysis","data","near-zero variance issues","it","the proposed framework","near-zero variance data","other existing algorithms.graphical abstract","abstractwith","quantum mechanics"]},{"title":"Deep learning-based network intrusion detection in smart healthcare enterprise systems","authors":["Vinayakumar Ravi"],"abstract":"Network-based intrusion detection (N-IDS) is an essential system inside an organization in a smart healthcare enterprise system to prevent the system and its networks from network attacks. A survey of the literature shows that in recent days deep learning approaches are employed successfully for N-IDS using network connections. However, finding the right features from a network connection is a daunting task. This work proposes a multidimensional attention-based deep learning approach for N-IDS that extracts the optimal features for intrusion detection using network payload. The proposed approach includes an embedding that transforms every word in the payload into a 100-dimensional feature vector representation and embedding follows deep learning layers such as a convolutional neural network (CNN) and long short-term memory (LSTM) with attention to extracting optimal features for attack classification. Next, the features of CNN and LSTM layers are concatenated and passed into fully connected layers for intrusion detection. The proposed approach showed 99% accuracy on the KISTI enterprise network payload dataset. In addition, the proposed approach showed 98% accuracy and 99% accuracy on network-based datasets such as KDDCup-99, CICIDS-2017, and WSN-DS and UNSW-NB15 respectively. The good experimental results on various network-based datasets suggest that the proposed N-IDS in smart healthcare enterprise systems is robust and generalizable to detect attacks from different network environments. The proposed approach performed better in all the experiments than the other deep learning-based methods. The model showed a 5% accuracy performance improvement compared to the existing study using the KISTI dataset. In addition, the proposed model has shown similar performances on the other intrusion datasets. The proposed approach serves as a network monitoring tool for efficient and accurate detection of attacks inside an organization on a healthcare enterprise network system.","doi":"10.1007\/s11042-023-17300-x","cleaned_title":"deep learning-based network intrusion detection in smart healthcare enterprise systems","cleaned_abstract":"network-based intrusion detection (n-ids) is an essential system inside an organization in a smart healthcare enterprise system to prevent the system and its networks from network attacks. a survey of the literature shows that in recent days deep learning approaches are employed successfully for n-ids using network connections. however, finding the right features from a network connection is a daunting task. this work proposes a multidimensional attention-based deep learning approach for n-ids that extracts the optimal features for intrusion detection using network payload. the proposed approach includes an embedding that transforms every word in the payload into a 100-dimensional feature vector representation and embedding follows deep learning layers such as a convolutional neural network (cnn) and long short-term memory (lstm) with attention to extracting optimal features for attack classification. next, the features of cnn and lstm layers are concatenated and passed into fully connected layers for intrusion detection. the proposed approach showed 99% accuracy on the kisti enterprise network payload dataset. in addition, the proposed approach showed 98% accuracy and 99% accuracy on network-based datasets such as kddcup-99, cicids-2017, and wsn-ds and unsw-nb15 respectively. the good experimental results on various network-based datasets suggest that the proposed n-ids in smart healthcare enterprise systems is robust and generalizable to detect attacks from different network environments. the proposed approach performed better in all the experiments than the other deep learning-based methods. the model showed a 5% accuracy performance improvement compared to the existing study using the kisti dataset. in addition, the proposed model has shown similar performances on the other intrusion datasets. the proposed approach serves as a network monitoring tool for efficient and accurate detection of attacks inside an organization on a healthcare enterprise network system.","key_phrases":["network-based intrusion detection","-ids","an essential system","an organization","a smart healthcare enterprise system","the system","its networks","network attacks","a survey","the literature","recent days","deep learning approaches","n-ids","network connections","the right features","a network connection","a daunting task","this work","a multidimensional attention-based deep learning approach","n-ids","that","the optimal features","intrusion detection","network payload","the proposed approach","that","every word","the payload","a 100-dimensional feature vector representation","embedding","deep learning layers","a convolutional neural network","cnn","long short-term memory","lstm","attention","optimal features","attack classification","the features","cnn","lstm layers","fully connected layers","intrusion detection","the proposed approach","99% accuracy","the kisti enterprise network payload dataset","addition","the proposed approach","98% accuracy","99% accuracy","network-based datasets","kddcup-99","cicids-2017","wsn-ds and unsw-nb15","the good experimental results","various network-based datasets","the proposed n-ids","smart healthcare enterprise systems","attacks","different network environments","the proposed approach","all the experiments","the other deep learning-based methods","the model","a 5% accuracy performance improvement","the existing study","the kisti dataset","addition","the proposed model","similar performances","the other intrusion datasets","the proposed approach","a network monitoring tool","efficient and accurate detection","attacks","an organization","a healthcare enterprise network system","smart healthcare","recent days","100","cnn","cnn","99%","98%","99%","kddcup-99","cicids-2017","smart healthcare enterprise","5%"]},{"title":"Implementing Deep Learning-Based Intelligent Inspection for Investment Castings","authors":["Nabhan Yousef","Amit Sata"],"abstract":"In this study, a user-friendly intelligent inspection device was developed employing deep learning techniques to effectively detect and characterize surface defects in investment castings. The inspection techniques encompassed a range from basic visual inspection to more specialized methods like liquid penetration and magnetic particle tests. The developed device leveraged convolutional neural networks (CNN), residual neural networks (ResNet), and recurrent neural networks (R-CNN) models, with a dataset of 3600 images depicting industrial castings, both defective and defect-free. A majority of the dataset was allocated for training (approximately 80%), while the remainder was used for testing. Among the deep learning models considered, ResNet exhibited superior accuracy in defect inspection, trailed by CNN and R-CNN. The resultant intelligent inspection device was successfully implemented in an industrial setting, enabling efficient defect identification in investment castings through the trained ResNet model.","doi":"10.1007\/s13369-023-08240-7","cleaned_title":"implementing deep learning-based intelligent inspection for investment castings","cleaned_abstract":"in this study, a user-friendly intelligent inspection device was developed employing deep learning techniques to effectively detect and characterize surface defects in investment castings. the inspection techniques encompassed a range from basic visual inspection to more specialized methods like liquid penetration and magnetic particle tests. the developed device leveraged convolutional neural networks (cnn), residual neural networks (resnet), and recurrent neural networks (r-cnn) models, with a dataset of 3600 images depicting industrial castings, both defective and defect-free. a majority of the dataset was allocated for training (approximately 80%), while the remainder was used for testing. among the deep learning models considered, resnet exhibited superior accuracy in defect inspection, trailed by cnn and r-cnn. the resultant intelligent inspection device was successfully implemented in an industrial setting, enabling efficient defect identification in investment castings through the trained resnet model.","key_phrases":["this study","a user-friendly intelligent inspection device","deep learning techniques","surface defects","investment castings","the inspection techniques","a range","basic visual inspection","more specialized methods","liquid penetration","magnetic particle tests","the developed device leveraged convolutional neural networks","cnn","residual neural networks","resnet","recurrent neural networks","r-cnn) models","a dataset","3600 images","industrial castings","a majority","the dataset","training","approximately 80%","the remainder","testing","the deep learning models","resnet","superior accuracy","defect inspection","cnn","r-cnn","the resultant intelligent inspection device","an industrial setting","efficient defect identification","investment castings","the trained resnet model","cnn","3600","approximately 80%","cnn"]},{"title":"Discovery of a structural class of antibiotics with explainable deep learning","authors":["Felix Wong","Erica J. Zheng","Jacqueline A. Valeri","Nina M. Donghia","Melis N. Anahtar","Satotaka Omori","Alicia Li","Andres Cubillos-Ruiz","Aarti Krishnan","Wengong Jin","Abigail L. Manson","Jens Friedrichs","Ralf Helbig","Behnoush Hajian","Dawid K. Fiejtek","Florence F. Wagner","Holly H. Soutter","Ashlee M. Earl","Jonathan M. Stokes","Lars D. Renner","James J. Collins"],"abstract":"The discovery of novel structural classes of antibiotics is urgently needed to address the ongoing antibiotic resistance crisis1,2,3,4,5,6,7,8,9. Deep learning approaches have aided in exploring chemical spaces1,10,11,12,13,14,15; these typically use black box models and do not provide chemical insights. Here we reasoned that the chemical substructures associated with antibiotic activity learned by neural network models can be identified and used to predict structural classes of antibiotics. We tested this hypothesis by developing an explainable, substructure-based approach for the efficient, deep learning-guided exploration of chemical spaces. We determined the antibiotic activities and human cell cytotoxicity profiles of 39,312 compounds and applied ensembles of graph neural networks to predict antibiotic activity and cytotoxicity for 12,076,365 compounds. Using explainable graph algorithms, we identified substructure-based rationales for compounds with high predicted antibiotic activity and low predicted cytotoxicity. We empirically tested 283 compounds and found that compounds exhibiting antibiotic activity against Staphylococcus aureus were enriched in putative structural classes arising from rationales. Of these structural classes of compounds, one is selective against methicillin-resistant S. aureus (MRSA) and vancomycin-resistant enterococci, evades substantial resistance, and reduces bacterial titres in mouse models of MRSA skin and systemic thigh infection. Our approach enables the deep learning-guided discovery of structural classes of antibiotics and demonstrates that machine learning models in drug discovery can be explainable, providing insights into the chemical substructures that underlie selective antibiotic activity.","doi":"10.1038\/s41586-023-06887-8","cleaned_title":"discovery of a structural class of antibiotics with explainable deep learning","cleaned_abstract":"the discovery of novel structural classes of antibiotics is urgently needed to address the ongoing antibiotic resistance crisis1,2,3,4,5,6,7,8,9. deep learning approaches have aided in exploring chemical spaces1,10,11,12,13,14,15; these typically use black box models and do not provide chemical insights. here we reasoned that the chemical substructures associated with antibiotic activity learned by neural network models can be identified and used to predict structural classes of antibiotics. we tested this hypothesis by developing an explainable, substructure-based approach for the efficient, deep learning-guided exploration of chemical spaces. we determined the antibiotic activities and human cell cytotoxicity profiles of 39,312 compounds and applied ensembles of graph neural networks to predict antibiotic activity and cytotoxicity for 12,076,365 compounds. using explainable graph algorithms, we identified substructure-based rationales for compounds with high predicted antibiotic activity and low predicted cytotoxicity. we empirically tested 283 compounds and found that compounds exhibiting antibiotic activity against staphylococcus aureus were enriched in putative structural classes arising from rationales. of these structural classes of compounds, one is selective against methicillin-resistant s. aureus (mrsa) and vancomycin-resistant enterococci, evades substantial resistance, and reduces bacterial titres in mouse models of mrsa skin and systemic thigh infection. our approach enables the deep learning-guided discovery of structural classes of antibiotics and demonstrates that machine learning models in drug discovery can be explainable, providing insights into the chemical substructures that underlie selective antibiotic activity.","key_phrases":["the discovery","novel structural classes","antibiotics","the ongoing antibiotic resistance crisis1,2,3,4,5,6,7,8,9","deep learning approaches","chemical spaces1,10,11,12,13,14,15","these","black box models","chemical insights","we","the chemical substructures","antibiotic activity","neural network models","structural classes","antibiotics","we","this hypothesis","an explainable, substructure-based approach","the efficient, deep learning-guided exploration","chemical spaces","we","the antibiotic activities","human cell cytotoxicity profiles","39,312 compounds","applied ensembles","graph neural networks","antibiotic activity","cytotoxicity","12,076,365 compounds","explainable graph algorithms","we","substructure-based rationales","compounds","high predicted antibiotic activity","cytotoxicity","we","283 compounds","compounds","antibiotic activity","staphylococcus aureus","putative structural classes","rationales","these structural classes","compounds","methicillin-resistant s. aureus","(mrsa","vancomycin-resistant enterococci","substantial resistance","bacterial titres","mouse models","mrsa skin","systemic thigh infection","our approach","the deep learning-guided discovery","structural classes","antibiotics","machine learning models","drug discovery","insights","the chemical substructures","that","selective antibiotic activity","deep learning","39,312","12,076,365","283","methicillin","s. aureus"]},{"title":"Incremental\u2013decremental data transformation based ensemble deep learning model (IDT-eDL) for temperature prediction","authors":["Vipin Kumar","Rana Kumar"],"abstract":"Human life heavily depends on weather conditions, which affect the necessary operations like agriculture, aviation, tourism, industries, etc., where the temperature plays a vital role in deciding the weather conditions along with other meteorological variables. Therefore, temperature forecasting has drawn considerable attention from researchers because of its significant effect on daily life activities and the ever-challenging forecasting task. These research objectives are to investigate the transformation of data based on incremental and decremental approaches and to find the practical ensemble approach over proposed models for effective temperature prediction, where the proposed model is called the Incremental\u2013Decremental Data Transformation-Based Ensemble Deep Learning Model (IDT-eDL). The temperature dataset from Delhi, India, has been utilized to compare proposed and traditional deep learning models over various performance measures. The proposed IDT-eDL with BiLSTM deep learning model (i.e., IDT-eDL_BiLSTM ) has performed the best among the proposed models and traditional deep learning model and achieved Performance over measures MSE: 1.36, RMSE: 1.16, MAE: 0.89, MAPE: 4.13 and \\(R^2\\):0.999. Additionally, non-parametric statistical analysis of Friedman ranking is also performed to validate the effectiveness of the proposed IDT-eDL model, which also shows a higher ranking of the proposed model than the traditional deep learning models.","doi":"10.1007\/s40808-024-01953-0","cleaned_title":"incremental\u2013decremental data transformation based ensemble deep learning model (idt-edl) for temperature prediction","cleaned_abstract":"human life heavily depends on weather conditions, which affect the necessary operations like agriculture, aviation, tourism, industries, etc., where the temperature plays a vital role in deciding the weather conditions along with other meteorological variables. therefore, temperature forecasting has drawn considerable attention from researchers because of its significant effect on daily life activities and the ever-challenging forecasting task. these research objectives are to investigate the transformation of data based on incremental and decremental approaches and to find the practical ensemble approach over proposed models for effective temperature prediction, where the proposed model is called the incremental\u2013decremental data transformation-based ensemble deep learning model (idt-edl). the temperature dataset from delhi, india, has been utilized to compare proposed and traditional deep learning models over various performance measures. the proposed idt-edl with bilstm deep learning model (i.e., idt-edl_bilstm ) has performed the best among the proposed models and traditional deep learning model and achieved performance over measures mse: 1.36, rmse: 1.16, mae: 0.89, mape: 4.13 and \\(r^2\\):0.999. additionally, non-parametric statistical analysis of friedman ranking is also performed to validate the effectiveness of the proposed idt-edl model, which also shows a higher ranking of the proposed model than the traditional deep learning models.","key_phrases":["human life","weather conditions","which","the necessary operations","agriculture","aviation","tourism","industries","the temperature","a vital role","the weather conditions","other meteorological variables","temperature forecasting","considerable attention","researchers","its significant effect","daily life activities","the ever-challenging forecasting task","these research objectives","the transformation","data","incremental and decremental approaches","the practical ensemble approach","proposed models","effective temperature prediction","the proposed model","the incremental\u2013decremental data transformation-based ensemble deep learning model","idt-edl","the temperature","delhi","india","proposed and traditional deep learning models","various performance measures","the proposed idt-edl","bilstm deep learning model","i.e., idt-edl_bilstm","the proposed models","traditional deep learning model","performance","measures mse","rmse","mape","\\(r^2\\):0.999","non-parametric statistical analysis","friedman ranking","the effectiveness","the proposed idt-edl model","which","a higher ranking","the proposed model","the traditional deep learning models","delhi","india","1.36","rmse","1.16","0.89","4.13"]},{"title":"Deep learning systems for forecasting the prices of crude oil and precious metals","authors":["Parisa Foroutan","Salim Lahmiri"],"abstract":"Commodity markets, such as crude oil and precious metals, play a strategic role in the economic development of nations, with crude oil prices influencing geopolitical relations and the global economy. Moreover, gold and silver are argued to hedge the stock and cryptocurrency markets during market downsides. Therefore, accurate forecasting of crude oil and precious metals prices is critical. Nevertheless, due to the nonlinear nature, substantial fluctuations, and irregular cycles of crude oil and precious metals, predicting their prices is a challenging task. Our study contributes to the commodity market price forecasting literature by implementing and comparing advanced deep-learning models. We address this gap by including silver alongside gold in our analysis, offering a more comprehensive understanding of the precious metal markets. This research expands existing knowledge and provides valuable insights into predicting commodity prices. In this study, we implemented 16 deep- and machine-learning models to forecast the daily price of the West Texas Intermediate (WTI), Brent, gold, and silver markets. The employed deep-learning models are long short-term memory (LSTM), BiLSTM, gated recurrent unit (GRU), bidirectional gated recurrent units (BiGRU), T2V-BiLSTM, T2V-BiGRU, convolutional neural networks (CNN), CNN-BiLSTM, CNN-BiGRU, temporal convolutional network (TCN), TCN-BiLSTM, and TCN-BiGRU. We compared the forecasting performance of deep-learning models with the baseline random forest, LightGBM, support vector regression, and k-nearest neighborhood models using mean absolute error (MAE), mean absolute percentage error, and root mean squared error as evaluation criteria. By considering different sliding window lengths, we examine the forecasting performance of our models. Our results reveal that the TCN model outperforms the others for WTI, Brent, and silver, achieving the lowest MAE values of 1.444, 1.295, and 0.346, respectively. The BiGRU model performs best for gold, with an MAE of 15.188 using a 30-day input sequence. Furthermore, LightGBM exhibits comparable performance to TCN and is the best-performing machine-learning model overall. These findings are critical for investors, policymakers, mining companies, and governmental agencies to effectively anticipate market trends, mitigate risk, manage uncertainty, and make timely decisions and strategies regarding crude oil, gold, and silver markets.","doi":"10.1186\/s40854-024-00637-z","cleaned_title":"deep learning systems for forecasting the prices of crude oil and precious metals","cleaned_abstract":"commodity markets, such as crude oil and precious metals, play a strategic role in the economic development of nations, with crude oil prices influencing geopolitical relations and the global economy. moreover, gold and silver are argued to hedge the stock and cryptocurrency markets during market downsides. therefore, accurate forecasting of crude oil and precious metals prices is critical. nevertheless, due to the nonlinear nature, substantial fluctuations, and irregular cycles of crude oil and precious metals, predicting their prices is a challenging task. our study contributes to the commodity market price forecasting literature by implementing and comparing advanced deep-learning models. we address this gap by including silver alongside gold in our analysis, offering a more comprehensive understanding of the precious metal markets. this research expands existing knowledge and provides valuable insights into predicting commodity prices. in this study, we implemented 16 deep- and machine-learning models to forecast the daily price of the west texas intermediate (wti), brent, gold, and silver markets. the employed deep-learning models are long short-term memory (lstm), bilstm, gated recurrent unit (gru), bidirectional gated recurrent units (bigru), t2v-bilstm, t2v-bigru, convolutional neural networks (cnn), cnn-bilstm, cnn-bigru, temporal convolutional network (tcn), tcn-bilstm, and tcn-bigru. we compared the forecasting performance of deep-learning models with the baseline random forest, lightgbm, support vector regression, and k-nearest neighborhood models using mean absolute error (mae), mean absolute percentage error, and root mean squared error as evaluation criteria. by considering different sliding window lengths, we examine the forecasting performance of our models. our results reveal that the tcn model outperforms the others for wti, brent, and silver, achieving the lowest mae values of 1.444, 1.295, and 0.346, respectively. the bigru model performs best for gold, with an mae of 15.188 using a 30-day input sequence. furthermore, lightgbm exhibits comparable performance to tcn and is the best-performing machine-learning model overall. these findings are critical for investors, policymakers, mining companies, and governmental agencies to effectively anticipate market trends, mitigate risk, manage uncertainty, and make timely decisions and strategies regarding crude oil, gold, and silver markets.","key_phrases":["commodity markets","crude oil","precious metals","a strategic role","the economic development","nations","crude oil prices","geopolitical relations","the global economy","gold","silver","the stock and cryptocurrency markets","accurate forecasting","crude oil","precious metals prices","the nonlinear nature","substantial fluctuations","irregular cycles","crude oil","precious metals","their prices","a challenging task","our study","the commodity market price forecasting literature","advanced deep-learning models","we","this gap","silver","gold","our analysis","a more comprehensive understanding","the precious metal markets","this research","existing knowledge","valuable insights","commodity prices","this study","we","16 deep- and machine-learning models","the daily price","the west texas intermediate","(wti","brent","gold","silver markets","the employed deep-learning models","long short-term memory","lstm","bilstm","gated recurrent unit","gru","bidirectional gated recurrent units","bigru","t2v-bilstm","t2v-bigru","convolutional neural networks","cnn","cnn-bilstm","cnn-bigru, temporal convolutional network","tcn","tcn-bilstm","tcn-bigru","we","the forecasting performance","deep-learning models","the baseline random forest","lightgbm","vector regression","k-nearest neighborhood models","mean absolute error","mae","absolute percentage error","root mean squared error","evaluation criteria","different sliding window lengths","we","the forecasting performance","our models","our results","the tcn model","the others","wti","brent","silver","the lowest mae values","the bigru model","gold","an mae","a 30-day input sequence","lightgbm","comparable performance","the best-performing machine-learning model","these findings","investors","policymakers","mining companies","governmental agencies","market trends","mitigate risk","uncertainty","timely decisions","strategies","crude oil","gold","silver markets","16","daily","west texas","cnn","cnn","cnn","1.444","1.295","0.346","15.188","30-day"]},{"title":"Deep learning-based weld defect classification using VGG16 transfer learning adaptive fine-tuning","authors":["Samuel Kumaresan","K. S. Jai Aultrin","S. S. Kumar","M. Dev Anand"],"abstract":"Welding is a vital joining process; however, occurrences of weld defects often degrade the quality of the welded part. The risk of occurrence of a variety of defects has led to the development of advanced weld defects detection systems such as automated weld defects detection and classification. The present work is a novel approach that proposes and investigates a unique image-centered method based on a deep learning model trained by a small X-ray image dataset. A data augmentation method able to process images on the go was used to offset the limitation of the small X-ray dataset. Fine-tuned transfer learning techniques were used to train two convolutional neural network based architectures with VGG16 and ResNet50 as the base models for the augmented sets. Out of the networks we fine-tuned, VGG16 based model performed well with a relatively higher average accuracy of 90%. Even though the small dataset was spread across 15 different classes in an unbalanced way, the learning curves showed acceptable model generalization characteristics.","doi":"10.1007\/s12008-023-01327-3","cleaned_title":"deep learning-based weld defect classification using vgg16 transfer learning adaptive fine-tuning","cleaned_abstract":"welding is a vital joining process; however, occurrences of weld defects often degrade the quality of the welded part. the risk of occurrence of a variety of defects has led to the development of advanced weld defects detection systems such as automated weld defects detection and classification. the present work is a novel approach that proposes and investigates a unique image-centered method based on a deep learning model trained by a small x-ray image dataset. a data augmentation method able to process images on the go was used to offset the limitation of the small x-ray dataset. fine-tuned transfer learning techniques were used to train two convolutional neural network based architectures with vgg16 and resnet50 as the base models for the augmented sets. out of the networks we fine-tuned, vgg16 based model performed well with a relatively higher average accuracy of 90%. even though the small dataset was spread across 15 different classes in an unbalanced way, the learning curves showed acceptable model generalization characteristics.","key_phrases":["welding","a vital joining process","occurrences","weld defects","the quality","the welded part","the risk","occurrence","a variety","defects","the development","advanced weld defects detection systems","automated weld defects detection","classification","the present work","a novel approach","that","a unique image-centered method","a deep learning model","a small x-ray image dataset","a data augmentation method","images","the go","the limitation","the small x-ray dataset","fine-tuned transfer learning techniques","two convolutional neural network based architectures","vgg16","resnet50","the base models","the augmented sets","the networks","we","vgg16 based model","a relatively higher average accuracy","90%","the small dataset","15 different classes","an unbalanced way","the learning curves","acceptable model generalization characteristics","two","resnet50","90%","15"]},{"title":"Transfer learning for emotion detection in conversational text: a hybrid deep learning approach with pre-trained embeddings","authors":["Sheetal Kusal","Shruti Patil","Jyoti Choudrie","Ketan Kotecha","Deepali Vora"],"abstract":"Understanding the emotions and sentiments from conversations has relevance in many application areas. Specifically, conversational agents, question-answering systems, or areas where natural language inference is used. Therefore, techniques to detect emotions from conversations have become the need of the moment. The convolutional network and recurrent networks have shown different capabilities in text representation. This work proposes a hybrid deep learning network based on the convolutional-recurrent network used to detect the emotions of people based on conversational text. A convolutional network has the ability to capture local patterns and relationships and is inherently shift-invariant. At the same time, the recurrent network captures long-range dependencies in sequential information. This work also utilises the power of transfer learning by employing pre-trained embeddings from Neural Network Language Model models. These pre-trained representations, generated from vast text corpora, encode rich semantic information about words. This study investigates a novel approach towards text-based emotion detection using pre-trained Neural Network Language Model embeddings with hybrid convolutional-recurrent architecture. The proposed hybrid experimental setup has been evaluated on the Empathetic Dialogues dataset and contrasted with the state-of-the-art works. A comparative analysis reveals that the proposed Convolutional Neural Network with a Bidirectional Gated Recurrent Unit hybrid approach with Neural Network Language Model embeddings achieves superior performance and accuracy.","doi":"10.1007\/s41870-024-02027-1","cleaned_title":"transfer learning for emotion detection in conversational text: a hybrid deep learning approach with pre-trained embeddings","cleaned_abstract":"understanding the emotions and sentiments from conversations has relevance in many application areas. specifically, conversational agents, question-answering systems, or areas where natural language inference is used. therefore, techniques to detect emotions from conversations have become the need of the moment. the convolutional network and recurrent networks have shown different capabilities in text representation. this work proposes a hybrid deep learning network based on the convolutional-recurrent network used to detect the emotions of people based on conversational text. a convolutional network has the ability to capture local patterns and relationships and is inherently shift-invariant. at the same time, the recurrent network captures long-range dependencies in sequential information. this work also utilises the power of transfer learning by employing pre-trained embeddings from neural network language model models. these pre-trained representations, generated from vast text corpora, encode rich semantic information about words. this study investigates a novel approach towards text-based emotion detection using pre-trained neural network language model embeddings with hybrid convolutional-recurrent architecture. the proposed hybrid experimental setup has been evaluated on the empathetic dialogues dataset and contrasted with the state-of-the-art works. a comparative analysis reveals that the proposed convolutional neural network with a bidirectional gated recurrent unit hybrid approach with neural network language model embeddings achieves superior performance and accuracy.","key_phrases":["the emotions","sentiments","conversations","relevance","many application areas","specifically, conversational agents","question-answering systems","areas","natural language inference","techniques","emotions","conversations","the need","the moment","the convolutional network","recurrent networks","different capabilities","text representation","this work","a hybrid deep learning network","the convolutional-recurrent network","the emotions","people","conversational text","a convolutional network","the ability","local patterns","relationships","the same time","the recurrent network","long-range dependencies","sequential information","this work","the power","pre-trained embeddings","neural network language model models","these pre-trained representations","vast text corpora","encode rich semantic information","words","this study","a novel approach","text-based emotion detection","pre-trained neural network language model embeddings","hybrid convolutional-recurrent architecture","the proposed hybrid experimental setup","the empathetic dialogues","the-art","a comparative analysis","the proposed convolutional neural network","a bidirectional gated recurrent unit hybrid approach","neural network language model embeddings","superior performance","accuracy","corpora","encode rich semantic information"]},{"title":"Quantum mechanics-based deep learning framework considering near-zero variance data","authors":["Eunseo Oh","Hyunsoo Lee"],"abstract":"AbstractWith the development of automation technology, big data is collected during operation processes, and among various machine learning analysis techniques using such data, deep neural network (DNN) has high analysis performance. However, most industrial data has low-variance or near-zero variance data from the refined processes in the collected data itself. This reduces deep learning analysis performance, which is affected by data quality. To overcome this, in this study, the weight learning pattern of an applied DNN is modeled as a stochastic differential equation (SDE) based on quantum mechanics. Through the drift and diffuse terms of quantum mechanics, the patterns of the DNN and data are quickly acquired, and the data with near-zero variance is effectively analyzed simultaneously. To demonstrate the superiority of the proposed framework, DNN analysis was performed using data with near-zero variance issues, and it was proved that the proposed framework is effective in processing near-zero variance data compared with other existing algorithms.Graphical abstract","doi":"10.1007\/s10489-024-05465-3","cleaned_title":"quantum mechanics-based deep learning framework considering near-zero variance data","cleaned_abstract":"abstractwith the development of automation technology, big data is collected during operation processes, and among various machine learning analysis techniques using such data, deep neural network (dnn) has high analysis performance. however, most industrial data has low-variance or near-zero variance data from the refined processes in the collected data itself. this reduces deep learning analysis performance, which is affected by data quality. to overcome this, in this study, the weight learning pattern of an applied dnn is modeled as a stochastic differential equation (sde) based on quantum mechanics. through the drift and diffuse terms of quantum mechanics, the patterns of the dnn and data are quickly acquired, and the data with near-zero variance is effectively analyzed simultaneously. to demonstrate the superiority of the proposed framework, dnn analysis was performed using data with near-zero variance issues, and it was proved that the proposed framework is effective in processing near-zero variance data compared with other existing algorithms.graphical abstract","key_phrases":["the development","automation technology","big data","operation processes","various machine","analysis techniques","such data","deep neural network","dnn","high analysis performance","most industrial data","low-variance or near-zero variance data","the refined processes","the collected data","itself","this","deep learning analysis performance","which","data quality","this","this study","the weight learning pattern","an applied dnn","a stochastic differential equation","sde","quantum mechanics","the drift","diffuse","terms","quantum mechanics","the patterns","the dnn","data","the data","near-zero variance","the superiority","the proposed framework","dnn analysis","data","near-zero variance issues","it","the proposed framework","near-zero variance data","other existing algorithms.graphical abstract","abstractwith","quantum mechanics"]},{"title":"Deep learning-based network intrusion detection in smart healthcare enterprise systems","authors":["Vinayakumar Ravi"],"abstract":"Network-based intrusion detection (N-IDS) is an essential system inside an organization in a smart healthcare enterprise system to prevent the system and its networks from network attacks. A survey of the literature shows that in recent days deep learning approaches are employed successfully for N-IDS using network connections. However, finding the right features from a network connection is a daunting task. This work proposes a multidimensional attention-based deep learning approach for N-IDS that extracts the optimal features for intrusion detection using network payload. The proposed approach includes an embedding that transforms every word in the payload into a 100-dimensional feature vector representation and embedding follows deep learning layers such as a convolutional neural network (CNN) and long short-term memory (LSTM) with attention to extracting optimal features for attack classification. Next, the features of CNN and LSTM layers are concatenated and passed into fully connected layers for intrusion detection. The proposed approach showed 99% accuracy on the KISTI enterprise network payload dataset. In addition, the proposed approach showed 98% accuracy and 99% accuracy on network-based datasets such as KDDCup-99, CICIDS-2017, and WSN-DS and UNSW-NB15 respectively. The good experimental results on various network-based datasets suggest that the proposed N-IDS in smart healthcare enterprise systems is robust and generalizable to detect attacks from different network environments. The proposed approach performed better in all the experiments than the other deep learning-based methods. The model showed a 5% accuracy performance improvement compared to the existing study using the KISTI dataset. In addition, the proposed model has shown similar performances on the other intrusion datasets. The proposed approach serves as a network monitoring tool for efficient and accurate detection of attacks inside an organization on a healthcare enterprise network system.","doi":"10.1007\/s11042-023-17300-x","cleaned_title":"deep learning-based network intrusion detection in smart healthcare enterprise systems","cleaned_abstract":"network-based intrusion detection (n-ids) is an essential system inside an organization in a smart healthcare enterprise system to prevent the system and its networks from network attacks. a survey of the literature shows that in recent days deep learning approaches are employed successfully for n-ids using network connections. however, finding the right features from a network connection is a daunting task. this work proposes a multidimensional attention-based deep learning approach for n-ids that extracts the optimal features for intrusion detection using network payload. the proposed approach includes an embedding that transforms every word in the payload into a 100-dimensional feature vector representation and embedding follows deep learning layers such as a convolutional neural network (cnn) and long short-term memory (lstm) with attention to extracting optimal features for attack classification. next, the features of cnn and lstm layers are concatenated and passed into fully connected layers for intrusion detection. the proposed approach showed 99% accuracy on the kisti enterprise network payload dataset. in addition, the proposed approach showed 98% accuracy and 99% accuracy on network-based datasets such as kddcup-99, cicids-2017, and wsn-ds and unsw-nb15 respectively. the good experimental results on various network-based datasets suggest that the proposed n-ids in smart healthcare enterprise systems is robust and generalizable to detect attacks from different network environments. the proposed approach performed better in all the experiments than the other deep learning-based methods. the model showed a 5% accuracy performance improvement compared to the existing study using the kisti dataset. in addition, the proposed model has shown similar performances on the other intrusion datasets. the proposed approach serves as a network monitoring tool for efficient and accurate detection of attacks inside an organization on a healthcare enterprise network system.","key_phrases":["network-based intrusion detection","-ids","an essential system","an organization","a smart healthcare enterprise system","the system","its networks","network attacks","a survey","the literature","recent days","deep learning approaches","n-ids","network connections","the right features","a network connection","a daunting task","this work","a multidimensional attention-based deep learning approach","n-ids","that","the optimal features","intrusion detection","network payload","the proposed approach","that","every word","the payload","a 100-dimensional feature vector representation","embedding","deep learning layers","a convolutional neural network","cnn","long short-term memory","lstm","attention","optimal features","attack classification","the features","cnn","lstm layers","fully connected layers","intrusion detection","the proposed approach","99% accuracy","the kisti enterprise network payload dataset","addition","the proposed approach","98% accuracy","99% accuracy","network-based datasets","kddcup-99","cicids-2017","wsn-ds and unsw-nb15","the good experimental results","various network-based datasets","the proposed n-ids","smart healthcare enterprise systems","attacks","different network environments","the proposed approach","all the experiments","the other deep learning-based methods","the model","a 5% accuracy performance improvement","the existing study","the kisti dataset","addition","the proposed model","similar performances","the other intrusion datasets","the proposed approach","a network monitoring tool","efficient and accurate detection","attacks","an organization","a healthcare enterprise network system","smart healthcare","recent days","100","cnn","cnn","99%","98%","99%","kddcup-99","cicids-2017","smart healthcare enterprise","5%"]},{"title":"Deep-learning model for evaluating histopathology of acute renal tubular injury","authors":["Thi Thuy Uyen Nguyen","Anh-Tien Nguyen","Hyeongwan Kim","Yu Jin Jung","Woong Park","Kyoung Min Kim","Ilwoo Park","Won Kim"],"abstract":"Tubular injury is the most common cause of acute kidney injury. Histopathological diagnosis may help distinguish between the different types of acute kidney injury and aid in treatment. To date, a limited number of study has used deep-learning models to assist in the histopathological diagnosis of acute kidney injury. This study aimed to perform histopathological segmentation to identify the four structures of acute renal tubular injury using deep-learning models. A segmentation model was used to classify tubule-specific injuries following cisplatin treatment. A total of 45 whole-slide images with 400 generated patches were used in the segmentation model, and 27,478 annotations were created for four classes: glomerulus, healthy tubules, necrotic tubules, and tubules with casts. A segmentation model was developed using the DeepLabV3 architecture with a MobileNetv3-Large backbone to accurately identify the four histopathological structures associated with acute renal tubular injury in PAS-stained mouse samples. In the segmentation model for four structures, the highest Intersection over Union and the Dice coefficient were obtained for the segmentation of the \u201cglomerulus\u201d class, followed by \u201cnecrotic tubules,\u201d \u201chealthy tubules,\u201d and \u201ctubules with cast\u201d classes. The overall performance of the segmentation algorithm for all classes in the test set included an Intersection over Union of 0.7968 and a Dice coefficient of 0.8772. The Dice scores for the glomerulus, healthy tubules, necrotic tubules, and tubules with cast are 91.78\u2009\u00b1\u200911.09, 87.37\u2009\u00b1\u20094.02, 88.08\u2009\u00b1\u20096.83, and 83.64\u2009\u00b1\u200920.39%, respectively. The utilization of deep learning in a predictive model has demonstrated promising performance in accurately identifying the degree of injured renal tubules. These results may provide new opportunities for the application of the proposed methods to evaluate renal pathology more effectively.","doi":"10.1038\/s41598-024-58506-9","cleaned_title":"deep-learning model for evaluating histopathology of acute renal tubular injury","cleaned_abstract":"tubular injury is the most common cause of acute kidney injury. histopathological diagnosis may help distinguish between the different types of acute kidney injury and aid in treatment. to date, a limited number of study has used deep-learning models to assist in the histopathological diagnosis of acute kidney injury. this study aimed to perform histopathological segmentation to identify the four structures of acute renal tubular injury using deep-learning models. a segmentation model was used to classify tubule-specific injuries following cisplatin treatment. a total of 45 whole-slide images with 400 generated patches were used in the segmentation model, and 27,478 annotations were created for four classes: glomerulus, healthy tubules, necrotic tubules, and tubules with casts. a segmentation model was developed using the deeplabv3 architecture with a mobilenetv3-large backbone to accurately identify the four histopathological structures associated with acute renal tubular injury in pas-stained mouse samples. in the segmentation model for four structures, the highest intersection over union and the dice coefficient were obtained for the segmentation of the \u201cglomerulus\u201d class, followed by \u201cnecrotic tubules,\u201d \u201chealthy tubules,\u201d and \u201ctubules with cast\u201d classes. the overall performance of the segmentation algorithm for all classes in the test set included an intersection over union of 0.7968 and a dice coefficient of 0.8772. the dice scores for the glomerulus, healthy tubules, necrotic tubules, and tubules with cast are 91.78 \u00b1 11.09, 87.37 \u00b1 4.02, 88.08 \u00b1 6.83, and 83.64 \u00b1 20.39%, respectively. the utilization of deep learning in a predictive model has demonstrated promising performance in accurately identifying the degree of injured renal tubules. these results may provide new opportunities for the application of the proposed methods to evaluate renal pathology more effectively.","key_phrases":["tubular injury","the most common cause","acute kidney injury","histopathological diagnosis","the different types","acute kidney injury","aid","treatment","date","a limited number","study","deep-learning models","the histopathological diagnosis","acute kidney injury","this study","histopathological segmentation","the four structures","acute renal tubular injury","deep-learning models","a segmentation model","tubule-specific injuries","cisplatin treatment","a total","45 whole-slide images","400 generated patches","the segmentation model","27,478 annotations","four classes","healthy tubules","necrotic tubules","tubules","casts","a segmentation model","the deeplabv3 architecture","a mobilenetv3-large backbone","the four histopathological structures","acute renal tubular injury","pas-stained mouse samples","the segmentation model","four structures","the highest intersection","union","the dice coefficient","the segmentation","the \u201cglomerulus\u201d class","necrotic tubules","\u201d \u201chealthy tubules","\u201ctubules","cast\u201d classes","the overall performance","the segmentation algorithm","all classes","an intersection","union","a dice coefficient","the dice scores","the glomerulus, healthy tubules","necrotic tubules","tubules","cast","91.78 \u00b1","\u00b1","\u00b1","83.64 \u00b1","20.39%","the utilization","deep learning","a predictive model","promising performance","the degree","injured renal tubules","these results","new opportunities","the application","the proposed methods","renal pathology","four","45","400","27,478","four","mobilenetv3","four","four","0.7968","0.8772","91.78","11.09","87.37","4.02","88.08","6.83","83.64","20.39%"]},{"title":"Deep Learning\u2013Assisted Identification of Femoroacetabular Impingement (FAI) on Routine Pelvic Radiographs","authors":["Michael K. Hoy","Vishal Desai","Simukayi Mutasa","Robert C. Hoy","Richard Gorniak","Jeffrey A. Belair"],"abstract":"To use a novel deep learning system to localize the hip joints and detect findings of cam-type femoroacetabular impingement (FAI). A retrospective search of hip\/pelvis radiographs obtained in patients to evaluate for FAI yielded 3050 total studies. Each hip was classified separately by the original interpreting radiologist in the following manner: 724 hips had severe cam-type FAI morphology, 962 moderate cam-type FAI morphology, 846 mild cam-type FAI morphology, and 518 hips were normal. The anteroposterior (AP) view from each study was anonymized and extracted. After localization of the hip joints by a novel convolutional neural network (CNN) based on the focal loss principle, a second CNN classified the images of the hip as cam positive, or no FAI. Accuracy was 74% for diagnosing normal vs. abnormal cam-type FAI morphology, with aggregate sensitivity and specificity of 0.821 and 0.669, respectively, at the chosen operating point. The aggregate AUC was 0.736. A deep learning system can be applied to detect FAI-related changes on single view pelvic radiographs. Deep learning is useful for quickly identifying and categorizing pathology on imaging, which may aid the interpreting radiologist.","doi":"10.1007\/s10278-023-00920-y","cleaned_title":"deep learning\u2013assisted identification of femoroacetabular impingement (fai) on routine pelvic radiographs","cleaned_abstract":"to use a novel deep learning system to localize the hip joints and detect findings of cam-type femoroacetabular impingement (fai). a retrospective search of hip\/pelvis radiographs obtained in patients to evaluate for fai yielded 3050 total studies. each hip was classified separately by the original interpreting radiologist in the following manner: 724 hips had severe cam-type fai morphology, 962 moderate cam-type fai morphology, 846 mild cam-type fai morphology, and 518 hips were normal. the anteroposterior (ap) view from each study was anonymized and extracted. after localization of the hip joints by a novel convolutional neural network (cnn) based on the focal loss principle, a second cnn classified the images of the hip as cam positive, or no fai. accuracy was 74% for diagnosing normal vs. abnormal cam-type fai morphology, with aggregate sensitivity and specificity of 0.821 and 0.669, respectively, at the chosen operating point. the aggregate auc was 0.736. a deep learning system can be applied to detect fai-related changes on single view pelvic radiographs. deep learning is useful for quickly identifying and categorizing pathology on imaging, which may aid the interpreting radiologist.","key_phrases":["a novel deep learning system","the hip joints","findings","cam-type femoroacetabular impingement (fai","a retrospective search","hip\/pelvis radiographs","patients","fai","3050 total studies","each hip","the original interpreting radiologist","the following manner","724 hips","severe cam-type fai morphology","962 moderate cam-type fai morphology","846 mild cam-type fai morphology","518 hips","the anteroposterior (ap) view","each study","localization","the hip joints","a novel convolutional neural network","cnn","the focal loss principle","a second cnn","the images","the hip","cam","no fai. accuracy","74%","abnormal cam-type fai morphology","aggregate sensitivity","specificity","the chosen operating point","the aggregate auc","a deep learning system","fai-related changes","single view pelvic radiographs","deep learning","categorizing pathology","imaging","which","the interpreting radiologist","3050","724","962","846","518","cnn","second","cnn","accuracy","74%","abnormal cam-type","0.821","0.669","0.736"]},{"title":"A deep learning-based framework for road traffic prediction","authors":["Redouane Benabdallah Benarmas","Kadda Beghdad Bey"],"abstract":"Due to the exponential rise in the number of vehicles and road segments in cities, traffic prediction becomes more difficult, necessitating the application of sophisticated algorithms such as deep learning (DL). The models used in the literature provide accurate predictions for specific cases when the data flow is properly prepared. However, in complex situations, these approaches fail, and thus, the prediction must be developed through a process rather than a prediction calculation method. In addition to using a pure and robust DL prediction model, an efficient approach could be built by taking into account two other factors, namely the relationships between road segments and the amount and quality of the training data. The main goal of our research is to develop a three-stage framework for road traffic prediction based on statistical and deep learning modules. First, a cross-correlation prediction with a Long Short-Term Memory model (LSTM) is implemented to predict the influential road segments; second, a deep generative model (DGM)-based data augmentation is used to improve the data of the related segments; and third, we adapt a Neural Basis Expansion Analysis for interpretable Time Series (N-BEATS) architecture, to the resulting data to implement the prediction module. The framework components are trained and validated using the 6th Beijing road traffic dataset.","doi":"10.1007\/s11227-023-05718-x","cleaned_title":"a deep learning-based framework for road traffic prediction","cleaned_abstract":"due to the exponential rise in the number of vehicles and road segments in cities, traffic prediction becomes more difficult, necessitating the application of sophisticated algorithms such as deep learning (dl). the models used in the literature provide accurate predictions for specific cases when the data flow is properly prepared. however, in complex situations, these approaches fail, and thus, the prediction must be developed through a process rather than a prediction calculation method. in addition to using a pure and robust dl prediction model, an efficient approach could be built by taking into account two other factors, namely the relationships between road segments and the amount and quality of the training data. the main goal of our research is to develop a three-stage framework for road traffic prediction based on statistical and deep learning modules. first, a cross-correlation prediction with a long short-term memory model (lstm) is implemented to predict the influential road segments; second, a deep generative model (dgm)-based data augmentation is used to improve the data of the related segments; and third, we adapt a neural basis expansion analysis for interpretable time series (n-beats) architecture, to the resulting data to implement the prediction module. the framework components are trained and validated using the 6th beijing road traffic dataset.","key_phrases":["the exponential rise","the number","vehicles","road segments","cities","traffic prediction","the application","sophisticated algorithms","deep learning","dl","the models","the literature","accurate predictions","specific cases","the data flow","complex situations","these approaches","the prediction","a process","a prediction calculation method","addition","a pure and robust dl prediction model","an efficient approach","account","two other factors","namely the relationships","road segments","the amount","quality","the training data","the main goal","our research","a three-stage framework","road traffic prediction","statistical and deep learning modules","first, a cross-correlation prediction","a long short-term memory model","lstm","the influential road segments","second, a deep generative model","dgm)-based data augmentation","the data","the related segments","we","a neural basis expansion analysis","interpretable time series (n-beats) architecture","the resulting data","the prediction module","the framework components","the 6th beijing road traffic dataset","two","three","first","second","third","6th"]},{"title":"Deep learning for Arabic healthcare: MedicalBot","authors":["Mohammed Abdelhay","Ammar Mohammed","Hesham A. Hefny"],"abstract":"Since the COVID-19 pandemic, healthcare services, particularly remote and automated healthcare consultations, have gained increased attention. Medical bots, which provide medical advice and support, are becoming increasingly popular. They offer numerous benefits, including 24\/7 access to medical counseling, reduced appointment wait times by providing quick answers to common questions or concerns, and cost savings associated with fewer visits or tests required for diagnosis and treatment plans. The success of medical bots depends on the quality of their learning, which in turn depends on the appropriate corpus within the domain of interest. Arabic is one of the most commonly used languages for sharing users\u2019 internet content. However, implementing medical bots in Arabic faces several challenges, including the language\u2019s morphological composition, the diversity of dialects, and the need for an appropriate and large enough corpus in the medical domain. To address this gap, this paper introduces the largest Arabic Healthcare Q &A dataset, called MAQA, consisting of over 430,000 questions distributed across 20 medical specializations. Furthermore, this paper adopts three deep learning models, namely LSTM, Bi-LSTM, and Transformers, for experimenting and benchmarking the proposed corpus MAQA. The experimental results demonstrate that the recent Transformer model outperforms the traditional deep learning models, achieving an average cosine similarity of 80.81% and a BLeU score of 58%.","doi":"10.1007\/s13278-023-01077-w","cleaned_title":"deep learning for arabic healthcare: medicalbot","cleaned_abstract":"since the covid-19 pandemic, healthcare services, particularly remote and automated healthcare consultations, have gained increased attention. medical bots, which provide medical advice and support, are becoming increasingly popular. they offer numerous benefits, including 24\/7 access to medical counseling, reduced appointment wait times by providing quick answers to common questions or concerns, and cost savings associated with fewer visits or tests required for diagnosis and treatment plans. the success of medical bots depends on the quality of their learning, which in turn depends on the appropriate corpus within the domain of interest. arabic is one of the most commonly used languages for sharing users\u2019 internet content. however, implementing medical bots in arabic faces several challenges, including the language\u2019s morphological composition, the diversity of dialects, and the need for an appropriate and large enough corpus in the medical domain. to address this gap, this paper introduces the largest arabic healthcare q &a dataset, called maqa, consisting of over 430,000 questions distributed across 20 medical specializations. furthermore, this paper adopts three deep learning models, namely lstm, bi-lstm, and transformers, for experimenting and benchmarking the proposed corpus maqa. the experimental results demonstrate that the recent transformer model outperforms the traditional deep learning models, achieving an average cosine similarity of 80.81% and a bleu score of 58%.","key_phrases":["the covid-19 pandemic, healthcare services","particularly remote and automated healthcare consultations","increased attention","medical bots","which","medical advice","support","they","numerous benefits","24\/7 access","medical counseling","reduced appointment","quick answers","common questions","concerns","cost savings","fewer visits","tests","diagnosis and treatment plans","the success","medical bots","the quality","their learning","which","turn","the appropriate corpus","the domain","interest","the most commonly used languages","users\u2019 internet content","medical bots","several challenges","the language\u2019s morphological composition","the diversity","dialects","the need","an appropriate and large enough corpus","the medical domain","this gap","this paper","the largest arabic healthcare q","a dataset","maqa","over 430,000 questions","20 medical specializations","this paper","three deep learning models","namely lstm","bi","-","lstm","transformers","the proposed corpus maqa","the experimental results","the recent transformer model","the traditional deep learning models","an average cosine similarity","80.81%","a bleu score","58%","covid-19","24\/7","arabic","one","arabic","arabic","healthcare q &a","over 430,000","20","three","80.81%","58%"]},{"title":"Optimization of news dissemination push mode by intelligent edge computing technology for deep learning","authors":["JiLe DeGe","Sina Sang"],"abstract":"The Internet era is an era of information explosion. By 2022, the global Internet users have reached more than 4 billion, and the social media users have exceeded 3 billion. People face a lot of news content every day, and it is almost impossible to get interesting information by browsing all the news content. Under this background, personalized news recommendation technology has been widely used, but it still needs to be further optimized and improved. In order to better push the news content of interest to different readers, users' satisfaction with major news websites should be further improved. This study proposes a new recommendation algorithm based on deep learning and reinforcement learning. Firstly, the RL algorithm is introduced based on deep learning. Deep learning is excellent in processing large-scale data and complex pattern recognition, but it often faces the challenge of low sample efficiency when it comes to complex decision-making and sequential tasks. While reinforcement learning (RL) emphasizes learning optimization strategies through continuous trial and error through interactive learning with the environment. Compared with deep learning, RL is more suitable for scenes that need long-term decision-making and trial-and-error learning. By feeding back the reward signal of the action, the system can better adapt to the unknown environment and complex tasks, which makes up for the relative shortcomings of deep learning in these aspects. A scenario is applied to an action to solve the sequential decision problem in the news dissemination process. In order to enable the news recommendation system to consider the dynamic changes in users' interest in news content, the Deep Deterministic Policy Gradient algorithm is applied to the news recommendation scenario. Opposing learning complements and combines Deep Q-network with the strategic network. On the basis of fully summarizing and thinking, this paper puts forward the mode of intelligent news dissemination and push. The push process of news communication information based on edge computing technology is proposed. Finally, based on Area Under Curve a Q-Leaning Area Under Curve for RL models is proposed. This indicator can measure the strengths and weaknesses of RL models efficiently and facilitates comparing models and evaluating offline experiments. The results show that the DDPG algorithm improves the click-through rate by 2.586% compared with the conventional recommendation algorithm. It shows that the algorithm designed in this paper has more obvious advantages in accurate recommendation by users. This paper effectively improves the efficiency of news dissemination by optimizing the push mode of intelligent news dissemination. In addition, the paper also deeply studies the innovative application of intelligent edge technology in news communication, which brings new ideas and practices to promote the development of news communication methods. Optimizing the push mode of intelligent news dissemination not only improves the user experience, but also provides strong support for the application of intelligent edge technology in this field, which has important practical application prospects.","doi":"10.1038\/s41598-024-53859-7","cleaned_title":"optimization of news dissemination push mode by intelligent edge computing technology for deep learning","cleaned_abstract":"the internet era is an era of information explosion. by 2022, the global internet users have reached more than 4 billion, and the social media users have exceeded 3 billion. people face a lot of news content every day, and it is almost impossible to get interesting information by browsing all the news content. under this background, personalized news recommendation technology has been widely used, but it still needs to be further optimized and improved. in order to better push the news content of interest to different readers, users' satisfaction with major news websites should be further improved. this study proposes a new recommendation algorithm based on deep learning and reinforcement learning. firstly, the rl algorithm is introduced based on deep learning. deep learning is excellent in processing large-scale data and complex pattern recognition, but it often faces the challenge of low sample efficiency when it comes to complex decision-making and sequential tasks. while reinforcement learning (rl) emphasizes learning optimization strategies through continuous trial and error through interactive learning with the environment. compared with deep learning, rl is more suitable for scenes that need long-term decision-making and trial-and-error learning. by feeding back the reward signal of the action, the system can better adapt to the unknown environment and complex tasks, which makes up for the relative shortcomings of deep learning in these aspects. a scenario is applied to an action to solve the sequential decision problem in the news dissemination process. in order to enable the news recommendation system to consider the dynamic changes in users' interest in news content, the deep deterministic policy gradient algorithm is applied to the news recommendation scenario. opposing learning complements and combines deep q-network with the strategic network. on the basis of fully summarizing and thinking, this paper puts forward the mode of intelligent news dissemination and push. the push process of news communication information based on edge computing technology is proposed. finally, based on area under curve a q-leaning area under curve for rl models is proposed. this indicator can measure the strengths and weaknesses of rl models efficiently and facilitates comparing models and evaluating offline experiments. the results show that the ddpg algorithm improves the click-through rate by 2.586% compared with the conventional recommendation algorithm. it shows that the algorithm designed in this paper has more obvious advantages in accurate recommendation by users. this paper effectively improves the efficiency of news dissemination by optimizing the push mode of intelligent news dissemination. in addition, the paper also deeply studies the innovative application of intelligent edge technology in news communication, which brings new ideas and practices to promote the development of news communication methods. optimizing the push mode of intelligent news dissemination not only improves the user experience, but also provides strong support for the application of intelligent edge technology in this field, which has important practical application prospects.","key_phrases":["the internet era","an era","information explosion","the global internet users","the social media users","people","a lot","news content","it","interesting information","all the news content","this background","personalized news recommendation technology","it","order","the news content","interest","different readers","users' satisfaction","major news websites","this study","a new recommendation algorithm","deep learning","reinforcement learning","the rl algorithm","deep learning","deep learning","large-scale data","complex pattern recognition","it","the challenge","low sample efficiency","it","complex decision-making","sequential tasks","reinforcement learning","(rl","optimization strategies","continuous trial","error","interactive learning","the environment","deep learning","rl","scenes","that","long-term decision-making and trial-and-error learning","the reward signal","the action","the system","the unknown environment","complex tasks","which","the relative shortcomings","deep learning","these aspects","a scenario","an action","the sequential decision problem","the news dissemination process","order","the news recommendation system","the dynamic changes","users' interest","news content","the deep deterministic policy gradient algorithm","the news recommendation scenario","complements","deep q-network","the strategic network","the basis","thinking","this paper","the mode","intelligent news dissemination","push","the push process","news communication information","edge computing technology","area","curve","a q-leaning area","curve","rl models","this indicator","the strengths","weaknesses","rl models","models","offline experiments","the results","the ddpg algorithm","the click-through rate","2.586%","the conventional recommendation algorithm","it","the algorithm","this paper","more obvious advantages","accurate recommendation","users","this paper","the efficiency","news dissemination","the push mode","intelligent news dissemination","addition","the paper","the innovative application","intelligent edge technology","news communication","which","new ideas","practices","the development","news communication methods","the push mode","intelligent news dissemination","the user experience","strong support","the application","intelligent edge technology","this field","which","important practical application prospects","2022","more than 4 billion","3 billion","firstly","2.586%"]},{"title":"Deep learning pose detection model for sow locomotion","authors":["Tauana Maria Carlos Guimar\u00e3es de Paula","Rafael Vieira de Sousa","Marisol Parada Sarmiento","Ton Kramer","Edson Jos\u00e9 de Souza Sardinha","Leandro Sabei","J\u00falia Silvestrini Machado","Mirela Vilioti","Adroaldo Jos\u00e9 Zanella"],"abstract":"Lameness affects animal mobility, causing pain and discomfort. Lameness in early stages often goes undetected due to a lack of observation, precision, and reliability. Automated and non-invasive systems offer precision and detection ease and may improve animal welfare. This study was conducted to create a repository of images and videos of sows with different locomotion scores. Our goal is to develop a computer vision model for automatically identifying specific points on the sow's body. The automatic identification and ability to track specific body areas, will allow us to conduct kinematic studies with the aim of facilitating the detection of lameness using deep learning. The video database was collected on a pig farm with a scenario built to allow filming of sows in locomotion with different lameness scores. Two stereo cameras were used to record 2D videos images. Thirteen locomotion experts assessed the videos using the Locomotion Score System developed by Zinpro Corporation. From this annotated repository, computational models were trained and tested using the open-source deep learning-based animal pose tracking framework SLEAP (Social LEAP Estimates Animal Poses). The top-performing models were constructed using the LEAP architecture to accurately track 6 (lateral view) and 10 (dorsal view) skeleton keypoints. The architecture achieved average precisions values of 0.90 and 0.72, average distances of 6.83 and 11.37 in pixel, and similarities of 0.94 and 0.86 for the lateral and dorsal views, respectively. These computational models are proposed as a Precision Livestock Farming tool and method for identifying and estimating postures in pigs automatically and objectively. The 2D video image repository with different pig locomotion scores can be used as a tool for teaching and research. Based on our skeleton keypoint classification results, an automatic system could be developed. This could contribute to the objective assessment of locomotion scores in sows, improving their welfare.","doi":"10.1038\/s41598-024-62151-7","cleaned_title":"deep learning pose detection model for sow locomotion","cleaned_abstract":"lameness affects animal mobility, causing pain and discomfort. lameness in early stages often goes undetected due to a lack of observation, precision, and reliability. automated and non-invasive systems offer precision and detection ease and may improve animal welfare. this study was conducted to create a repository of images and videos of sows with different locomotion scores. our goal is to develop a computer vision model for automatically identifying specific points on the sow's body. the automatic identification and ability to track specific body areas, will allow us to conduct kinematic studies with the aim of facilitating the detection of lameness using deep learning. the video database was collected on a pig farm with a scenario built to allow filming of sows in locomotion with different lameness scores. two stereo cameras were used to record 2d videos images. thirteen locomotion experts assessed the videos using the locomotion score system developed by zinpro corporation. from this annotated repository, computational models were trained and tested using the open-source deep learning-based animal pose tracking framework sleap (social leap estimates animal poses). the top-performing models were constructed using the leap architecture to accurately track 6 (lateral view) and 10 (dorsal view) skeleton keypoints. the architecture achieved average precisions values of 0.90 and 0.72, average distances of 6.83 and 11.37 in pixel, and similarities of 0.94 and 0.86 for the lateral and dorsal views, respectively. these computational models are proposed as a precision livestock farming tool and method for identifying and estimating postures in pigs automatically and objectively. the 2d video image repository with different pig locomotion scores can be used as a tool for teaching and research. based on our skeleton keypoint classification results, an automatic system could be developed. this could contribute to the objective assessment of locomotion scores in sows, improving their welfare.","key_phrases":["animal mobility","pain","discomfort","early stages","a lack","observation","precision","reliability","automated and non-invasive systems","precision and detection ease","animal welfare","this study","a repository","images","videos","sows","different locomotion scores","our goal","a computer vision model","specific points","the sow's body","the automatic identification","ability","specific body areas","us","kinematic studies","the aim","the detection","lameness","deep learning","the video database","a pig farm","a scenario","filming","sows","locomotion","different lameness scores","two stereo cameras","2d videos images","thirteen locomotion experts","the videos","the locomotion score system","zinpro corporation","this annotated repository","computational models","the open-source deep learning-based animal","framework sleap","social leap","animal poses","the top-performing models","the leap architecture","6 (lateral view","10 (dorsal view) skeleton keypoints","the architecture","average precisions values","average distances","pixel","similarities","the lateral and dorsal views","these computational models","a precision livestock farming tool","method","postures","pigs","the 2d video image repository","different pig locomotion scores","a tool","teaching","research","our skeleton keypoint classification results","an automatic system","this","the objective assessment","locomotion scores","sows","their welfare","two","2d","thirteen","zinpro corporation","6","10","0.90","0.72","6.83","11.37","0.94","0.86","2d"]},{"title":"Deep learning automatically assesses 2-\u00b5m laser-induced skin damage OCT images","authors":["Changke Wang","Qiong Ma","Yu Wei","Qi Liu","Yuqing Wang","Chenliang Xu","Caihui Li","Qingyu Cai","Haiyang Sun","Xiaoan Tang","Hongxiang Kang"],"abstract":"The present study proposed a noninvasive, automated, in vivo assessment method based on optical coherence tomography (OCT) and deep learning techniques to qualitatively and quantitatively analyze the biological effects of 2-\u00b5m laser-induced skin damage at different irradiation doses. Different doses of 2-\u00b5m laser irradiation established a mouse skin damage model, after which the skin-damaged tissues were imaged non-invasively in vivo using OCT. The acquired images were preprocessed to construct the dataset required for deep learning. The deep learning models used were U-Net, DeepLabV3+, PSP-Net, and HR-Net, and the trained models were used to segment the damage images and further quantify the damage volume of mouse skin under different irradiation doses. The comparison of the qualitative and quantitative results of the four network models showed that HR-Net had the best performance, the highest agreement between the segmentation results and real values, and the smallest error in the quantitative assessment of the damage volume. Based on HR-Net to segment the damage image and quantify the damage volume, the irradiation doses 5.41, 9.55, 13.05, 20.85, 32.71, 52.92, 76.71, and 97.24\u00a0J\/cm\u00b2 corresponded to a damage volume of 4.58, 12.56, 16.74, 20.88, 24.52, 30.75, 34.13, and 37.32\u00a0mm\u00b3. The damage volume increased in a radiation dose-dependent manner.","doi":"10.1007\/s10103-024-04053-8","cleaned_title":"deep learning automatically assesses 2-\u00b5m laser-induced skin damage oct images","cleaned_abstract":"the present study proposed a noninvasive, automated, in vivo assessment method based on optical coherence tomography (oct) and deep learning techniques to qualitatively and quantitatively analyze the biological effects of 2-\u00b5m laser-induced skin damage at different irradiation doses. different doses of 2-\u00b5m laser irradiation established a mouse skin damage model, after which the skin-damaged tissues were imaged non-invasively in vivo using oct. the acquired images were preprocessed to construct the dataset required for deep learning. the deep learning models used were u-net, deeplabv3+, psp-net, and hr-net, and the trained models were used to segment the damage images and further quantify the damage volume of mouse skin under different irradiation doses. the comparison of the qualitative and quantitative results of the four network models showed that hr-net had the best performance, the highest agreement between the segmentation results and real values, and the smallest error in the quantitative assessment of the damage volume. based on hr-net to segment the damage image and quantify the damage volume, the irradiation doses 5.41, 9.55, 13.05, 20.85, 32.71, 52.92, 76.71, and 97.24 j\/cm\u00b2 corresponded to a damage volume of 4.58, 12.56, 16.74, 20.88, 24.52, 30.75, 34.13, and 37.32 mm\u00b3. the damage volume increased in a radiation dose-dependent manner.","key_phrases":["the present study","vivo assessment method","optical coherence tomography","oct","techniques","the biological effects","2-\u00b5m laser-induced skin damage","different irradiation doses","different doses","2-\u00b5m laser irradiation","a mouse skin damage model","which","the skin-damaged tissues","vivo","oct","the acquired images","the dataset","deep learning","the deep learning models","u","-","net","psp-net","hr-net","the trained models","the damage images","the damage volume","mouse skin","different irradiation doses","the comparison","the qualitative and quantitative results","the four network models","hr-net","the best performance","the highest agreement","the segmentation results","real values","the smallest error","the quantitative assessment","the damage volume","hr-net","the damage image","the damage volume","the irradiation","97.24 j\/cm\u00b2","a damage volume","37.32 mm\u00b3.","the damage volume","a radiation dose-dependent manner","2-\u00b5m","2-\u00b5m","oct","four","5.41","9.55","13.05","20.85","32.71","52.92","76.71","97.24","4.58","12.56","16.74","20.88","24.52","30.75","34.13","37.32"]},{"title":"Low Dose CT Image Reconstruction Using Deep Convolutional Residual Learning Network","authors":["Shalini Ramanathan","Mohan Ramasundaram"],"abstract":"Image reconstruction from computed tomography measurement is formulated as a thought-provoking statistical inverse problem. Deep learning algorithms are best for ill-posed statistical inverse problems that presently achieve state-of-art reconstruction results. The challenging task is to lower the potentially harmful radiation a patient is exposed to during the CT scan. In recently available CT Scanners, Low-Dose CT (LDCT) reconstruction is presented with a post-processing approach, which uses deep learning-based medical image reconstruction methods to reduce the dose level without compromising the image quality. Therefore, this paper proposes a deep learning-based post-processing method called Deep Convolutional Neural Network with Residual Learning (DCNN-RL). The method trains the network on a newly available low-dose CT benchmark dataset (LoDoPaB-CT). It also enables to compare with other benchmark CT datasets such as AAPM LDCT and COVIDx-CT. The proposed architecture optimizes the filtering part to minimize the error function. It learns the parameters of the residual network via numerous training to maximize the efficiency of production. This paper compares noise methods on DCNN-RL using various LDCT datasets of the same domain (human being's chest CT scan) to analyze the image quality. The experiment findings suggest that the Adagrad optimizer is the best for LDCT images. Gaussian noise with a minor variance outperforms the medical image reconstruction task. Here, it has been demonstrated that this approach with these benchmark datasets drastically improves the medical CT image quality, shown through qualitative and quantitative outcomes.","doi":"10.1007\/s42979-023-02210-4","cleaned_title":"low dose ct image reconstruction using deep convolutional residual learning network","cleaned_abstract":"image reconstruction from computed tomography measurement is formulated as a thought-provoking statistical inverse problem. deep learning algorithms are best for ill-posed statistical inverse problems that presently achieve state-of-art reconstruction results. the challenging task is to lower the potentially harmful radiation a patient is exposed to during the ct scan. in recently available ct scanners, low-dose ct (ldct) reconstruction is presented with a post-processing approach, which uses deep learning-based medical image reconstruction methods to reduce the dose level without compromising the image quality. therefore, this paper proposes a deep learning-based post-processing method called deep convolutional neural network with residual learning (dcnn-rl). the method trains the network on a newly available low-dose ct benchmark dataset (lodopab-ct). it also enables to compare with other benchmark ct datasets such as aapm ldct and covidx-ct. the proposed architecture optimizes the filtering part to minimize the error function. it learns the parameters of the residual network via numerous training to maximize the efficiency of production. this paper compares noise methods on dcnn-rl using various ldct datasets of the same domain (human being's chest ct scan) to analyze the image quality. the experiment findings suggest that the adagrad optimizer is the best for ldct images. gaussian noise with a minor variance outperforms the medical image reconstruction task. here, it has been demonstrated that this approach with these benchmark datasets drastically improves the medical ct image quality, shown through qualitative and quantitative outcomes.","key_phrases":["image reconstruction","computed tomography measurement","a thought-provoking statistical inverse problem","deep learning algorithms","ill-posed statistical inverse problems","that","art","the challenging task","the potentially harmful radiation","a patient","the ct scan","recently available ct scanners","low-dose ct (ldct) reconstruction","a post-processing approach","which","deep learning-based medical image reconstruction methods","the dose level","the image quality","this paper","a deep learning-based post-processing method","deep convolutional neural network","residual learning","dcnn-rl","the method","the network","a newly available low-dose ct benchmark dataset","lodopab-ct","it","other benchmark ct datasets","aapm ldct","covidx-ct","the proposed architecture","the filtering part","the error function","it","the parameters","the residual network","numerous training","the efficiency","production","this paper","noise methods","dcnn-rl","various ldct datasets","the same domain","human being's chest","ct scan","the image quality","the experiment findings","the adagrad optimizer","ldct images","gaussian noise","a minor variance","the medical image reconstruction task","it","this approach","these benchmark datasets","the medical ct image quality","qualitative and quantitative outcomes","gaussian"]},{"title":"Transforming clinical virology with AI, machine learning and deep learning: a comprehensive review and outlook","authors":["Abhishek Padhi","Ashwini Agarwal","Shailendra K. Saxena","C. D. S. Katoch"],"abstract":"In the rapidly evolving field of clinical virology, technological advancements have always played a pivotal role in driving transformative changes. This comprehensive review delves into the burgeoning integration of artificial intelligence (AI), machine learning, and deep learning into virological research and practice. As we elucidate, these computational tools have significantly enhanced diagnostic precision, therapeutic interventions, and epidemiological monitoring. Through in-depth analyses of notable case studies, we showcase how algorithms can optimize viral genome sequencing, accelerate drug discovery, and offer predictive insights into viral outbreaks. However, with these advancements come inherent challenges, particularly in data security, algorithmic biases, and ethical considerations. Addressing these challenges head-on, we discuss potential remedial measures and underscore the significance of interdisciplinary collaboration between virologists, data scientists, and ethicists. Conclusively, this review posits an outlook that anticipates a symbiotic relationship between AI-driven tools and virology, heralding a new era of proactive and personalized patient care.","doi":"10.1007\/s13337-023-00841-y","cleaned_title":"transforming clinical virology with ai, machine learning and deep learning: a comprehensive review and outlook","cleaned_abstract":"in the rapidly evolving field of clinical virology, technological advancements have always played a pivotal role in driving transformative changes. this comprehensive review delves into the burgeoning integration of artificial intelligence (ai), machine learning, and deep learning into virological research and practice. as we elucidate, these computational tools have significantly enhanced diagnostic precision, therapeutic interventions, and epidemiological monitoring. through in-depth analyses of notable case studies, we showcase how algorithms can optimize viral genome sequencing, accelerate drug discovery, and offer predictive insights into viral outbreaks. however, with these advancements come inherent challenges, particularly in data security, algorithmic biases, and ethical considerations. addressing these challenges head-on, we discuss potential remedial measures and underscore the significance of interdisciplinary collaboration between virologists, data scientists, and ethicists. conclusively, this review posits an outlook that anticipates a symbiotic relationship between ai-driven tools and virology, heralding a new era of proactive and personalized patient care.","key_phrases":["the rapidly evolving field","clinical virology","technological advancements","a pivotal role","transformative changes","this comprehensive review","the burgeoning integration","artificial intelligence","ai","machine learning","deep learning","virological research","practice","we","these computational tools","diagnostic precision","therapeutic interventions","epidemiological monitoring","-depth","notable case studies","we","algorithms","drug discovery","predictive insights","viral outbreaks","these advancements","inherent challenges","data security","algorithmic biases","ethical considerations","these challenges","we","potential remedial measures","the significance","interdisciplinary collaboration","virologists","data scientists","ethicists","this review","an outlook","that","a symbiotic relationship","ai-driven tools","virology","a new era","proactive and personalized patient care"]},{"title":"A secured deep learning based smart home automation system","authors":["Chitukula Sanjay","Konda Jahnavi","Shyam Karanth"],"abstract":"With the expansion of modern technologies and the Internet of Things (IoT), the concept of smart homes has gained tremendous popularity with a view to making people\u2019s lives easier by ensuring a secured environment. Several home automation systems have been developed to report suspicious activities by capturing the movements of residents. However, these systems are associated with challenges such as weak security, lack of interoperability and integration with IoT devices, timely reporting of suspicious movements, etc. Therefore, the given paper proposes a novel smart home automation framework for controlling home appliances by integrating with sensors, IoT devices, and microcontrollers, which would in turn monitor the movements and send notifications about suspicious movements on the resident\u2019s smartphone. The proposed framework makes use of convolutional neural networks (CNNs) for motion detection and classification based on pre-processing of images. The images related to the movements of residents are captured by a spy camera installed in the system. It helps in identification of outsiders based on differentiation of motion patterns. The performance of the framework is compared with existing deep learning models used in recent studies based on evaluation metrics such as accuracy (%), precision (%), recall (%), and f-1 measure (%). The results show that the proposed framework attains the highest accuracy (98.67%), thereby surpassing the existing deep learning models used in smart home automation systems.","doi":"10.1007\/s41870-024-02097-1","cleaned_title":"a secured deep learning based smart home automation system","cleaned_abstract":"with the expansion of modern technologies and the internet of things (iot), the concept of smart homes has gained tremendous popularity with a view to making people\u2019s lives easier by ensuring a secured environment. several home automation systems have been developed to report suspicious activities by capturing the movements of residents. however, these systems are associated with challenges such as weak security, lack of interoperability and integration with iot devices, timely reporting of suspicious movements, etc. therefore, the given paper proposes a novel smart home automation framework for controlling home appliances by integrating with sensors, iot devices, and microcontrollers, which would in turn monitor the movements and send notifications about suspicious movements on the resident\u2019s smartphone. the proposed framework makes use of convolutional neural networks (cnns) for motion detection and classification based on pre-processing of images. the images related to the movements of residents are captured by a spy camera installed in the system. it helps in identification of outsiders based on differentiation of motion patterns. the performance of the framework is compared with existing deep learning models used in recent studies based on evaluation metrics such as accuracy (%), precision (%), recall (%), and f-1 measure (%). the results show that the proposed framework attains the highest accuracy (98.67%), thereby surpassing the existing deep learning models used in smart home automation systems.","key_phrases":["the expansion","modern technologies","the internet","things","iot","the concept","smart homes","tremendous popularity","a view","people","a secured environment","several home automation systems","suspicious activities","the movements","residents","these systems","challenges","weak security","lack","interoperability","integration","iot devices","timely reporting","suspicious movements","the given paper","a novel smart home automation framework","home appliances","sensors","iot devices","microcontrollers","which","turn","the movements","notifications","suspicious movements","the resident\u2019s smartphone","the proposed framework","use","convolutional neural networks","cnns","motion detection","classification","pre","processing","images","the images","the movements","residents","a spy camera","the system","it","identification","outsiders","differentiation","motion patterns","the performance","the framework","existing deep learning models","recent studies","evaluation metrics","accuracy","precision","recall","f-1 measure","the results","the proposed framework","the highest accuracy","98.67%","the existing deep learning models","smart home automation systems","98.67%"]},{"title":"A hyperspectral deep learning attention model for predicting lettuce chlorophyll content","authors":["Ziran Ye","Xiangfeng Tan","Mengdi Dai","Xuting Chen","Yuanxiang Zhong","Yi Zhang","Yunjie Ruan","Dedong Kong"],"abstract":"BackgroundThe phenotypic traits of leaves are the direct reflection of the agronomic traits in the growth process of leafy vegetables, which plays a vital role in the selection of high-quality leafy vegetable varieties. The current image-based phenotypic traits extraction research mainly focuses on the morphological and structural traits of plants or leaves, and there are few studies on the phenotypes of physiological traits of leaves. The current research has developed a deep learning model aimed at predicting the total chlorophyll of greenhouse lettuce directly from the full spectrum of hyperspectral images.ResultsA CNN-based one-dimensional deep learning model with spectral attention module was utilized for the estimate of the total chlorophyll of greenhouse lettuce from the full spectrum of hyperspectral images. Experimental results demonstrate that the deep neural network with spectral attention module outperformed the existing standard approaches, including partial least squares regression (PLSR) and random forest (RF), with an average R2 of 0.746 and an average RMSE of 2.018.ConclusionsThis study unveils the capability of leveraging deep attention networks and hyperspectral imaging for estimating lettuce chlorophyll levels. This approach offers a convenient, non-destructive, and effective estimation method for the automatic monitoring and production management of leafy vegetables.","doi":"10.1186\/s13007-024-01148-9","cleaned_title":"a hyperspectral deep learning attention model for predicting lettuce chlorophyll content","cleaned_abstract":"backgroundthe phenotypic traits of leaves are the direct reflection of the agronomic traits in the growth process of leafy vegetables, which plays a vital role in the selection of high-quality leafy vegetable varieties. the current image-based phenotypic traits extraction research mainly focuses on the morphological and structural traits of plants or leaves, and there are few studies on the phenotypes of physiological traits of leaves. the current research has developed a deep learning model aimed at predicting the total chlorophyll of greenhouse lettuce directly from the full spectrum of hyperspectral images.resultsa cnn-based one-dimensional deep learning model with spectral attention module was utilized for the estimate of the total chlorophyll of greenhouse lettuce from the full spectrum of hyperspectral images. experimental results demonstrate that the deep neural network with spectral attention module outperformed the existing standard approaches, including partial least squares regression (plsr) and random forest (rf), with an average r2 of 0.746 and an average rmse of 2.018.conclusionsthis study unveils the capability of leveraging deep attention networks and hyperspectral imaging for estimating lettuce chlorophyll levels. this approach offers a convenient, non-destructive, and effective estimation method for the automatic monitoring and production management of leafy vegetables.","key_phrases":["backgroundthe phenotypic traits","leaves","the direct reflection","the agronomic traits","the growth process","leafy vegetables","which","a vital role","the selection","high-quality leafy vegetable varieties","the current image-based phenotypic traits extraction research","the morphological and structural traits","plants","leaves","few studies","the phenotypes","physiological traits","leaves","the current research","a deep learning model","the total chlorophyll","greenhouse lettuce","the full spectrum","hyperspectral","images.resultsa cnn-based one-dimensional deep learning model","spectral attention module","the estimate","the total chlorophyll","greenhouse lettuce","the full spectrum","hyperspectral images","experimental results","the deep neural network","spectral attention module","the existing standard approaches","partial least squares regression","plsr","random forest","rf","an average r2","an average rmse","2.018.conclusionsthis study","the capability","deep attention networks","hyperspectral imaging","lettuce chlorophyll levels","this approach","destructive, and effective estimation method","the automatic monitoring","production management","leafy vegetables","cnn","0.746"]},{"title":"Low Dose CT Image Reconstruction Using Deep Convolutional Residual Learning Network","authors":["Shalini Ramanathan","Mohan Ramasundaram"],"abstract":"Image reconstruction from computed tomography measurement is formulated as a thought-provoking statistical inverse problem. Deep learning algorithms are best for ill-posed statistical inverse problems that presently achieve state-of-art reconstruction results. The challenging task is to lower the potentially harmful radiation a patient is exposed to during the CT scan. In recently available CT Scanners, Low-Dose CT (LDCT) reconstruction is presented with a post-processing approach, which uses deep learning-based medical image reconstruction methods to reduce the dose level without compromising the image quality. Therefore, this paper proposes a deep learning-based post-processing method called Deep Convolutional Neural Network with Residual Learning (DCNN-RL). The method trains the network on a newly available low-dose CT benchmark dataset (LoDoPaB-CT). It also enables to compare with other benchmark CT datasets such as AAPM LDCT and COVIDx-CT. The proposed architecture optimizes the filtering part to minimize the error function. It learns the parameters of the residual network via numerous training to maximize the efficiency of production. This paper compares noise methods on DCNN-RL using various LDCT datasets of the same domain (human being's chest CT scan) to analyze the image quality. The experiment findings suggest that the Adagrad optimizer is the best for LDCT images. Gaussian noise with a minor variance outperforms the medical image reconstruction task. Here, it has been demonstrated that this approach with these benchmark datasets drastically improves the medical CT image quality, shown through qualitative and quantitative outcomes.","doi":"10.1007\/s42979-023-02210-4","cleaned_title":"low dose ct image reconstruction using deep convolutional residual learning network","cleaned_abstract":"image reconstruction from computed tomography measurement is formulated as a thought-provoking statistical inverse problem. deep learning algorithms are best for ill-posed statistical inverse problems that presently achieve state-of-art reconstruction results. the challenging task is to lower the potentially harmful radiation a patient is exposed to during the ct scan. in recently available ct scanners, low-dose ct (ldct) reconstruction is presented with a post-processing approach, which uses deep learning-based medical image reconstruction methods to reduce the dose level without compromising the image quality. therefore, this paper proposes a deep learning-based post-processing method called deep convolutional neural network with residual learning (dcnn-rl). the method trains the network on a newly available low-dose ct benchmark dataset (lodopab-ct). it also enables to compare with other benchmark ct datasets such as aapm ldct and covidx-ct. the proposed architecture optimizes the filtering part to minimize the error function. it learns the parameters of the residual network via numerous training to maximize the efficiency of production. this paper compares noise methods on dcnn-rl using various ldct datasets of the same domain (human being's chest ct scan) to analyze the image quality. the experiment findings suggest that the adagrad optimizer is the best for ldct images. gaussian noise with a minor variance outperforms the medical image reconstruction task. here, it has been demonstrated that this approach with these benchmark datasets drastically improves the medical ct image quality, shown through qualitative and quantitative outcomes.","key_phrases":["image reconstruction","computed tomography measurement","a thought-provoking statistical inverse problem","deep learning algorithms","ill-posed statistical inverse problems","that","art","the challenging task","the potentially harmful radiation","a patient","the ct scan","recently available ct scanners","low-dose ct (ldct) reconstruction","a post-processing approach","which","deep learning-based medical image reconstruction methods","the dose level","the image quality","this paper","a deep learning-based post-processing method","deep convolutional neural network","residual learning","dcnn-rl","the method","the network","a newly available low-dose ct benchmark dataset","lodopab-ct","it","other benchmark ct datasets","aapm ldct","covidx-ct","the proposed architecture","the filtering part","the error function","it","the parameters","the residual network","numerous training","the efficiency","production","this paper","noise methods","dcnn-rl","various ldct datasets","the same domain","human being's chest","ct scan","the image quality","the experiment findings","the adagrad optimizer","ldct images","gaussian noise","a minor variance","the medical image reconstruction task","it","this approach","these benchmark datasets","the medical ct image quality","qualitative and quantitative outcomes","gaussian"]},{"title":"Meta-heuristic-based hybrid deep learning model for vulnerability detection and prevention in software system","authors":["Lijin Shaji","R. Suji Pramila"],"abstract":"Software vulnerabilities are flaws that may be exploited to cause loss or harm. Various automated machine-learning techniques have been developed in preceding studies to detect software vulnerabilities. This work tries to develop a technique for securing the software on the basis of their vulnerabilities that are already known, by developing a hybrid deep learning model to detect those vulnerabilities. Moreover, certain countermeasures are suggested based on the types of vulnerability to prevent the attack further. For different software projects taken as the dataset, feature fusion is done by utilizing canonical correlation analysis together with Deep Residual Network (DRN). A hybrid deep learning technique trained using AdamW-Rat Swarm Optimizer (AdamW-RSO) is designed to detect software vulnerability. Hybrid deep learning makes use of the Deep Belief Network (DBN) and Generative Adversarial Network (GAN). For every vulnerability, its location of occurrence within the software development procedures and techniques of alleviation via implementation level or design level activities are described. Thus, it helps in understanding the appearance of vulnerabilities, suggesting the use of various countermeasures during the initial phases of software design, and therefore, assures software security. Evaluating the performance of vulnerability detection by the proposed technique regarding recall, precision, and f-measure, it is found to be more effective than the existing methods.","doi":"10.1007\/s10878-024-01185-z","cleaned_title":"meta-heuristic-based hybrid deep learning model for vulnerability detection and prevention in software system","cleaned_abstract":"software vulnerabilities are flaws that may be exploited to cause loss or harm. various automated machine-learning techniques have been developed in preceding studies to detect software vulnerabilities. this work tries to develop a technique for securing the software on the basis of their vulnerabilities that are already known, by developing a hybrid deep learning model to detect those vulnerabilities. moreover, certain countermeasures are suggested based on the types of vulnerability to prevent the attack further. for different software projects taken as the dataset, feature fusion is done by utilizing canonical correlation analysis together with deep residual network (drn). a hybrid deep learning technique trained using adamw-rat swarm optimizer (adamw-rso) is designed to detect software vulnerability. hybrid deep learning makes use of the deep belief network (dbn) and generative adversarial network (gan). for every vulnerability, its location of occurrence within the software development procedures and techniques of alleviation via implementation level or design level activities are described. thus, it helps in understanding the appearance of vulnerabilities, suggesting the use of various countermeasures during the initial phases of software design, and therefore, assures software security. evaluating the performance of vulnerability detection by the proposed technique regarding recall, precision, and f-measure, it is found to be more effective than the existing methods.","key_phrases":["software vulnerabilities","flaws","that","loss","harm","various automated machine-learning techniques","studies","software vulnerabilities","this work","a technique","the software","the basis","their vulnerabilities","that","a hybrid deep learning model","those vulnerabilities","certain countermeasures","the types","vulnerability","the attack","different software projects","the dataset","feature fusion","canonical correlation analysis","deep residual network","drn","a hybrid deep learning technique","adamw-rat swarm optimizer","adamw-rso","software vulnerability","hybrid deep learning","use","the deep belief network","dbn","adversarial network","gan","every vulnerability","its location","occurrence","the software development procedures","techniques","alleviation","implementation level or design level activities","it","the appearance","vulnerabilities","the use","various countermeasures","the initial phases","software design","therefore, assures software security","the performance","vulnerability detection","the proposed technique","recall","precision","f-measure","it","the existing methods"]},{"title":"Deep learning pose detection model for sow locomotion","authors":["Tauana Maria Carlos Guimar\u00e3es de Paula","Rafael Vieira de Sousa","Marisol Parada Sarmiento","Ton Kramer","Edson Jos\u00e9 de Souza Sardinha","Leandro Sabei","J\u00falia Silvestrini Machado","Mirela Vilioti","Adroaldo Jos\u00e9 Zanella"],"abstract":"Lameness affects animal mobility, causing pain and discomfort. Lameness in early stages often goes undetected due to a lack of observation, precision, and reliability. Automated and non-invasive systems offer precision and detection ease and may improve animal welfare. This study was conducted to create a repository of images and videos of sows with different locomotion scores. Our goal is to develop a computer vision model for automatically identifying specific points on the sow's body. The automatic identification and ability to track specific body areas, will allow us to conduct kinematic studies with the aim of facilitating the detection of lameness using deep learning. The video database was collected on a pig farm with a scenario built to allow filming of sows in locomotion with different lameness scores. Two stereo cameras were used to record 2D videos images. Thirteen locomotion experts assessed the videos using the Locomotion Score System developed by Zinpro Corporation. From this annotated repository, computational models were trained and tested using the open-source deep learning-based animal pose tracking framework SLEAP (Social LEAP Estimates Animal Poses). The top-performing models were constructed using the LEAP architecture to accurately track 6 (lateral view) and 10 (dorsal view) skeleton keypoints. The architecture achieved average precisions values of 0.90 and 0.72, average distances of 6.83 and 11.37 in pixel, and similarities of 0.94 and 0.86 for the lateral and dorsal views, respectively. These computational models are proposed as a Precision Livestock Farming tool and method for identifying and estimating postures in pigs automatically and objectively. The 2D video image repository with different pig locomotion scores can be used as a tool for teaching and research. Based on our skeleton keypoint classification results, an automatic system could be developed. This could contribute to the objective assessment of locomotion scores in sows, improving their welfare.","doi":"10.1038\/s41598-024-62151-7","cleaned_title":"deep learning pose detection model for sow locomotion","cleaned_abstract":"lameness affects animal mobility, causing pain and discomfort. lameness in early stages often goes undetected due to a lack of observation, precision, and reliability. automated and non-invasive systems offer precision and detection ease and may improve animal welfare. this study was conducted to create a repository of images and videos of sows with different locomotion scores. our goal is to develop a computer vision model for automatically identifying specific points on the sow's body. the automatic identification and ability to track specific body areas, will allow us to conduct kinematic studies with the aim of facilitating the detection of lameness using deep learning. the video database was collected on a pig farm with a scenario built to allow filming of sows in locomotion with different lameness scores. two stereo cameras were used to record 2d videos images. thirteen locomotion experts assessed the videos using the locomotion score system developed by zinpro corporation. from this annotated repository, computational models were trained and tested using the open-source deep learning-based animal pose tracking framework sleap (social leap estimates animal poses). the top-performing models were constructed using the leap architecture to accurately track 6 (lateral view) and 10 (dorsal view) skeleton keypoints. the architecture achieved average precisions values of 0.90 and 0.72, average distances of 6.83 and 11.37 in pixel, and similarities of 0.94 and 0.86 for the lateral and dorsal views, respectively. these computational models are proposed as a precision livestock farming tool and method for identifying and estimating postures in pigs automatically and objectively. the 2d video image repository with different pig locomotion scores can be used as a tool for teaching and research. based on our skeleton keypoint classification results, an automatic system could be developed. this could contribute to the objective assessment of locomotion scores in sows, improving their welfare.","key_phrases":["animal mobility","pain","discomfort","early stages","a lack","observation","precision","reliability","automated and non-invasive systems","precision and detection ease","animal welfare","this study","a repository","images","videos","sows","different locomotion scores","our goal","a computer vision model","specific points","the sow's body","the automatic identification","ability","specific body areas","us","kinematic studies","the aim","the detection","lameness","deep learning","the video database","a pig farm","a scenario","filming","sows","locomotion","different lameness scores","two stereo cameras","2d videos images","thirteen locomotion experts","the videos","the locomotion score system","zinpro corporation","this annotated repository","computational models","the open-source deep learning-based animal","framework sleap","social leap","animal poses","the top-performing models","the leap architecture","6 (lateral view","10 (dorsal view) skeleton keypoints","the architecture","average precisions values","average distances","pixel","similarities","the lateral and dorsal views","these computational models","a precision livestock farming tool","method","postures","pigs","the 2d video image repository","different pig locomotion scores","a tool","teaching","research","our skeleton keypoint classification results","an automatic system","this","the objective assessment","locomotion scores","sows","their welfare","two","2d","thirteen","zinpro corporation","6","10","0.90","0.72","6.83","11.37","0.94","0.86","2d"]},{"title":"Sentiment analysis of Canadian maritime case law: a sentiment case law and deep learning approach","authors":["Bola Abimbola","Qing Tan","Enrique A. De La Cal Mar\u00edn"],"abstract":"Historical information in the Canadian Maritime Judiciary increases with time because of the need to archive data to be utilized in case references and for later application when determining verdicts for similar cases. However, such data are typically stored in multiple systems, making its reachability technical. Utilizing technologies like deep learning and sentiment analysis provides chances to facilitate faster access to court records. Such practice enhances impartial verdicts, minimizing workloads for court employees, and decreases the time used in legal proceedings for claims during maritime contracts such as shipping disputes between parties. This paper seeks to develop a sentiment analysis framework that uses deep learning, distributed learning, and machine learning to improve access to statutes, laws, and cases used by maritime judges in making judgments to back their claims. The suggested approach uses deep learning models, including convolutional neural networks (CNNs), deep neural networks, long short-term memory (LSTM), and recurrent neural networks. It extracts court records having crucial sentiments or statements for maritime court verdicts. The suggested approach has been used successfully during sentiment analysis by emphasizing feature selection from a legal repository. The LSTM\u2009+\u2009CNN model has shown promising results in obtaining sentiments and records from multiple devices and sufficiently proposing practical guidance to judicial personnel regarding the regulations applicable to various situations.","doi":"10.1007\/s41870-024-01820-2","cleaned_title":"sentiment analysis of canadian maritime case law: a sentiment case law and deep learning approach","cleaned_abstract":"historical information in the canadian maritime judiciary increases with time because of the need to archive data to be utilized in case references and for later application when determining verdicts for similar cases. however, such data are typically stored in multiple systems, making its reachability technical. utilizing technologies like deep learning and sentiment analysis provides chances to facilitate faster access to court records. such practice enhances impartial verdicts, minimizing workloads for court employees, and decreases the time used in legal proceedings for claims during maritime contracts such as shipping disputes between parties. this paper seeks to develop a sentiment analysis framework that uses deep learning, distributed learning, and machine learning to improve access to statutes, laws, and cases used by maritime judges in making judgments to back their claims. the suggested approach uses deep learning models, including convolutional neural networks (cnns), deep neural networks, long short-term memory (lstm), and recurrent neural networks. it extracts court records having crucial sentiments or statements for maritime court verdicts. the suggested approach has been used successfully during sentiment analysis by emphasizing feature selection from a legal repository. the lstm + cnn model has shown promising results in obtaining sentiments and records from multiple devices and sufficiently proposing practical guidance to judicial personnel regarding the regulations applicable to various situations.","key_phrases":["historical information","the canadian maritime judiciary","time","the need","data","case references","later application","verdicts","similar cases","such data","multiple systems","its reachability","technologies","deep learning","sentiment analysis","chances","faster access","court records","such practice","impartial verdicts","workloads","court employees","the time","legal proceedings","claims","maritime contracts","shipping disputes","parties","this paper","a sentiment analysis framework","that","deep learning","learning","access","statutes","laws","cases","maritime judges","judgments","their claims","the suggested approach","deep learning models","convolutional neural networks","cnns","deep neural networks","long short-term memory","lstm","neural networks","it","court records","crucial sentiments","statements","maritime court verdicts","the suggested approach","sentiment analysis","feature selection","a legal repository","the lstm + cnn model","promising results","sentiments","records","multiple devices","practical guidance","judicial personnel","the regulations","various situations","canadian"]},{"title":"A structural reliability analysis method under non-parameterized P-box based on double-loop deep learning models","authors":["Hao Hu","Minya Deng","Weichuan Sun","Jinwen Li","Huichao Xie","Haibo Liu"],"abstract":"Structural reliability analysis, when accounting for non-parameterized probability box (P-box) uncertainty, typically entails multiple calls to performance functions and poses significant computational hurdles, largely attributable to its inherently nested double-loop structure. Therefore, this paper proposes a new reliability analysis method tailored for structures with uncertain parameters represented using non-parameterized P-boxes. This method leverages double-loop deep learning models to efficiently calculate both the upper and lower bounds of the failure probability. In the development phase of the double-loop deep learning model, an active learning function is devised that integrates the local prediction uncertainty of the deep learning model, based on the K-fold cross-validation principle with the proximity of training samples to candidate sample points. Different stopping criteria are formulated at distinct stages of the model construction process. Firstly, within the inner loop, a deep learning model is established to represent the original performance function in relation to the input parameters. Secondly, based on the inner-loop deep learning model for the performance function, an outer-loop deep learning model is established for the auxiliary response function corresponding to the P-box bound curves of the performance function response with respect to standard uniform distribution variables. Thirdly, utilizing the outer-loop deep learning approximate model, the Monte Carlo simulation technique is employed to compute the upper and lower bounds of the structural failure probability. Finally, the effectiveness of the proposed method is validated through the investigation of two numerical examples and a practical engineering problem. The influence of parameters in the active learning function, threshold values for stopping criteria, and the number of sample points on the computational results is deliberated.","doi":"10.1007\/s00158-024-03854-3","cleaned_title":"a structural reliability analysis method under non-parameterized p-box based on double-loop deep learning models","cleaned_abstract":"structural reliability analysis, when accounting for non-parameterized probability box (p-box) uncertainty, typically entails multiple calls to performance functions and poses significant computational hurdles, largely attributable to its inherently nested double-loop structure. therefore, this paper proposes a new reliability analysis method tailored for structures with uncertain parameters represented using non-parameterized p-boxes. this method leverages double-loop deep learning models to efficiently calculate both the upper and lower bounds of the failure probability. in the development phase of the double-loop deep learning model, an active learning function is devised that integrates the local prediction uncertainty of the deep learning model, based on the k-fold cross-validation principle with the proximity of training samples to candidate sample points. different stopping criteria are formulated at distinct stages of the model construction process. firstly, within the inner loop, a deep learning model is established to represent the original performance function in relation to the input parameters. secondly, based on the inner-loop deep learning model for the performance function, an outer-loop deep learning model is established for the auxiliary response function corresponding to the p-box bound curves of the performance function response with respect to standard uniform distribution variables. thirdly, utilizing the outer-loop deep learning approximate model, the monte carlo simulation technique is employed to compute the upper and lower bounds of the structural failure probability. finally, the effectiveness of the proposed method is validated through the investigation of two numerical examples and a practical engineering problem. the influence of parameters in the active learning function, threshold values for stopping criteria, and the number of sample points on the computational results is deliberated.","key_phrases":["structural reliability analysis","p-box","multiple calls","performance functions","significant computational hurdles","its inherently nested double-loop structure","this paper","a new reliability analysis method","structures","uncertain parameters","non-parameterized p-boxes","this method","double-loop deep learning models","both the upper and lower bounds","the failure probability","the development phase","the double-loop deep learning model","an active learning function","that","the local prediction uncertainty","the deep learning model","the k-fold cross-validation principle","the proximity","training samples","sample points","different stopping criteria","distinct stages","the model construction process","the inner loop","a deep learning model","the original performance function","relation","the input parameters","the inner-loop deep learning model","the performance function","an outer-loop deep learning model","the auxiliary response function","the p-box bound curves","the performance function response","respect","standard uniform distribution variables","approximate model","the monte carlo simulation technique","the upper and lower bounds","the structural failure probability","the effectiveness","the proposed method","the investigation","two numerical examples","a practical engineering problem","the influence","parameters","the active learning function","criteria","the number","sample points","the computational results","the k-fold","firstly","secondly","thirdly","two"]},{"title":"A comprehensive survey on deep-learning-based visual captioning","authors":["Bowen Xin","Ning Xu","Yingchen Zhai","Tingting Zhang","Zimu Lu","Jing Liu","Weizhi Nie","Xuanya Li","An-An Liu"],"abstract":"Generating a description for an image\/video is termed as the visual captioning task. It requires the model to capture the semantic information of visual content and translate them into syntactically and semantically human language. Connecting both research communities of computer vision (CV) and natural language processing (NLP), visual captioning presents the big challenge to bridge the gap between low-level visual features and high-level language information. Thanks to recent advances in deep learning, which are widely applied to the fields of visual and language modeling, the visual captioning methods depending on the deep neural networks has demonstrated state-of-the-art performances. In this paper, we aim to present a comprehensive survey of existing deep learning-based visual captioning methods. Relying on the adopted mechanism and technique to narrow the semantic gap, we divide visual captioning methods into various groups. Representative categories in each group are summarized, and their strengths and limitations are discussed. The quantitative evaluations of state-of-the-art approaches on popular benchmark datasets are also presented and analyzed. Furthermore, we provide the discussions on future research directions.","doi":"10.1007\/s00530-023-01175-x","cleaned_title":"a comprehensive survey on deep-learning-based visual captioning","cleaned_abstract":"generating a description for an image\/video is termed as the visual captioning task. it requires the model to capture the semantic information of visual content and translate them into syntactically and semantically human language. connecting both research communities of computer vision (cv) and natural language processing (nlp), visual captioning presents the big challenge to bridge the gap between low-level visual features and high-level language information. thanks to recent advances in deep learning, which are widely applied to the fields of visual and language modeling, the visual captioning methods depending on the deep neural networks has demonstrated state-of-the-art performances. in this paper, we aim to present a comprehensive survey of existing deep learning-based visual captioning methods. relying on the adopted mechanism and technique to narrow the semantic gap, we divide visual captioning methods into various groups. representative categories in each group are summarized, and their strengths and limitations are discussed. the quantitative evaluations of state-of-the-art approaches on popular benchmark datasets are also presented and analyzed. furthermore, we provide the discussions on future research directions.","key_phrases":["a description","an image\/video","the visual captioning task","it","the model","the semantic information","visual content","them","syntactically and semantically human language","both research communities","computer vision","cv","natural language processing","nlp","visual captioning","the big challenge","the gap","low-level visual features","high-level language information","recent advances","deep learning","which","the fields","visual and language modeling","the visual captioning methods","the deep neural networks","the-art","this paper","we","a comprehensive survey","existing deep learning-based visual captioning methods","the adopted mechanism","technique","the semantic gap","we","visual captioning methods","various groups","representative categories","each group","their strengths","limitations","the quantitative evaluations","the-art","popular benchmark datasets","we","the discussions","future research directions"]},{"title":"Thermoplastic waste segregation classification system using deep learning techniques","authors":["M. Monica Subashini","R. S. Vignesh"],"abstract":"This research proposes a deep learning-based system, named deep CNN architecture, for the automated classification of the plastic resin in plastic waste. The system aims to detect and recognize objects such as drinking water bottles, detergent bottles, squeezable bottles, and plastic plates, and segregate them into PET, PE-HD, PE-LD, and other resin categories. The process involves capturing input images through a camera and using deep learning or traditional algorithms to detect and recognize the objects by comparing them with a trained database containing labeled objects. Unrecognized objects are dynamically trained, labeled, and updated in the database. The proposed system is implemented using Python, a versatile open-source programming language. Python\u2019s functional and aspect-oriented programming paradigms are leveraged to develop the models. The performance of the proposed architecture is evaluated against existing works, demonstrating a classification accuracy of 92.66% according to experimental results.","doi":"10.1007\/s11042-023-16237-5","cleaned_title":"thermoplastic waste segregation classification system using deep learning techniques","cleaned_abstract":"this research proposes a deep learning-based system, named deep cnn architecture, for the automated classification of the plastic resin in plastic waste. the system aims to detect and recognize objects such as drinking water bottles, detergent bottles, squeezable bottles, and plastic plates, and segregate them into pet, pe-hd, pe-ld, and other resin categories. the process involves capturing input images through a camera and using deep learning or traditional algorithms to detect and recognize the objects by comparing them with a trained database containing labeled objects. unrecognized objects are dynamically trained, labeled, and updated in the database. the proposed system is implemented using python, a versatile open-source programming language. python\u2019s functional and aspect-oriented programming paradigms are leveraged to develop the models. the performance of the proposed architecture is evaluated against existing works, demonstrating a classification accuracy of 92.66% according to experimental results.","key_phrases":["this research","a deep learning-based system","deep cnn architecture","the automated classification","the plastic resin","plastic waste","the system","objects","drinking water bottles","detergent bottles","squeezable bottles","plastic plates","them","other resin categories","the process","input images","a camera","deep learning","traditional algorithms","the objects","them","a trained database","labeled objects","unrecognized objects","the database","the proposed system","python","python\u2019s functional and aspect-oriented programming paradigms","the models","the performance","the proposed architecture","existing works","a classification accuracy","92.66%","experimental results","cnn","92.66%"]},{"title":"Multi-step framework for glaucoma diagnosis in retinal fundus images using deep learning","authors":["Sanli Yi","Lingxiang Zhou"],"abstract":"Glaucoma is one of the most common causes of blindness in the world. Screening glaucoma from retinal fundus images based on deep learning is a common method at present. In the diagnosis of glaucoma based on deep learning, the blood vessels within the optic disc interfere with the diagnosis, and there is also some pathological information outside the optic disc in fundus images. Therefore, integrating the original fundus image with the vessel-removed optic disc image can improve diagnostic efficiency. In this paper, we propose a novel multi-step framework named MSGC-CNN that can better diagnose glaucoma. In the framework, (1) we combine glaucoma pathological knowledge with deep learning model, fuse the features of original fundus image and optic disc region in which the interference of blood vessel is specifically removed by U-Net, and make glaucoma diagnosis based on the fused features. (2) Aiming at the characteristics of glaucoma fundus images, such as small amount of data, high resolution, and rich feature information, we design a new feature extraction network RA-ResNet and combined it with transfer learning. In order to verify our method, we conduct binary classification experiments on three public datasets, Drishti-GS, RIM-ONE-R3, and ACRIMA, with accuracy of 92.01%, 93.75%, and 97.87%. The results demonstrate a significant improvement over earlier results.Graphical abstract","doi":"10.1007\/s11517-024-03172-2","cleaned_title":"multi-step framework for glaucoma diagnosis in retinal fundus images using deep learning","cleaned_abstract":"glaucoma is one of the most common causes of blindness in the world. screening glaucoma from retinal fundus images based on deep learning is a common method at present. in the diagnosis of glaucoma based on deep learning, the blood vessels within the optic disc interfere with the diagnosis, and there is also some pathological information outside the optic disc in fundus images. therefore, integrating the original fundus image with the vessel-removed optic disc image can improve diagnostic efficiency. in this paper, we propose a novel multi-step framework named msgc-cnn that can better diagnose glaucoma. in the framework, (1) we combine glaucoma pathological knowledge with deep learning model, fuse the features of original fundus image and optic disc region in which the interference of blood vessel is specifically removed by u-net, and make glaucoma diagnosis based on the fused features. (2) aiming at the characteristics of glaucoma fundus images, such as small amount of data, high resolution, and rich feature information, we design a new feature extraction network ra-resnet and combined it with transfer learning. in order to verify our method, we conduct binary classification experiments on three public datasets, drishti-gs, rim-one-r3, and acrima, with accuracy of 92.01%, 93.75%, and 97.87%. the results demonstrate a significant improvement over earlier results.graphical abstract","key_phrases":["glaucoma","the most common causes","blindness","the world","glaucoma","retinal fundus images","deep learning","a common method","present","the diagnosis","glaucoma","deep learning","the blood vessels","the optic disc","the diagnosis","some pathological information","the optic disc","fundus images","the original fundus image","the vessel-removed optic disc image","diagnostic efficiency","this paper","we","a novel multi-step framework","msgc-cnn","that","glaucoma","the framework","we","glaucoma pathological knowledge","deep learning model","the features","original fundus image","optic disc region","which","the interference","blood vessel","u","-","net","glaucoma diagnosis","the fused features","the characteristics","glaucoma fundus images","small amount","data","high resolution","rich feature information","we","a new feature extraction network","ra-resnet","it","transfer learning","order","our method","we","binary classification experiments","three public datasets","drishti-gs","rim-one-r3","acrima","accuracy","92.01%","93.75%","97.87%","the results","a significant improvement","glaucoma","msgc-cnn","1","2","three","one-r3","92.01%","93.75%","97.87%"]},{"title":"Deep learning-based power usage effectiveness optimization for IoT-enabled data center","authors":["Yu Sun","Yanyi Wang","Gaoxiang Jiang","Bo Cheng","Haibo Zhou"],"abstract":"The proliferation of data centers is driving increased energy consumption, leading to environmentally unacceptable carbon emissions. As the use of Internet-of-Things (IoT) techniques for extensive data collection in data centers continues to grow, deep learning-based solutions have emerged as attractive alternatives to suboptimal traditional methods. However, existing approaches suffer from unsatisfactory performance, unrealistic assumptions, and an inability to address practical data center optimization. In this paper, we focus on power usage effectiveness (PUE) optimization in IoT-enabled data centers using deep learning algorithms. We first develop a deep learning-based PUE optimization framework tailored to IoT-enabled data centers. We then formulate the general PUE optimization problem, simplifying and specifying it for the minimization of long-term energy consumption in chiller cooling systems. Additionally, we introduce a transformer-based prediction network designed for energy consumption forecasting. Subsequently, we transform this formulation into a Markov decision process (MDP) and present the branching double dueling deep Q-network. This approach effectively tackles the challenges posed by enormous action spaces within MDP by branching actions into sub-actions. Extensive experiments conducted on real-world datasets demonstrate the exceptional performance of our algorithms, excelling in prediction precision, optimization convergence, and optimality while effectively managing a substantial number of actions on the order of \\(10^{13}\\).","doi":"10.1007\/s12083-024-01663-5","cleaned_title":"deep learning-based power usage effectiveness optimization for iot-enabled data center","cleaned_abstract":"the proliferation of data centers is driving increased energy consumption, leading to environmentally unacceptable carbon emissions. as the use of internet-of-things (iot) techniques for extensive data collection in data centers continues to grow, deep learning-based solutions have emerged as attractive alternatives to suboptimal traditional methods. however, existing approaches suffer from unsatisfactory performance, unrealistic assumptions, and an inability to address practical data center optimization. in this paper, we focus on power usage effectiveness (pue) optimization in iot-enabled data centers using deep learning algorithms. we first develop a deep learning-based pue optimization framework tailored to iot-enabled data centers. we then formulate the general pue optimization problem, simplifying and specifying it for the minimization of long-term energy consumption in chiller cooling systems. additionally, we introduce a transformer-based prediction network designed for energy consumption forecasting. subsequently, we transform this formulation into a markov decision process (mdp) and present the branching double dueling deep q-network. this approach effectively tackles the challenges posed by enormous action spaces within mdp by branching actions into sub-actions. extensive experiments conducted on real-world datasets demonstrate the exceptional performance of our algorithms, excelling in prediction precision, optimization convergence, and optimality while effectively managing a substantial number of actions on the order of \\(10^{13}\\).","key_phrases":["the proliferation","data centers","increased energy consumption","environmentally unacceptable carbon emissions","the use","things","iot","extensive data collection","data centers","deep learning-based solutions","attractive alternatives","suboptimal traditional methods","existing approaches","unsatisfactory performance","unrealistic assumptions","an inability","practical data center optimization","this paper","we","power usage effectiveness","(pue) optimization","iot-enabled data centers","deep learning algorithms","we","a deep learning-based pue optimization framework","iot-enabled data centers","we","the general pue optimization problem","it","the minimization","long-term energy consumption","chiller cooling systems","we","a transformer-based prediction network","energy consumption forecasting","we","this formulation","a markov decision process","mdp","the branching","deep q-network","this approach","the challenges","enormous action spaces","mdp","actions","sub","-","actions","extensive experiments","real-world datasets","the exceptional performance","our algorithms","prediction precision","optimization convergence","optimality","a substantial number","actions","the order","\\(10^{13}\\","first"]},{"title":"Question classification task based on deep learning models with self-attention mechanism","authors":["Subhash Mondal","Manas Barman","Amitava Nag"],"abstract":"Question classification (QC) is a process that involves classifying questions based on their type to enable systems to provide accurate responses by matching the question type with relevant information. To understand and respond to natural language questions posed by humans, machine learning models or systems must comprehend the type of information requested, which can often be inferred from the structure and wording of the question. The high dimensionality and sparse nature of text data lead to challenges for text classification. These tasks can be improved using deep learning (DL) approaches to process complex patterns and features within input data. By training on large amounts of labeled data, deep learning algorithms can automatically extract relevant features and representations from text, resulting in more accurate and robust classification. This study utilizes a dataset comprising 5452 instances of questions and six output labels and uses two different word embedding techniques, like GloVe and Word2Vec, tested on the dataset using three deep learning models, LSTM, BiLSTM, and GRU, followed by a convolution layer. Additionally, a self-attention layer is included, which helps the model to focus on more relevant information when making predictions. Finally, an analytical discussion of the proposed models and their performance results provide insight into how GloVe and Word2Vec perform on the above-mentioned models. The GloVe embedding outperforms by achieving 97.68% accuracy and a moderate loss of 16.98 with the GRU model.","doi":"10.1007\/s11042-024-19239-z","cleaned_title":"question classification task based on deep learning models with self-attention mechanism","cleaned_abstract":"question classification (qc) is a process that involves classifying questions based on their type to enable systems to provide accurate responses by matching the question type with relevant information. to understand and respond to natural language questions posed by humans, machine learning models or systems must comprehend the type of information requested, which can often be inferred from the structure and wording of the question. the high dimensionality and sparse nature of text data lead to challenges for text classification. these tasks can be improved using deep learning (dl) approaches to process complex patterns and features within input data. by training on large amounts of labeled data, deep learning algorithms can automatically extract relevant features and representations from text, resulting in more accurate and robust classification. this study utilizes a dataset comprising 5452 instances of questions and six output labels and uses two different word embedding techniques, like glove and word2vec, tested on the dataset using three deep learning models, lstm, bilstm, and gru, followed by a convolution layer. additionally, a self-attention layer is included, which helps the model to focus on more relevant information when making predictions. finally, an analytical discussion of the proposed models and their performance results provide insight into how glove and word2vec perform on the above-mentioned models. the glove embedding outperforms by achieving 97.68% accuracy and a moderate loss of 16.98 with the gru model.","key_phrases":["question classification","qc","a process","that","questions","their type","systems","accurate responses","the question type","relevant information","natural language questions","humans","machine learning models","systems","the type","information","which","the structure","wording","the question","the high dimensionality","sparse nature","text data","challenges","text classification","these tasks","dl","complex patterns","features","input data","training","large amounts","labeled data","deep learning algorithms","relevant features","representations","text","more accurate and robust classification","this study","a dataset","5452 instances","questions","six output labels","two different word","techniques","glove","word2vec","the dataset","three deep learning models","lstm","bilstm","gru","a convolution layer","a self-attention layer","which","the model","more relevant information","predictions","an analytical discussion","the proposed models","their performance results","insight","how glove","the above-mentioned models","the glove","outperforms","97.68% accuracy","a moderate loss","the gru model","5452","six","two","three","97.68%","16.98"]},{"title":"A new content-aware image resizing based on R\u00e9nyi entropy and deep learning","authors":["Jila Ayubi","Mehdi Chehel Amirani","Morteza Valizadeh"],"abstract":"One of the most popular techniques for changing the purpose of an image or resizing a digital image with content awareness is the seam-carving method. The performance of image resizing algorithms based on seam machining shows that these algorithms are highly dependent on the extraction of importance map techniques and the detection of salient objects. So far, various algorithms have been proposed to extract the importance map. In this paper, a new method based on R\u00e9nyi entropy is proposed to extract the importance map. Also, a deep learning network has been used to detect salient objects. The simulator results showed that combining R\u00e9nyi\u2019s importance map with a deep network of salient object detection performed better than classical seam-carving and other extended seam-carving algorithms based on deep learning.","doi":"10.1007\/s00521-024-09517-0","cleaned_title":"a new content-aware image resizing based on r\u00e9nyi entropy and deep learning","cleaned_abstract":"one of the most popular techniques for changing the purpose of an image or resizing a digital image with content awareness is the seam-carving method. the performance of image resizing algorithms based on seam machining shows that these algorithms are highly dependent on the extraction of importance map techniques and the detection of salient objects. so far, various algorithms have been proposed to extract the importance map. in this paper, a new method based on r\u00e9nyi entropy is proposed to extract the importance map. also, a deep learning network has been used to detect salient objects. the simulator results showed that combining r\u00e9nyi\u2019s importance map with a deep network of salient object detection performed better than classical seam-carving and other extended seam-carving algorithms based on deep learning.","key_phrases":["the most popular techniques","the purpose","an image","a digital image","content awareness","the seam-carving method","the performance","image resizing algorithms","seam machining","these algorithms","the extraction","importance map techniques","the detection","salient objects","various algorithms","the importance map","this paper","a new method","r\u00e9nyi entropy","the importance map","a deep learning network","salient objects","the simulator results","r\u00e9nyi\u2019s importance map","a deep network","salient object detection","classical seam-carving","other extended seam-carving algorithms","deep learning","one"]},{"title":"A deep learning framework for students' academic performance analysis","authors":["Sumati Pathak","Hiral Raja","Sumit Srivastava","Neelam Sahu","Rohit Raja","Amit Kumar Dewangan"],"abstract":"Students Performance (SP) analysis is regarded as one of the most important steps in the educational system for supporting students' academic success and the institutions' overall outcomes. Nevertheless, it is tremendously challenging due to the numerous details that many students have. Data Mining (DM) is the most widely used approach for SP prediction that extracts imperative information from a bigger raw data set. Even though there are various DM-centered performance prediction approaches, they all have low accuracy and high training time and don't produce the desired output. This paper proposes a hybrid deep learning framework using Deer Hunting Optimization based Deep Learning Neural Networks (DH-DLNN). A self-structured questionnaire covers all aspects of using information and communication technology, including increased access, knowledge building, learning, performance, motivation, classroom management and interaction, collaborative learning, and satisfaction. Data Cleaning and data conversion preprocess the dataset. The prediction of the student's level is then performed by extracting imperative features from the preprocessed data, followed by feature ranking using entropy calculations. The obtained entropy values are inputted into the DH-DLNN, which predicts the students' academic performance. Finally, the accuracy of the proposed system is evaluated using K-fold cross-validation. The experiment results revealed that DH-DLNN outperforms the other classification approaches with an accuracy of 96.33%.","doi":"10.1007\/s40012-023-00388-9","cleaned_title":"a deep learning framework for students' academic performance analysis","cleaned_abstract":"students performance (sp) analysis is regarded as one of the most important steps in the educational system for supporting students' academic success and the institutions' overall outcomes. nevertheless, it is tremendously challenging due to the numerous details that many students have. data mining (dm) is the most widely used approach for sp prediction that extracts imperative information from a bigger raw data set. even though there are various dm-centered performance prediction approaches, they all have low accuracy and high training time and don't produce the desired output. this paper proposes a hybrid deep learning framework using deer hunting optimization based deep learning neural networks (dh-dlnn). a self-structured questionnaire covers all aspects of using information and communication technology, including increased access, knowledge building, learning, performance, motivation, classroom management and interaction, collaborative learning, and satisfaction. data cleaning and data conversion preprocess the dataset. the prediction of the student's level is then performed by extracting imperative features from the preprocessed data, followed by feature ranking using entropy calculations. the obtained entropy values are inputted into the dh-dlnn, which predicts the students' academic performance. finally, the accuracy of the proposed system is evaluated using k-fold cross-validation. the experiment results revealed that dh-dlnn outperforms the other classification approaches with an accuracy of 96.33%.","key_phrases":["students","performance (sp) analysis","the most important steps","the educational system","students' academic success","the institutions' overall outcomes","it","the numerous details","many students","data mining","(dm","the most widely used approach","sp prediction","that","imperative information","a bigger raw data set","various dm-centered performance prediction approaches","they","all","low accuracy","high training time","the desired output","this paper","a hybrid deep learning framework","deer hunting optimization","neural networks","dh-dlnn","a self-structured questionnaire","all aspects","information and communication technology","increased access","knowledge building","learning","performance","motivation","classroom management","interaction","collaborative learning","satisfaction","data cleaning","data conversion","the dataset","the prediction","the student's level","imperative features","the preprocessed data","feature","entropy calculations","the obtained entropy values","the dh-dlnn","which","the students' academic performance","the accuracy","the proposed system","k","fold cross","-","validation","the experiment results","dh-dlnn","the other classification approaches","an accuracy","96.33%","one","96.33%"]},{"title":"Deep multi-metric training: the need of multi-metric curve evaluation to avoid weak learning","authors":["Michail Mamalakis","Abhirup Banerjee","Surajit Ray","Craig Wilkie","Richard H. Clayton","Andrew J. Swift","George Panoutsos","Bart Vorselaars"],"abstract":"The development and application of artificial intelligence-based computer vision systems in medicine, environment, and industry are playing an increasingly prominent role. Hence, the need for optimal and efficient hyperparameter tuning strategies is more than crucial to deliver the highest performance of the deep learning networks in large and demanding datasets. In our study, we have developed and evaluated a new training methodology named deep multi-metric training (DMMT) for enhanced training performance. The DMMT delivers a state of robust learning for deep networks using a new important criterion of multi-metric performance evaluation. We have tested the DMMT methodology in multi-class (three, four, and ten), multi-vendors (different X-ray imaging devices), and multi-size (large, medium, and small) datasets. The validity of the DMMT methodology has been tested in three different classification problems: (i) medical disease classification, (ii) environmental classification, and (iii) ecological classification. For disease classification, we have used two large COVID-19 chest X-rays datasets, namely the BIMCV COVID-19+ and Sheffield hospital datasets. The environmental application is related to the classification of weather images in cloudy, rainy, shine or sunrise conditions. The ecological classification task involves a classification of three animal species (cat, dog, wild) and a classification of ten animals and transportation vehicles categories (CIFAR-10). We have used state-of-the-art networks of DenseNet-121, ResNet-50, VGG-16, VGG-19, and DenResCov-19 (DenRes-131) to verify that our novel methodology is applicable in a variety of different deep learning networks. To the best of our knowledge, this is the first work that proposes a training methodology to deliver robust learning, over a variety of deep learning networks and multi-field classification problems.","doi":"10.1007\/s00521-024-10182-6","cleaned_title":"deep multi-metric training: the need of multi-metric curve evaluation to avoid weak learning","cleaned_abstract":"the development and application of artificial intelligence-based computer vision systems in medicine, environment, and industry are playing an increasingly prominent role. hence, the need for optimal and efficient hyperparameter tuning strategies is more than crucial to deliver the highest performance of the deep learning networks in large and demanding datasets. in our study, we have developed and evaluated a new training methodology named deep multi-metric training (dmmt) for enhanced training performance. the dmmt delivers a state of robust learning for deep networks using a new important criterion of multi-metric performance evaluation. we have tested the dmmt methodology in multi-class (three, four, and ten), multi-vendors (different x-ray imaging devices), and multi-size (large, medium, and small) datasets. the validity of the dmmt methodology has been tested in three different classification problems: (i) medical disease classification, (ii) environmental classification, and (iii) ecological classification. for disease classification, we have used two large covid-19 chest x-rays datasets, namely the bimcv covid-19+ and sheffield hospital datasets. the environmental application is related to the classification of weather images in cloudy, rainy, shine or sunrise conditions. the ecological classification task involves a classification of three animal species (cat, dog, wild) and a classification of ten animals and transportation vehicles categories (cifar-10). we have used state-of-the-art networks of densenet-121, resnet-50, vgg-16, vgg-19, and denrescov-19 (denres-131) to verify that our novel methodology is applicable in a variety of different deep learning networks. to the best of our knowledge, this is the first work that proposes a training methodology to deliver robust learning, over a variety of deep learning networks and multi-field classification problems.","key_phrases":["the development","application","artificial intelligence-based computer vision systems","medicine","environment","industry","an increasingly prominent role","the need","optimal and efficient hyperparameter tuning strategies","the highest performance","the deep learning networks","large and demanding datasets","our study","we","a new training methodology","deep multi-metric training","dmmt","enhanced training performance","the dmmt","a state","robust learning","deep networks","a new important criterion","multi-metric performance evaluation","we","the dmmt methodology","-","-","vendors (different x-ray imaging devices","multi-size (large, medium, and small) datasets","the validity","the dmmt methodology","three different classification problems","(i) medical disease classification","(ii) environmental classification","(iii) ecological classification","disease classification","we","two large covid-19 chest x-rays datasets","namely the bimcv covid-19","hospital datasets","the environmental application","the classification","weather images","shine","sunrise conditions","the ecological classification task","a classification","three animal species","a classification","ten animals","transportation vehicles categories","cifar-10","we","the-art","resnet-50","vgg-16","vgg-19","denrescov-19","(denres-131","our novel methodology","a variety","different deep learning networks","our knowledge","this","the first work","that","a training methodology","robust learning","a variety","deep learning networks","multi-field classification problems","three","four","ten","three","two","covid-19","covid-19+","three","ten","cifar-10","densenet-121, resnet-50","vgg-16","vgg-19","denrescov-19","first"]},{"title":"Deep Reinforcement Learning Model for Stock Portfolio Management Based on Data Fusion","authors":["Haifeng Li","Mo Hai"],"abstract":"Deep reinforcement learning (DRL) can be used to extract deep features that can be incorporated into reinforcement learning systems to enable improved decision-making; DRL can therefore also be used for managing stock portfolios. Traditional methods cannot fully exploit the advantages of DRL because they are generally based on real-time stock quotes, which do not have sufficient features for making comprehensive decisions. In this study, in addition to stock quotes, we introduced stock financial indices as additional stock features. Moreover, we used Markowitz mean-variance theory for determining stock correlation. A three-agent deep reinforcement learning model called Collaborative Multi-agent reinforcement learning-based stock Portfolio management System (CMPS) was designed and trained based on fused data. In CMPS, each agent was implemented with a deep Q-network to obtain the features of time-series stock data, and a self-attention network was used to combine the output of each agent. We added a risk-free asset strategy to CMPS to prevent risks and referred to this model as CMPS-Risk Free (CMPS-RF). We conducted experiments under different market conditions using the stock data of China Shanghai Stock Exchange 50 and compared our model with the state-of-the-art models. The results showed that CMPS could obtain better profits than the compared benchmark models, and CMPS-RF was able to accurately recognize the market risk and achieved the best Sharpe and Calmar ratios. The study findings are expected to aid in the development of an efficient investment-trading strategy.","doi":"10.1007\/s11063-024-11582-4","cleaned_title":"deep reinforcement learning model for stock portfolio management based on data fusion","cleaned_abstract":"deep reinforcement learning (drl) can be used to extract deep features that can be incorporated into reinforcement learning systems to enable improved decision-making; drl can therefore also be used for managing stock portfolios. traditional methods cannot fully exploit the advantages of drl because they are generally based on real-time stock quotes, which do not have sufficient features for making comprehensive decisions. in this study, in addition to stock quotes, we introduced stock financial indices as additional stock features. moreover, we used markowitz mean-variance theory for determining stock correlation. a three-agent deep reinforcement learning model called collaborative multi-agent reinforcement learning-based stock portfolio management system (cmps) was designed and trained based on fused data. in cmps, each agent was implemented with a deep q-network to obtain the features of time-series stock data, and a self-attention network was used to combine the output of each agent. we added a risk-free asset strategy to cmps to prevent risks and referred to this model as cmps-risk free (cmps-rf). we conducted experiments under different market conditions using the stock data of china shanghai stock exchange 50 and compared our model with the state-of-the-art models. the results showed that cmps could obtain better profits than the compared benchmark models, and cmps-rf was able to accurately recognize the market risk and achieved the best sharpe and calmar ratios. the study findings are expected to aid in the development of an efficient investment-trading strategy.","key_phrases":["deep reinforcement learning","drl","deep features","that","reinforcement learning systems","improved decision-making","drl","stock portfolios","traditional methods","the advantages","drl","they","real-time stock quotes","which","sufficient features","comprehensive decisions","this study","addition","stock quotes","we","stock financial indices","additional stock features","we","markowitz mean-variance theory","stock correlation","a three-agent deep reinforcement learning model","collaborative multi-agent reinforcement learning-based stock portfolio management system","cmps","fused data","cmps","each agent","a deep q-network","the features","time-series stock data","a self-attention network","the output","each agent","we","a risk-free asset strategy","cmps","risks","this model","cmps-risk free (cmps","we","experiments","different market conditions","the stock data","china shanghai stock exchange","our model","the-art","the results","cmps","better profits","the compared benchmark models","cmps-rf","the market risk","the best sharpe","calmar ratios","the study findings","the development","an efficient investment-trading strategy","three","china shanghai stock exchange","50","sharpe"]},{"title":"Prediction of non-muscle invasive bladder cancer recurrence using deep learning of pathology image","authors":["Guang-Yue Wang","Jing-Fei Zhu","Qi-Chao Wang","Jia-Xin Qin","Xin-Lei Wang","Xing Liu","Xin-Yu Liu","Jun-Zhi Chen","Jie-Fei Zhu","Shi-Chao Zhuo","Di Wu","Na Li","Liu Chao","Fan-Lai Meng","Hao Lu","Zhen-Duo Shi","Zhi-Gang Jia","Cong-Hui Han"],"abstract":"We aimed to build a deep learning-based pathomics model to predict the early recurrence of non-muscle-infiltrating bladder cancer (NMIBC) in this work. A total of 147 patients from Xuzhou Central Hospital were enrolled as the training cohort, and 63 patients from Suqian Affiliated Hospital of Xuzhou Medical University were enrolled as the test cohort. Based on two consecutive phases of patch level prediction and WSI-level predictione, we built a pathomics model, with the initial model developed in the training cohort and subjected to transfer learning, and then the test cohort was validated for generalization. The features extracted from the visualization model were used for model interpretation. After migration learning, the area under the receiver operating characteristic curve for the deep learning-based pathomics model in the test cohort was 0.860 (95% CI 0.752\u20130.969), with good agreement between the migration training cohort and the test cohort in predicting recurrence, and the predicted values matched well with the observed values, with p values of 0.667766 and 0.140233 for the Hosmer\u2013Lemeshow test, respectively. The good clinical application was observed using a decision curve analysis method. We developed a deep learning-based pathomics model showed promising performance in predicting recurrence within one year in NMIBC patients. Including 10 state prediction NMIBC recurrence group pathology features be visualized, which may be used to facilitate personalized management of NMIBC patients to avoid ineffective or unnecessary treatment for the benefit of patients.","doi":"10.1038\/s41598-024-66870-9","cleaned_title":"prediction of non-muscle invasive bladder cancer recurrence using deep learning of pathology image","cleaned_abstract":"we aimed to build a deep learning-based pathomics model to predict the early recurrence of non-muscle-infiltrating bladder cancer (nmibc) in this work. a total of 147 patients from xuzhou central hospital were enrolled as the training cohort, and 63 patients from suqian affiliated hospital of xuzhou medical university were enrolled as the test cohort. based on two consecutive phases of patch level prediction and wsi-level predictione, we built a pathomics model, with the initial model developed in the training cohort and subjected to transfer learning, and then the test cohort was validated for generalization. the features extracted from the visualization model were used for model interpretation. after migration learning, the area under the receiver operating characteristic curve for the deep learning-based pathomics model in the test cohort was 0.860 (95% ci 0.752\u20130.969), with good agreement between the migration training cohort and the test cohort in predicting recurrence, and the predicted values matched well with the observed values, with p values of 0.667766 and 0.140233 for the hosmer\u2013lemeshow test, respectively. the good clinical application was observed using a decision curve analysis method. we developed a deep learning-based pathomics model showed promising performance in predicting recurrence within one year in nmibc patients. including 10 state prediction nmibc recurrence group pathology features be visualized, which may be used to facilitate personalized management of nmibc patients to avoid ineffective or unnecessary treatment for the benefit of patients.","key_phrases":["we","a deep learning-based pathomics model","the early recurrence","non-muscle-infiltrating bladder cancer","nmibc","this work","a total","147 patients","xuzhou central hospital","the training cohort","63 patients","suqian affiliated hospital","xuzhou medical university","the test cohort","two consecutive phases","patch level prediction","wsi-level predictione","we","a pathomics model","the initial model","the training cohort","learning","the test cohort","generalization","the features","the visualization model","model interpretation","migration learning","the area","the receiver operating characteristic curve","the deep learning-based pathomics model","the test cohort","(95%","good agreement","the migration training cohort","the test cohort","recurrence","the predicted values","the observed values","p values","the hosmer","lemeshow test","the good clinical application","a decision curve analysis method","we","a deep learning-based pathomics model","promising performance","recurrence","one year","nmibc patients","10 state prediction nmibc recurrence group pathology features","which","personalized management","nmibc patients","ineffective or unnecessary treatment","the benefit","patients","147","xuzhou","63","xuzhou medical university","two","0.860","95%","0.667766","0.140233","one year","10"]},{"title":"Building trust in deep learning-based immune response predictors with interpretable explanations","authors":["Piyush Borole","Ajitha Rajan"],"abstract":"The ability to predict whether a peptide will get presented on Major Histocompatibility Complex (MHC) class I molecules has profound implications in designing vaccines. Numerous deep learning-based predictors for peptide presentation on MHC class I molecules exist with high levels of accuracy. However, these MHC class I predictors are treated as black-box functions, providing little insight into their decision making. To build turst in these predictors, it is crucial to understand the rationale behind their decisions with human-interpretable explanations. We present MHCXAI, eXplainable AI (XAI) techniques to help interpret the outputs from MHC class I predictors in terms of input peptide features. In our experiments, we explain the outputs of four state-of-the-art MHC class I predictors over a large dataset of peptides and MHC alleles. Additionally, we evaluate the reliability of the explanations by comparing against ground truth and checking their robustness. MHCXAI seeks to increase understanding of deep learning-based predictors in the immune response domain and build trust with validated explanations.","doi":"10.1038\/s42003-024-05968-2","cleaned_title":"building trust in deep learning-based immune response predictors with interpretable explanations","cleaned_abstract":"the ability to predict whether a peptide will get presented on major histocompatibility complex (mhc) class i molecules has profound implications in designing vaccines. numerous deep learning-based predictors for peptide presentation on mhc class i molecules exist with high levels of accuracy. however, these mhc class i predictors are treated as black-box functions, providing little insight into their decision making. to build turst in these predictors, it is crucial to understand the rationale behind their decisions with human-interpretable explanations. we present mhcxai, explainable ai (xai) techniques to help interpret the outputs from mhc class i predictors in terms of input peptide features. in our experiments, we explain the outputs of four state-of-the-art mhc class i predictors over a large dataset of peptides and mhc alleles. additionally, we evaluate the reliability of the explanations by comparing against ground truth and checking their robustness. mhcxai seeks to increase understanding of deep learning-based predictors in the immune response domain and build trust with validated explanations.","key_phrases":["the ability","a peptide","major histocompatibility complex (mhc) class","i molecules","profound implications","vaccines","numerous deep learning-based predictors","peptide presentation","mhc class","i molecules","high levels","accuracy","these mhc class","i predictors","black-box functions","little insight","their decision making","turst","these predictors","it","the rationale","their decisions","human-interpretable explanations","we","mhcxai, explainable ai (xai) techniques","the outputs","mhc class","i","terms","input peptide features","our experiments","we","the outputs","the-art","i","a large dataset","peptides","mhc alleles","we","the reliability","the explanations","ground truth","their robustness","mhcxai","understanding","deep learning-based predictors","the immune response domain","trust","validated explanations","four"]},{"title":"Deep learning approach to detect cyberbullying on twitter","authors":["\u00c7inare O\u011fuz Aliyeva","Mete Ya\u011fano\u011flu"],"abstract":"In recent years, especially children and adolescents have shown increased interest in social media, making them a potential risk group for cyberbullying. Cyberbullying posts spread very quickly, often taking a long time to be deleted and sometimes remaining online indefinitely. Cyberbullying can have severe mental, psychological, and emotional effects on children and adolescents, and in extreme cases, it can lead to suicide. Turkey is among the top 10 countries with the highest number of children who are victims of cyberbullying. However, there are very few studies conducted in the Turkish language on this topic. This study aims to identify cyberbullying in Turkish Twitter posts. The Multi-Layer Detection (MLP) based model was evaluated using a dataset of 5000 tweets. The model was trained using both social media features and textual features extracted from the dataset. Textual features were obtained using various feature extraction methods such as Bag of Words (BOW), Term Frequency-Inverse Term Frequency (TF-IDF), Hashing Vectorizer, N-gram, and word embedding. These features were utilized in training the model, and their effectiveness was evaluated. The experiments revealed that the features obtained from TF-IDF and unigram methods significantly improved the model\u2019s performance. Subsequently, unnecessary features were eliminated using the Chi-Square feature selection method. The proposed model achieved a higher accuracy of 93.2% compared to machine learning (ML) methods used in previous studies on the same dataset. Additionally, the proposed model was compared with popular deep learning models in the literature, such as LSTM, BLSTM, and CNN, demonstrating promising results.","doi":"10.1007\/s11042-024-19869-3","cleaned_title":"deep learning approach to detect cyberbullying on twitter","cleaned_abstract":"in recent years, especially children and adolescents have shown increased interest in social media, making them a potential risk group for cyberbullying. cyberbullying posts spread very quickly, often taking a long time to be deleted and sometimes remaining online indefinitely. cyberbullying can have severe mental, psychological, and emotional effects on children and adolescents, and in extreme cases, it can lead to suicide. turkey is among the top 10 countries with the highest number of children who are victims of cyberbullying. however, there are very few studies conducted in the turkish language on this topic. this study aims to identify cyberbullying in turkish twitter posts. the multi-layer detection (mlp) based model was evaluated using a dataset of 5000 tweets. the model was trained using both social media features and textual features extracted from the dataset. textual features were obtained using various feature extraction methods such as bag of words (bow), term frequency-inverse term frequency (tf-idf), hashing vectorizer, n-gram, and word embedding. these features were utilized in training the model, and their effectiveness was evaluated. the experiments revealed that the features obtained from tf-idf and unigram methods significantly improved the model\u2019s performance. subsequently, unnecessary features were eliminated using the chi-square feature selection method. the proposed model achieved a higher accuracy of 93.2% compared to machine learning (ml) methods used in previous studies on the same dataset. additionally, the proposed model was compared with popular deep learning models in the literature, such as lstm, blstm, and cnn, demonstrating promising results.","key_phrases":["recent years","especially children","adolescents","increased interest","social media","them","cyberbullying posts","a long time","cyberbullying","severe mental, psychological, and emotional effects","children","adolescents","extreme cases","it","suicide","turkey","the top 10 countries","the highest number","children","who","victims","very few studies","the turkish language","this topic","this study","turkish twitter posts","the multi-layer detection","mlp) based model","a dataset","5000 tweets","the model","both social media features","textual features","the dataset","textual features","various feature extraction methods","bag","words","bow","tf-idf","vectorizer","word","these features","the model","their effectiveness","the experiments","the features","tf-idf and unigram methods","the model\u2019s performance","unnecessary features","the chi-square feature selection method","the proposed model","a higher accuracy","93.2%","machine learning (ml) methods","previous studies","the same dataset","the proposed model","popular deep learning models","the literature","lstm","blstm","cnn","promising results","recent years","turkey","10","5000","n-gram","93.2%","cnn"]},{"title":"A comparative study: prediction of parkinson\u2019s disease using machine learning, deep learning and nature inspired algorithm","authors":["Pankaj Kumar Keserwani","Suman Das","Nairita Sarkar"],"abstract":"Parkinson\u2019s Disease (PD) is a degenerative and progressive neurological disorder worsens over time. This disease initially affects people over 55 years old. Patients with PD often exhibit a variety of non-motor and motor symptoms and are diagnosed based on those motor and non-motor symptoms as well as numerous clinical indicators. Advancement in medical science has produced medicines for many diseases but till now no significant remedies are discovered for Parkinson disease. It is very necessary to detect PD at early phase to take precautions accordingly to reduce its harmful impact and improve the patient\u2019s life style to a considerable level. In this direction Artificial Intelligence (AI) based approaches have recently attracted many researchers to work accordingly as AI can handle vast amounts of data and generate accurate statistical predictions. Addressing this imperative, researchers have turned their focus toward Artificial Intelligence (AI) as a promising avenue. AI\u2019s capacity to manage vast datasets and generate precise statistical predictions makes it an invaluable tool for PD detection. This article aims to provide a comprehensive survey and in-depth analysis of various AI-based approaches. Leveraging machine learning (ML), deep learning (DL), and meta-heuristic algorithms, these approaches contribute to the prediction of PD. Additionally, the article delves into current research directions. As the pursuit of advancements continues, the integration of AI holds promise in revolutionizing early detection methods and subsequently improving the lives of individuals grappling with Parkinson\u2019s disease.","doi":"10.1007\/s11042-024-18186-z","cleaned_title":"a comparative study: prediction of parkinson\u2019s disease using machine learning, deep learning and nature inspired algorithm","cleaned_abstract":"parkinson\u2019s disease (pd) is a degenerative and progressive neurological disorder worsens over time. this disease initially affects people over 55 years old. patients with pd often exhibit a variety of non-motor and motor symptoms and are diagnosed based on those motor and non-motor symptoms as well as numerous clinical indicators. advancement in medical science has produced medicines for many diseases but till now no significant remedies are discovered for parkinson disease. it is very necessary to detect pd at early phase to take precautions accordingly to reduce its harmful impact and improve the patient\u2019s life style to a considerable level. in this direction artificial intelligence (ai) based approaches have recently attracted many researchers to work accordingly as ai can handle vast amounts of data and generate accurate statistical predictions. addressing this imperative, researchers have turned their focus toward artificial intelligence (ai) as a promising avenue. ai\u2019s capacity to manage vast datasets and generate precise statistical predictions makes it an invaluable tool for pd detection. this article aims to provide a comprehensive survey and in-depth analysis of various ai-based approaches. leveraging machine learning (ml), deep learning (dl), and meta-heuristic algorithms, these approaches contribute to the prediction of pd. additionally, the article delves into current research directions. as the pursuit of advancements continues, the integration of ai holds promise in revolutionizing early detection methods and subsequently improving the lives of individuals grappling with parkinson\u2019s disease.","key_phrases":["parkinson\u2019s disease","pd","a degenerative and progressive neurological disorder","time","this disease","people","patients","pd","a variety","non-motor and motor symptoms","those motor and non-motor symptoms","numerous clinical indicators","advancement","medical science","medicines","many diseases","no significant remedies","parkinson disease","it","pd","early phase","precautions","its harmful impact","the patient\u2019s life style","a considerable level","this direction","artificial intelligence","ai) based approaches","many researchers","ai","vast amounts","data","accurate statistical predictions","this imperative","researchers","their focus","artificial intelligence","(ai","a promising avenue","ai\u2019s capacity","vast datasets","precise statistical predictions","it","pd detection","this article","a comprehensive survey","-depth","various ai-based approaches","machine learning","ml","deep learning","dl","meta-heuristic algorithms","these approaches","the prediction","pd","the article","current research directions","the pursuit","advancements","the integration","ai","promise","early detection methods","the lives","individuals","parkinson\u2019s disease","55 years old"]},{"title":"A hybrid approach to detecting Parkinson's disease using spectrogram and deep learning CNN-LSTM network","authors":["V. Shibina","T. M. Thasleema"],"abstract":"Parkinson\u2019s disease (PD) is a common illness that affects brain neurons. Medical practitioners and caregivers face challenges in detecting Parkinson's disease promptly, either in its early or late stages. There is an urgent need for non-invasive PD diagnostic technologies because timely diagnosis substantially impacts patient outcomes. This research aims to provide an efficient way of\u00a0identifying Parkinson's disease by transforming voice inputs into spectrograms using Short Term Fourier Transform and applying\u00a0deep learning algorithms. The identification of Parkinson's disease can be done by leveraging the deep learning architectures such as Convolutional Neural Networks and Long Short-Term Memory networks. The experiment produced positive findings, with 95.67% accuracy, 97.62% precision, 94.67% recall, and an F1-score of 95.91%. The outcomes indicate that the suggested deep learning method is more successful in PD identification, surpassing the results of traditional classification methods.","doi":"10.1007\/s10772-024-10128-2","cleaned_title":"a hybrid approach to detecting parkinson's disease using spectrogram and deep learning cnn-lstm network","cleaned_abstract":"parkinson\u2019s disease (pd) is a common illness that affects brain neurons. medical practitioners and caregivers face challenges in detecting parkinson's disease promptly, either in its early or late stages. there is an urgent need for non-invasive pd diagnostic technologies because timely diagnosis substantially impacts patient outcomes. this research aims to provide an efficient way of identifying parkinson's disease by transforming voice inputs into spectrograms using short term fourier transform and applying deep learning algorithms. the identification of parkinson's disease can be done by leveraging the deep learning architectures such as convolutional neural networks and long short-term memory networks. the experiment produced positive findings, with 95.67% accuracy, 97.62% precision, 94.67% recall, and an f1-score of 95.91%. the outcomes indicate that the suggested deep learning method is more successful in pd identification, surpassing the results of traditional classification methods.","key_phrases":["parkinson\u2019s disease","pd","a common illness","that","brain neurons","medical practitioners","caregivers","challenges","parkinson's disease","its early or late stages","an urgent need","non-invasive pd diagnostic technologies","timely diagnosis","patient outcomes","this research","an efficient way","parkinson's disease","voice inputs","spectrograms","short term fourier transform","deep learning algorithms","the identification","parkinson's disease","the deep learning architectures","convolutional neural networks","long short-term memory networks","the experiment","positive findings","95.67% accuracy","97.62% precision","94.67% recall","an f1-score","95.91%","the outcomes","the suggested deep learning method","pd identification","the results","traditional classification methods","95.67%","97.62%","94.67%","95.91%"]},{"title":"Thermoplastic waste segregation classification system using deep learning techniques","authors":["M. Monica Subashini","R. S. Vignesh"],"abstract":"This research proposes a deep learning-based system, named deep CNN architecture, for the automated classification of the plastic resin in plastic waste. The system aims to detect and recognize objects such as drinking water bottles, detergent bottles, squeezable bottles, and plastic plates, and segregate them into PET, PE-HD, PE-LD, and other resin categories. The process involves capturing input images through a camera and using deep learning or traditional algorithms to detect and recognize the objects by comparing them with a trained database containing labeled objects. Unrecognized objects are dynamically trained, labeled, and updated in the database. The proposed system is implemented using Python, a versatile open-source programming language. Python\u2019s functional and aspect-oriented programming paradigms are leveraged to develop the models. The performance of the proposed architecture is evaluated against existing works, demonstrating a classification accuracy of 92.66% according to experimental results.","doi":"10.1007\/s11042-023-16237-5","cleaned_title":"thermoplastic waste segregation classification system using deep learning techniques","cleaned_abstract":"this research proposes a deep learning-based system, named deep cnn architecture, for the automated classification of the plastic resin in plastic waste. the system aims to detect and recognize objects such as drinking water bottles, detergent bottles, squeezable bottles, and plastic plates, and segregate them into pet, pe-hd, pe-ld, and other resin categories. the process involves capturing input images through a camera and using deep learning or traditional algorithms to detect and recognize the objects by comparing them with a trained database containing labeled objects. unrecognized objects are dynamically trained, labeled, and updated in the database. the proposed system is implemented using python, a versatile open-source programming language. python\u2019s functional and aspect-oriented programming paradigms are leveraged to develop the models. the performance of the proposed architecture is evaluated against existing works, demonstrating a classification accuracy of 92.66% according to experimental results.","key_phrases":["this research","a deep learning-based system","deep cnn architecture","the automated classification","the plastic resin","plastic waste","the system","objects","drinking water bottles","detergent bottles","squeezable bottles","plastic plates","them","other resin categories","the process","input images","a camera","deep learning","traditional algorithms","the objects","them","a trained database","labeled objects","unrecognized objects","the database","the proposed system","python","python\u2019s functional and aspect-oriented programming paradigms","the models","the performance","the proposed architecture","existing works","a classification accuracy","92.66%","experimental results","cnn","92.66%"]},{"title":"Deep learning model for detection of hotspots using infrared thermographic images of electrical installations","authors":["Ezechukwu Kalu Ukiwe","Steve A. Adeshina","Tsado Jacob","Bukola Babatunde Adetokun"],"abstract":"Hotspots in electrical power equipment or installations are a major issue whenever it occurs within the power system. Factors responsible for this phenomenon are many, sometimes inter-related and other times they are isolated. Electrical hotspots caused by poor connections are common. Deep learning models have become popular for diagnosing anomalies in physical and biological systems, by the instrumentality of feature extraction of images in convolutional neural networks. In this work, a VGG-16 deep neural network model is applied for identifying electrical hotspots by means of transfer learning. This model was achieved by first augmenting the acquired infrared thermographic images, using the pre-trained ImageNet weights of the VGG-16 algorithm with additional global average pooling in place of conventional fully connected layers and a softmax layer at the output. With the categorical cross-entropy loss function, the model was implemented using the Adam optimizer at learning rate of 0.0001 as well as some variants of the Adam optimization algorithm. On evaluation, with a test IRT image dataset, and a comparison with similar works, the research showed that a better accuracy of 99.98% in identification of electrical hotspots was achieved. The model shows good score in performance metrics like accuracy, precision, recall, and F1-score. The obtained results proved the potential of deep learning using computer vision parameters for infrared thermographic identification of electrical hotspots in power system installations. Also, there is need for careful selection of the IR sensor\u2019s thermal range during image acquisition, and suitable choice of color palette would make for easy hotspot isolation, reduce the pixel to pixel temperature differential across any of the images, and easily highlight the critical region of interest with high pixel values. However, it makes edge detection difficult for human visual perception which computer vision-based deep learning model could overcome.","doi":"10.1186\/s43067-024-00148-y","cleaned_title":"deep learning model for detection of hotspots using infrared thermographic images of electrical installations","cleaned_abstract":"hotspots in electrical power equipment or installations are a major issue whenever it occurs within the power system. factors responsible for this phenomenon are many, sometimes inter-related and other times they are isolated. electrical hotspots caused by poor connections are common. deep learning models have become popular for diagnosing anomalies in physical and biological systems, by the instrumentality of feature extraction of images in convolutional neural networks. in this work, a vgg-16 deep neural network model is applied for identifying electrical hotspots by means of transfer learning. this model was achieved by first augmenting the acquired infrared thermographic images, using the pre-trained imagenet weights of the vgg-16 algorithm with additional global average pooling in place of conventional fully connected layers and a softmax layer at the output. with the categorical cross-entropy loss function, the model was implemented using the adam optimizer at learning rate of 0.0001 as well as some variants of the adam optimization algorithm. on evaluation, with a test irt image dataset, and a comparison with similar works, the research showed that a better accuracy of 99.98% in identification of electrical hotspots was achieved. the model shows good score in performance metrics like accuracy, precision, recall, and f1-score. the obtained results proved the potential of deep learning using computer vision parameters for infrared thermographic identification of electrical hotspots in power system installations. also, there is need for careful selection of the ir sensor\u2019s thermal range during image acquisition, and suitable choice of color palette would make for easy hotspot isolation, reduce the pixel to pixel temperature differential across any of the images, and easily highlight the critical region of interest with high pixel values. however, it makes edge detection difficult for human visual perception which computer vision-based deep learning model could overcome.","key_phrases":["hotspots","electrical power equipment","installations","a major issue","it","the power system","factors","this phenomenon","many, sometimes inter-related and other times","they","electrical hotspots","poor connections","deep learning models","anomalies","physical and biological systems","the instrumentality","feature extraction","images","convolutional neural networks","this work","a vgg-16 deep neural network model","electrical hotspots","means","transfer learning","this model","the acquired infrared thermographic images","the pre-trained imagenet weights","the vgg-16 algorithm","additional global average pooling","place","conventional fully connected layers","a softmax layer","the output","the categorical cross-entropy loss function","the model","the adam optimizer","rate","some variants","the adam optimization algorithm","evaluation","a test irt image dataset","a comparison","similar works","the research","a better accuracy","99.98%","identification","electrical hotspots","the model","good score","performance metrics","accuracy","precision","recall","f1-score","the obtained results","the potential","deep learning","computer vision parameters","infrared thermographic identification","electrical hotspots","power system installations","need","careful selection","the ir sensor\u2019s thermal range","image acquisition","suitable choice","color palette","easy hotspot isolation","the pixel","temperature differential","any","the images","the critical region","interest","high pixel values","it","edge detection","human visual perception","which computer vision-based deep learning model","first","vgg-16","0.0001","99.98%"]},{"title":"Automatic retinoblastoma screening and surveillance using deep learning","authors":["Ruiheng Zhang","Li Dong","Ruyue Li","Kai Zhang","Yitong Li","Hongshu Zhao","Jitong Shi","Xin Ge","Xiaolin Xu","Libin Jiang","Xuhan Shi","Chuan Zhang","Wenda Zhou","Liangyuan Xu","Haotian Wu","Heyan Li","Chuyao Yu","Jing Li","Jianmin Ma","Wenbin Wei"],"abstract":"BackgroundRetinoblastoma is the most common intraocular malignancy in childhood. With the advanced management strategy, the globe salvage and overall survival have significantly improved, which proposes subsequent challenges regarding long-term surveillance and offspring screening. This study aimed to apply a deep learning algorithm to reduce the burden of follow-up and offspring screening.MethodsThis cohort study includes retinoblastoma patients who visited Beijing Tongren Hospital from\u00a0March 2018 to January 2022 for deep learning algorism development. Clinical-suspected and treated retinoblastoma patients from February 2022 to June 2022 were prospectively collected for prospective validation. Images from the posterior pole and peripheral retina were collected, and reference standards were made according to the consensus of the multidisciplinary management team. A deep learning algorithm was trained to identify \u201cnormal fundus\u201d, \u201cstable retinoblastoma\u201d in which specific treatment is not required, and \u201cactive retinoblastoma\u201d in which specific treatment is required. The performance of each classifier included sensitivity, specificity, accuracy, and cost-utility.ResultsA total of 36,623 images were included for developing the Deep Learning Assistant for Retinoblastoma Monitoring (DLA-RB) algorithm. In internal fivefold cross-validation, DLA-RB achieved an area under curve (AUC) of 0.998 (95% confidence interval [CI] 0.986\u20131.000) in distinguishing normal fundus and active retinoblastoma, and 0.940 (95%\u00a0CI 0.851\u20130.996) in distinguishing stable and active retinoblastoma. From February 2022 to June 2022, 139 eyes of 103 patients were prospectively collected. In identifying active retinoblastoma tumours from all clinical-suspected patients and active retinoblastoma from all treated retinoblastoma patients, the AUC of DLA-RB reached 0.991 (95% CI 0.970\u20131.000), and 0.962 (95% CI 0.915\u20131.000), respectively. The combination between ophthalmologists and DLA-RB significantly improved the accuracy of competent ophthalmologists and residents regarding both binary tasks. Cost-utility analysis revealed DLA-RB-based diagnosis mode is cost-effective in both retinoblastoma diagnosis and active retinoblastoma identification.ConclusionsDLA-RB achieved high accuracy and sensitivity in identifying active retinoblastoma from the normal and stable retinoblastoma fundus. It can be used to surveil the activity of retinoblastoma during follow-up and screen high-risk offspring. Compared with referral procedures to ophthalmologic centres, DLA-RB-based screening and surveillance is cost-effective and can be incorporated within telemedicine programs.Clinical Trial RegistrationThis study was registered on ClinicalTrials.gov (NCT05308043).","doi":"10.1038\/s41416-023-02320-z","cleaned_title":"automatic retinoblastoma screening and surveillance using deep learning","cleaned_abstract":"backgroundretinoblastoma is the most common intraocular malignancy in childhood. with the advanced management strategy, the globe salvage and overall survival have significantly improved, which proposes subsequent challenges regarding long-term surveillance and offspring screening. this study aimed to apply a deep learning algorithm to reduce the burden of follow-up and offspring screening.methodsthis cohort study includes retinoblastoma patients who visited beijing tongren hospital from march 2018 to january 2022 for deep learning algorism development. clinical-suspected and treated retinoblastoma patients from february 2022 to june 2022 were prospectively collected for prospective validation. images from the posterior pole and peripheral retina were collected, and reference standards were made according to the consensus of the multidisciplinary management team. a deep learning algorithm was trained to identify \u201cnormal fundus\u201d, \u201cstable retinoblastoma\u201d in which specific treatment is not required, and \u201cactive retinoblastoma\u201d in which specific treatment is required. the performance of each classifier included sensitivity, specificity, accuracy, and cost-utility.resultsa total of 36,623 images were included for developing the deep learning assistant for retinoblastoma monitoring (dla-rb) algorithm. in internal fivefold cross-validation, dla-rb achieved an area under curve (auc) of 0.998 (95% confidence interval [ci] 0.986\u20131.000) in distinguishing normal fundus and active retinoblastoma, and 0.940 (95% ci 0.851\u20130.996) in distinguishing stable and active retinoblastoma. from february 2022 to june 2022, 139 eyes of 103 patients were prospectively collected. in identifying active retinoblastoma tumours from all clinical-suspected patients and active retinoblastoma from all treated retinoblastoma patients, the auc of dla-rb reached 0.991 (95% ci 0.970\u20131.000), and 0.962 (95% ci 0.915\u20131.000), respectively. the combination between ophthalmologists and dla-rb significantly improved the accuracy of competent ophthalmologists and residents regarding both binary tasks. cost-utility analysis revealed dla-rb-based diagnosis mode is cost-effective in both retinoblastoma diagnosis and active retinoblastoma identification.conclusionsdla-rb achieved high accuracy and sensitivity in identifying active retinoblastoma from the normal and stable retinoblastoma fundus. it can be used to surveil the activity of retinoblastoma during follow-up and screen high-risk offspring. compared with referral procedures to ophthalmologic centres, dla-rb-based screening and surveillance is cost-effective and can be incorporated within telemedicine programs.clinical trial registrationthis study was registered on clinicaltrials.gov (nct05308043).","key_phrases":["backgroundretinoblastoma","the most common intraocular malignancy","childhood","the advanced management strategy","the globe salvage","overall survival","which","subsequent challenges","long-term surveillance","offspring screening","this study","a deep learning algorithm","the burden","follow-up","screening.methodsthis cohort study","retinoblastoma patients","who","beijing tongren hospital","march","january","deep learning algorism development","clinical-suspected and treated retinoblastoma patients","february","june","prospective validation","images","the posterior pole","peripheral retina","reference standards","the consensus","the multidisciplinary management team","a deep learning algorithm","normal fundus\u201d, \u201cstable retinoblastoma","which","specific treatment","\u201cactive retinoblastoma","which","specific treatment","the performance","each classifier","sensitivity","specificity","accuracy","cost-utility.resultsa total","36,623 images","the deep learning assistant","retinoblastoma monitoring","dla-rb","-","dla-rb","an area","curve","auc","(95% confidence interval","normal fundus","active retinoblastoma","0.940 (95%","ci 0.851\u20130.996","stable and active retinoblastoma","february","june","139 eyes","103 patients","active retinoblastoma tumours","all clinical-suspected patients","active retinoblastoma","all treated retinoblastoma patients","the auc","dla-rb","0.991 (95%","ci 0.970\u20131.000","ci 0.915\u20131.000","the combination","ophthalmologists","dla-rb","the accuracy","competent ophthalmologists","residents","both binary tasks","cost-utility analysis","dla-rb-based diagnosis mode","both retinoblastoma diagnosis","active retinoblastoma","high accuracy","sensitivity","active retinoblastoma","the normal and stable retinoblastoma fundus","it","the activity","retinoblastoma","follow-up and screen high-risk offspring","referral procedures","ophthalmologic centres","dla-rb-based screening","surveillance","telemedicine programs.clinical trial registrationthis study","nct05308043","march 2018 to","january 2022","february 2022 to june 2022","36,623","0.998","95%","0.940","95%","0.851\u20130.996","february 2022 to june 2022","139","103","0.991","95%","0.962","95%","0.915\u20131.000"]},{"title":"Curriculum learning and evolutionary optimization into deep learning for text classification","authors":["Alfredo Arturo El\u00edas-Miranda","Daniel Vallejo-Aldana","Fernando S\u00e1nchez-Vega","A. Pastor L\u00f3pez-Monroy","Alejandro Rosales-P\u00e9rez","Victor Mu\u00f1iz-Sanchez"],"abstract":"The exponential growth of social networks has given rise to a wide variety of content. Some social content violates the integrity and dignity of users, therefore, this task has become challenging. The need to deal with short texts, poorly written language, unbalanced classes, and non-thematic aspects. These can lead to overfitting in deep neural network (DNN) models used for classification tasks. Empirical evidence in previous studies indicates that some of these problems can be overcome by improving the optimization process of the DNN weights to avoid overfitting. Moreover, a well-defined learning process in the input examples could improve the order of the patterns learned throughout the optimization process. In this paper, we propose four Curriculum Learning strategies and a new Hybrid Genetic\u2013Gradient Algorithm that proved to improve the performance of DNN models detecting the class of interest even in highly imbalanced datasets.","doi":"10.1007\/s00521-023-08632-8","cleaned_title":"curriculum learning and evolutionary optimization into deep learning for text classification","cleaned_abstract":"the exponential growth of social networks has given rise to a wide variety of content. some social content violates the integrity and dignity of users, therefore, this task has become challenging. the need to deal with short texts, poorly written language, unbalanced classes, and non-thematic aspects. these can lead to overfitting in deep neural network (dnn) models used for classification tasks. empirical evidence in previous studies indicates that some of these problems can be overcome by improving the optimization process of the dnn weights to avoid overfitting. moreover, a well-defined learning process in the input examples could improve the order of the patterns learned throughout the optimization process. in this paper, we propose four curriculum learning strategies and a new hybrid genetic\u2013gradient algorithm that proved to improve the performance of dnn models detecting the class of interest even in highly imbalanced datasets.","key_phrases":["the exponential growth","social networks","rise","a wide variety","content","some social content","the integrity","dignity","users","this task","the need","short texts","poorly written language","unbalanced classes","non-thematic aspects","these","deep neural network (dnn) models","classification tasks","empirical evidence","previous studies","some","these problems","the optimization process","the dnn weights","a well-defined learning process","the input examples","the order","the patterns","the optimization process","this paper","we","four curriculum","strategies","a new hybrid genetic\u2013gradient algorithm","that","the performance","dnn models","the class","interest","highly imbalanced datasets","four"]},{"title":"Deep learning-based power usage effectiveness optimization for IoT-enabled data center","authors":["Yu Sun","Yanyi Wang","Gaoxiang Jiang","Bo Cheng","Haibo Zhou"],"abstract":"The proliferation of data centers is driving increased energy consumption, leading to environmentally unacceptable carbon emissions. As the use of Internet-of-Things (IoT) techniques for extensive data collection in data centers continues to grow, deep learning-based solutions have emerged as attractive alternatives to suboptimal traditional methods. However, existing approaches suffer from unsatisfactory performance, unrealistic assumptions, and an inability to address practical data center optimization. In this paper, we focus on power usage effectiveness (PUE) optimization in IoT-enabled data centers using deep learning algorithms. We first develop a deep learning-based PUE optimization framework tailored to IoT-enabled data centers. We then formulate the general PUE optimization problem, simplifying and specifying it for the minimization of long-term energy consumption in chiller cooling systems. Additionally, we introduce a transformer-based prediction network designed for energy consumption forecasting. Subsequently, we transform this formulation into a Markov decision process (MDP) and present the branching double dueling deep Q-network. This approach effectively tackles the challenges posed by enormous action spaces within MDP by branching actions into sub-actions. Extensive experiments conducted on real-world datasets demonstrate the exceptional performance of our algorithms, excelling in prediction precision, optimization convergence, and optimality while effectively managing a substantial number of actions on the order of \\(10^{13}\\).","doi":"10.1007\/s12083-024-01663-5","cleaned_title":"deep learning-based power usage effectiveness optimization for iot-enabled data center","cleaned_abstract":"the proliferation of data centers is driving increased energy consumption, leading to environmentally unacceptable carbon emissions. as the use of internet-of-things (iot) techniques for extensive data collection in data centers continues to grow, deep learning-based solutions have emerged as attractive alternatives to suboptimal traditional methods. however, existing approaches suffer from unsatisfactory performance, unrealistic assumptions, and an inability to address practical data center optimization. in this paper, we focus on power usage effectiveness (pue) optimization in iot-enabled data centers using deep learning algorithms. we first develop a deep learning-based pue optimization framework tailored to iot-enabled data centers. we then formulate the general pue optimization problem, simplifying and specifying it for the minimization of long-term energy consumption in chiller cooling systems. additionally, we introduce a transformer-based prediction network designed for energy consumption forecasting. subsequently, we transform this formulation into a markov decision process (mdp) and present the branching double dueling deep q-network. this approach effectively tackles the challenges posed by enormous action spaces within mdp by branching actions into sub-actions. extensive experiments conducted on real-world datasets demonstrate the exceptional performance of our algorithms, excelling in prediction precision, optimization convergence, and optimality while effectively managing a substantial number of actions on the order of \\(10^{13}\\).","key_phrases":["the proliferation","data centers","increased energy consumption","environmentally unacceptable carbon emissions","the use","things","iot","extensive data collection","data centers","deep learning-based solutions","attractive alternatives","suboptimal traditional methods","existing approaches","unsatisfactory performance","unrealistic assumptions","an inability","practical data center optimization","this paper","we","power usage effectiveness","(pue) optimization","iot-enabled data centers","deep learning algorithms","we","a deep learning-based pue optimization framework","iot-enabled data centers","we","the general pue optimization problem","it","the minimization","long-term energy consumption","chiller cooling systems","we","a transformer-based prediction network","energy consumption forecasting","we","this formulation","a markov decision process","mdp","the branching","deep q-network","this approach","the challenges","enormous action spaces","mdp","actions","sub","-","actions","extensive experiments","real-world datasets","the exceptional performance","our algorithms","prediction precision","optimization convergence","optimality","a substantial number","actions","the order","\\(10^{13}\\","first"]},{"title":"Question classification task based on deep learning models with self-attention mechanism","authors":["Subhash Mondal","Manas Barman","Amitava Nag"],"abstract":"Question classification (QC) is a process that involves classifying questions based on their type to enable systems to provide accurate responses by matching the question type with relevant information. To understand and respond to natural language questions posed by humans, machine learning models or systems must comprehend the type of information requested, which can often be inferred from the structure and wording of the question. The high dimensionality and sparse nature of text data lead to challenges for text classification. These tasks can be improved using deep learning (DL) approaches to process complex patterns and features within input data. By training on large amounts of labeled data, deep learning algorithms can automatically extract relevant features and representations from text, resulting in more accurate and robust classification. This study utilizes a dataset comprising 5452 instances of questions and six output labels and uses two different word embedding techniques, like GloVe and Word2Vec, tested on the dataset using three deep learning models, LSTM, BiLSTM, and GRU, followed by a convolution layer. Additionally, a self-attention layer is included, which helps the model to focus on more relevant information when making predictions. Finally, an analytical discussion of the proposed models and their performance results provide insight into how GloVe and Word2Vec perform on the above-mentioned models. The GloVe embedding outperforms by achieving 97.68% accuracy and a moderate loss of 16.98 with the GRU model.","doi":"10.1007\/s11042-024-19239-z","cleaned_title":"question classification task based on deep learning models with self-attention mechanism","cleaned_abstract":"question classification (qc) is a process that involves classifying questions based on their type to enable systems to provide accurate responses by matching the question type with relevant information. to understand and respond to natural language questions posed by humans, machine learning models or systems must comprehend the type of information requested, which can often be inferred from the structure and wording of the question. the high dimensionality and sparse nature of text data lead to challenges for text classification. these tasks can be improved using deep learning (dl) approaches to process complex patterns and features within input data. by training on large amounts of labeled data, deep learning algorithms can automatically extract relevant features and representations from text, resulting in more accurate and robust classification. this study utilizes a dataset comprising 5452 instances of questions and six output labels and uses two different word embedding techniques, like glove and word2vec, tested on the dataset using three deep learning models, lstm, bilstm, and gru, followed by a convolution layer. additionally, a self-attention layer is included, which helps the model to focus on more relevant information when making predictions. finally, an analytical discussion of the proposed models and their performance results provide insight into how glove and word2vec perform on the above-mentioned models. the glove embedding outperforms by achieving 97.68% accuracy and a moderate loss of 16.98 with the gru model.","key_phrases":["question classification","qc","a process","that","questions","their type","systems","accurate responses","the question type","relevant information","natural language questions","humans","machine learning models","systems","the type","information","which","the structure","wording","the question","the high dimensionality","sparse nature","text data","challenges","text classification","these tasks","dl","complex patterns","features","input data","training","large amounts","labeled data","deep learning algorithms","relevant features","representations","text","more accurate and robust classification","this study","a dataset","5452 instances","questions","six output labels","two different word","techniques","glove","word2vec","the dataset","three deep learning models","lstm","bilstm","gru","a convolution layer","a self-attention layer","which","the model","more relevant information","predictions","an analytical discussion","the proposed models","their performance results","insight","how glove","the above-mentioned models","the glove","outperforms","97.68% accuracy","a moderate loss","the gru model","5452","six","two","three","97.68%","16.98"]},{"title":"Deep multi-metric training: the need of multi-metric curve evaluation to avoid weak learning","authors":["Michail Mamalakis","Abhirup Banerjee","Surajit Ray","Craig Wilkie","Richard H. Clayton","Andrew J. Swift","George Panoutsos","Bart Vorselaars"],"abstract":"The development and application of artificial intelligence-based computer vision systems in medicine, environment, and industry are playing an increasingly prominent role. Hence, the need for optimal and efficient hyperparameter tuning strategies is more than crucial to deliver the highest performance of the deep learning networks in large and demanding datasets. In our study, we have developed and evaluated a new training methodology named deep multi-metric training (DMMT) for enhanced training performance. The DMMT delivers a state of robust learning for deep networks using a new important criterion of multi-metric performance evaluation. We have tested the DMMT methodology in multi-class (three, four, and ten), multi-vendors (different X-ray imaging devices), and multi-size (large, medium, and small) datasets. The validity of the DMMT methodology has been tested in three different classification problems: (i) medical disease classification, (ii) environmental classification, and (iii) ecological classification. For disease classification, we have used two large COVID-19 chest X-rays datasets, namely the BIMCV COVID-19+ and Sheffield hospital datasets. The environmental application is related to the classification of weather images in cloudy, rainy, shine or sunrise conditions. The ecological classification task involves a classification of three animal species (cat, dog, wild) and a classification of ten animals and transportation vehicles categories (CIFAR-10). We have used state-of-the-art networks of DenseNet-121, ResNet-50, VGG-16, VGG-19, and DenResCov-19 (DenRes-131) to verify that our novel methodology is applicable in a variety of different deep learning networks. To the best of our knowledge, this is the first work that proposes a training methodology to deliver robust learning, over a variety of deep learning networks and multi-field classification problems.","doi":"10.1007\/s00521-024-10182-6","cleaned_title":"deep multi-metric training: the need of multi-metric curve evaluation to avoid weak learning","cleaned_abstract":"the development and application of artificial intelligence-based computer vision systems in medicine, environment, and industry are playing an increasingly prominent role. hence, the need for optimal and efficient hyperparameter tuning strategies is more than crucial to deliver the highest performance of the deep learning networks in large and demanding datasets. in our study, we have developed and evaluated a new training methodology named deep multi-metric training (dmmt) for enhanced training performance. the dmmt delivers a state of robust learning for deep networks using a new important criterion of multi-metric performance evaluation. we have tested the dmmt methodology in multi-class (three, four, and ten), multi-vendors (different x-ray imaging devices), and multi-size (large, medium, and small) datasets. the validity of the dmmt methodology has been tested in three different classification problems: (i) medical disease classification, (ii) environmental classification, and (iii) ecological classification. for disease classification, we have used two large covid-19 chest x-rays datasets, namely the bimcv covid-19+ and sheffield hospital datasets. the environmental application is related to the classification of weather images in cloudy, rainy, shine or sunrise conditions. the ecological classification task involves a classification of three animal species (cat, dog, wild) and a classification of ten animals and transportation vehicles categories (cifar-10). we have used state-of-the-art networks of densenet-121, resnet-50, vgg-16, vgg-19, and denrescov-19 (denres-131) to verify that our novel methodology is applicable in a variety of different deep learning networks. to the best of our knowledge, this is the first work that proposes a training methodology to deliver robust learning, over a variety of deep learning networks and multi-field classification problems.","key_phrases":["the development","application","artificial intelligence-based computer vision systems","medicine","environment","industry","an increasingly prominent role","the need","optimal and efficient hyperparameter tuning strategies","the highest performance","the deep learning networks","large and demanding datasets","our study","we","a new training methodology","deep multi-metric training","dmmt","enhanced training performance","the dmmt","a state","robust learning","deep networks","a new important criterion","multi-metric performance evaluation","we","the dmmt methodology","-","-","vendors (different x-ray imaging devices","multi-size (large, medium, and small) datasets","the validity","the dmmt methodology","three different classification problems","(i) medical disease classification","(ii) environmental classification","(iii) ecological classification","disease classification","we","two large covid-19 chest x-rays datasets","namely the bimcv covid-19","hospital datasets","the environmental application","the classification","weather images","shine","sunrise conditions","the ecological classification task","a classification","three animal species","a classification","ten animals","transportation vehicles categories","cifar-10","we","the-art","resnet-50","vgg-16","vgg-19","denrescov-19","(denres-131","our novel methodology","a variety","different deep learning networks","our knowledge","this","the first work","that","a training methodology","robust learning","a variety","deep learning networks","multi-field classification problems","three","four","ten","three","two","covid-19","covid-19+","three","ten","cifar-10","densenet-121, resnet-50","vgg-16","vgg-19","denrescov-19","first"]},{"title":"Deep Reinforcement Learning Model for Stock Portfolio Management Based on Data Fusion","authors":["Haifeng Li","Mo Hai"],"abstract":"Deep reinforcement learning (DRL) can be used to extract deep features that can be incorporated into reinforcement learning systems to enable improved decision-making; DRL can therefore also be used for managing stock portfolios. Traditional methods cannot fully exploit the advantages of DRL because they are generally based on real-time stock quotes, which do not have sufficient features for making comprehensive decisions. In this study, in addition to stock quotes, we introduced stock financial indices as additional stock features. Moreover, we used Markowitz mean-variance theory for determining stock correlation. A three-agent deep reinforcement learning model called Collaborative Multi-agent reinforcement learning-based stock Portfolio management System (CMPS) was designed and trained based on fused data. In CMPS, each agent was implemented with a deep Q-network to obtain the features of time-series stock data, and a self-attention network was used to combine the output of each agent. We added a risk-free asset strategy to CMPS to prevent risks and referred to this model as CMPS-Risk Free (CMPS-RF). We conducted experiments under different market conditions using the stock data of China Shanghai Stock Exchange 50 and compared our model with the state-of-the-art models. The results showed that CMPS could obtain better profits than the compared benchmark models, and CMPS-RF was able to accurately recognize the market risk and achieved the best Sharpe and Calmar ratios. The study findings are expected to aid in the development of an efficient investment-trading strategy.","doi":"10.1007\/s11063-024-11582-4","cleaned_title":"deep reinforcement learning model for stock portfolio management based on data fusion","cleaned_abstract":"deep reinforcement learning (drl) can be used to extract deep features that can be incorporated into reinforcement learning systems to enable improved decision-making; drl can therefore also be used for managing stock portfolios. traditional methods cannot fully exploit the advantages of drl because they are generally based on real-time stock quotes, which do not have sufficient features for making comprehensive decisions. in this study, in addition to stock quotes, we introduced stock financial indices as additional stock features. moreover, we used markowitz mean-variance theory for determining stock correlation. a three-agent deep reinforcement learning model called collaborative multi-agent reinforcement learning-based stock portfolio management system (cmps) was designed and trained based on fused data. in cmps, each agent was implemented with a deep q-network to obtain the features of time-series stock data, and a self-attention network was used to combine the output of each agent. we added a risk-free asset strategy to cmps to prevent risks and referred to this model as cmps-risk free (cmps-rf). we conducted experiments under different market conditions using the stock data of china shanghai stock exchange 50 and compared our model with the state-of-the-art models. the results showed that cmps could obtain better profits than the compared benchmark models, and cmps-rf was able to accurately recognize the market risk and achieved the best sharpe and calmar ratios. the study findings are expected to aid in the development of an efficient investment-trading strategy.","key_phrases":["deep reinforcement learning","drl","deep features","that","reinforcement learning systems","improved decision-making","drl","stock portfolios","traditional methods","the advantages","drl","they","real-time stock quotes","which","sufficient features","comprehensive decisions","this study","addition","stock quotes","we","stock financial indices","additional stock features","we","markowitz mean-variance theory","stock correlation","a three-agent deep reinforcement learning model","collaborative multi-agent reinforcement learning-based stock portfolio management system","cmps","fused data","cmps","each agent","a deep q-network","the features","time-series stock data","a self-attention network","the output","each agent","we","a risk-free asset strategy","cmps","risks","this model","cmps-risk free (cmps","we","experiments","different market conditions","the stock data","china shanghai stock exchange","our model","the-art","the results","cmps","better profits","the compared benchmark models","cmps-rf","the market risk","the best sharpe","calmar ratios","the study findings","the development","an efficient investment-trading strategy","three","china shanghai stock exchange","50","sharpe"]},{"title":"Sub-trajectory clustering with deep reinforcement learning","authors":["Anqi Liang","Bin Yao","Bo Wang","Yinpei Liu","Zhida Chen","Jiong Xie","Feifei Li"],"abstract":"Sub-trajectory clustering is a fundamental problem in many trajectory applications. Existing approaches usually divide the clustering procedure into two phases: segmenting trajectories into sub-trajectories and then clustering these sub-trajectories. However, researchers need to develop complex human-crafted segmentation rules for specific applications, making the clustering results sensitive to the segmentation rules and lacking in generality. To solve this problem, we propose a novel algorithm using the clustering results to guide the segmentation, which is based on reinforcement learning (RL). The novelty is that the segmentation and clustering components cooperate closely and improve each other continuously to yield better clustering results. To devise our RL-based algorithm, we model the procedure of trajectory segmentation as a Markov decision process (MDP). We apply Deep-Q-Network (DQN) learning to train an RL model for the segmentation and achieve excellent clustering results. Experimental results on real datasets demonstrate the superior performance of the proposed RL-based approach over state-of-the-art methods.","doi":"10.1007\/s00778-023-00833-w","cleaned_title":"sub-trajectory clustering with deep reinforcement learning","cleaned_abstract":"sub-trajectory clustering is a fundamental problem in many trajectory applications. existing approaches usually divide the clustering procedure into two phases: segmenting trajectories into sub-trajectories and then clustering these sub-trajectories. however, researchers need to develop complex human-crafted segmentation rules for specific applications, making the clustering results sensitive to the segmentation rules and lacking in generality. to solve this problem, we propose a novel algorithm using the clustering results to guide the segmentation, which is based on reinforcement learning (rl). the novelty is that the segmentation and clustering components cooperate closely and improve each other continuously to yield better clustering results. to devise our rl-based algorithm, we model the procedure of trajectory segmentation as a markov decision process (mdp). we apply deep-q-network (dqn) learning to train an rl model for the segmentation and achieve excellent clustering results. experimental results on real datasets demonstrate the superior performance of the proposed rl-based approach over state-of-the-art methods.","key_phrases":["sub-trajectory clustering","a fundamental problem","many trajectory applications","existing approaches","the clustering procedure","two phases","trajectories","sub","-","trajectories","these sub","-","trajectories","researchers","complex human-crafted segmentation rules","specific applications","the clustering results","the segmentation rules","generality","this problem","we","a novel algorithm","the clustering results","the segmentation","which","reinforcement learning","rl","the novelty","the segmentation","clustering components","better clustering results","our rl-based algorithm","we","the procedure","trajectory segmentation","a markov decision process","mdp","we","q","dqn","an rl model","the segmentation","excellent clustering results","experimental results","real datasets","the superior performance","the proposed rl-based approach","the-art","two"]},{"title":"Prediction of non-muscle invasive bladder cancer recurrence using deep learning of pathology image","authors":["Guang-Yue Wang","Jing-Fei Zhu","Qi-Chao Wang","Jia-Xin Qin","Xin-Lei Wang","Xing Liu","Xin-Yu Liu","Jun-Zhi Chen","Jie-Fei Zhu","Shi-Chao Zhuo","Di Wu","Na Li","Liu Chao","Fan-Lai Meng","Hao Lu","Zhen-Duo Shi","Zhi-Gang Jia","Cong-Hui Han"],"abstract":"We aimed to build a deep learning-based pathomics model to predict the early recurrence of non-muscle-infiltrating bladder cancer (NMIBC) in this work. A total of 147 patients from Xuzhou Central Hospital were enrolled as the training cohort, and 63 patients from Suqian Affiliated Hospital of Xuzhou Medical University were enrolled as the test cohort. Based on two consecutive phases of patch level prediction and WSI-level predictione, we built a pathomics model, with the initial model developed in the training cohort and subjected to transfer learning, and then the test cohort was validated for generalization. The features extracted from the visualization model were used for model interpretation. After migration learning, the area under the receiver operating characteristic curve for the deep learning-based pathomics model in the test cohort was 0.860 (95% CI 0.752\u20130.969), with good agreement between the migration training cohort and the test cohort in predicting recurrence, and the predicted values matched well with the observed values, with p values of 0.667766 and 0.140233 for the Hosmer\u2013Lemeshow test, respectively. The good clinical application was observed using a decision curve analysis method. We developed a deep learning-based pathomics model showed promising performance in predicting recurrence within one year in NMIBC patients. Including 10 state prediction NMIBC recurrence group pathology features be visualized, which may be used to facilitate personalized management of NMIBC patients to avoid ineffective or unnecessary treatment for the benefit of patients.","doi":"10.1038\/s41598-024-66870-9","cleaned_title":"prediction of non-muscle invasive bladder cancer recurrence using deep learning of pathology image","cleaned_abstract":"we aimed to build a deep learning-based pathomics model to predict the early recurrence of non-muscle-infiltrating bladder cancer (nmibc) in this work. a total of 147 patients from xuzhou central hospital were enrolled as the training cohort, and 63 patients from suqian affiliated hospital of xuzhou medical university were enrolled as the test cohort. based on two consecutive phases of patch level prediction and wsi-level predictione, we built a pathomics model, with the initial model developed in the training cohort and subjected to transfer learning, and then the test cohort was validated for generalization. the features extracted from the visualization model were used for model interpretation. after migration learning, the area under the receiver operating characteristic curve for the deep learning-based pathomics model in the test cohort was 0.860 (95% ci 0.752\u20130.969), with good agreement between the migration training cohort and the test cohort in predicting recurrence, and the predicted values matched well with the observed values, with p values of 0.667766 and 0.140233 for the hosmer\u2013lemeshow test, respectively. the good clinical application was observed using a decision curve analysis method. we developed a deep learning-based pathomics model showed promising performance in predicting recurrence within one year in nmibc patients. including 10 state prediction nmibc recurrence group pathology features be visualized, which may be used to facilitate personalized management of nmibc patients to avoid ineffective or unnecessary treatment for the benefit of patients.","key_phrases":["we","a deep learning-based pathomics model","the early recurrence","non-muscle-infiltrating bladder cancer","nmibc","this work","a total","147 patients","xuzhou central hospital","the training cohort","63 patients","suqian affiliated hospital","xuzhou medical university","the test cohort","two consecutive phases","patch level prediction","wsi-level predictione","we","a pathomics model","the initial model","the training cohort","learning","the test cohort","generalization","the features","the visualization model","model interpretation","migration learning","the area","the receiver operating characteristic curve","the deep learning-based pathomics model","the test cohort","(95%","good agreement","the migration training cohort","the test cohort","recurrence","the predicted values","the observed values","p values","the hosmer","lemeshow test","the good clinical application","a decision curve analysis method","we","a deep learning-based pathomics model","promising performance","recurrence","one year","nmibc patients","10 state prediction nmibc recurrence group pathology features","which","personalized management","nmibc patients","ineffective or unnecessary treatment","the benefit","patients","147","xuzhou","63","xuzhou medical university","two","0.860","95%","0.667766","0.140233","one year","10"]},{"title":"Pre-operative lung ablation prediction using deep learning","authors":["Krishna Nand Keshavamurthy","Carsten Eickhoff","Etay Ziv"],"abstract":"ObjectiveMicrowave lung ablation (MWA) is a minimally invasive and inexpensive alternative cancer treatment for patients who are not candidates for surgery\/radiotherapy. However, a major challenge for MWA is its relatively high tumor recurrence rates, due to incomplete treatment as a result of inaccurate planning. We introduce a patient-specific, deep-learning model to accurately predict post-treatment ablation zones to\u00a0aid\u00a0planning and enable effective treatments.Materials and methodsOur IRB-approved retrospective study consisted of ablations with a single applicator\/burn\/vendor between 01\/2015 and 01\/2019. The input data included pre-procedure computerized tomography (CT), ablation power\/time, and applicator position. The ground truth ablation zone was segmented from follow-up CT post-treatment. Novel deformable image registration optimized for ablation scans and an applicator-centric co-ordinate system for data analysis were applied. Our prediction model was based on the U-net architecture. The registrations were evaluated using target registration error (TRE) and predictions using Bland-Altman plots, Dice co-efficient, precision, and recall, compared against the applicator vendor\u2019s estimates.ResultsThe data included 113 unique ablations from 72 patients (median age 57, interquartile range (IQR) (49\u201367); 41 women). We obtained a TRE\u2009\u2264\u20092\u2009mm on 52 ablations. Our prediction had no bias from ground truth ablation volumes (p\u2009=\u20090.169) unlike the vendor\u2019s estimate (p\u2009<\u20090.001) and had smaller limits of agreement (p\u2009<\u20090.001). An 11% improvement was achieved in the Dice score. The ability to account for patient-specific in-vivo anatomical effects due to vessels, chest wall, heart, lung boundaries, and fissures was shown.ConclusionsWe demonstrated a patient-specific deep-learning model to predict the ablation treatment effect prior to the procedure, with the potential for improved planning, achieving\u00a0complete treatments, and reduce tumor recurrence.Clinical relevance statementOur method addresses the current lack of reliable tools to estimate ablation extents, required for ensuring successful ablation treatments. The potential clinical implications include improved treatment planning, ensuring complete treatments, and reducing tumor recurrence.","doi":"10.1007\/s00330-024-10767-8","cleaned_title":"pre-operative lung ablation prediction using deep learning","cleaned_abstract":"objectivemicrowave lung ablation (mwa) is a minimally invasive and inexpensive alternative cancer treatment for patients who are not candidates for surgery\/radiotherapy. however, a major challenge for mwa is its relatively high tumor recurrence rates, due to incomplete treatment as a result of inaccurate planning. we introduce a patient-specific, deep-learning model to accurately predict post-treatment ablation zones to aid planning and enable effective treatments.materials and methodsour irb-approved retrospective study consisted of ablations with a single applicator\/burn\/vendor between 01\/2015 and 01\/2019. the input data included pre-procedure computerized tomography (ct), ablation power\/time, and applicator position. the ground truth ablation zone was segmented from follow-up ct post-treatment. novel deformable image registration optimized for ablation scans and an applicator-centric co-ordinate system for data analysis were applied. our prediction model was based on the u-net architecture. the registrations were evaluated using target registration error (tre) and predictions using bland-altman plots, dice co-efficient, precision, and recall, compared against the applicator vendor\u2019s estimates.resultsthe data included 113 unique ablations from 72 patients (median age 57, interquartile range (iqr) (49\u201367); 41 women). we obtained a tre \u2264 2 mm on 52 ablations. our prediction had no bias from ground truth ablation volumes (p = 0.169) unlike the vendor\u2019s estimate (p < 0.001) and had smaller limits of agreement (p < 0.001). an 11% improvement was achieved in the dice score. the ability to account for patient-specific in-vivo anatomical effects due to vessels, chest wall, heart, lung boundaries, and fissures was shown.conclusionswe demonstrated a patient-specific deep-learning model to predict the ablation treatment effect prior to the procedure, with the potential for improved planning, achieving complete treatments, and reduce tumor recurrence.clinical relevance statementour method addresses the current lack of reliable tools to estimate ablation extents, required for ensuring successful ablation treatments. the potential clinical implications include improved treatment planning, ensuring complete treatments, and reducing tumor recurrence.","key_phrases":["objectivemicrowave lung ablation","mwa","a minimally invasive and inexpensive alternative cancer treatment","patients","who","candidates","surgery\/radiotherapy","a major challenge","mwa","its relatively high tumor recurrence rates","incomplete treatment","a result","inaccurate planning","we","a patient-specific, deep-learning model","post-treatment ablation zones","planning","effective treatments.materials","methodsour irb-approved retrospective study","ablations","a single applicator\/burn\/vendor","01\/2015","01\/2019","the input data","pre-procedure computerized tomography","ablation power\/time","applicator position","the ground truth ablation zone","follow-up","-","treatment","novel deformable image registration","ablation scans","an applicator-centric co-ordinate system","data analysis","our prediction model","the u-net architecture","the registrations","target registration error","tre","predictions","bland-altman plots","dice co","precision","recall","the applicator vendor\u2019s estimates.resultsthe data","113 unique ablations","72 patients","median age","interquartile range","iqr","41 women","we","a tre \u2264","2 mm","52 ablations","our prediction","no bias","ground truth ablation volumes","the vendor\u2019s estimate","p","smaller limits","agreement","p","an 11% improvement","the dice score","the ability","patient-specific in-vivo anatomical effects","vessels","chest wall","heart","lung boundaries","fissures","shown.conclusionswe","a patient-specific deep-learning model","the ablation treatment effect","the procedure","the potential","improved planning","complete treatments","the current lack","reliable tools","ablation extents","successful ablation treatments","the potential clinical implications","improved treatment planning","complete treatments","tumor recurrence","mwa","mwa","methodsour irb-","113","72","age 57","49\u201367","41","52","0.169","p < 0.001","11%"]},{"title":"Building trust in deep learning-based immune response predictors with interpretable explanations","authors":["Piyush Borole","Ajitha Rajan"],"abstract":"The ability to predict whether a peptide will get presented on Major Histocompatibility Complex (MHC) class I molecules has profound implications in designing vaccines. Numerous deep learning-based predictors for peptide presentation on MHC class I molecules exist with high levels of accuracy. However, these MHC class I predictors are treated as black-box functions, providing little insight into their decision making. To build turst in these predictors, it is crucial to understand the rationale behind their decisions with human-interpretable explanations. We present MHCXAI, eXplainable AI (XAI) techniques to help interpret the outputs from MHC class I predictors in terms of input peptide features. In our experiments, we explain the outputs of four state-of-the-art MHC class I predictors over a large dataset of peptides and MHC alleles. Additionally, we evaluate the reliability of the explanations by comparing against ground truth and checking their robustness. MHCXAI seeks to increase understanding of deep learning-based predictors in the immune response domain and build trust with validated explanations.","doi":"10.1038\/s42003-024-05968-2","cleaned_title":"building trust in deep learning-based immune response predictors with interpretable explanations","cleaned_abstract":"the ability to predict whether a peptide will get presented on major histocompatibility complex (mhc) class i molecules has profound implications in designing vaccines. numerous deep learning-based predictors for peptide presentation on mhc class i molecules exist with high levels of accuracy. however, these mhc class i predictors are treated as black-box functions, providing little insight into their decision making. to build turst in these predictors, it is crucial to understand the rationale behind their decisions with human-interpretable explanations. we present mhcxai, explainable ai (xai) techniques to help interpret the outputs from mhc class i predictors in terms of input peptide features. in our experiments, we explain the outputs of four state-of-the-art mhc class i predictors over a large dataset of peptides and mhc alleles. additionally, we evaluate the reliability of the explanations by comparing against ground truth and checking their robustness. mhcxai seeks to increase understanding of deep learning-based predictors in the immune response domain and build trust with validated explanations.","key_phrases":["the ability","a peptide","major histocompatibility complex (mhc) class","i molecules","profound implications","vaccines","numerous deep learning-based predictors","peptide presentation","mhc class","i molecules","high levels","accuracy","these mhc class","i predictors","black-box functions","little insight","their decision making","turst","these predictors","it","the rationale","their decisions","human-interpretable explanations","we","mhcxai, explainable ai (xai) techniques","the outputs","mhc class","i","terms","input peptide features","our experiments","we","the outputs","the-art","i","a large dataset","peptides","mhc alleles","we","the reliability","the explanations","ground truth","their robustness","mhcxai","understanding","deep learning-based predictors","the immune response domain","trust","validated explanations","four"]},{"title":"Deep learning approach to detect cyberbullying on twitter","authors":["\u00c7inare O\u011fuz Aliyeva","Mete Ya\u011fano\u011flu"],"abstract":"In recent years, especially children and adolescents have shown increased interest in social media, making them a potential risk group for cyberbullying. Cyberbullying posts spread very quickly, often taking a long time to be deleted and sometimes remaining online indefinitely. Cyberbullying can have severe mental, psychological, and emotional effects on children and adolescents, and in extreme cases, it can lead to suicide. Turkey is among the top 10 countries with the highest number of children who are victims of cyberbullying. However, there are very few studies conducted in the Turkish language on this topic. This study aims to identify cyberbullying in Turkish Twitter posts. The Multi-Layer Detection (MLP) based model was evaluated using a dataset of 5000 tweets. The model was trained using both social media features and textual features extracted from the dataset. Textual features were obtained using various feature extraction methods such as Bag of Words (BOW), Term Frequency-Inverse Term Frequency (TF-IDF), Hashing Vectorizer, N-gram, and word embedding. These features were utilized in training the model, and their effectiveness was evaluated. The experiments revealed that the features obtained from TF-IDF and unigram methods significantly improved the model\u2019s performance. Subsequently, unnecessary features were eliminated using the Chi-Square feature selection method. The proposed model achieved a higher accuracy of 93.2% compared to machine learning (ML) methods used in previous studies on the same dataset. Additionally, the proposed model was compared with popular deep learning models in the literature, such as LSTM, BLSTM, and CNN, demonstrating promising results.","doi":"10.1007\/s11042-024-19869-3","cleaned_title":"deep learning approach to detect cyberbullying on twitter","cleaned_abstract":"in recent years, especially children and adolescents have shown increased interest in social media, making them a potential risk group for cyberbullying. cyberbullying posts spread very quickly, often taking a long time to be deleted and sometimes remaining online indefinitely. cyberbullying can have severe mental, psychological, and emotional effects on children and adolescents, and in extreme cases, it can lead to suicide. turkey is among the top 10 countries with the highest number of children who are victims of cyberbullying. however, there are very few studies conducted in the turkish language on this topic. this study aims to identify cyberbullying in turkish twitter posts. the multi-layer detection (mlp) based model was evaluated using a dataset of 5000 tweets. the model was trained using both social media features and textual features extracted from the dataset. textual features were obtained using various feature extraction methods such as bag of words (bow), term frequency-inverse term frequency (tf-idf), hashing vectorizer, n-gram, and word embedding. these features were utilized in training the model, and their effectiveness was evaluated. the experiments revealed that the features obtained from tf-idf and unigram methods significantly improved the model\u2019s performance. subsequently, unnecessary features were eliminated using the chi-square feature selection method. the proposed model achieved a higher accuracy of 93.2% compared to machine learning (ml) methods used in previous studies on the same dataset. additionally, the proposed model was compared with popular deep learning models in the literature, such as lstm, blstm, and cnn, demonstrating promising results.","key_phrases":["recent years","especially children","adolescents","increased interest","social media","them","cyberbullying posts","a long time","cyberbullying","severe mental, psychological, and emotional effects","children","adolescents","extreme cases","it","suicide","turkey","the top 10 countries","the highest number","children","who","victims","very few studies","the turkish language","this topic","this study","turkish twitter posts","the multi-layer detection","mlp) based model","a dataset","5000 tweets","the model","both social media features","textual features","the dataset","textual features","various feature extraction methods","bag","words","bow","tf-idf","vectorizer","word","these features","the model","their effectiveness","the experiments","the features","tf-idf and unigram methods","the model\u2019s performance","unnecessary features","the chi-square feature selection method","the proposed model","a higher accuracy","93.2%","machine learning (ml) methods","previous studies","the same dataset","the proposed model","popular deep learning models","the literature","lstm","blstm","cnn","promising results","recent years","turkey","10","5000","n-gram","93.2%","cnn"]},{"title":"Impact of log parsing on deep learning-based anomaly detection","authors":["Zanis Ali Khan","Donghwan Shin","Domenico Bianculli","Lionel C. Briand"],"abstract":"Software systems log massive amounts of data, recording important runtime information. Such logs are used, for example, for log-based anomaly detection, which aims to automatically detect abnormal behaviors of the system under analysis by processing the information recorded in its logs. Many log-based anomaly detection techniques based on deep learning models include a pre-processing step called log parsing. However, understanding the impact of log parsing on the accuracy of anomaly detection techniques has received surprisingly little attention so far. Investigating what are the key properties log parsing techniques should ideally have to help anomaly detection is therefore warranted. In this paper, we report on a comprehensive empirical study on the impact of log parsing on anomaly detection accuracy, using 13 log parsing techniques, seven anomly detection techniques (five based on deep learning and two based on traditional machine learning) on three publicly available log datasets. Our empirical results show that, despite what is widely assumed, there is no strong correlation between log parsing accuracy and anomaly detection accuracy, regardless of the metric used for measuring log parsing accuracy. Moreover, we experimentally confirm existing theoretical results showing that it is a property that we refer to as distinguishability in log parsing results\u2014as opposed to their accuracy\u2014that plays an essential role in achieving accurate anomaly detection.","doi":"10.1007\/s10664-024-10533-w","cleaned_title":"impact of log parsing on deep learning-based anomaly detection","cleaned_abstract":"software systems log massive amounts of data, recording important runtime information. such logs are used, for example, for log-based anomaly detection, which aims to automatically detect abnormal behaviors of the system under analysis by processing the information recorded in its logs. many log-based anomaly detection techniques based on deep learning models include a pre-processing step called log parsing. however, understanding the impact of log parsing on the accuracy of anomaly detection techniques has received surprisingly little attention so far. investigating what are the key properties log parsing techniques should ideally have to help anomaly detection is therefore warranted. in this paper, we report on a comprehensive empirical study on the impact of log parsing on anomaly detection accuracy, using 13 log parsing techniques, seven anomly detection techniques (five based on deep learning and two based on traditional machine learning) on three publicly available log datasets. our empirical results show that, despite what is widely assumed, there is no strong correlation between log parsing accuracy and anomaly detection accuracy, regardless of the metric used for measuring log parsing accuracy. moreover, we experimentally confirm existing theoretical results showing that it is a property that we refer to as distinguishability in log parsing results\u2014as opposed to their accuracy\u2014that plays an essential role in achieving accurate anomaly detection.","key_phrases":["software systems","massive amounts","data","important runtime information","such logs","example","log-based anomaly detection","which","abnormal behaviors","the system","analysis","the information","its logs","many log-based anomaly detection techniques","deep learning models","a pre-processing step","the impact","log","the accuracy","anomaly detection techniques","surprisingly little attention","what","the key properties","techniques","anomaly detection","this paper","we","a comprehensive empirical study","the impact","log","anomaly detection accuracy","13 log parsing techniques","seven anomly detection techniques","deep learning","traditional machine learning","three publicly available log datasets","our empirical results","what","no strong correlation","log","accuracy","anomaly detection accuracy","the metric","log","accuracy","we","existing theoretical results","it","a property","that","we","distinguishability","log parsing results","their accuracy","that","an essential role","accurate anomaly detection","software systems","anomaly detection","anomaly detection","anomaly","anomaly detection","anomaly detection accuracy","13","seven","five","two","three","anomaly detection"]},{"title":"Chest X-ray Images for Lung Disease Detection Using Deep Learning Techniques: A Comprehensive Survey","authors":["Mohammed A. A. Al-qaness","Jie Zhu","Dalal AL-Alimi","Abdelghani Dahou","Saeed Hamood Alsamhi","Mohamed Abd Elaziz","Ahmed A. Ewees"],"abstract":"In medical imaging, the last decade has witnessed a remarkable increase in the availability and diversity of chest X-ray (CXR) datasets. Concurrently, there has been a significant advancement in deep learning techniques, noted for their escalating accuracy. These developments have catalyzed a surge in the application of deep learning in various medical studies, particularly in detecting and classifying lung diseases. This study delves into an extensive compilation of over 200 studies from the recent five years (2018\u20132023), employing advanced machine learning, including deep learning methodologies to analyze CXR images. Our exploration is twofold: it categorizes these studies based on the methods used and the types of lung diseases addressed. It also presents an in-depth examination of the current limitations and prospective trajectories in this rapidly evolving field. Our findings underscore the transformative impact and continual progress of deep learning models in enhancing the accuracy and efficiency of lung disease detection using CXR images. This survey culminates by emphasizing the critical need for further technological advancement in this domain, aiming to bridge gaps in healthcare provision and improve patient outcomes. The overarching goal is to pave the way for more precise, efficient, and accessible diagnostic tools in the battle against lung diseases, reinforcing the indispensable role of technology in modern healthcare.","doi":"10.1007\/s11831-024-10081-y","cleaned_title":"chest x-ray images for lung disease detection using deep learning techniques: a comprehensive survey","cleaned_abstract":"in medical imaging, the last decade has witnessed a remarkable increase in the availability and diversity of chest x-ray (cxr) datasets. concurrently, there has been a significant advancement in deep learning techniques, noted for their escalating accuracy. these developments have catalyzed a surge in the application of deep learning in various medical studies, particularly in detecting and classifying lung diseases. this study delves into an extensive compilation of over 200 studies from the recent five years (2018\u20132023), employing advanced machine learning, including deep learning methodologies to analyze cxr images. our exploration is twofold: it categorizes these studies based on the methods used and the types of lung diseases addressed. it also presents an in-depth examination of the current limitations and prospective trajectories in this rapidly evolving field. our findings underscore the transformative impact and continual progress of deep learning models in enhancing the accuracy and efficiency of lung disease detection using cxr images. this survey culminates by emphasizing the critical need for further technological advancement in this domain, aiming to bridge gaps in healthcare provision and improve patient outcomes. the overarching goal is to pave the way for more precise, efficient, and accessible diagnostic tools in the battle against lung diseases, reinforcing the indispensable role of technology in modern healthcare.","key_phrases":["medical imaging","the last decade","a remarkable increase","the availability","diversity","chest x","-ray (cxr) datasets","a significant advancement","deep learning techniques","their escalating accuracy","these developments","a surge","the application","deep learning","various medical studies","lung diseases","this study","an extensive compilation","over 200 studies","the recent five years","advanced machine learning","deep learning methodologies","cxr images","our exploration","it","these studies","the methods","the types","lung diseases","it","-depth","the current limitations","prospective trajectories","this rapidly evolving field","our findings","the transformative impact","continual progress","deep learning models","the accuracy","efficiency","lung disease detection","cxr images","this survey","the critical need","further technological advancement","this domain","bridge gaps","healthcare provision","patient outcomes","the overarching goal","the way","the battle","lung diseases","the indispensable role","technology","modern healthcare","the last decade","over 200","the recent five years","2018\u20132023"]},{"title":"Comparative analysis of deep learning models for dysarthric speech detection","authors":["P. Shanmugapriya","V. Mohan"],"abstract":"Dysarthria is a speech communication disorder that is associated with neurological impairments. To detect this disorder from speech, we present an experimental comparison of deep models developed based on frequency domain features. A comparative analysis of deep models is performed in the detection of dysarthria using scalogram of dysarthric speech. Also, it can assist physicians, specialists, and doctors based on the results of its detection. Since dysarthric speech signals have segments of breathy and semi-whispery, experiments are performed only on the frequency-domain representation of speech signals. Time-domain speech signal is transformed into a 2-D scalogram image through wavelet transformation. Then, the scalogram images are applied to pre-trained convolutional neural networks. The layers of pre-trained networks are tuned for our scalogram images through transfer learning. The proposed method of applying the scalogram images as input to pre-trained CNNs is evaluated on the TORGO database and the classification performance of these networks is compared. In this work, AlexNet, GoogLeNet, ResNet 50 and two pre-trained sound CNNs, namely VGGish and YAMNET are considered deep models of pre-trained convolutional neural networks. The proposed method of using pre-trained and transfer learned CNN with scalogram image feature achieved better accuracy when compared to other machine learning models in the dysarthria detection system.","doi":"10.1007\/s00500-023-09302-6","cleaned_title":"comparative analysis of deep learning models for dysarthric speech detection","cleaned_abstract":"dysarthria is a speech communication disorder that is associated with neurological impairments. to detect this disorder from speech, we present an experimental comparison of deep models developed based on frequency domain features. a comparative analysis of deep models is performed in the detection of dysarthria using scalogram of dysarthric speech. also, it can assist physicians, specialists, and doctors based on the results of its detection. since dysarthric speech signals have segments of breathy and semi-whispery, experiments are performed only on the frequency-domain representation of speech signals. time-domain speech signal is transformed into a 2-d scalogram image through wavelet transformation. then, the scalogram images are applied to pre-trained convolutional neural networks. the layers of pre-trained networks are tuned for our scalogram images through transfer learning. the proposed method of applying the scalogram images as input to pre-trained cnns is evaluated on the torgo database and the classification performance of these networks is compared. in this work, alexnet, googlenet, resnet 50 and two pre-trained sound cnns, namely vggish and yamnet are considered deep models of pre-trained convolutional neural networks. the proposed method of using pre-trained and transfer learned cnn with scalogram image feature achieved better accuracy when compared to other machine learning models in the dysarthria detection system.","key_phrases":["dysarthria","a speech communication disorder","that","neurological impairments","this disorder","speech","we","an experimental comparison","deep models","frequency domain features","a comparative analysis","deep models","the detection","dysarthria","scalogram","dysarthric speech","it","physicians","specialists","doctors","the results","its detection","dysarthric speech signals","segments","-","whispery","experiments","the frequency-domain representation","speech signals","time-domain speech signal","a 2-d scalogram image","wavelet transformation","the scalogram images","pre-trained convolutional neural networks","the layers","pre-trained networks","our scalogram images","transfer learning","the proposed method","the scalogram images","input","pre-trained cnns","the torgo database","the classification performance","these networks","this work","yamnet","deep models","pre-trained convolutional neural networks","the proposed method","pre-trained and transfer","cnn","scalogram image feature","better accuracy","other machine learning models","the dysarthria detection system","dysarthria","2","50","two","cnn"]},{"title":"Deep Learning Radiomics Analysis of CT Imaging for Differentiating Between Crohn\u2019s Disease and Intestinal Tuberculosis","authors":["Ming Cheng","Hanyue Zhang","Wenpeng Huang","Fei Li","Jianbo Gao"],"abstract":"This study aimed to develop and evaluate a CT-based deep learning radiomics model for differentiating between Crohn\u2019s disease (CD) and intestinal tuberculosis (ITB). A total of 330 patients with pathologically confirmed as CD or ITB from the First Affiliated Hospital of Zhengzhou University were divided into the validation dataset one (CD: 167; ITB: 57) and validation dataset two (CD: 78; ITB: 28). Based on the validation dataset one, the synthetic minority oversampling technique (SMOTE) was adopted to create balanced dataset as training data for feature selection and model construction. The handcrafted and deep learning (DL) radiomics features were extracted from the arterial and venous phases images, respectively. The interobserver consistency analysis, Spearman\u2019s correlation, univariate analysis, and the least absolute shrinkage and selection operator (LASSO) regression were used to select features. Based on extracted multi-phase radiomics features, six logistic regression models were finally constructed. The diagnostic performances of different models were compared using ROC analysis and Delong test. The arterial-venous combined deep learning radiomics model for differentiating between CD and ITB showed a high prediction quality with AUCs of 0.885, 0.877, and 0.800 in SMOTE dataset, validation dataset one, and validation dataset two, respectively. Moreover, the deep learning radiomics model outperformed the handcrafted radiomics model in same phase images. In validation dataset one, the Delong test results indicated that there was a significant difference in the AUC of the arterial models (p\u2009=\u20090.037), while not in venous and arterial-venous combined models (p\u2009=\u20090.398 and p\u2009=\u20090.265) as comparing deep learning radiomics models and handcrafted radiomics models. In our study, the arterial-venous combined model based on deep learning radiomics analysis exhibited good performance in differentiating between CD and ITB.","doi":"10.1007\/s10278-024-01059-0","cleaned_title":"deep learning radiomics analysis of ct imaging for differentiating between crohn\u2019s disease and intestinal tuberculosis","cleaned_abstract":"this study aimed to develop and evaluate a ct-based deep learning radiomics model for differentiating between crohn\u2019s disease (cd) and intestinal tuberculosis (itb). a total of 330 patients with pathologically confirmed as cd or itb from the first affiliated hospital of zhengzhou university were divided into the validation dataset one (cd: 167; itb: 57) and validation dataset two (cd: 78; itb: 28). based on the validation dataset one, the synthetic minority oversampling technique (smote) was adopted to create balanced dataset as training data for feature selection and model construction. the handcrafted and deep learning (dl) radiomics features were extracted from the arterial and venous phases images, respectively. the interobserver consistency analysis, spearman\u2019s correlation, univariate analysis, and the least absolute shrinkage and selection operator (lasso) regression were used to select features. based on extracted multi-phase radiomics features, six logistic regression models were finally constructed. the diagnostic performances of different models were compared using roc analysis and delong test. the arterial-venous combined deep learning radiomics model for differentiating between cd and itb showed a high prediction quality with aucs of 0.885, 0.877, and 0.800 in smote dataset, validation dataset one, and validation dataset two, respectively. moreover, the deep learning radiomics model outperformed the handcrafted radiomics model in same phase images. in validation dataset one, the delong test results indicated that there was a significant difference in the auc of the arterial models (p = 0.037), while not in venous and arterial-venous combined models (p = 0.398 and p = 0.265) as comparing deep learning radiomics models and handcrafted radiomics models. in our study, the arterial-venous combined model based on deep learning radiomics analysis exhibited good performance in differentiating between cd and itb.","key_phrases":["this study","a ct-based deep learning radiomics model","crohn\u2019s disease","cd","intestinal tuberculosis","itb","a total","330 patients","cd","itb","the first affiliated hospital","zhengzhou university","the validation","itb","validation","(cd","itb","the validation","the synthetic minority oversampling technique","smote","balanced dataset","training data","feature selection","model construction","the handcrafted and deep learning","(dl) radiomics features","the arterial and venous phases images","the interobserver consistency analysis","spearman\u2019s correlation","analysis","the least absolute shrinkage and selection operator","lasso) regression","features","extracted multi-phase radiomics features","six logistic regression models","the diagnostic performances","different models","roc analysis","delong test","the arterial-venous combined deep learning radiomics model","cd","itb","a high prediction quality","aucs","smote dataset","validation","one","validation","the deep learning radiomics model","the handcrafted radiomics model","same phase images","validation","the delong test results","a significant difference","the auc","the arterial models","venous and arterial-venous combined models","p =","deep learning radiomics models","handcrafted radiomics models","our study","the arterial-venous combined model","deep learning radiomics analysis","good performance","cd","itb","itb","330","itb","first","167","itb","57","two","78","28","six","roc","itb","0.885","0.877","0.800","one","two","0.037","0.398","0.265","itb"]},{"title":"Guarding Against the Unknown: Deep Transfer Learning for Hardware Image-Based Malware Detection","authors":["Zhangying He","Houman Homayoun","Hossein Sayadi"],"abstract":"Malware is increasingly becoming a significant threat to computing systems, and detecting zero-day (unknown) malware is crucial to ensure the security of modern systems. These attacks exploit software security vulnerabilities that are not documented or known in the detection mechanism\u2019s database, making it particularly a pressing challenge to address. In recent times, there has been a shift in focus by security researchers toward the architecture of underlying processors. They have suggested implementing hardware-based malware detection (HMD) countermeasures to address the shortcomings of software-based detection methods. HMD techniques involve applying standard machine learning (ML) algorithms to low-level events of processors that are gathered from hardware performance counter (HPC) registers. While these techniques have shown promising results for detecting known malware, accurately recognizing zero-day malware remains an unsolved issue in the existing HPC-based detection methods. Our comprehensive analysis has revealed that standard ML classifiers are ineffective in identifying zero-day malware traces using HPC events. In response, we propose Deep-HMD, a multi-level intelligent and flexible approach based on deep neural network and transfer learning, for accurate zero-day malware detection using image-based hardware events.\u00a0Deep-HMD first converts HPC-based malware and benign data into images, and subsequently employs a lightweight deep transfer learning methodology to obtain a high malware detection performance for both known and unknown test scenarios. To conduct a thorough analysis, three deep learning-based and nine standard ML algorithms are implemented and evaluated for hardware-based malware detection. The experimental results indicate that our proposed image-based malware detection solution achieves superior performance compared to all other methods, with a 97% detection performance (measured by F-measure and area under the curve) for run-time zero-day malware detection utilizing soley the top four performance counter events. Specifically, our novel approach outperforms the binarized MLP by 16% and the best classical ML algorithm by 18% in F-measure, while maintaining a minimal false positive rate and without incurring any hardware redesign overhead.","doi":"10.1007\/s41635-024-00146-6","cleaned_title":"guarding against the unknown: deep transfer learning for hardware image-based malware detection","cleaned_abstract":"malware is increasingly becoming a significant threat to computing systems, and detecting zero-day (unknown) malware is crucial to ensure the security of modern systems. these attacks exploit software security vulnerabilities that are not documented or known in the detection mechanism\u2019s database, making it particularly a pressing challenge to address. in recent times, there has been a shift in focus by security researchers toward the architecture of underlying processors. they have suggested implementing hardware-based malware detection (hmd) countermeasures to address the shortcomings of software-based detection methods. hmd techniques involve applying standard machine learning (ml) algorithms to low-level events of processors that are gathered from hardware performance counter (hpc) registers. while these techniques have shown promising results for detecting known malware, accurately recognizing zero-day malware remains an unsolved issue in the existing hpc-based detection methods. our comprehensive analysis has revealed that standard ml classifiers are ineffective in identifying zero-day malware traces using hpc events. in response, we propose deep-hmd, a multi-level intelligent and flexible approach based on deep neural network and transfer learning, for accurate zero-day malware detection using image-based hardware events. deep-hmd first converts hpc-based malware and benign data into images, and subsequently employs a lightweight deep transfer learning methodology to obtain a high malware detection performance for both known and unknown test scenarios. to conduct a thorough analysis, three deep learning-based and nine standard ml algorithms are implemented and evaluated for hardware-based malware detection. the experimental results indicate that our proposed image-based malware detection solution achieves superior performance compared to all other methods, with a 97% detection performance (measured by f-measure and area under the curve) for run-time zero-day malware detection utilizing soley the top four performance counter events. specifically, our novel approach outperforms the binarized mlp by 16% and the best classical ml algorithm by 18% in f-measure, while maintaining a minimal false positive rate and without incurring any hardware redesign overhead.","key_phrases":["malware","a significant threat","computing systems","zero-day (unknown) malware","the security","modern systems","these attacks","software security vulnerabilities","that","the detection mechanism\u2019s database","it","recent times","a shift","focus","security researchers","the architecture","underlying processors","they","hardware-based malware detection","hmd","the shortcomings","software-based detection methods","hmd techniques","standard machine learning","ml","low-level events","processors","that","hardware performance counter (hpc) registers","these techniques","promising results","known malware","zero-day malware","an unsolved issue","the existing hpc-based detection methods","our comprehensive analysis","standard ml classifiers","hpc events","response","we","deep-hmd","a multi-level intelligent and flexible approach","deep neural network","transfer","learning","accurate zero-day malware detection","image-based hardware events","deep-hmd","hpc-based malware","benign data","images","a lightweight deep transfer","methodology","a high malware detection performance","both known and unknown test scenarios","a thorough analysis","three deep learning-based and nine standard ml algorithms","hardware-based malware detection","the experimental results","our proposed image-based malware detection solution","superior performance","all other methods","a 97% detection performance","f-measure","area","the curve","run-time zero-day malware detection","soley","the top four performance counter events","our novel approach","the binarized mlp","16%","the best classical ml algorithm","18%","f-measure","a minimal false positive rate","any hardware redesign","malware","zero-day","zero-day","zero-day","zero-day","first","three","nine","97%","zero-day","four","16%","18%"]},{"title":"Deep learning of causal structures in high dimensions under data limitations","authors":["Kai Lagemann","Christian Lagemann","Bernd Taschler","Sach Mukherjee"],"abstract":"Causal learning is a key challenge in scientific artificial intelligence as it allows researchers to go beyond purely correlative or predictive analyses towards learning underlying cause-and-effect relationships, which are important for scientific understanding as well as for a wide range of downstream tasks. Here, motivated by emerging biomedical questions, we propose a deep neural architecture for learning causal relationships between variables from a combination of high-dimensional data and prior causal knowledge. We combine convolutional and graph neural networks within a causal risk framework to provide an approach that is demonstrably effective under the conditions of high dimensionality, noise and data limitations that are characteristic of many applications, including in large-scale biology. In experiments, we find that the proposed learners can effectively identify novel causal relationships across thousands of variables. Results include extensive (linear and nonlinear) simulations (where the ground truth is known and can be directly compared against), as well as real biological examples where the models are applied to high-dimensional molecular data and their outputs compared against entirely unseen validation experiments. These results support the notion that deep learning approaches can be used to learn causal networks at large scale.","doi":"10.1038\/s42256-023-00744-z","cleaned_title":"deep learning of causal structures in high dimensions under data limitations","cleaned_abstract":"causal learning is a key challenge in scientific artificial intelligence as it allows researchers to go beyond purely correlative or predictive analyses towards learning underlying cause-and-effect relationships, which are important for scientific understanding as well as for a wide range of downstream tasks. here, motivated by emerging biomedical questions, we propose a deep neural architecture for learning causal relationships between variables from a combination of high-dimensional data and prior causal knowledge. we combine convolutional and graph neural networks within a causal risk framework to provide an approach that is demonstrably effective under the conditions of high dimensionality, noise and data limitations that are characteristic of many applications, including in large-scale biology. in experiments, we find that the proposed learners can effectively identify novel causal relationships across thousands of variables. results include extensive (linear and nonlinear) simulations (where the ground truth is known and can be directly compared against), as well as real biological examples where the models are applied to high-dimensional molecular data and their outputs compared against entirely unseen validation experiments. these results support the notion that deep learning approaches can be used to learn causal networks at large scale.","key_phrases":["causal learning","a key challenge","scientific artificial intelligence","it","researchers","purely correlative or predictive analyses","underlying cause-and-effect relationships","which","scientific understanding","a wide range","downstream tasks","biomedical questions","we","a deep neural architecture","causal relationships","variables","a combination","high-dimensional data","prior causal knowledge","we","convolutional and graph neural networks","a causal risk framework","an approach","that","the conditions","high dimensionality, noise and data limitations","that","many applications","large-scale biology","experiments","we","the proposed learners","novel causal relationships","thousands","variables","results","extensive (linear and nonlinear) simulations","the ground truth","real biological examples","the models","high-dimensional molecular data","their outputs","entirely unseen validation experiments","these results","the notion","deep learning approaches","causal networks","large scale","thousands","linear"]},{"title":"Deep learning-based biometric cryptographic key generation with post-quantum security","authors":["Oleksandr Kuznetsov","Dmytro Zakharov","Emanuele Frontoni"],"abstract":"In contemporary digital security systems, the generation and management of cryptographic keys, such as passwords and pin codes, often rely on stochastic random processes and intricate mathematical transformations. While these keys ensure robust security, their storage and distribution necessitate sophisticated and costly mechanisms. This study explores an alternative approach that leverages biometric data for generating cryptographic keys, thereby eliminating the need for complex storage and distribution processes. The paper investigates biometric key generation technologies based on deep learning models, specifically utilizing convolutional neural networks to extract biometric features from human facial images. Subsequently, code-based cryptographic extractors are employed to process the primary extracted features. The performance of various deep learning models and the extractor is evaluated by considering Type 1 and Type 2 errors. The optimized algorithm parameters yield an error rate of less than \\(10\\%\\), rendering the generated keys suitable for biometric authentication. Additionally, this study demonstrates that the application of code-based cryptographic extractors provides a post-quantum level of security, further enhancing the practicality and effectiveness of biometric key generation technologies in modern information security systems. This research contributes to the ongoing efforts towards secure, efficient, and user-friendly authentication and encryption methods, harnessing the power of biometric data and deep learning techniques.","doi":"10.1007\/s11042-023-17714-7","cleaned_title":"deep learning-based biometric cryptographic key generation with post-quantum security","cleaned_abstract":"in contemporary digital security systems, the generation and management of cryptographic keys, such as passwords and pin codes, often rely on stochastic random processes and intricate mathematical transformations. while these keys ensure robust security, their storage and distribution necessitate sophisticated and costly mechanisms. this study explores an alternative approach that leverages biometric data for generating cryptographic keys, thereby eliminating the need for complex storage and distribution processes. the paper investigates biometric key generation technologies based on deep learning models, specifically utilizing convolutional neural networks to extract biometric features from human facial images. subsequently, code-based cryptographic extractors are employed to process the primary extracted features. the performance of various deep learning models and the extractor is evaluated by considering type 1 and type 2 errors. the optimized algorithm parameters yield an error rate of less than \\(10\\%\\), rendering the generated keys suitable for biometric authentication. additionally, this study demonstrates that the application of code-based cryptographic extractors provides a post-quantum level of security, further enhancing the practicality and effectiveness of biometric key generation technologies in modern information security systems. this research contributes to the ongoing efforts towards secure, efficient, and user-friendly authentication and encryption methods, harnessing the power of biometric data and deep learning techniques.","key_phrases":["contemporary digital security systems","the generation","management","cryptographic keys","passwords","pin codes","stochastic random processes","intricate mathematical transformations","these keys","robust security","their storage and distribution necessitate sophisticated and costly mechanisms","this study","an alternative approach","that","biometric data","cryptographic keys","the need","complex storage and distribution processes","the paper investigates","key generation technologies","deep learning models","convolutional neural networks","biometric features","human facial images","code-based cryptographic extractors","the primary extracted features","the performance","various deep learning models","the extractor","type","2 errors","the optimized algorithm parameters","an error rate","the generated keys","biometric authentication","this study","the application","code-based cryptographic extractors","a post-quantum level","security","the practicality","effectiveness","biometric key generation technologies","modern information security systems","this research","the ongoing efforts","user-friendly authentication and encryption methods","the power","biometric data","deep learning techniques","1","2"]},{"title":"Connecting national flags \u2013 a deep learning approach","authors":["Theofanis Kalampokas","Dimitrios Mentizis","Eleni Vrochidou","George A. Papakostas"],"abstract":"National flags are the most recognizable symbols of the identity of a country. Similarities between flags may be observed due to cultural, historical, or ethical connections between nations, because they may be originated from the same group of people, or due to unrelated sharing of common symbols and colors. Although the fact that similar flags exist is indisputable, this has never been quantified. Quantifying flags\u2019 similarities could provide a useful body of knowledge for vexillologists and historians. To this end, this work aims to develop a supporting tool for the scientific study of nations\u2019 history and symbolisms, through the quantification of the varying degrees of similarity between their flags, by considering three initially stated hypotheses and by using a novel feature inclusion (FI) measure. The proposed FI measure aims to objectively quantify the overall similarity between flags based on optical multi-scaled features extracted from flag images. State-of-the-art deep learning models built for other applications tested their capability for the first time for the problem under study by using transfer learning, towards calculating the FI measure. More specifically, FI was quantified by six deep learning models: Yolo (V4 and V5), SSD, RetinaNet, Fast R-CNN, FCOS and CornerNet. Flags\u2019 images dataset included flags of 195 nations officially recognized by the United Nations. Experimental results reported maximum feature inclusion between flags of up to 99%. The extracted degrees of similarity were subsequently justified with the help of the Vexillology scientific domain, to support research findings and to raise questions for further investigation. Experimental results reveal that the proposed approach and FI measure are reliable and able to serve as a supporting tool to social sciences for knowledge extraction and quantification.","doi":"10.1007\/s11042-023-15056-y","cleaned_title":"connecting national flags \u2013 a deep learning approach","cleaned_abstract":"national flags are the most recognizable symbols of the identity of a country. similarities between flags may be observed due to cultural, historical, or ethical connections between nations, because they may be originated from the same group of people, or due to unrelated sharing of common symbols and colors. although the fact that similar flags exist is indisputable, this has never been quantified. quantifying flags\u2019 similarities could provide a useful body of knowledge for vexillologists and historians. to this end, this work aims to develop a supporting tool for the scientific study of nations\u2019 history and symbolisms, through the quantification of the varying degrees of similarity between their flags, by considering three initially stated hypotheses and by using a novel feature inclusion (fi) measure. the proposed fi measure aims to objectively quantify the overall similarity between flags based on optical multi-scaled features extracted from flag images. state-of-the-art deep learning models built for other applications tested their capability for the first time for the problem under study by using transfer learning, towards calculating the fi measure. more specifically, fi was quantified by six deep learning models: yolo (v4 and v5), ssd, retinanet, fast r-cnn, fcos and cornernet. flags\u2019 images dataset included flags of 195 nations officially recognized by the united nations. experimental results reported maximum feature inclusion between flags of up to 99%. the extracted degrees of similarity were subsequently justified with the help of the vexillology scientific domain, to support research findings and to raise questions for further investigation. experimental results reveal that the proposed approach and fi measure are reliable and able to serve as a supporting tool to social sciences for knowledge extraction and quantification.","key_phrases":["national flags","the most recognizable symbols","the identity","a country","similarities","flags","cultural, historical, or ethical connections","nations","they","the same group","people","unrelated sharing","common symbols","colors","the fact","similar flags","this","flags","a useful body","knowledge","vexillologists","historians","this end","this work","a supporting tool","the scientific study","nations\u2019 history","symbolisms","the quantification","the varying degrees","similarity","their flags","three initially stated hypotheses","a novel feature inclusion (fi) measure","the proposed fi measure","the overall similarity","flags","optical multi-scaled features","flag images","the-art","other applications","their capability","the first time","the problem","study","transfer learning","the fi measure","fi","six deep learning models","v4","v5","flags","flags","195 nations","the united nations","experimental results","maximum feature inclusion","flags","up to 99%","the extracted degrees","similarity","the help","the vexillology scientific domain","research findings","questions","further investigation","experimental results","the proposed approach","fi measure","a supporting tool","social sciences","knowledge extraction","quantification","three","first","six","fcos","195","the united nations","up to 99%"]},{"title":"Secure Communications with THz Reconfigurable Intelligent Surfaces and Deep Learning in 6G Systems","authors":["Ajmeera Kiran","Abhilash Sonker","Sachin Jadhav","Makarand Mohan Jadhav","Janjhyam Venkata Naga Ramesh","Elangovan Muniyandy"],"abstract":"In anticipation of the 6G era, this paper explores the integration of terahertz (THz) communications with Reconfigurable Intelligent Surfaces (RIS) and deep learning to establish a secure wireless network capable of ultra-high data rates. Addressing the non-convex challenge of maximizing secure energy efficiency, we introduce a novel deep learning framework that employs a variety of neural network architectures for optimizing RIS reflection and beamforming. Our simulations, set against scenarios with varying eavesdropper cooperation, confirm the efficacy of the proposed solution, achieving 97% of the optimal performance benchmarked against a genie-aided model. This research underlines a significant advancement in 6G network security, potentially influencing future standards and laying the groundwork for practical deployment, thereby marking a milestone in the convergence of THz technology, intelligent surfaces, and AI for future-proof secure communications.","doi":"10.1007\/s11277-024-11163-7","cleaned_title":"secure communications with thz reconfigurable intelligent surfaces and deep learning in 6g systems","cleaned_abstract":"in anticipation of the 6g era, this paper explores the integration of terahertz (thz) communications with reconfigurable intelligent surfaces (ris) and deep learning to establish a secure wireless network capable of ultra-high data rates. addressing the non-convex challenge of maximizing secure energy efficiency, we introduce a novel deep learning framework that employs a variety of neural network architectures for optimizing ris reflection and beamforming. our simulations, set against scenarios with varying eavesdropper cooperation, confirm the efficacy of the proposed solution, achieving 97% of the optimal performance benchmarked against a genie-aided model. this research underlines a significant advancement in 6g network security, potentially influencing future standards and laying the groundwork for practical deployment, thereby marking a milestone in the convergence of thz technology, intelligent surfaces, and ai for future-proof secure communications.","key_phrases":["anticipation","the 6g era","this paper","the integration","thz","reconfigurable intelligent surfaces","ris","deep learning","a secure wireless network","ultra-high data rates","the non-convex challenge","secure energy efficiency","we","a novel deep learning framework","that","a variety","neural network","ris reflection","beamforming","our simulations","scenarios","varying eavesdropper cooperation","the efficacy","the proposed solution","97%","the optimal performance","a genie-aided model","this research","a significant advancement","6g network security","future standards","the groundwork","practical deployment","a milestone","the convergence","thz technology","intelligent surfaces","future-proof secure communications","6","97%","genie","6"]},{"title":"A deep learning-based automated diagnosis system for SPECT myocardial perfusion imaging","authors":["Dai Kusumoto","Takumi Akiyama","Masahiro Hashimoto","Yu Iwabuchi","Toshiomi Katsuki","Mai Kimura","Yohei Akiba","Hiromune Sawada","Taku Inohara","Shinsuke Yuasa","Keiichi Fukuda","Masahiro Jinzaki","Masaki Ieda"],"abstract":"Images obtained from single-photon emission computed tomography for myocardial perfusion imaging (MPI SPECT) contain noises and artifacts, making cardiovascular disease diagnosis difficult. We developed a deep learning-based diagnosis support system using MPI SPECT images. Single-center datasets of MPI SPECT images (n\u2009=\u20095443) were obtained and labeled as healthy or coronary artery disease based on diagnosis reports. Three axes of four-dimensional datasets, resting, and stress conditions of three-dimensional reconstruction data, were reconstructed, and an AI model was trained to classify them. The trained convolutional neural network showed high performance [area under the curve (AUC) of the ROC curve: approximately 0.91; area under the recall precision curve: 0.87]. Additionally, using unsupervised learning and the Grad-CAM method, diseased lesions were successfully visualized. The AI-based automated diagnosis system had the highest performance (88%), followed by cardiologists with AI-guided diagnosis (80%) and cardiologists alone (65%). Furthermore, diagnosis time was shorter for AI-guided diagnosis (12\u00a0min) than for cardiologists alone (31\u00a0min). Our high-quality deep learning-based diagnosis support system may benefit cardiologists by improving diagnostic accuracy and reducing working hours.","doi":"10.1038\/s41598-024-64445-2","cleaned_title":"a deep learning-based automated diagnosis system for spect myocardial perfusion imaging","cleaned_abstract":"images obtained from single-photon emission computed tomography for myocardial perfusion imaging (mpi spect) contain noises and artifacts, making cardiovascular disease diagnosis difficult. we developed a deep learning-based diagnosis support system using mpi spect images. single-center datasets of mpi spect images (n = 5443) were obtained and labeled as healthy or coronary artery disease based on diagnosis reports. three axes of four-dimensional datasets, resting, and stress conditions of three-dimensional reconstruction data, were reconstructed, and an ai model was trained to classify them. the trained convolutional neural network showed high performance [area under the curve (auc) of the roc curve: approximately 0.91; area under the recall precision curve: 0.87]. additionally, using unsupervised learning and the grad-cam method, diseased lesions were successfully visualized. the ai-based automated diagnosis system had the highest performance (88%), followed by cardiologists with ai-guided diagnosis (80%) and cardiologists alone (65%). furthermore, diagnosis time was shorter for ai-guided diagnosis (12 min) than for cardiologists alone (31 min). our high-quality deep learning-based diagnosis support system may benefit cardiologists by improving diagnostic accuracy and reducing working hours.","key_phrases":["images","single-photon emission","tomography","myocardial perfusion imaging","(mpi spect","noises","artifacts","cardiovascular disease diagnosis","we","a deep learning-based diagnosis support system","mpi spect images","single-center datasets","mpi spect images","healthy or coronary artery disease","diagnosis reports","three axes","four-dimensional datasets","stress conditions","three-dimensional reconstruction data","an ai model","them","the trained convolutional neural network","high performance","[area","the curve","auc","the roc curve","area","the recall precision curve","unsupervised learning","the grad-cam method","diseased lesions","the ai-based automated diagnosis system","the highest performance","88%","cardiologists","ai-guided diagnosis","80%","cardiologists","65%","diagnosis time","ai-guided diagnosis","12 min","cardiologists","31 min","our high-quality deep learning-based diagnosis support system","cardiologists","diagnostic accuracy","working hours","5443","three","four","three","roc","approximately 0.91","0.87","88%","80%","65%","12","31","working hours"]},{"title":"Deep learning based water leakage detection for shield tunnel lining","authors":["Shichang Liu","Xu Xu","Gwanggil Jeon","Junxin Chen","Ben-Guo He"],"abstract":"Shield tunnel lining is prone to water leakage, which may further bring about corrosion and structural damage to the walls, potentially leading to dangerous accidents. To avoid tedious and inefficient manual inspection, many projects use artificial intelligence (AI) to detect cracks and water leakage. A novel method for water leakage inspection in shield tunnel lining that utilizes deep learning is introduced in this paper. Our proposal includes a ConvNeXt-S backbone, deconvolutional-feature pyramid network (D-FPN), spatial attention module (SPAM). and a detection head. It can extract representative features of leaking areas to aid inspection processes. To further improve the model\u2019s robustness, we innovatively use an inversed low-light enhancement method to convert normally illuminated images to low light ones and introduce them into the training samples. Validation experiments are performed, achieving the average precision (AP) score of 56.8%, which outperforms previous work by a margin of 5.7%. Visualization illustrations also support our method\u2019s practical effectiveness.","doi":"10.1007\/s11709-024-1071-5","cleaned_title":"deep learning based water leakage detection for shield tunnel lining","cleaned_abstract":"shield tunnel lining is prone to water leakage, which may further bring about corrosion and structural damage to the walls, potentially leading to dangerous accidents. to avoid tedious and inefficient manual inspection, many projects use artificial intelligence (ai) to detect cracks and water leakage. a novel method for water leakage inspection in shield tunnel lining that utilizes deep learning is introduced in this paper. our proposal includes a convnext-s backbone, deconvolutional-feature pyramid network (d-fpn), spatial attention module (spam). and a detection head. it can extract representative features of leaking areas to aid inspection processes. to further improve the model\u2019s robustness, we innovatively use an inversed low-light enhancement method to convert normally illuminated images to low light ones and introduce them into the training samples. validation experiments are performed, achieving the average precision (ap) score of 56.8%, which outperforms previous work by a margin of 5.7%. visualization illustrations also support our method\u2019s practical effectiveness.","key_phrases":["shield tunnel lining","water leakage","which","corrosion","structural damage","the walls","dangerous accidents","tedious and inefficient manual inspection","many projects","artificial intelligence","(ai","cracks","water leakage","a novel method","water leakage inspection","shield tunnel lining","that","deep learning","this paper","our proposal","a convnext-s backbone, deconvolutional-feature pyramid network","d","fpn","spatial attention module","spam","and a detection head","it","representative features","leaking areas","inspection processes","the model\u2019s robustness","we","an inversed low-light enhancement method","normally illuminated images","low light ones","them","the training samples","validation experiments","the average precision","ap",") score","56.8%","which","previous work","a margin","5.7%","visualization illustrations","our method\u2019s practical effectiveness","56.8%","5.7%"]},{"title":"HCCNet Fusion: a synergistic approach for accurate hepatocellular carcinoma staging using deep learning paradigm","authors":["Devi Rajeev","S. Remya","Anand Nayyar"],"abstract":"Hepatocellular carcinoma (HCC) stands as the second most prevalent cancer and a leading cause of cancer-related mortality globally, necessitating precise diagnostic and prognostic methodologies. The study introduces an innovative approach centered around the HCCNet Fusion model; a robust integration of advanced deep-learning techniques designed to elevate the accuracy of HCC stage recognition. Leveraging the synergies between the VGG16 architecture and U-Net and incorporating sophisticated data pre-processing methods such as Otsu\u2019s binary thresholding and marker-based watershed segmentation, this approach aims to strengthen the precision of HCC stage identification. Furthermore, transfer learning plays a pivotal role in HCCNet Fusion, enabling the models to integrate knowledge from diverse medical image settings through pre-trained weights from VGG16 and U-Net architectures. This strategic integration demonstrates the efficacy of advanced deep learning strategies in addressing intricate medical challenges, and outperforms conventional methods with a remarkable accuracy rate of 95%, underscoring the potential of cutting-edge deep learning techniques in medical diagnostics. The evaluation and validation of the proposed HCCNet Fusion model demonstrate its strong performance across many metrics like AUC ROC, Loss, Accuracy, Precision, Recall, and F1. Additionally, a comparative study was done against well-known methods, including CNN, Inception ResNetV2, VGG16, Inception V3, EfficientNet-B0, and ResNet50 and the results state that the proposed system not only advances HCC detection but also sets a pattern for leveraging state-of-the-art methodologies in addressing complex medical issues.","doi":"10.1007\/s11042-024-19446-8","cleaned_title":"hccnet fusion: a synergistic approach for accurate hepatocellular carcinoma staging using deep learning paradigm","cleaned_abstract":"hepatocellular carcinoma (hcc) stands as the second most prevalent cancer and a leading cause of cancer-related mortality globally, necessitating precise diagnostic and prognostic methodologies. the study introduces an innovative approach centered around the hccnet fusion model; a robust integration of advanced deep-learning techniques designed to elevate the accuracy of hcc stage recognition. leveraging the synergies between the vgg16 architecture and u-net and incorporating sophisticated data pre-processing methods such as otsu\u2019s binary thresholding and marker-based watershed segmentation, this approach aims to strengthen the precision of hcc stage identification. furthermore, transfer learning plays a pivotal role in hccnet fusion, enabling the models to integrate knowledge from diverse medical image settings through pre-trained weights from vgg16 and u-net architectures. this strategic integration demonstrates the efficacy of advanced deep learning strategies in addressing intricate medical challenges, and outperforms conventional methods with a remarkable accuracy rate of 95%, underscoring the potential of cutting-edge deep learning techniques in medical diagnostics. the evaluation and validation of the proposed hccnet fusion model demonstrate its strong performance across many metrics like auc roc, loss, accuracy, precision, recall, and f1. additionally, a comparative study was done against well-known methods, including cnn, inception resnetv2, vgg16, inception v3, efficientnet-b0, and resnet50 and the results state that the proposed system not only advances hcc detection but also sets a pattern for leveraging state-of-the-art methodologies in addressing complex medical issues.","key_phrases":["hepatocellular carcinoma","hcc","the second most prevalent cancer","a leading cause","cancer-related mortality","precise diagnostic and prognostic methodologies","the study","an innovative approach","the hccnet fusion model","a robust integration","advanced deep-learning techniques","the accuracy","hcc stage recognition","the synergies","the vgg16 architecture","u","-","net","sophisticated data pre-processing methods","otsu\u2019s binary thresholding and marker-based watershed segmentation","this approach","the precision","hcc stage identification","transfer learning","a pivotal role","hccnet fusion","the models","knowledge","diverse medical image settings","pre-trained weights","vgg16 and u-net architectures","this strategic integration","the efficacy","advanced deep learning strategies","intricate medical challenges","conventional methods","a remarkable accuracy rate","95%","the potential","cutting-edge deep learning techniques","medical diagnostics","the evaluation","validation","the proposed hccnet fusion model","its strong performance","many metrics","auc roc","loss","accuracy","precision","recall","f1","a comparative study","well-known methods","cnn","inception resnetv2","vgg16","inception v3","efficientnet-b0","resnet50","the results","the proposed system","hcc detection","a pattern","the-art","complex medical issues","second","95%","roc","cnn","resnetv2","v3","resnet50"]},{"title":"Visual sentiment analysis using data-augmented deep transfer learning techniques","authors":["Haoran Hong","Waneeza Zaheer","Aamir Wali"],"abstract":"The use of visual content to express emotions on social media platforms has become increasingly popular. Visual sentiment analysis can be used to understand the sentiment conveyed by the users using images. Compared to text, visual sentiment analysis is a challenging task since images are a more condensed form of data, have ambiguity and do not have explicit textual clues. Recently, a few studies used deep transfer learning techniques for visual sentiment analysis but the results reported can be significantly improved. In this research paper, we introduce a novel architecture that combines data augmentation and transfer learning. Our approach involves feature fusion of the pretrained VGG16 and MobileNetV1 models, followed by fine-tuning using an SVM classifier using augmented training data. For evaluation, we used two image datasets. To augment these datasets, we apply various techniques. The proposed model is also compared with three other transfer techniques, as well as four machine learning models (excluding SVM). VGG16+MobileNetV1-SVM has the best accuracy of 96% and recall of 99% for both datasets. Compared to other studies that also employed the same dataset, the proposed model produced the best results.","doi":"10.1007\/s00530-024-01308-w","cleaned_title":"visual sentiment analysis using data-augmented deep transfer learning techniques","cleaned_abstract":"the use of visual content to express emotions on social media platforms has become increasingly popular. visual sentiment analysis can be used to understand the sentiment conveyed by the users using images. compared to text, visual sentiment analysis is a challenging task since images are a more condensed form of data, have ambiguity and do not have explicit textual clues. recently, a few studies used deep transfer learning techniques for visual sentiment analysis but the results reported can be significantly improved. in this research paper, we introduce a novel architecture that combines data augmentation and transfer learning. our approach involves feature fusion of the pretrained vgg16 and mobilenetv1 models, followed by fine-tuning using an svm classifier using augmented training data. for evaluation, we used two image datasets. to augment these datasets, we apply various techniques. the proposed model is also compared with three other transfer techniques, as well as four machine learning models (excluding svm). vgg16+mobilenetv1-svm has the best accuracy of 96% and recall of 99% for both datasets. compared to other studies that also employed the same dataset, the proposed model produced the best results.","key_phrases":["the use","visual content","emotions","social media platforms","visual sentiment analysis","the sentiment","the users","images","text","visual sentiment analysis","a challenging task","images","a more condensed form","data","ambiguity","explicit textual clues","a few studies","techniques","visual sentiment analysis","the results","this research paper","we","a novel architecture","that","data augmentation","transfer learning","our approach","feature fusion","the pretrained vgg16 and mobilenetv1 models","fine-tuning","an svm classifier","augmented training data","evaluation","we","two image datasets","these datasets","we","various techniques","the proposed model","three other transfer techniques","four machine learning models","svm","vgg16+mobilenetv1-svm","the best accuracy","96%","recall","99%","both datasets","other studies","that","the same dataset","the proposed model","the best results","mobilenetv1","two","three","four","96%","99%"]},{"title":"Deep Learning Methods for Binding Site Prediction in Protein Structures","authors":["E. P. Geraseva"],"abstract":"AbstractThis work is an overview of deep machine learning methods aimed at predicting binding sites in protein structures. Several classes of methods are selected: prediction of binding sites for small molecules, proteins, and nucleic acids. For each class, various approaches to prediction are considered (prediction of binding atoms, residues, surfaces, pockets). Specifics of feature selection and neural network architectures inherent to each class and approach are highlighted, and an attempt is made to explain these specifics and foresee the further direction of their development.","doi":"10.1134\/S1990750823600498","cleaned_title":"deep learning methods for binding site prediction in protein structures","cleaned_abstract":"abstractthis work is an overview of deep machine learning methods aimed at predicting binding sites in protein structures. several classes of methods are selected: prediction of binding sites for small molecules, proteins, and nucleic acids. for each class, various approaches to prediction are considered (prediction of binding atoms, residues, surfaces, pockets). specifics of feature selection and neural network architectures inherent to each class and approach are highlighted, and an attempt is made to explain these specifics and foresee the further direction of their development.","key_phrases":["abstractthis work","an overview","deep machine learning methods","binding sites","protein structures","several classes","methods","prediction","binding sites","small molecules","proteins","nucleic acids","each class","various approaches","prediction","(prediction","binding atoms","residues","surfaces","pockets","specifics","feature selection","neural network","each class","approach","an attempt","these specifics","the further direction","their development"]},{"title":"Deep learning for tumor margin identification in electromagnetic imaging","authors":["Amir Mirbeik","Negar Ebadi"],"abstract":"In this work, a novel method for tumor margin identification in electromagnetic imaging is proposed to optimize the tumor removal surgery. This capability will enable the visualization of the border of the cancerous tissue for the surgeon prior or during the excision surgery. To this end, the border between the normal and tumor parts needs to be identified. Therefore, the images need to be segmented into tumor and normal areas. We propose a deep learning technique which divides the electromagnetic images into two regions: tumor and normal, with high accuracy. We formulate deep learning from a perspective relevant to electromagnetic image reconstruction. A recurrent auto-encoder network architecture (termed here DeepTMI) is presented. The effectiveness of the algorithm is demonstrated by segmenting the reconstructed images of an experimental tissue-mimicking phantom. The structure similarity measure (SSIM) and mean-square-error (MSE) average of normalized reconstructed results by the DeepTMI method are about 0.94 and 0.04 respectively, while that average obtained from the conventional backpropagation (BP) method can hardly overcome 0.35 and 0.41 respectively.","doi":"10.1038\/s41598-023-42625-w","cleaned_title":"deep learning for tumor margin identification in electromagnetic imaging","cleaned_abstract":"in this work, a novel method for tumor margin identification in electromagnetic imaging is proposed to optimize the tumor removal surgery. this capability will enable the visualization of the border of the cancerous tissue for the surgeon prior or during the excision surgery. to this end, the border between the normal and tumor parts needs to be identified. therefore, the images need to be segmented into tumor and normal areas. we propose a deep learning technique which divides the electromagnetic images into two regions: tumor and normal, with high accuracy. we formulate deep learning from a perspective relevant to electromagnetic image reconstruction. a recurrent auto-encoder network architecture (termed here deeptmi) is presented. the effectiveness of the algorithm is demonstrated by segmenting the reconstructed images of an experimental tissue-mimicking phantom. the structure similarity measure (ssim) and mean-square-error (mse) average of normalized reconstructed results by the deeptmi method are about 0.94 and 0.04 respectively, while that average obtained from the conventional backpropagation (bp) method can hardly overcome 0.35 and 0.41 respectively.","key_phrases":["this work","a novel method","tumor margin identification","electromagnetic imaging","the tumor removal surgery","this capability","the visualization","the border","the cancerous tissue","the surgeon","the excision surgery","this end","the border","the normal and tumor parts","the images","tumor and normal areas","we","a deep learning technique","which","the electromagnetic images","two regions","tumor","high accuracy","we","deep learning","a perspective","electromagnetic image reconstruction","a recurrent auto-encoder network architecture","deeptmi","the effectiveness","the algorithm","the reconstructed images","an experimental tissue-mimicking phantom","the structure similarity measure","ssim","mean-square-error (mse) average","normalized reconstructed results","the deeptmi method","that average","the conventional backpropagation (bp) method","two","about 0.94","0.04","0.35","0.41"]},{"title":"High-dimensional stochastic control models for newsvendor problems and deep learning resolution","authors":["Jingtang Ma","Shan Yang"],"abstract":"This paper studies continuous-time models for newsvendor problems with dynamic replenishment, financial hedging and Stackelberg competition. These factors are considered simultaneously and the high-dimensional stochastic control models are established. High-dimensional Hamilton-Jacobi-Bellman (HJB) equations are derived for the value functions. To circumvent the curse of dimensionality, a deep learning algorithm is proposed to solve the HJB equations. A projection is introduced in the algorithm to avoid the gradient explosion during the training phase. The deep learning algorithm is implemented for HJB equations derived from the newsvendor models with dimensions up to six. Numerical outcomes validate the algorithm\u2019s accuracy and demonstrate that the high-dimensional stochastic control models can successfully mitigate the risk.","doi":"10.1007\/s10479-024-05872-2","cleaned_title":"high-dimensional stochastic control models for newsvendor problems and deep learning resolution","cleaned_abstract":"this paper studies continuous-time models for newsvendor problems with dynamic replenishment, financial hedging and stackelberg competition. these factors are considered simultaneously and the high-dimensional stochastic control models are established. high-dimensional hamilton-jacobi-bellman (hjb) equations are derived for the value functions. to circumvent the curse of dimensionality, a deep learning algorithm is proposed to solve the hjb equations. a projection is introduced in the algorithm to avoid the gradient explosion during the training phase. the deep learning algorithm is implemented for hjb equations derived from the newsvendor models with dimensions up to six. numerical outcomes validate the algorithm\u2019s accuracy and demonstrate that the high-dimensional stochastic control models can successfully mitigate the risk.","key_phrases":["this paper","newsvendor problems","dynamic replenishment","financial hedging","stackelberg competition","these factors","the high-dimensional stochastic control models","hjb","the value functions","the curse","dimensionality","a deep learning algorithm","the hjb equations","a projection","the algorithm","the gradient explosion","the training phase","the deep learning algorithm","hjb equations","the newsvendor models","dimensions","numerical outcomes","the algorithm\u2019s accuracy","the high-dimensional stochastic control models","the risk","hamilton-jacobi-bellman","six"]},{"title":"An improved federated deep learning for plant leaf disease detection","authors":["Pragya Hari","Maheshwari Prasad Singh","Amit Kumar Singh"],"abstract":"Leaf diseases are hazardous to the yield and quality of food crops. There are many deep learning algorithms developed for their detection that may require large computational resources to train a single model with voluminous amounts of data. In the agriculture domain, various plant diseases are found across the country. Accumulating such a huge dataset from various regions is a tedious task, and training them with a single model can also be challenging. This paper proposes Federated Deep Learning (FDL) for Plant Leaf Disease Detection. This concept allows multiple local models to get trained with their region-based datasets and share their knowledge with siblings through the parent, instead of sharing complete datasets. Knowledge transfer significantly reduces the computational costs. This paper formulates the federated dataset using PlantVillage, to simulate the configuration of the FDL. A lightweight and efficient Hierarchical Convolutional Neural Network (H-CNN) is proposed for parent model and child models with 0.09 million parameters and 0.35MB model size. The simulation results show that the proposed FDL attains 93% testing accuracy, outperforming state-of-the-art methods namely, FedAdam and FedAvg with 86.8% and 87.7% respectively. In addition to this, the proposed FDL achieves 95.7%, 95.4% and 95.3% of weighted precision, weighted recall and weighted F1-score respectively for first local model and 92.1%, 90.8% and 91.2% respectively of weighted precision, weighted recall and weighted F1-score respectively for second local model.","doi":"10.1007\/s11042-024-18867-9","cleaned_title":"an improved federated deep learning for plant leaf disease detection","cleaned_abstract":"leaf diseases are hazardous to the yield and quality of food crops. there are many deep learning algorithms developed for their detection that may require large computational resources to train a single model with voluminous amounts of data. in the agriculture domain, various plant diseases are found across the country. accumulating such a huge dataset from various regions is a tedious task, and training them with a single model can also be challenging. this paper proposes federated deep learning (fdl) for plant leaf disease detection. this concept allows multiple local models to get trained with their region-based datasets and share their knowledge with siblings through the parent, instead of sharing complete datasets. knowledge transfer significantly reduces the computational costs. this paper formulates the federated dataset using plantvillage, to simulate the configuration of the fdl. a lightweight and efficient hierarchical convolutional neural network (h-cnn) is proposed for parent model and child models with 0.09 million parameters and 0.35mb model size. the simulation results show that the proposed fdl attains 93% testing accuracy, outperforming state-of-the-art methods namely, fedadam and fedavg with 86.8% and 87.7% respectively. in addition to this, the proposed fdl achieves 95.7%, 95.4% and 95.3% of weighted precision, weighted recall and weighted f1-score respectively for first local model and 92.1%, 90.8% and 91.2% respectively of weighted precision, weighted recall and weighted f1-score respectively for second local model.","key_phrases":["leaf diseases","the yield","quality","food crops","many deep learning algorithms","their detection","that","large computational resources","a single model","voluminous amounts","data","the agriculture domain","various plant diseases","the country","such a huge dataset","various regions","a tedious task","them","a single model","deep learning","(fdl","plant leaf disease detection","this concept","multiple local models","their region-based datasets","their knowledge","siblings","the parent","complete datasets","knowledge transfer","the computational costs","this paper","the federated dataset","plantvillage","the configuration","the fdl","a lightweight and efficient hierarchical convolutional neural network","h-cnn","parent model","child models","0.09 million parameters","0.35mb model size","the simulation results","93% testing accuracy","the-art","86.8%","87.7%","addition","this","the proposed fdl","95.7%","95.4%","95.3%","weighted precision","recall","f1-score","first local model","92.1%","90.8%","91.2%","weighted precision","recall","f1-score","second local model","0.09 million","0.35","93%","86.8%","87.7%","95.7%","95.4%","95.3%","first","92.1%","90.8%","91.2%","second"]},{"title":"Deep learning-based power usage effectiveness optimization for IoT-enabled data center","authors":["Yu Sun","Yanyi Wang","Gaoxiang Jiang","Bo Cheng","Haibo Zhou"],"abstract":"The proliferation of data centers is driving increased energy consumption, leading to environmentally unacceptable carbon emissions. As the use of Internet-of-Things (IoT) techniques for extensive data collection in data centers continues to grow, deep learning-based solutions have emerged as attractive alternatives to suboptimal traditional methods. However, existing approaches suffer from unsatisfactory performance, unrealistic assumptions, and an inability to address practical data center optimization. In this paper, we focus on power usage effectiveness (PUE) optimization in IoT-enabled data centers using deep learning algorithms. We first develop a deep learning-based PUE optimization framework tailored to IoT-enabled data centers. We then formulate the general PUE optimization problem, simplifying and specifying it for the minimization of long-term energy consumption in chiller cooling systems. Additionally, we introduce a transformer-based prediction network designed for energy consumption forecasting. Subsequently, we transform this formulation into a Markov decision process (MDP) and present the branching double dueling deep Q-network. This approach effectively tackles the challenges posed by enormous action spaces within MDP by branching actions into sub-actions. Extensive experiments conducted on real-world datasets demonstrate the exceptional performance of our algorithms, excelling in prediction precision, optimization convergence, and optimality while effectively managing a substantial number of actions on the order of \\(10^{13}\\).","doi":"10.1007\/s12083-024-01663-5","cleaned_title":"deep learning-based power usage effectiveness optimization for iot-enabled data center","cleaned_abstract":"the proliferation of data centers is driving increased energy consumption, leading to environmentally unacceptable carbon emissions. as the use of internet-of-things (iot) techniques for extensive data collection in data centers continues to grow, deep learning-based solutions have emerged as attractive alternatives to suboptimal traditional methods. however, existing approaches suffer from unsatisfactory performance, unrealistic assumptions, and an inability to address practical data center optimization. in this paper, we focus on power usage effectiveness (pue) optimization in iot-enabled data centers using deep learning algorithms. we first develop a deep learning-based pue optimization framework tailored to iot-enabled data centers. we then formulate the general pue optimization problem, simplifying and specifying it for the minimization of long-term energy consumption in chiller cooling systems. additionally, we introduce a transformer-based prediction network designed for energy consumption forecasting. subsequently, we transform this formulation into a markov decision process (mdp) and present the branching double dueling deep q-network. this approach effectively tackles the challenges posed by enormous action spaces within mdp by branching actions into sub-actions. extensive experiments conducted on real-world datasets demonstrate the exceptional performance of our algorithms, excelling in prediction precision, optimization convergence, and optimality while effectively managing a substantial number of actions on the order of \\(10^{13}\\).","key_phrases":["the proliferation","data centers","increased energy consumption","environmentally unacceptable carbon emissions","the use","things","iot","extensive data collection","data centers","deep learning-based solutions","attractive alternatives","suboptimal traditional methods","existing approaches","unsatisfactory performance","unrealistic assumptions","an inability","practical data center optimization","this paper","we","power usage effectiveness","(pue) optimization","iot-enabled data centers","deep learning algorithms","we","a deep learning-based pue optimization framework","iot-enabled data centers","we","the general pue optimization problem","it","the minimization","long-term energy consumption","chiller cooling systems","we","a transformer-based prediction network","energy consumption forecasting","we","this formulation","a markov decision process","mdp","the branching","deep q-network","this approach","the challenges","enormous action spaces","mdp","actions","sub","-","actions","extensive experiments","real-world datasets","the exceptional performance","our algorithms","prediction precision","optimization convergence","optimality","a substantial number","actions","the order","\\(10^{13}\\","first"]},{"title":"Deep reinforcement learning imbalanced credit risk of SMEs in supply chain finance","authors":["Wen Zhang","Shaoshan Yan","Jian Li","Rui Peng","Xin Tian"],"abstract":"It is crucial to predict the credit risk of small and medium-sized enterprises (SMEs) accurately for the success of supply chain finance (SCF). However, most of the existing research ignore the fact that the data distribution is usually imbalanced, that is, the proportion of default SMEs is much smaller than that of non-default SMEs. To fill this research gap, we propose a novel approach called DRL-Risk to deal with the imbalanced credit risk prediction (ICRP) of SMEs in SCF with deep reinforcement learning (DRL). Specifically, we formulate the ICRP problem as a Markov decision process and suggest an instance-based reward function to incorporate financial loss into the reward function with consideration of the actual loss caused by misclassification in the ICRP of SMEs. Then, we recommend a deep dueling neural network for decision policy to predict the credit risk of SMEs. With deep reinforcement learning, the DRL-Risk approach can prioritize the learning on the SMEs that would lead to great financial losses. Experimental results demonstrate that the DRL-Risk approach can significantly improve the performance of credit risk prediction of SMEs in SCF compared with the baseline methods in recall, G-mean, and financial loss. We have also identified management implications for the decision-makers participating in SCF.","doi":"10.1007\/s10479-024-05921-w","cleaned_title":"deep reinforcement learning imbalanced credit risk of smes in supply chain finance","cleaned_abstract":"it is crucial to predict the credit risk of small and medium-sized enterprises (smes) accurately for the success of supply chain finance (scf). however, most of the existing research ignore the fact that the data distribution is usually imbalanced, that is, the proportion of default smes is much smaller than that of non-default smes. to fill this research gap, we propose a novel approach called drl-risk to deal with the imbalanced credit risk prediction (icrp) of smes in scf with deep reinforcement learning (drl). specifically, we formulate the icrp problem as a markov decision process and suggest an instance-based reward function to incorporate financial loss into the reward function with consideration of the actual loss caused by misclassification in the icrp of smes. then, we recommend a deep dueling neural network for decision policy to predict the credit risk of smes. with deep reinforcement learning, the drl-risk approach can prioritize the learning on the smes that would lead to great financial losses. experimental results demonstrate that the drl-risk approach can significantly improve the performance of credit risk prediction of smes in scf compared with the baseline methods in recall, g-mean, and financial loss. we have also identified management implications for the decision-makers participating in scf.","key_phrases":["it","the credit risk","small and medium-sized enterprises","smes","the success","supply chain finance","scf","the existing research","the fact","the data distribution","the proportion","default smes","that","non-default smes","this research gap","we","a novel approach","drl-risk","the imbalanced credit risk prediction","icrp","smes","scf","deep reinforcement learning","drl","we","the icrp problem","a markov decision process","an instance-based reward function","financial loss","the reward function","consideration","the actual loss","misclassification","the icrp","smes","we","a deep dueling neural network","decision policy","the credit risk","smes","deep reinforcement learning","the drl-risk approach","the learning","the smes","that","great financial losses","experimental results","the drl-risk approach","the performance","credit risk prediction","smes","scf","the baseline methods","recall, g-mean, and financial loss","we","management implications","the decision-makers","scf"]},{"title":"Deep learning with autoencoders and LSTM for ENSO forecasting","authors":["Chibuike Chiedozie Ibebuchi","Michael B. Richman"],"abstract":"El Ni\u00f1o Southern Oscillation (ENSO) is the prominent recurrent climatic pattern in the tropical Pacific Ocean with global impacts on regional climates. This study utilizes deep learning to predict the Ni\u00f1o 3.4 index by encoding non-linear sea surface temperature patterns in the tropical Pacific using an autoencoder neural network. The resulting encoded patterns identify crucial centers of action in the Pacific that serve as predictors of the ENSO mode. These patterns are utilized as predictors for forecasting the Ni\u00f1o 3.4 index with a lead time of at least 6 months using the Long Short-Term Memory (LSTM) deep learning model. The analysis uncovers multiple non-linear dipole patterns in the tropical Pacific, with anomalies that are both regionalized and latitudinally oriented that should support a single inter-tropical convergence zone for modeling efforts. Leveraging these encoded patterns as predictors, the LSTM - trained on monthly data from 1950 to 2007 and tested from 2008 to 2022 - shows fidelity in predicting the Ni\u00f1o 3.4 index. The encoded patterns captured the annual cycle of ENSO with a 0.94 correlation between the actual and predicted Ni\u00f1o 3.4 index for lag 12 and 0.91 for lags 6 and 18. Additionally, the 6-month lag predictions excel in detecting extreme ENSO events, achieving an 85% hit rate, outperforming the 70% hit rate at lag 12 and 55% hit rate at lag 18. The prediction accuracy peaks from November to March, with correlations ranging from 0.94 to 0.96. The average correlations in the boreal spring were as large as 0.84, indicating the method has the capability to decrease the spring predictability barrier.","doi":"10.1007\/s00382-024-07180-8","cleaned_title":"deep learning with autoencoders and lstm for enso forecasting","cleaned_abstract":"el ni\u00f1o southern oscillation (enso) is the prominent recurrent climatic pattern in the tropical pacific ocean with global impacts on regional climates. this study utilizes deep learning to predict the ni\u00f1o 3.4 index by encoding non-linear sea surface temperature patterns in the tropical pacific using an autoencoder neural network. the resulting encoded patterns identify crucial centers of action in the pacific that serve as predictors of the enso mode. these patterns are utilized as predictors for forecasting the ni\u00f1o 3.4 index with a lead time of at least 6 months using the long short-term memory (lstm) deep learning model. the analysis uncovers multiple non-linear dipole patterns in the tropical pacific, with anomalies that are both regionalized and latitudinally oriented that should support a single inter-tropical convergence zone for modeling efforts. leveraging these encoded patterns as predictors, the lstm - trained on monthly data from 1950 to 2007 and tested from 2008 to 2022 - shows fidelity in predicting the ni\u00f1o 3.4 index. the encoded patterns captured the annual cycle of enso with a 0.94 correlation between the actual and predicted ni\u00f1o 3.4 index for lag 12 and 0.91 for lags 6 and 18. additionally, the 6-month lag predictions excel in detecting extreme enso events, achieving an 85% hit rate, outperforming the 70% hit rate at lag 12 and 55% hit rate at lag 18. the prediction accuracy peaks from november to march, with correlations ranging from 0.94 to 0.96. the average correlations in the boreal spring were as large as 0.84, indicating the method has the capability to decrease the spring predictability barrier.","key_phrases":["el ni\u00f1o southern oscillation","enso","the prominent recurrent climatic pattern","the tropical pacific ocean","global impacts","regional climates","this study","deep learning","the ni\u00f1o 3.4 index","non-linear sea surface temperature patterns","the tropical pacific","an autoencoder neural network","the resulting encoded patterns","crucial centers","action","the pacific","that","predictors","the enso mode","these patterns","predictors","the ni\u00f1o 3.4 index","a lead time","at least 6 months","the long short-term memory","lstm","deep learning model","the analysis uncovers","non-linear dipole patterns","the tropical pacific","anomalies","that","that","a single inter-tropical convergence zone","modeling efforts","these encoded patterns","predictors","monthly data","fidelity","the ni\u00f1o 3.4 index","the encoded patterns","the annual cycle","enso","a 0.94 correlation","3.4 index","lag","lags","the 6-month lag predictions","extreme enso events","an 85% hit rate","the 70% hit rate","lag","hit rate","lag","the prediction accuracy peaks","november","march","correlations","the average correlations","the boreal spring","the method","the capability","the spring predictability barrier","el ni\u00f1o southern oscillation","pacific","3.4","non-linear","3.4","at least 6 months","monthly","1950","2007","2008","2022","3.4","annual","0.94","3.4","12","0.91","6","6-month","85%","70%","55%","november to march","0.94","0.96","as large as 0.84"]},{"title":"Deep multi-metric training: the need of multi-metric curve evaluation to avoid weak learning","authors":["Michail Mamalakis","Abhirup Banerjee","Surajit Ray","Craig Wilkie","Richard H. Clayton","Andrew J. Swift","George Panoutsos","Bart Vorselaars"],"abstract":"The development and application of artificial intelligence-based computer vision systems in medicine, environment, and industry are playing an increasingly prominent role. Hence, the need for optimal and efficient hyperparameter tuning strategies is more than crucial to deliver the highest performance of the deep learning networks in large and demanding datasets. In our study, we have developed and evaluated a new training methodology named deep multi-metric training (DMMT) for enhanced training performance. The DMMT delivers a state of robust learning for deep networks using a new important criterion of multi-metric performance evaluation. We have tested the DMMT methodology in multi-class (three, four, and ten), multi-vendors (different X-ray imaging devices), and multi-size (large, medium, and small) datasets. The validity of the DMMT methodology has been tested in three different classification problems: (i) medical disease classification, (ii) environmental classification, and (iii) ecological classification. For disease classification, we have used two large COVID-19 chest X-rays datasets, namely the BIMCV COVID-19+ and Sheffield hospital datasets. The environmental application is related to the classification of weather images in cloudy, rainy, shine or sunrise conditions. The ecological classification task involves a classification of three animal species (cat, dog, wild) and a classification of ten animals and transportation vehicles categories (CIFAR-10). We have used state-of-the-art networks of DenseNet-121, ResNet-50, VGG-16, VGG-19, and DenResCov-19 (DenRes-131) to verify that our novel methodology is applicable in a variety of different deep learning networks. To the best of our knowledge, this is the first work that proposes a training methodology to deliver robust learning, over a variety of deep learning networks and multi-field classification problems.","doi":"10.1007\/s00521-024-10182-6","cleaned_title":"deep multi-metric training: the need of multi-metric curve evaluation to avoid weak learning","cleaned_abstract":"the development and application of artificial intelligence-based computer vision systems in medicine, environment, and industry are playing an increasingly prominent role. hence, the need for optimal and efficient hyperparameter tuning strategies is more than crucial to deliver the highest performance of the deep learning networks in large and demanding datasets. in our study, we have developed and evaluated a new training methodology named deep multi-metric training (dmmt) for enhanced training performance. the dmmt delivers a state of robust learning for deep networks using a new important criterion of multi-metric performance evaluation. we have tested the dmmt methodology in multi-class (three, four, and ten), multi-vendors (different x-ray imaging devices), and multi-size (large, medium, and small) datasets. the validity of the dmmt methodology has been tested in three different classification problems: (i) medical disease classification, (ii) environmental classification, and (iii) ecological classification. for disease classification, we have used two large covid-19 chest x-rays datasets, namely the bimcv covid-19+ and sheffield hospital datasets. the environmental application is related to the classification of weather images in cloudy, rainy, shine or sunrise conditions. the ecological classification task involves a classification of three animal species (cat, dog, wild) and a classification of ten animals and transportation vehicles categories (cifar-10). we have used state-of-the-art networks of densenet-121, resnet-50, vgg-16, vgg-19, and denrescov-19 (denres-131) to verify that our novel methodology is applicable in a variety of different deep learning networks. to the best of our knowledge, this is the first work that proposes a training methodology to deliver robust learning, over a variety of deep learning networks and multi-field classification problems.","key_phrases":["the development","application","artificial intelligence-based computer vision systems","medicine","environment","industry","an increasingly prominent role","the need","optimal and efficient hyperparameter tuning strategies","the highest performance","the deep learning networks","large and demanding datasets","our study","we","a new training methodology","deep multi-metric training","dmmt","enhanced training performance","the dmmt","a state","robust learning","deep networks","a new important criterion","multi-metric performance evaluation","we","the dmmt methodology","-","-","vendors (different x-ray imaging devices","multi-size (large, medium, and small) datasets","the validity","the dmmt methodology","three different classification problems","(i) medical disease classification","(ii) environmental classification","(iii) ecological classification","disease classification","we","two large covid-19 chest x-rays datasets","namely the bimcv covid-19","hospital datasets","the environmental application","the classification","weather images","shine","sunrise conditions","the ecological classification task","a classification","three animal species","a classification","ten animals","transportation vehicles categories","cifar-10","we","the-art","resnet-50","vgg-16","vgg-19","denrescov-19","(denres-131","our novel methodology","a variety","different deep learning networks","our knowledge","this","the first work","that","a training methodology","robust learning","a variety","deep learning networks","multi-field classification problems","three","four","ten","three","two","covid-19","covid-19+","three","ten","cifar-10","densenet-121, resnet-50","vgg-16","vgg-19","denrescov-19","first"]},{"title":"HCR-Net: a deep learning based script independent handwritten character recognition network","authors":["Vinod Kumar Chauhan","Sukhdeep Singh","Anuj Sharma"],"abstract":"Handwritten character recognition (HCR) remains a challenging pattern recognition problem despite decades of research, and lacks research on script independent recognition techniques. This is mainly because of similar character structures, different handwriting styles, diverse scripts, handcrafted feature extraction techniques, unavailability of data and code, and the development of script-specific deep learning techniques. To address these limitations, we have proposed a script independent deep learning network for HCR research, called HCR-Net, that sets a new research direction for the field. HCR-Net is based on a novel transfer learning approach for HCR, which partly utilizes feature extraction layers of a pre-trained network. Due to transfer learning and image augmentation, HCR-Net provides faster and computationally efficient training, better performance and generalizations, and can work with small datasets. HCR-Net is extensively evaluated on 40 publicly available datasets of Bangla, Punjabi, Hindi, English, Swedish, Urdu, Farsi, Tibetan, Kannada, Malayalam, Telugu, Marathi, Nepali and Arabic languages, and established 26 new benchmark results while performed close to the best results in the rest cases. HCR-Net showed performance improvements up to 11% against the existing results and achieved a fast convergence rate showing up to 99% of final performance in the very first epoch. HCR-Net significantly outperformed the state-of-the-art transfer learning techniques and also reduced the number of trainable parameters by 34% as compared with the corresponding pre-trained network. To facilitate reproducibility and further advancements of HCR research, the complete code is publicly released at https:\/\/github.com\/jmdvinodjmd\/HCR-Net.","doi":"10.1007\/s11042-024-18655-5","cleaned_title":"hcr-net: a deep learning based script independent handwritten character recognition network","cleaned_abstract":"handwritten character recognition (hcr) remains a challenging pattern recognition problem despite decades of research, and lacks research on script independent recognition techniques. this is mainly because of similar character structures, different handwriting styles, diverse scripts, handcrafted feature extraction techniques, unavailability of data and code, and the development of script-specific deep learning techniques. to address these limitations, we have proposed a script independent deep learning network for hcr research, called hcr-net, that sets a new research direction for the field. hcr-net is based on a novel transfer learning approach for hcr, which partly utilizes feature extraction layers of a pre-trained network. due to transfer learning and image augmentation, hcr-net provides faster and computationally efficient training, better performance and generalizations, and can work with small datasets. hcr-net is extensively evaluated on 40 publicly available datasets of bangla, punjabi, hindi, english, swedish, urdu, farsi, tibetan, kannada, malayalam, telugu, marathi, nepali and arabic languages, and established 26 new benchmark results while performed close to the best results in the rest cases. hcr-net showed performance improvements up to 11% against the existing results and achieved a fast convergence rate showing up to 99% of final performance in the very first epoch. hcr-net significantly outperformed the state-of-the-art transfer learning techniques and also reduced the number of trainable parameters by 34% as compared with the corresponding pre-trained network. to facilitate reproducibility and further advancements of hcr research, the complete code is publicly released at https:\/\/github.com\/jmdvinodjmd\/hcr-net.","key_phrases":["handwritten character recognition","hcr","a challenging pattern recognition problem","decades","research","research","script independent recognition techniques","this","similar character structures","different handwriting styles","diverse scripts","handcrafted feature extraction techniques","unavailability","data","code","the development","script-specific deep learning techniques","these limitations","we","a script independent deep learning network","hcr research","hcr","-","net","that","a new research direction","the field","hcr","net","a novel transfer learning approach","hcr","which","feature extraction layers","a pre-trained network","learning","image augmentation","hcr-net","faster and computationally efficient training","better performance","generalizations","small datasets","hcr","net","40 publicly available datasets","bangla","punjabi","hindi","english","urdu","farsi","tibetan","kannada","malayalam","telugu","marathi","nepali","arabic languages","26 new benchmark results","the best results","the rest cases","net","performance improvements","the existing results","a fast convergence rate","up to 99%","final performance","the very first epoch","hcr","net","the-art","techniques","the number","trainable parameters","34%","the corresponding pre-trained network","reproducibility","further advancements","hcr research","the complete code","https:\/\/github.com\/jmdvinodjmd\/hcr-net","decades","40","english","swedish","tibetan","kannada","malayalam","arabic","26","11%","up to 99%","first","34%"]},{"title":"Prediction of non-muscle invasive bladder cancer recurrence using deep learning of pathology image","authors":["Guang-Yue Wang","Jing-Fei Zhu","Qi-Chao Wang","Jia-Xin Qin","Xin-Lei Wang","Xing Liu","Xin-Yu Liu","Jun-Zhi Chen","Jie-Fei Zhu","Shi-Chao Zhuo","Di Wu","Na Li","Liu Chao","Fan-Lai Meng","Hao Lu","Zhen-Duo Shi","Zhi-Gang Jia","Cong-Hui Han"],"abstract":"We aimed to build a deep learning-based pathomics model to predict the early recurrence of non-muscle-infiltrating bladder cancer (NMIBC) in this work. A total of 147 patients from Xuzhou Central Hospital were enrolled as the training cohort, and 63 patients from Suqian Affiliated Hospital of Xuzhou Medical University were enrolled as the test cohort. Based on two consecutive phases of patch level prediction and WSI-level predictione, we built a pathomics model, with the initial model developed in the training cohort and subjected to transfer learning, and then the test cohort was validated for generalization. The features extracted from the visualization model were used for model interpretation. After migration learning, the area under the receiver operating characteristic curve for the deep learning-based pathomics model in the test cohort was 0.860 (95% CI 0.752\u20130.969), with good agreement between the migration training cohort and the test cohort in predicting recurrence, and the predicted values matched well with the observed values, with p values of 0.667766 and 0.140233 for the Hosmer\u2013Lemeshow test, respectively. The good clinical application was observed using a decision curve analysis method. We developed a deep learning-based pathomics model showed promising performance in predicting recurrence within one year in NMIBC patients. Including 10 state prediction NMIBC recurrence group pathology features be visualized, which may be used to facilitate personalized management of NMIBC patients to avoid ineffective or unnecessary treatment for the benefit of patients.","doi":"10.1038\/s41598-024-66870-9","cleaned_title":"prediction of non-muscle invasive bladder cancer recurrence using deep learning of pathology image","cleaned_abstract":"we aimed to build a deep learning-based pathomics model to predict the early recurrence of non-muscle-infiltrating bladder cancer (nmibc) in this work. a total of 147 patients from xuzhou central hospital were enrolled as the training cohort, and 63 patients from suqian affiliated hospital of xuzhou medical university were enrolled as the test cohort. based on two consecutive phases of patch level prediction and wsi-level predictione, we built a pathomics model, with the initial model developed in the training cohort and subjected to transfer learning, and then the test cohort was validated for generalization. the features extracted from the visualization model were used for model interpretation. after migration learning, the area under the receiver operating characteristic curve for the deep learning-based pathomics model in the test cohort was 0.860 (95% ci 0.752\u20130.969), with good agreement between the migration training cohort and the test cohort in predicting recurrence, and the predicted values matched well with the observed values, with p values of 0.667766 and 0.140233 for the hosmer\u2013lemeshow test, respectively. the good clinical application was observed using a decision curve analysis method. we developed a deep learning-based pathomics model showed promising performance in predicting recurrence within one year in nmibc patients. including 10 state prediction nmibc recurrence group pathology features be visualized, which may be used to facilitate personalized management of nmibc patients to avoid ineffective or unnecessary treatment for the benefit of patients.","key_phrases":["we","a deep learning-based pathomics model","the early recurrence","non-muscle-infiltrating bladder cancer","nmibc","this work","a total","147 patients","xuzhou central hospital","the training cohort","63 patients","suqian affiliated hospital","xuzhou medical university","the test cohort","two consecutive phases","patch level prediction","wsi-level predictione","we","a pathomics model","the initial model","the training cohort","learning","the test cohort","generalization","the features","the visualization model","model interpretation","migration learning","the area","the receiver operating characteristic curve","the deep learning-based pathomics model","the test cohort","(95%","good agreement","the migration training cohort","the test cohort","recurrence","the predicted values","the observed values","p values","the hosmer","lemeshow test","the good clinical application","a decision curve analysis method","we","a deep learning-based pathomics model","promising performance","recurrence","one year","nmibc patients","10 state prediction nmibc recurrence group pathology features","which","personalized management","nmibc patients","ineffective or unnecessary treatment","the benefit","patients","147","xuzhou","63","xuzhou medical university","two","0.860","95%","0.667766","0.140233","one year","10"]},{"title":"A predictive analytics framework for sensor data using time series and deep learning techniques","authors":["Hend A. Selmy","Hoda K. Mohamed","Walaa Medhat"],"abstract":"IoT devices convert billions of objects into data-generating entities, enabling them to report status and interact with their surroundings. This data comes in various formats, like structured, semi-structured, or unstructured. In addition, it can be collected in batches or in real time. The problem now is how to benefit from all of this data gathered by sensing and monitoring changes like temperature, light, and position. In this paper, we propose a predictive analytics framework constructed on top of open-source technologies such as Apache Spark and Kafka. The framework focuses on forecasting temperature time series data using traditional and deep learning predictive analytics methods. The analysis and prediction tasks were performed using Autoregressive Integrated Moving Average (ARIMA), Seasonal Autoregressive Integrated Moving Average (SARIMA), Long Short-Term Memory (LSTM), and a novel hybrid model based on Convolution Neural Network (CNN) and LSTM. The purpose of this paper is to determine whether and how recently developed deep learning-based models outperform traditional algorithms in the prediction of time series data.\u00a0The empirical studies conducted and reported in this paper demonstrate that deep learning-based models, specifically LSTM and CNN-LSTM, exhibit superior performance compared to traditional-based algorithms, ARIMA and SARIMA. More specifically, the average reduction in error rates obtained by LSTM and CNN-LSTM models were substantial when compared to other models indicating the superiority of deep learning. Moreover, the CNN-LSTM-based deep learning model exhibits a higher degree of closeness to the actual values when compared to the LSTM-based model.","doi":"10.1007\/s00521-023-09398-9","cleaned_title":"a predictive analytics framework for sensor data using time series and deep learning techniques","cleaned_abstract":"iot devices convert billions of objects into data-generating entities, enabling them to report status and interact with their surroundings. this data comes in various formats, like structured, semi-structured, or unstructured. in addition, it can be collected in batches or in real time. the problem now is how to benefit from all of this data gathered by sensing and monitoring changes like temperature, light, and position. in this paper, we propose a predictive analytics framework constructed on top of open-source technologies such as apache spark and kafka. the framework focuses on forecasting temperature time series data using traditional and deep learning predictive analytics methods. the analysis and prediction tasks were performed using autoregressive integrated moving average (arima), seasonal autoregressive integrated moving average (sarima), long short-term memory (lstm), and a novel hybrid model based on convolution neural network (cnn) and lstm. the purpose of this paper is to determine whether and how recently developed deep learning-based models outperform traditional algorithms in the prediction of time series data. the empirical studies conducted and reported in this paper demonstrate that deep learning-based models, specifically lstm and cnn-lstm, exhibit superior performance compared to traditional-based algorithms, arima and sarima. more specifically, the average reduction in error rates obtained by lstm and cnn-lstm models were substantial when compared to other models indicating the superiority of deep learning. moreover, the cnn-lstm-based deep learning model exhibits a higher degree of closeness to the actual values when compared to the lstm-based model.","key_phrases":["iot devices","billions","objects","data-generating entities","them","status","their surroundings","this data","various formats","addition","it","batches","real time","the problem","all","this data","changes","temperature","light","position","this paper","we","a predictive analytics framework","top","open-source technologies","apache spark","the framework","temperature time series data","traditional and deep learning predictive analytics methods","the analysis and prediction tasks","arima","sarima","lstm","convolution neural network","cnn","lstm","the purpose","this paper","how recently developed deep learning-based models","traditional algorithms","the prediction","time series data","the empirical studies","this paper demonstrate","deep learning-based models","specifically lstm","cnn-lstm","superior performance","traditional-based algorithms","arima","sarima","the average reduction","error rates","lstm","cnn-lstm models","other models","the superiority","deep learning","the cnn-lstm-based deep learning model","a higher degree","closeness","the actual values","the lstm-based model","billions","kafka","cnn","cnn-lstm","cnn","cnn"]},{"title":"Deep learning-based comprehensive review on pulmonary tuberculosis","authors":["Twinkle Bansal","Sheifali Gupta","Neeru Jindal"],"abstract":"In areas with high tuberculosis (TB) prevalence, high mortality rate has significantly increased over the past few decades. Even though tuberculosis can be treated, areas with high disease burden continue to have insufficient screening tools, leading to diagnostic delays and incorrect diagnoses. As a result of these challenges, a computer-aided diagnostics (CAD) system has been developed that can automatically detect tuberculosis. There are few different methods that can be used to screen for tuberculosis; however, chest X-ray (CXR) is most commonly used and strongly suggested because it is so effective in identifying lung irregularities. Over past ten years, we have seen a meteoric rise in amount of research conducted into application of machine learning strategies to examination of chest X-ray images for screening regarding pulmonary abnormalities. Particularly, we have also noticed significant interest in testing for TB. This attentiveness has increased in tandem with phenomenal progress that has been made in deep learning (DL), which is predominately founded on convolutional neural networks (CNNs). Because of these advancements, significant research contributions have been made in field of DL techniques for TB screening by utilizing CXR images. The main focus of this paper is to emphasize favorable methods and data collection, as well as methodological contributions, identify data collections, and identify challenges.","doi":"10.1007\/s00521-023-09381-4","cleaned_title":"deep learning-based comprehensive review on pulmonary tuberculosis","cleaned_abstract":"in areas with high tuberculosis (tb) prevalence, high mortality rate has significantly increased over the past few decades. even though tuberculosis can be treated, areas with high disease burden continue to have insufficient screening tools, leading to diagnostic delays and incorrect diagnoses. as a result of these challenges, a computer-aided diagnostics (cad) system has been developed that can automatically detect tuberculosis. there are few different methods that can be used to screen for tuberculosis; however, chest x-ray (cxr) is most commonly used and strongly suggested because it is so effective in identifying lung irregularities. over past ten years, we have seen a meteoric rise in amount of research conducted into application of machine learning strategies to examination of chest x-ray images for screening regarding pulmonary abnormalities. particularly, we have also noticed significant interest in testing for tb. this attentiveness has increased in tandem with phenomenal progress that has been made in deep learning (dl), which is predominately founded on convolutional neural networks (cnns). because of these advancements, significant research contributions have been made in field of dl techniques for tb screening by utilizing cxr images. the main focus of this paper is to emphasize favorable methods and data collection, as well as methodological contributions, identify data collections, and identify challenges.","key_phrases":["areas","high tuberculosis","tb","prevalence","high mortality rate","the past few decades","tuberculosis","areas","high disease burden","insufficient screening tools","diagnostic delays","incorrect diagnoses","a result","these challenges","a computer-aided diagnostics","cad) system","that","tuberculosis","few different methods","that","tuberculosis","chest x","-","(cxr","it","lung irregularities","past ten years","we","a meteoric rise","amount","research","application","machine learning strategies","examination","chest x-ray images","pulmonary abnormalities","we","significant interest","testing","tb","this attentiveness","tandem","phenomenal progress","that","deep learning","dl","which","convolutional neural networks","cnns","these advancements","significant research contributions","field","dl techniques","tb","cxr images","the main focus","this paper","favorable methods","data collection","methodological contributions","data collections","challenges","the past few decades"]},{"title":"Adversarial defence by learning differentiated feature representation in deep ensemble","authors":["Xi Chen","Wei Huang","Wei Guo","Fan Zhang","Jiayu Du","Zhizhong Zhou"],"abstract":"Deep learning models have been shown to be vulnerable to critical attacks under adversarial conditions. Attackers are able to generate powerful adversarial examples by searching for adversarial perturbations, without interfering with model training or directly modifying the model. This phenomenon indicates an endogenous problem in existing deep learning frameworks. Therefore, optimizing individual models for defense is often limited and can always be defeated by new attack methods. Ensemble defense has been shown to be effective in defending against adversarial attacks by combining diverse models. However, the problem of insufficient differentiation among existing models persists. Active defense in cyberspace security has successfully defended against unknown vulnerabilities by integrating subsystems with multiple different implementations to achieve a unified mission objective. Inspired by this, we propose exploring the feasibility of achieving model differentiation by changing the data features used in training individual models, as they are the core factor of functional implementation. We utilize several feature extraction methods to preprocess the data and train differentiated models based on these features. By generating adversarial perturbations to attack different models, we demonstrate that the feature representation of the data is highly resistant to adversarial perturbations. The entire ensemble is able to operate normally in an error-bearing environment.","doi":"10.1007\/s00138-024-01571-x","cleaned_title":"adversarial defence by learning differentiated feature representation in deep ensemble","cleaned_abstract":"deep learning models have been shown to be vulnerable to critical attacks under adversarial conditions. attackers are able to generate powerful adversarial examples by searching for adversarial perturbations, without interfering with model training or directly modifying the model. this phenomenon indicates an endogenous problem in existing deep learning frameworks. therefore, optimizing individual models for defense is often limited and can always be defeated by new attack methods. ensemble defense has been shown to be effective in defending against adversarial attacks by combining diverse models. however, the problem of insufficient differentiation among existing models persists. active defense in cyberspace security has successfully defended against unknown vulnerabilities by integrating subsystems with multiple different implementations to achieve a unified mission objective. inspired by this, we propose exploring the feasibility of achieving model differentiation by changing the data features used in training individual models, as they are the core factor of functional implementation. we utilize several feature extraction methods to preprocess the data and train differentiated models based on these features. by generating adversarial perturbations to attack different models, we demonstrate that the feature representation of the data is highly resistant to adversarial perturbations. the entire ensemble is able to operate normally in an error-bearing environment.","key_phrases":["deep learning models","critical attacks","adversarial conditions","attackers","powerful adversarial examples","adversarial perturbations","model training","the model","this phenomenon","an endogenous problem","existing deep learning frameworks","individual models","defense","new attack methods","ensemble defense","adversarial attacks","diverse models","the problem","insufficient differentiation","existing models","active defense","cyberspace security","unknown vulnerabilities","subsystems","multiple different implementations","a unified mission objective","this","we","the feasibility","model differentiation","the data features","individual models","they","the core factor","functional implementation","we","several feature extraction methods","the data","differentiated models","these features","adversarial perturbations","different models","we","the feature representation","the data","adversarial perturbations","the entire ensemble","an error-bearing environment"]},{"title":"Scaling deep learning for materials discovery","authors":["Amil Merchant","Simon Batzner","Samuel S. Schoenholz","Muratahan Aykol","Gowoon Cheon","Ekin Dogus Cubuk"],"abstract":"Novel functional materials enable fundamental breakthroughs across technological applications from clean energy to information processing1,2,3,4,5,6,7,8,9,10,11. From microchips to batteries and photovoltaics, discovery of inorganic crystals has been bottlenecked by expensive trial-and-error approaches. Concurrently, deep-learning models for language, vision and biology have showcased emergent predictive capabilities with increasing data and computation12,13,14. Here we show that graph networks trained at scale can reach unprecedented levels of generalization, improving the efficiency of materials discovery by an order of magnitude. Building on 48,000 stable crystals identified in continuing studies15,16,17, improved efficiency enables the discovery of 2.2 million structures below the current convex hull, many of which escaped previous human chemical intuition. Our work represents an order-of-magnitude expansion in stable materials known to humanity. Stable discoveries that are on the final convex hull will be made available to screen for technological applications, as we demonstrate for layered materials and solid-electrolyte candidates. Of the stable structures, 736 have already been independently experimentally realized. The scale and diversity of hundreds of millions of first-principles calculations also unlock modelling capabilities for downstream applications, leading in particular to highly accurate and robust learned interatomic potentials that can be used in condensed-phase molecular-dynamics simulations and high-fidelity zero-shot prediction of ionic conductivity.","doi":"10.1038\/s41586-023-06735-9","cleaned_title":"scaling deep learning for materials discovery","cleaned_abstract":"novel functional materials enable fundamental breakthroughs across technological applications from clean energy to information processing1,2,3,4,5,6,7,8,9,10,11. from microchips to batteries and photovoltaics, discovery of inorganic crystals has been bottlenecked by expensive trial-and-error approaches. concurrently, deep-learning models for language, vision and biology have showcased emergent predictive capabilities with increasing data and computation12,13,14. here we show that graph networks trained at scale can reach unprecedented levels of generalization, improving the efficiency of materials discovery by an order of magnitude. building on 48,000 stable crystals identified in continuing studies15,16,17, improved efficiency enables the discovery of 2.2 million structures below the current convex hull, many of which escaped previous human chemical intuition. our work represents an order-of-magnitude expansion in stable materials known to humanity. stable discoveries that are on the final convex hull will be made available to screen for technological applications, as we demonstrate for layered materials and solid-electrolyte candidates. of the stable structures, 736 have already been independently experimentally realized. the scale and diversity of hundreds of millions of first-principles calculations also unlock modelling capabilities for downstream applications, leading in particular to highly accurate and robust learned interatomic potentials that can be used in condensed-phase molecular-dynamics simulations and high-fidelity zero-shot prediction of ionic conductivity.","key_phrases":["novel functional materials","fundamental breakthroughs","technological applications","clean energy","information processing1,2,3,4,5,6,7,8,9,10,11","microchips","batteries","photovoltaics","discovery","inorganic crystals","expensive trial-and-error approaches","deep-learning models","language","vision","biology","emergent predictive capabilities","increasing data","computation12,13,14","we","graph networks","scale","unprecedented levels","generalization","the efficiency","materials discovery","an order","magnitude","48,000 stable crystals","continuing studies15,16,17","improved efficiency","the discovery","2.2 million structures","the current convex hull","which","previous human chemical intuition","our work","magnitude","stable materials","humanity","stable discoveries","that","the final convex hull","technological applications","we","layered materials","solid-electrolyte candidates","the stable structures","the scale","diversity","hundreds of millions","first-principles calculations","modelling capabilities","downstream applications","learned interatomic potentials","that","condensed-phase molecular-dynamics simulations","high-fidelity zero-shot prediction","ionic conductivity","48,000","2.2 million","736","hundreds of millions","first","zero"]},{"title":"Deep learning-aided 3D proxy-bridged region-growing framework for multi-organ segmentation","authors":["Zhihong Chen","Lisha Yao","Yue Liu","Xiaorui Han","Zhengze Gong","Jichao Luo","Jietong Zhao","Gang Fang"],"abstract":"Accurate multi-organ segmentation\u00a0in 3D CT images is imperative for enhancing computer-aided diagnosis and radiotherapy planning. However, current deep learning-based methods for 3D multi-organ segmentation face challenges such as the need for labor-intensive manual pixel-level annotations and high hardware resource demands, especially regarding GPU resources. To address these issues, we propose a 3D proxy-bridged region-growing framework specifically designed for the segmentation of the liver and spleen. Specifically, a key slice is selected from each 3D volume according to the corresponding intensity histogram. Subsequently, a deep learning model is employed to pinpoint the semantic central patch on this key slice, to calculate the growing seed. To counteract the impact of noise, segmentation of the liver and spleen is conducted on superpixel images created through proxy-bridging strategy. The segmentation process is then extended to adjacent slices by applying the same methodology iteratively, culminating in the comprehensive segmentation results. Experimental results demonstrate that the proposed framework accomplishes segmentation of the liver and spleen with an average Dice Similarity Coefficient of approximately 0.93 and a Jaccard Similarity Coefficient of around 0.88. These outcomes substantiate the framework's capability to achieve performance on par with that of deep learning methods, albeit requiring less guidance information and lower GPU resources.","doi":"10.1038\/s41598-024-60668-5","cleaned_title":"deep learning-aided 3d proxy-bridged region-growing framework for multi-organ segmentation","cleaned_abstract":"accurate multi-organ segmentation in 3d ct images is imperative for enhancing computer-aided diagnosis and radiotherapy planning. however, current deep learning-based methods for 3d multi-organ segmentation face challenges such as the need for labor-intensive manual pixel-level annotations and high hardware resource demands, especially regarding gpu resources. to address these issues, we propose a 3d proxy-bridged region-growing framework specifically designed for the segmentation of the liver and spleen. specifically, a key slice is selected from each 3d volume according to the corresponding intensity histogram. subsequently, a deep learning model is employed to pinpoint the semantic central patch on this key slice, to calculate the growing seed. to counteract the impact of noise, segmentation of the liver and spleen is conducted on superpixel images created through proxy-bridging strategy. the segmentation process is then extended to adjacent slices by applying the same methodology iteratively, culminating in the comprehensive segmentation results. experimental results demonstrate that the proposed framework accomplishes segmentation of the liver and spleen with an average dice similarity coefficient of approximately 0.93 and a jaccard similarity coefficient of around 0.88. these outcomes substantiate the framework's capability to achieve performance on par with that of deep learning methods, albeit requiring less guidance information and lower gpu resources.","key_phrases":["accurate multi-organ segmentation","3d ct images","computer-aided diagnosis","radiotherapy","planning","however, current deep learning-based methods","3d multi-organ segmentation face challenges","the need","labor-intensive manual pixel-level annotations","high hardware resource demands","gpu resources","these issues","we","a 3d proxy-bridged region-growing framework","the segmentation","the liver","spleen","a key slice","each 3d volume","the corresponding intensity histogram","a deep learning model","the semantic central patch","this key slice","the growing seed","the impact","noise","segmentation","the liver","spleen","superpixel images","proxy-bridging strategy","the segmentation process","adjacent slices","the same methodology","the comprehensive segmentation results","experimental results","the proposed framework","segmentation","the liver","an average dice similarity","a jaccard similarity","these outcomes","the framework's capability","performance","par","that","deep learning methods","less guidance information","lower gpu resources","3d","3d","3d","3d","approximately 0.93","around 0.88"]},{"title":"Biometrics recognition using deep learning: a survey","authors":["Shervin Minaee","Amirali Abdolrashidi","Hang Su","Mohammed Bennamoun","David Zhang"],"abstract":"In the past few years, deep learning-based models have been very successful in achieving state-of-the-art results in many tasks in computer vision, speech recognition, and natural language processing. These models seem to be a natural fit for handling the ever-increasing scale of biometric recognition problems, from cellphone authentication to airport security systems. Deep learning-based models have increasingly been leveraged to improve the accuracy of different biometric recognition systems in recent years. In this work, we provide a comprehensive survey of more than 150 promising works on biometric recognition (including face, fingerprint, iris, palmprint, ear, voice, signature, and gait recognition), which deploy deep learning models, and show their strengths and potentials in different applications. For each biometric, we first introduce the available datasets that are widely used in the literature and their characteristics. We will then talk about several promising deep learning works developed for that biometric, and show their performance on popular public benchmarks. We will also discuss some of the main challenges while using these models for biometric recognition, and possible future directions to which research in this area is headed.","doi":"10.1007\/s10462-022-10237-x","cleaned_title":"biometrics recognition using deep learning: a survey","cleaned_abstract":"in the past few years, deep learning-based models have been very successful in achieving state-of-the-art results in many tasks in computer vision, speech recognition, and natural language processing. these models seem to be a natural fit for handling the ever-increasing scale of biometric recognition problems, from cellphone authentication to airport security systems. deep learning-based models have increasingly been leveraged to improve the accuracy of different biometric recognition systems in recent years. in this work, we provide a comprehensive survey of more than 150 promising works on biometric recognition (including face, fingerprint, iris, palmprint, ear, voice, signature, and gait recognition), which deploy deep learning models, and show their strengths and potentials in different applications. for each biometric, we first introduce the available datasets that are widely used in the literature and their characteristics. we will then talk about several promising deep learning works developed for that biometric, and show their performance on popular public benchmarks. we will also discuss some of the main challenges while using these models for biometric recognition, and possible future directions to which research in this area is headed.","key_phrases":["the past few years","deep learning-based models","the-art","many tasks","computer vision","speech recognition","natural language processing","these models","a natural fit","the ever-increasing scale","biometric recognition problems","cellphone authentication","security systems","deep learning-based models","the accuracy","different biometric recognition systems","recent years","this work","we","a comprehensive survey","more than 150 promising works","biometric recognition","face","fingerprint","iris","ear","voice","signature","gait recognition","which","deep learning models","their strengths","potentials","different applications","each biometric","we","the available datasets","that","the literature","their characteristics","we","several promising deep learning works","their performance","popular public benchmarks","we","some","the main challenges","these models","biometric recognition","possible future directions","which research","this area","the past few years","recent years","more than 150","iris","first"]},{"title":"Harnessing the Power of 6G Connectivity for Advanced Big Data Analytics with Deep Learning","authors":["Maojin Sun","Luyi Sun"],"abstract":"The smart applications development worldwide demands for ultra-reliable data communication to assure the richness of data and processing in time. These smart applications create massive amounts of data to be processed in 6G networks with advanced technologies. 6G big data analytics become the demand for next-generation data communication and smart city applications. Traditional data analytics algorithms lag in efficiency while processing big data due to huge volume, data dependency and timely processing. A deep learning model called reinforcement learning is promising for processing big data in smart applications. The proposed study, advanced big data Analytics using Deep learning (ABDAS-DL), gives a pioneering approach that combines Deep Reinforcement learning (DRL) based Deep Q network (DQN) with long-term, short-term memory (LSTM) for harnessing the vast capacity of 6G connectivity within the domain of advanced big data analytics. This study utilises smart transport-based data for taxi route optimisation by analysing climatic and surrounding factors. The look of 6G connectivity guarantees incredible facts of data transmission speeds and tremendously low latency, taking off new horizons for managing large datasets in real time. The performance of the proposed model is measured in terms of processing time, network, reliability and scalability. The proposed model takes 30\u00a0s to process the data and fix the taxi route, while another traditional model consumes more than an hour.","doi":"10.1007\/s11277-024-11044-z","cleaned_title":"harnessing the power of 6g connectivity for advanced big data analytics with deep learning","cleaned_abstract":"the smart applications development worldwide demands for ultra-reliable data communication to assure the richness of data and processing in time. these smart applications create massive amounts of data to be processed in 6g networks with advanced technologies. 6g big data analytics become the demand for next-generation data communication and smart city applications. traditional data analytics algorithms lag in efficiency while processing big data due to huge volume, data dependency and timely processing. a deep learning model called reinforcement learning is promising for processing big data in smart applications. the proposed study, advanced big data analytics using deep learning (abdas-dl), gives a pioneering approach that combines deep reinforcement learning (drl) based deep q network (dqn) with long-term, short-term memory (lstm) for harnessing the vast capacity of 6g connectivity within the domain of advanced big data analytics. this study utilises smart transport-based data for taxi route optimisation by analysing climatic and surrounding factors. the look of 6g connectivity guarantees incredible facts of data transmission speeds and tremendously low latency, taking off new horizons for managing large datasets in real time. the performance of the proposed model is measured in terms of processing time, network, reliability and scalability. the proposed model takes 30 s to process the data and fix the taxi route, while another traditional model consumes more than an hour.","key_phrases":["the smart applications development","ultra-reliable data communication","the richness","data","processing","time","these smart applications","massive amounts","data","6g networks","advanced technologies","6g big data analytics","the demand","next-generation data communication","smart city applications","efficiency","big data","huge volume","data dependency","timely processing","a deep learning model","reinforcement learning","big data","smart applications","the proposed study","advanced big data analytics","deep learning","abdas","dl","a pioneering approach","that","deep reinforcement learning","drl","deep q network","dqn","long-term, short-term memory","lstm","the vast capacity","6g connectivity","the domain","advanced big data analytics","this study","smart transport-based data","taxi route optimisation","climatic and surrounding factors","the look","6g connectivity","incredible facts","data transmission speeds","tremendously low latency","new horizons","large datasets","real time","the performance","the proposed model","terms","processing time","network","reliability","scalability","the proposed model","30 s","the data","the taxi route","another traditional model","an hour","6","6","6","6","30","more than an hour"]},{"title":"Assessments of Data-Driven Deep Learning Models on One-Month Predictions of Pan-Arctic Sea Ice Thickness","authors":["Chentao Song","Jiang Zhu","Xichen Li"],"abstract":"In recent years, deep learning methods have gradually been applied to prediction tasks related to Arctic sea ice concentration, but relatively little research has been conducted for larger spatial and temporal scales, mainly due to the limited time coverage of observations and reanalysis data. Meanwhile, deep learning predictions of sea ice thickness (SIT) have yet to receive ample attention. In this study, two data-driven deep learning (DL) models are built based on the ConvLSTM and fully convolutional U-net (FC-Unet) algorithms and trained using CMIP6 historical simulations for transfer learning and fine-tuned using reanalysis\/observations. These models enable monthly predictions of Arctic SIT without considering the complex physical processes involved. Through comprehensive assessments of prediction skills by season and region, the results suggest that using a broader set of CMIP6 data for transfer learning, as well as incorporating multiple climate variables as predictors, contribute to better prediction results, although both DL models can effectively predict the spatiotemporal features of SIT anomalies. Regarding the predicted SIT anomalies of the FC-Unet model, the spatial correlations with reanalysis reach an average level of 89% over all months, while the temporal anomaly correlation coefficients are close to unity in most cases. The models also demonstrate robust performances in predicting SIT and SIE during extreme events. The effectiveness and reliability of the proposed deep transfer learning models in predicting Arctic SIT can facilitate more accurate pan-Arctic predictions, aiding climate change research and real-time business applications.","doi":"10.1007\/s00376-023-3259-3","cleaned_title":"assessments of data-driven deep learning models on one-month predictions of pan-arctic sea ice thickness","cleaned_abstract":"in recent years, deep learning methods have gradually been applied to prediction tasks related to arctic sea ice concentration, but relatively little research has been conducted for larger spatial and temporal scales, mainly due to the limited time coverage of observations and reanalysis data. meanwhile, deep learning predictions of sea ice thickness (sit) have yet to receive ample attention. in this study, two data-driven deep learning (dl) models are built based on the convlstm and fully convolutional u-net (fc-unet) algorithms and trained using cmip6 historical simulations for transfer learning and fine-tuned using reanalysis\/observations. these models enable monthly predictions of arctic sit without considering the complex physical processes involved. through comprehensive assessments of prediction skills by season and region, the results suggest that using a broader set of cmip6 data for transfer learning, as well as incorporating multiple climate variables as predictors, contribute to better prediction results, although both dl models can effectively predict the spatiotemporal features of sit anomalies. regarding the predicted sit anomalies of the fc-unet model, the spatial correlations with reanalysis reach an average level of 89% over all months, while the temporal anomaly correlation coefficients are close to unity in most cases. the models also demonstrate robust performances in predicting sit and sie during extreme events. the effectiveness and reliability of the proposed deep transfer learning models in predicting arctic sit can facilitate more accurate pan-arctic predictions, aiding climate change research and real-time business applications.","key_phrases":["recent years","deep learning methods","prediction tasks","arctic sea ice concentration","relatively little research","larger spatial and temporal scales","the limited time coverage","observations","reanalysis data","deep learning predictions","sea ice thickness","ample attention","this study","two data-driven deep learning (dl) models","the convlstm","fully convolutional u","net","fc-unet","algorithms","cmip6 historical simulations","transfer learning","reanalysis\/observations","these models","monthly predictions","the complex physical processes","comprehensive assessments","prediction skills","season","region","the results","a broader set","cmip6 data","transfer learning","multiple climate variables","predictors","better prediction results","both dl models","the spatiotemporal features","sit anomalies","the predicted sit anomalies","the fc-unet model","the spatial correlations","reanalysis","an average level","89%","all months","the temporal anomaly correlation coefficients","unity","most cases","the models","robust performances","sit","sie","extreme events","the effectiveness","reliability","the proposed deep transfer learning models","arctic sit","more accurate pan-arctic predictions","climate change research","real-time business applications","recent years","arctic sea ice concentration","two","cmip6","monthly","arctic","season","cmip6","89%","all months","pan-arctic"]},{"title":"Early\u00a0betel\u00a0leaf\u00a0disease\u00a0detection using vision transformer and deep learning algorithms","authors":["S. Kusuma","K. R. Jothi"],"abstract":"Betel leaves benefits include their high content of antioxidants, which can help protect against oxidative stress and promote overall health. Additionally, these leaves are known for their potential anti-inflammatory properties, making them valuable in traditional medicine practises. The biggest threat to the security of the food supply is posed by plant diseases, and it is difficult to detect them early enough to prevent potential economic harm. This crop loss not only affects the economy but also poses a threat to food security, as betel leaves are widely used in traditional cuisines and herbal remedies. By analysing large datasets of plant images, deep learning algorithms can quickly identify specific patterns and symptoms associated with various diseases. In this study, we evaluated how well four deep learning models\u2014VGG19, DenseNet201, ResNet152V2, and a Vision Transform model\u2014performed at detecting diseases that affect betel leaves. Both the ResNet152V2 and VIT models attained levels of accuracy, with testing accuracies of 98.42% and 97.83%, respectively. However, the VGG19 model had slightly lower accuracy, with a testing accuracy of 91%. Overall, these deep learning models showed promising results in detecting diseases affecting betel leaves, with the DenseNet201 model performing the best with a testing accuracy of 98.77%.","doi":"10.1007\/s41870-023-01647-3","cleaned_title":"early betel leaf disease detection using vision transformer and deep learning algorithms","cleaned_abstract":"betel leaves benefits include their high content of antioxidants, which can help protect against oxidative stress and promote overall health. additionally, these leaves are known for their potential anti-inflammatory properties, making them valuable in traditional medicine practises. the biggest threat to the security of the food supply is posed by plant diseases, and it is difficult to detect them early enough to prevent potential economic harm. this crop loss not only affects the economy but also poses a threat to food security, as betel leaves are widely used in traditional cuisines and herbal remedies. by analysing large datasets of plant images, deep learning algorithms can quickly identify specific patterns and symptoms associated with various diseases. in this study, we evaluated how well four deep learning models\u2014vgg19, densenet201, resnet152v2, and a vision transform model\u2014performed at detecting diseases that affect betel leaves. both the resnet152v2 and vit models attained levels of accuracy, with testing accuracies of 98.42% and 97.83%, respectively. however, the vgg19 model had slightly lower accuracy, with a testing accuracy of 91%. overall, these deep learning models showed promising results in detecting diseases affecting betel leaves, with the densenet201 model performing the best with a testing accuracy of 98.77%.","key_phrases":["betel","benefits","their high content","antioxidants","which","oxidative stress","overall health","these leaves","their potential anti-inflammatory properties","them","traditional medicine practises","the biggest threat","the security","the food supply","plant diseases","it","them","potential economic harm","this crop loss","the economy","a threat","food security","betel leaves","traditional cuisines","herbal remedies","large datasets","plant images","deep learning algorithms","specific patterns","symptoms","various diseases","this study","we","how well four deep learning models","vgg19","a vision transform model","diseases","that","betel leaves","both the resnet152v2 and vit models","levels","accuracy","testing","accuracies","98.42%","97.83%","the vgg19 model","slightly lower accuracy","a testing accuracy","91%","these deep learning models","promising results","diseases","betel leaves","the densenet201 model","a testing accuracy","98.77%","betel","four","vit","98.42%","97.83%","91%","98.77%"]},{"title":"Exploring the deep learning of artificial intelligence in nursing: a concept analysis with Walker and Avant\u2019s approach","authors":["Supichaya Wangpitipanit","Lininger Jiraporn","Nick Anderson"],"abstract":"BackgroundIn recent years, increased attention has been given to using deep learning (DL) of artificial intelligence (AI) in healthcare to address nursing challenges. The adoption of new technologies in nursing needs to be improved, and AI in nursing is still in its early stages. However, the current literature needs more clarity, which affects clinical practice, research, and theory development. This study aimed to clarify the meaning of deep learning and identify the defining attributes of artificial intelligence within nursing.MethodsWe conducted a concept analysis of the deep learning of AI in nursing care using Walker and Avant\u2019s 8-step approach. Our search strategy employed Boolean techniques and MeSH terms across databases, including BMC, CINAHL, ClinicalKey for Nursing, Embase, Ovid, Scopus, SpringerLink and Spinger Nature, ProQuest, PubMed, and Web of Science. By focusing on relevant keywords in titles and abstracts from articles published between 2018 and 2024, we initially found 571 sources.ResultsThirty-seven articles that met the inclusion criteria were analyzed in this study. The attributes of evidence included four themes: focus and immersion, coding and understanding, arranging layers and algorithms, and implementing within the process of use cases to modify recommendations. Antecedents, unclear systems and communication, insufficient data management knowledge and support, and compound challenges can lead to suffering and risky caregiving tasks. Applying deep learning techniques enables nurses to simulate scenarios, predict outcomes, and plan care more precisely. Embracing deep learning equipment allows nurses to make better decisions. It empowers them with enhanced knowledge while ensuring adequate support and resources essential for caregiver and patient well-being. Access to necessary equipment is vital for high-quality home healthcare.ConclusionThis study provides a clearer understanding of the use of deep learning in nursing and its implications for nursing practice. Future research should focus on exploring the impact of deep learning on healthcare operations management through quantitative and qualitative studies. Additionally, developing a framework to guide the integration of deep learning into nursing practice is recommended to facilitate its adoption and implementation.","doi":"10.1186\/s12912-024-02170-x","cleaned_title":"exploring the deep learning of artificial intelligence in nursing: a concept analysis with walker and avant\u2019s approach","cleaned_abstract":"backgroundin recent years, increased attention has been given to using deep learning (dl) of artificial intelligence (ai) in healthcare to address nursing challenges. the adoption of new technologies in nursing needs to be improved, and ai in nursing is still in its early stages. however, the current literature needs more clarity, which affects clinical practice, research, and theory development. this study aimed to clarify the meaning of deep learning and identify the defining attributes of artificial intelligence within nursing.methodswe conducted a concept analysis of the deep learning of ai in nursing care using walker and avant\u2019s 8-step approach. our search strategy employed boolean techniques and mesh terms across databases, including bmc, cinahl, clinicalkey for nursing, embase, ovid, scopus, springerlink and spinger nature, proquest, pubmed, and web of science. by focusing on relevant keywords in titles and abstracts from articles published between 2018 and 2024, we initially found 571 sources.resultsthirty-seven articles that met the inclusion criteria were analyzed in this study. the attributes of evidence included four themes: focus and immersion, coding and understanding, arranging layers and algorithms, and implementing within the process of use cases to modify recommendations. antecedents, unclear systems and communication, insufficient data management knowledge and support, and compound challenges can lead to suffering and risky caregiving tasks. applying deep learning techniques enables nurses to simulate scenarios, predict outcomes, and plan care more precisely. embracing deep learning equipment allows nurses to make better decisions. it empowers them with enhanced knowledge while ensuring adequate support and resources essential for caregiver and patient well-being. access to necessary equipment is vital for high-quality home healthcare.conclusionthis study provides a clearer understanding of the use of deep learning in nursing and its implications for nursing practice. future research should focus on exploring the impact of deep learning on healthcare operations management through quantitative and qualitative studies. additionally, developing a framework to guide the integration of deep learning into nursing practice is recommended to facilitate its adoption and implementation.","key_phrases":["increased attention","deep learning","dl","artificial intelligence","healthcare","nursing challenges","the adoption","new technologies","nursing needs","nursing","its early stages","the current literature","more clarity","which","clinical practice","research","theory development","this study","the meaning","deep learning","the defining attributes","artificial intelligence","a concept analysis","the deep learning","ai","nursing care","walker and avant\u2019s 8-step approach","our search strategy","boolean techniques","mesh","terms","databases","bmc","cinahl","clinicalkey","nursing","embase","ovid","scopus","springerlink","spinger nature","web","science","relevant keywords","titles","abstracts","articles","we","571 sources.resultsthirty-seven articles","that","the inclusion criteria","this study","the attributes","evidence","four themes","immersion","understanding","layers","algorithms","the process","use cases","recommendations","antecedents","unclear systems","communication","insufficient data management knowledge","support","compound challenges","suffering and risky caregiving tasks","deep learning techniques","nurses","scenarios","outcomes","plan","deep learning equipment","nurses","better decisions","it","them","enhanced knowledge","adequate support","resources","caregiver","well-being","access","necessary equipment","high-quality home","healthcare.conclusionthis study","a clearer understanding","the use","deep learning","nursing","its implications","nursing practice","future research","the impact","deep learning","healthcare operations management","quantitative and qualitative studies","a framework","the integration","deep learning","nursing practice","its adoption","implementation","backgroundin recent years","8","between 2018 and 2024","571","four"]},{"title":"Automatic retinoblastoma screening and surveillance using deep learning","authors":["Ruiheng Zhang","Li Dong","Ruyue Li","Kai Zhang","Yitong Li","Hongshu Zhao","Jitong Shi","Xin Ge","Xiaolin Xu","Libin Jiang","Xuhan Shi","Chuan Zhang","Wenda Zhou","Liangyuan Xu","Haotian Wu","Heyan Li","Chuyao Yu","Jing Li","Jianmin Ma","Wenbin Wei"],"abstract":"BackgroundRetinoblastoma is the most common intraocular malignancy in childhood. With the advanced management strategy, the globe salvage and overall survival have significantly improved, which proposes subsequent challenges regarding long-term surveillance and offspring screening. This study aimed to apply a deep learning algorithm to reduce the burden of follow-up and offspring screening.MethodsThis cohort study includes retinoblastoma patients who visited Beijing Tongren Hospital from\u00a0March 2018 to January 2022 for deep learning algorism development. Clinical-suspected and treated retinoblastoma patients from February 2022 to June 2022 were prospectively collected for prospective validation. Images from the posterior pole and peripheral retina were collected, and reference standards were made according to the consensus of the multidisciplinary management team. A deep learning algorithm was trained to identify \u201cnormal fundus\u201d, \u201cstable retinoblastoma\u201d in which specific treatment is not required, and \u201cactive retinoblastoma\u201d in which specific treatment is required. The performance of each classifier included sensitivity, specificity, accuracy, and cost-utility.ResultsA total of 36,623 images were included for developing the Deep Learning Assistant for Retinoblastoma Monitoring (DLA-RB) algorithm. In internal fivefold cross-validation, DLA-RB achieved an area under curve (AUC) of 0.998 (95% confidence interval [CI] 0.986\u20131.000) in distinguishing normal fundus and active retinoblastoma, and 0.940 (95%\u00a0CI 0.851\u20130.996) in distinguishing stable and active retinoblastoma. From February 2022 to June 2022, 139 eyes of 103 patients were prospectively collected. In identifying active retinoblastoma tumours from all clinical-suspected patients and active retinoblastoma from all treated retinoblastoma patients, the AUC of DLA-RB reached 0.991 (95% CI 0.970\u20131.000), and 0.962 (95% CI 0.915\u20131.000), respectively. The combination between ophthalmologists and DLA-RB significantly improved the accuracy of competent ophthalmologists and residents regarding both binary tasks. Cost-utility analysis revealed DLA-RB-based diagnosis mode is cost-effective in both retinoblastoma diagnosis and active retinoblastoma identification.ConclusionsDLA-RB achieved high accuracy and sensitivity in identifying active retinoblastoma from the normal and stable retinoblastoma fundus. It can be used to surveil the activity of retinoblastoma during follow-up and screen high-risk offspring. Compared with referral procedures to ophthalmologic centres, DLA-RB-based screening and surveillance is cost-effective and can be incorporated within telemedicine programs.Clinical Trial RegistrationThis study was registered on ClinicalTrials.gov (NCT05308043).","doi":"10.1038\/s41416-023-02320-z","cleaned_title":"automatic retinoblastoma screening and surveillance using deep learning","cleaned_abstract":"backgroundretinoblastoma is the most common intraocular malignancy in childhood. with the advanced management strategy, the globe salvage and overall survival have significantly improved, which proposes subsequent challenges regarding long-term surveillance and offspring screening. this study aimed to apply a deep learning algorithm to reduce the burden of follow-up and offspring screening.methodsthis cohort study includes retinoblastoma patients who visited beijing tongren hospital from march 2018 to january 2022 for deep learning algorism development. clinical-suspected and treated retinoblastoma patients from february 2022 to june 2022 were prospectively collected for prospective validation. images from the posterior pole and peripheral retina were collected, and reference standards were made according to the consensus of the multidisciplinary management team. a deep learning algorithm was trained to identify \u201cnormal fundus\u201d, \u201cstable retinoblastoma\u201d in which specific treatment is not required, and \u201cactive retinoblastoma\u201d in which specific treatment is required. the performance of each classifier included sensitivity, specificity, accuracy, and cost-utility.resultsa total of 36,623 images were included for developing the deep learning assistant for retinoblastoma monitoring (dla-rb) algorithm. in internal fivefold cross-validation, dla-rb achieved an area under curve (auc) of 0.998 (95% confidence interval [ci] 0.986\u20131.000) in distinguishing normal fundus and active retinoblastoma, and 0.940 (95% ci 0.851\u20130.996) in distinguishing stable and active retinoblastoma. from february 2022 to june 2022, 139 eyes of 103 patients were prospectively collected. in identifying active retinoblastoma tumours from all clinical-suspected patients and active retinoblastoma from all treated retinoblastoma patients, the auc of dla-rb reached 0.991 (95% ci 0.970\u20131.000), and 0.962 (95% ci 0.915\u20131.000), respectively. the combination between ophthalmologists and dla-rb significantly improved the accuracy of competent ophthalmologists and residents regarding both binary tasks. cost-utility analysis revealed dla-rb-based diagnosis mode is cost-effective in both retinoblastoma diagnosis and active retinoblastoma identification.conclusionsdla-rb achieved high accuracy and sensitivity in identifying active retinoblastoma from the normal and stable retinoblastoma fundus. it can be used to surveil the activity of retinoblastoma during follow-up and screen high-risk offspring. compared with referral procedures to ophthalmologic centres, dla-rb-based screening and surveillance is cost-effective and can be incorporated within telemedicine programs.clinical trial registrationthis study was registered on clinicaltrials.gov (nct05308043).","key_phrases":["backgroundretinoblastoma","the most common intraocular malignancy","childhood","the advanced management strategy","the globe salvage","overall survival","which","subsequent challenges","long-term surveillance","offspring screening","this study","a deep learning algorithm","the burden","follow-up","screening.methodsthis cohort study","retinoblastoma patients","who","beijing tongren hospital","march","january","deep learning algorism development","clinical-suspected and treated retinoblastoma patients","february","june","prospective validation","images","the posterior pole","peripheral retina","reference standards","the consensus","the multidisciplinary management team","a deep learning algorithm","normal fundus\u201d, \u201cstable retinoblastoma","which","specific treatment","\u201cactive retinoblastoma","which","specific treatment","the performance","each classifier","sensitivity","specificity","accuracy","cost-utility.resultsa total","36,623 images","the deep learning assistant","retinoblastoma monitoring","dla-rb","-","dla-rb","an area","curve","auc","(95% confidence interval","normal fundus","active retinoblastoma","0.940 (95%","ci 0.851\u20130.996","stable and active retinoblastoma","february","june","139 eyes","103 patients","active retinoblastoma tumours","all clinical-suspected patients","active retinoblastoma","all treated retinoblastoma patients","the auc","dla-rb","0.991 (95%","ci 0.970\u20131.000","ci 0.915\u20131.000","the combination","ophthalmologists","dla-rb","the accuracy","competent ophthalmologists","residents","both binary tasks","cost-utility analysis","dla-rb-based diagnosis mode","both retinoblastoma diagnosis","active retinoblastoma","high accuracy","sensitivity","active retinoblastoma","the normal and stable retinoblastoma fundus","it","the activity","retinoblastoma","follow-up and screen high-risk offspring","referral procedures","ophthalmologic centres","dla-rb-based screening","surveillance","telemedicine programs.clinical trial registrationthis study","nct05308043","march 2018 to","january 2022","february 2022 to june 2022","36,623","0.998","95%","0.940","95%","0.851\u20130.996","february 2022 to june 2022","139","103","0.991","95%","0.962","95%","0.915\u20131.000"]},{"title":"OSS reliability assessment method based on deep learning and independent Wiener data preprocessing","authors":["Yoshinobu Tamura","Shoichiro Miyamoto","Lei Zhou","Adarsh Anand","P. K. Kapur","Shigeru Yamada"],"abstract":"The fault big data sets of many open source software (OSS) are recorded on the bug tracking systems. In the past, we have proposed the effort assessment method under the assumption that the fault detection phenomenon depends on the maintenance effort, because the number of software fault is influenced by the effort expenditure. The past research in terms of the effort assessment method of OSS is based on the effort data sets. On the other hand, we propose the deep learning approach to the OSS fault big data. In the past, the existing method without Wiener process cannot estimate within the range of existing data only. The proposed method assumes that the fault detection process follows the Wiener process such as the imperfect debugging and Markov property. Thereby, the proposed method can estimate the exceeding values by adding the white noise based on the Wiener process. Then, the proposed method make it possible for the OSS managers to assess the values exceeding from the existing data. Then, we show several reliability assessment measures based on the fault modification time based on the deep learning. Moreover, several numerical illustrations based on the proposed deep learning model are shown in this paper.","doi":"10.1007\/s13198-024-02288-w","cleaned_title":"oss reliability assessment method based on deep learning and independent wiener data preprocessing","cleaned_abstract":"the fault big data sets of many open source software (oss) are recorded on the bug tracking systems. in the past, we have proposed the effort assessment method under the assumption that the fault detection phenomenon depends on the maintenance effort, because the number of software fault is influenced by the effort expenditure. the past research in terms of the effort assessment method of oss is based on the effort data sets. on the other hand, we propose the deep learning approach to the oss fault big data. in the past, the existing method without wiener process cannot estimate within the range of existing data only. the proposed method assumes that the fault detection process follows the wiener process such as the imperfect debugging and markov property. thereby, the proposed method can estimate the exceeding values by adding the white noise based on the wiener process. then, the proposed method make it possible for the oss managers to assess the values exceeding from the existing data. then, we show several reliability assessment measures based on the fault modification time based on the deep learning. moreover, several numerical illustrations based on the proposed deep learning model are shown in this paper.","key_phrases":["the fault big data sets","many open source software","oss","the bug tracking systems","the past","we","the effort assessment method","the assumption","the fault detection phenomenon","the maintenance effort","the number","software fault","the effort expenditure","the past research","terms","the effort assessment method","oss","the effort data sets","the other hand","we","the deep learning approach","the oss","big data","the past","the existing method","wiener process","the range","existing data","the proposed method","the fault detection process","the wiener process","the imperfect debugging","markov property","the proposed method","the exceeding values","the white noise","the wiener process","the proposed method","it","the oss managers","the values","the existing data","we","several reliability assessment measures","the fault modification time","the deep learning","several numerical illustrations","the proposed deep learning model","this paper"]},{"title":"A comparative study: prediction of parkinson\u2019s disease using machine learning, deep learning and nature inspired algorithm","authors":["Pankaj Kumar Keserwani","Suman Das","Nairita Sarkar"],"abstract":"Parkinson\u2019s Disease (PD) is a degenerative and progressive neurological disorder worsens over time. This disease initially affects people over 55 years old. Patients with PD often exhibit a variety of non-motor and motor symptoms and are diagnosed based on those motor and non-motor symptoms as well as numerous clinical indicators. Advancement in medical science has produced medicines for many diseases but till now no significant remedies are discovered for Parkinson disease. It is very necessary to detect PD at early phase to take precautions accordingly to reduce its harmful impact and improve the patient\u2019s life style to a considerable level. In this direction Artificial Intelligence (AI) based approaches have recently attracted many researchers to work accordingly as AI can handle vast amounts of data and generate accurate statistical predictions. Addressing this imperative, researchers have turned their focus toward Artificial Intelligence (AI) as a promising avenue. AI\u2019s capacity to manage vast datasets and generate precise statistical predictions makes it an invaluable tool for PD detection. This article aims to provide a comprehensive survey and in-depth analysis of various AI-based approaches. Leveraging machine learning (ML), deep learning (DL), and meta-heuristic algorithms, these approaches contribute to the prediction of PD. Additionally, the article delves into current research directions. As the pursuit of advancements continues, the integration of AI holds promise in revolutionizing early detection methods and subsequently improving the lives of individuals grappling with Parkinson\u2019s disease.","doi":"10.1007\/s11042-024-18186-z","cleaned_title":"a comparative study: prediction of parkinson\u2019s disease using machine learning, deep learning and nature inspired algorithm","cleaned_abstract":"parkinson\u2019s disease (pd) is a degenerative and progressive neurological disorder worsens over time. this disease initially affects people over 55 years old. patients with pd often exhibit a variety of non-motor and motor symptoms and are diagnosed based on those motor and non-motor symptoms as well as numerous clinical indicators. advancement in medical science has produced medicines for many diseases but till now no significant remedies are discovered for parkinson disease. it is very necessary to detect pd at early phase to take precautions accordingly to reduce its harmful impact and improve the patient\u2019s life style to a considerable level. in this direction artificial intelligence (ai) based approaches have recently attracted many researchers to work accordingly as ai can handle vast amounts of data and generate accurate statistical predictions. addressing this imperative, researchers have turned their focus toward artificial intelligence (ai) as a promising avenue. ai\u2019s capacity to manage vast datasets and generate precise statistical predictions makes it an invaluable tool for pd detection. this article aims to provide a comprehensive survey and in-depth analysis of various ai-based approaches. leveraging machine learning (ml), deep learning (dl), and meta-heuristic algorithms, these approaches contribute to the prediction of pd. additionally, the article delves into current research directions. as the pursuit of advancements continues, the integration of ai holds promise in revolutionizing early detection methods and subsequently improving the lives of individuals grappling with parkinson\u2019s disease.","key_phrases":["parkinson\u2019s disease","pd","a degenerative and progressive neurological disorder","time","this disease","people","patients","pd","a variety","non-motor and motor symptoms","those motor and non-motor symptoms","numerous clinical indicators","advancement","medical science","medicines","many diseases","no significant remedies","parkinson disease","it","pd","early phase","precautions","its harmful impact","the patient\u2019s life style","a considerable level","this direction","artificial intelligence","ai) based approaches","many researchers","ai","vast amounts","data","accurate statistical predictions","this imperative","researchers","their focus","artificial intelligence","(ai","a promising avenue","ai\u2019s capacity","vast datasets","precise statistical predictions","it","pd detection","this article","a comprehensive survey","-depth","various ai-based approaches","machine learning","ml","deep learning","dl","meta-heuristic algorithms","these approaches","the prediction","pd","the article","current research directions","the pursuit","advancements","the integration","ai","promise","early detection methods","the lives","individuals","parkinson\u2019s disease","55 years old"]},{"title":"A hybrid approach to detecting Parkinson's disease using spectrogram and deep learning CNN-LSTM network","authors":["V. Shibina","T. M. Thasleema"],"abstract":"Parkinson\u2019s disease (PD) is a common illness that affects brain neurons. Medical practitioners and caregivers face challenges in detecting Parkinson's disease promptly, either in its early or late stages. There is an urgent need for non-invasive PD diagnostic technologies because timely diagnosis substantially impacts patient outcomes. This research aims to provide an efficient way of\u00a0identifying Parkinson's disease by transforming voice inputs into spectrograms using Short Term Fourier Transform and applying\u00a0deep learning algorithms. The identification of Parkinson's disease can be done by leveraging the deep learning architectures such as Convolutional Neural Networks and Long Short-Term Memory networks. The experiment produced positive findings, with 95.67% accuracy, 97.62% precision, 94.67% recall, and an F1-score of 95.91%. The outcomes indicate that the suggested deep learning method is more successful in PD identification, surpassing the results of traditional classification methods.","doi":"10.1007\/s10772-024-10128-2","cleaned_title":"a hybrid approach to detecting parkinson's disease using spectrogram and deep learning cnn-lstm network","cleaned_abstract":"parkinson\u2019s disease (pd) is a common illness that affects brain neurons. medical practitioners and caregivers face challenges in detecting parkinson's disease promptly, either in its early or late stages. there is an urgent need for non-invasive pd diagnostic technologies because timely diagnosis substantially impacts patient outcomes. this research aims to provide an efficient way of identifying parkinson's disease by transforming voice inputs into spectrograms using short term fourier transform and applying deep learning algorithms. the identification of parkinson's disease can be done by leveraging the deep learning architectures such as convolutional neural networks and long short-term memory networks. the experiment produced positive findings, with 95.67% accuracy, 97.62% precision, 94.67% recall, and an f1-score of 95.91%. the outcomes indicate that the suggested deep learning method is more successful in pd identification, surpassing the results of traditional classification methods.","key_phrases":["parkinson\u2019s disease","pd","a common illness","that","brain neurons","medical practitioners","caregivers","challenges","parkinson's disease","its early or late stages","an urgent need","non-invasive pd diagnostic technologies","timely diagnosis","patient outcomes","this research","an efficient way","parkinson's disease","voice inputs","spectrograms","short term fourier transform","deep learning algorithms","the identification","parkinson's disease","the deep learning architectures","convolutional neural networks","long short-term memory networks","the experiment","positive findings","95.67% accuracy","97.62% precision","94.67% recall","an f1-score","95.91%","the outcomes","the suggested deep learning method","pd identification","the results","traditional classification methods","95.67%","97.62%","94.67%","95.91%"]},{"title":"A comprehensive review of deep learning approaches for group activity analysis","authors":["Gang Zhang","Yang Geng","Zhao G. Gong"],"abstract":"The study of group activity analysis has garnered significant attention. Group activity offers a unique perspective on the relationships between individuals, providing insights that individual and crowd activities may not reveal. This paper aims to contribute to the existing body of knowledge by providing a comprehensive review of the methods employed in utilizing deep learning for the analysis of group activity. The review encompasses an overview of various methodologies for group detection, segmentation, feature extraction, and description. Additionally, it delves into the classification, recognition, and prediction of group activities, including crowd trajectory prediction. The representation of crowd activity patterns and labels, along with an exploration of datasets for crowd activity analysis, is also included. Ultimately, the paper concludes with a discussion of potential future research directions in the field. By offering a comprehensive review of the advancements in group activity analysis through the lens of deep learning, this paper aims to provide researchers with a better understanding of the field's progress, thereby contributing to the continued development of this area of study.","doi":"10.1007\/s00371-024-03479-z","cleaned_title":"a comprehensive review of deep learning approaches for group activity analysis","cleaned_abstract":"the study of group activity analysis has garnered significant attention. group activity offers a unique perspective on the relationships between individuals, providing insights that individual and crowd activities may not reveal. this paper aims to contribute to the existing body of knowledge by providing a comprehensive review of the methods employed in utilizing deep learning for the analysis of group activity. the review encompasses an overview of various methodologies for group detection, segmentation, feature extraction, and description. additionally, it delves into the classification, recognition, and prediction of group activities, including crowd trajectory prediction. the representation of crowd activity patterns and labels, along with an exploration of datasets for crowd activity analysis, is also included. ultimately, the paper concludes with a discussion of potential future research directions in the field. by offering a comprehensive review of the advancements in group activity analysis through the lens of deep learning, this paper aims to provide researchers with a better understanding of the field's progress, thereby contributing to the continued development of this area of study.","key_phrases":["the study","group activity analysis","significant attention","group activity","a unique perspective","the relationships","individuals","insights","that","individual and crowd activities","this paper","the existing body","knowledge","a comprehensive review","the methods","deep learning","the analysis","group activity","the review","an overview","various methodologies","group detection","segmentation","feature extraction","description","it","the classification","recognition","prediction","group activities","crowd","trajectory prediction","the representation","crowd activity patterns","labels","an exploration","datasets","crowd activity analysis","the paper","a discussion","potential future research directions","the field","a comprehensive review","the advancements","group activity analysis","the lens","deep learning","this paper","researchers","a better understanding","the field's progress","the continued development","this area","study"]},{"title":"Curriculum learning and evolutionary optimization into deep learning for text classification","authors":["Alfredo Arturo El\u00edas-Miranda","Daniel Vallejo-Aldana","Fernando S\u00e1nchez-Vega","A. Pastor L\u00f3pez-Monroy","Alejandro Rosales-P\u00e9rez","Victor Mu\u00f1iz-Sanchez"],"abstract":"The exponential growth of social networks has given rise to a wide variety of content. Some social content violates the integrity and dignity of users, therefore, this task has become challenging. The need to deal with short texts, poorly written language, unbalanced classes, and non-thematic aspects. These can lead to overfitting in deep neural network (DNN) models used for classification tasks. Empirical evidence in previous studies indicates that some of these problems can be overcome by improving the optimization process of the DNN weights to avoid overfitting. Moreover, a well-defined learning process in the input examples could improve the order of the patterns learned throughout the optimization process. In this paper, we propose four Curriculum Learning strategies and a new Hybrid Genetic\u2013Gradient Algorithm that proved to improve the performance of DNN models detecting the class of interest even in highly imbalanced datasets.","doi":"10.1007\/s00521-023-08632-8","cleaned_title":"curriculum learning and evolutionary optimization into deep learning for text classification","cleaned_abstract":"the exponential growth of social networks has given rise to a wide variety of content. some social content violates the integrity and dignity of users, therefore, this task has become challenging. the need to deal with short texts, poorly written language, unbalanced classes, and non-thematic aspects. these can lead to overfitting in deep neural network (dnn) models used for classification tasks. empirical evidence in previous studies indicates that some of these problems can be overcome by improving the optimization process of the dnn weights to avoid overfitting. moreover, a well-defined learning process in the input examples could improve the order of the patterns learned throughout the optimization process. in this paper, we propose four curriculum learning strategies and a new hybrid genetic\u2013gradient algorithm that proved to improve the performance of dnn models detecting the class of interest even in highly imbalanced datasets.","key_phrases":["the exponential growth","social networks","rise","a wide variety","content","some social content","the integrity","dignity","users","this task","the need","short texts","poorly written language","unbalanced classes","non-thematic aspects","these","deep neural network (dnn) models","classification tasks","empirical evidence","previous studies","some","these problems","the optimization process","the dnn weights","a well-defined learning process","the input examples","the order","the patterns","the optimization process","this paper","we","four curriculum","strategies","a new hybrid genetic\u2013gradient algorithm","that","the performance","dnn models","the class","interest","highly imbalanced datasets","four"]},{"title":"Deep learning model for detection of hotspots using infrared thermographic images of electrical installations","authors":["Ezechukwu Kalu Ukiwe","Steve A. Adeshina","Tsado Jacob","Bukola Babatunde Adetokun"],"abstract":"Hotspots in electrical power equipment or installations are a major issue whenever it occurs within the power system. Factors responsible for this phenomenon are many, sometimes inter-related and other times they are isolated. Electrical hotspots caused by poor connections are common. Deep learning models have become popular for diagnosing anomalies in physical and biological systems, by the instrumentality of feature extraction of images in convolutional neural networks. In this work, a VGG-16 deep neural network model is applied for identifying electrical hotspots by means of transfer learning. This model was achieved by first augmenting the acquired infrared thermographic images, using the pre-trained ImageNet weights of the VGG-16 algorithm with additional global average pooling in place of conventional fully connected layers and a softmax layer at the output. With the categorical cross-entropy loss function, the model was implemented using the Adam optimizer at learning rate of 0.0001 as well as some variants of the Adam optimization algorithm. On evaluation, with a test IRT image dataset, and a comparison with similar works, the research showed that a better accuracy of 99.98% in identification of electrical hotspots was achieved. The model shows good score in performance metrics like accuracy, precision, recall, and F1-score. The obtained results proved the potential of deep learning using computer vision parameters for infrared thermographic identification of electrical hotspots in power system installations. Also, there is need for careful selection of the IR sensor\u2019s thermal range during image acquisition, and suitable choice of color palette would make for easy hotspot isolation, reduce the pixel to pixel temperature differential across any of the images, and easily highlight the critical region of interest with high pixel values. However, it makes edge detection difficult for human visual perception which computer vision-based deep learning model could overcome.","doi":"10.1186\/s43067-024-00148-y","cleaned_title":"deep learning model for detection of hotspots using infrared thermographic images of electrical installations","cleaned_abstract":"hotspots in electrical power equipment or installations are a major issue whenever it occurs within the power system. factors responsible for this phenomenon are many, sometimes inter-related and other times they are isolated. electrical hotspots caused by poor connections are common. deep learning models have become popular for diagnosing anomalies in physical and biological systems, by the instrumentality of feature extraction of images in convolutional neural networks. in this work, a vgg-16 deep neural network model is applied for identifying electrical hotspots by means of transfer learning. this model was achieved by first augmenting the acquired infrared thermographic images, using the pre-trained imagenet weights of the vgg-16 algorithm with additional global average pooling in place of conventional fully connected layers and a softmax layer at the output. with the categorical cross-entropy loss function, the model was implemented using the adam optimizer at learning rate of 0.0001 as well as some variants of the adam optimization algorithm. on evaluation, with a test irt image dataset, and a comparison with similar works, the research showed that a better accuracy of 99.98% in identification of electrical hotspots was achieved. the model shows good score in performance metrics like accuracy, precision, recall, and f1-score. the obtained results proved the potential of deep learning using computer vision parameters for infrared thermographic identification of electrical hotspots in power system installations. also, there is need for careful selection of the ir sensor\u2019s thermal range during image acquisition, and suitable choice of color palette would make for easy hotspot isolation, reduce the pixel to pixel temperature differential across any of the images, and easily highlight the critical region of interest with high pixel values. however, it makes edge detection difficult for human visual perception which computer vision-based deep learning model could overcome.","key_phrases":["hotspots","electrical power equipment","installations","a major issue","it","the power system","factors","this phenomenon","many, sometimes inter-related and other times","they","electrical hotspots","poor connections","deep learning models","anomalies","physical and biological systems","the instrumentality","feature extraction","images","convolutional neural networks","this work","a vgg-16 deep neural network model","electrical hotspots","means","transfer learning","this model","the acquired infrared thermographic images","the pre-trained imagenet weights","the vgg-16 algorithm","additional global average pooling","place","conventional fully connected layers","a softmax layer","the output","the categorical cross-entropy loss function","the model","the adam optimizer","rate","some variants","the adam optimization algorithm","evaluation","a test irt image dataset","a comparison","similar works","the research","a better accuracy","99.98%","identification","electrical hotspots","the model","good score","performance metrics","accuracy","precision","recall","f1-score","the obtained results","the potential","deep learning","computer vision parameters","infrared thermographic identification","electrical hotspots","power system installations","need","careful selection","the ir sensor\u2019s thermal range","image acquisition","suitable choice","color palette","easy hotspot isolation","the pixel","temperature differential","any","the images","the critical region","interest","high pixel values","it","edge detection","human visual perception","which computer vision-based deep learning model","first","vgg-16","0.0001","99.98%"]},{"title":"Fire Hawks Optimizer with hybrid deep learning driven fall detection on multimodal sensor data","authors":["K. Durga Bhavani","M. Ferni Ukrit"],"abstract":"Falls are the main factor contributing to nonfatal and fatal injuries among older persons. Fall detection on sensor data uses different sensors to identify when an individual has fallen. The data from these sensors can be examined to determine if a fall event has occurred. This technology is highly capable of improving the well-being and safety of persons, particularly older adults, by automatically identifying falls and alerting emergency services or caregivers. Deep learning-based fall detection on sensor data is a significant area of research and development that mainly focuses on utilizing deep learning methods for identifying and predicting falls based on data gathered from different sensors. This study presents a novel approach for fall detection utilizing the Fire Hawks Optimizer with hybrid deep learning technique on multimodal sensor data. The proposed methodology integrates deep learning algorithms with the Fire Hawks Optimizer metaheuristic to enhance fall detection accuracy. The technique preprocesses input data using standard scaling and employs a convolutional-recurrent Hopfield neural network model for fall detection and classification. Hyperparameter tuning using the Fire Hawks Optimizer further improves detection outcomes. Experimental evaluation conducted on the KFall dataset demonstrates the effectiveness of the proposed approach. Key metrics including accuracy, precision, recall, F-score, and Matthews correlation coefficient are utilized for evaluation. Results indicate superior performance of the method compared to existing fall detection approaches, with accuracies exceeding 99% on both training and testing sets. Visual representations such as confusion matrices, precision-recall curves, and ROC curves further validate the method's robustness. The proposed model offers significant advancements in fall detection accuracy, making it a promising solution for ensuring the safety and well-being of individuals, particularly older adults.","doi":"10.1007\/s11042-024-19970-7","cleaned_title":"fire hawks optimizer with hybrid deep learning driven fall detection on multimodal sensor data","cleaned_abstract":"falls are the main factor contributing to nonfatal and fatal injuries among older persons. fall detection on sensor data uses different sensors to identify when an individual has fallen. the data from these sensors can be examined to determine if a fall event has occurred. this technology is highly capable of improving the well-being and safety of persons, particularly older adults, by automatically identifying falls and alerting emergency services or caregivers. deep learning-based fall detection on sensor data is a significant area of research and development that mainly focuses on utilizing deep learning methods for identifying and predicting falls based on data gathered from different sensors. this study presents a novel approach for fall detection utilizing the fire hawks optimizer with hybrid deep learning technique on multimodal sensor data. the proposed methodology integrates deep learning algorithms with the fire hawks optimizer metaheuristic to enhance fall detection accuracy. the technique preprocesses input data using standard scaling and employs a convolutional-recurrent hopfield neural network model for fall detection and classification. hyperparameter tuning using the fire hawks optimizer further improves detection outcomes. experimental evaluation conducted on the kfall dataset demonstrates the effectiveness of the proposed approach. key metrics including accuracy, precision, recall, f-score, and matthews correlation coefficient are utilized for evaluation. results indicate superior performance of the method compared to existing fall detection approaches, with accuracies exceeding 99% on both training and testing sets. visual representations such as confusion matrices, precision-recall curves, and roc curves further validate the method's robustness. the proposed model offers significant advancements in fall detection accuracy, making it a promising solution for ensuring the safety and well-being of individuals, particularly older adults.","key_phrases":["falls","the main factor","nonfatal and fatal injuries","older persons","fall detection","sensor data","different sensors","an individual","the data","these sensors","a fall event","this technology","the well-being","safety","persons","particularly older adults","falls","emergency services","caregivers","deep learning-based fall detection","sensor data","a significant area","research","development","that","deep learning methods","falls","data","different sensors","this study","a novel approach","fall detection","the fire hawks optimizer","hybrid deep learning technique","multimodal sensor data","the proposed methodology","algorithms","the fire hawks optimizer metaheuristic","fall detection accuracy","the technique preprocesses","standard scaling","a convolutional-recurrent hopfield neural network model","fall detection","classification","hyperparameter","the fire hawks optimizer","detection outcomes","experimental evaluation","the kfall dataset","the effectiveness","the proposed approach","key metrics","accuracy","precision","recall","f-score","matthews correlation coefficient","evaluation","results","superior performance","the method","existing fall detection approaches","accuracies","99%","both training and testing sets","visual representations","confusion matrices","precision-recall curves","the method's robustness","the proposed model","significant advancements","fall detection accuracy","it","the safety","well-being","individuals","particularly older adults","hopfield neural","99%","roc"]},{"title":"Anthropogenic fingerprints in daily precipitation revealed by deep learning","authors":["Yoo-Geun Ham","Jeong-Hwan Kim","Seung-Ki Min","Daehyun Kim","Tim Li","Axel Timmermann","Malte F. Stuecker"],"abstract":"According to twenty-first century climate-model projections, greenhouse warming will intensify rainfall variability and extremes across the globe1,2,3,4. However, verifying this prediction using observations has remained a substantial challenge owing to large natural rainfall fluctuations at regional scales3,4. Here we show that deep learning successfully detects the emerging climate-change signals in daily precipitation fields during the observed record. We trained a convolutional neural network (CNN)5 with daily precipitation fields and annual global mean surface air temperature data obtained from an ensemble of present-day and future climate-model simulations6. After applying the algorithm to the observational record, we found that the daily precipitation data represented an excellent predictor for the observed planetary warming, as they showed a clear deviation from natural variability since the mid-2010s. Furthermore, we analysed the deep-learning model with an explainable framework and observed that the precipitation variability of the weather timescale (period less than 10\u2009days) over the tropical eastern Pacific and mid-latitude storm-track regions was most sensitive to anthropogenic warming. Our results highlight that, although the long-term shifts in annual mean precipitation remain indiscernible from the natural background variability, the impact of global warming on daily hydrological fluctuations has already emerged.","doi":"10.1038\/s41586-023-06474-x","cleaned_title":"anthropogenic fingerprints in daily precipitation revealed by deep learning","cleaned_abstract":"according to twenty-first century climate-model projections, greenhouse warming will intensify rainfall variability and extremes across the globe1,2,3,4. however, verifying this prediction using observations has remained a substantial challenge owing to large natural rainfall fluctuations at regional scales3,4. here we show that deep learning successfully detects the emerging climate-change signals in daily precipitation fields during the observed record. we trained a convolutional neural network (cnn)5 with daily precipitation fields and annual global mean surface air temperature data obtained from an ensemble of present-day and future climate-model simulations6. after applying the algorithm to the observational record, we found that the daily precipitation data represented an excellent predictor for the observed planetary warming, as they showed a clear deviation from natural variability since the mid-2010s. furthermore, we analysed the deep-learning model with an explainable framework and observed that the precipitation variability of the weather timescale (period less than 10 days) over the tropical eastern pacific and mid-latitude storm-track regions was most sensitive to anthropogenic warming. our results highlight that, although the long-term shifts in annual mean precipitation remain indiscernible from the natural background variability, the impact of global warming on daily hydrological fluctuations has already emerged.","key_phrases":["twenty-first century climate-model projections","greenhouse warming","rainfall variability","extremes","the globe1,2,3,4","this prediction","observations","a substantial challenge","large natural rainfall fluctuations","regional scales3,4","we","deep learning","the emerging climate-change signals","daily precipitation fields","the observed record","we","a convolutional neural network","cnn)5","daily precipitation fields","annual global mean surface air temperature data","an ensemble","present-day and future climate-model simulations6","the algorithm","the observational record","we","the daily precipitation data","an excellent predictor","the observed planetary warming","they","a clear deviation","natural variability","we","the deep-learning model","an explainable framework","the precipitation variability","the weather timescale","period","the tropical eastern pacific and mid-latitude storm-track regions","anthropogenic warming","our results","the long-term shifts","annual mean precipitation","the natural background variability","the impact","global warming","daily hydrological fluctuations","twenty-first century","daily","cnn)5","daily","present-day","daily","the mid-2010s","less than 10 days","annual","daily"]},{"title":"Optimization of deep learning models: benchmark and analysis","authors":["Rasheed Ahmad","Izzat Alsmadi","Mohammad Al-Ramahi"],"abstract":"Model optimization in deep learning (DL) and neural networks is concerned about how and why the model can be successfully trained towards one or more objective functions. The evolutionary learning or training process continuously considers the dynamic parameters of the model. Many researchers propose a deep learning-based solution by randomly selecting a single classifier model architecture. Such approaches generally overlook the hidden and complex nature of the model\u2019s internal working, producing biased results. Larger and deeper NN models bring many complexities and logistic challenges while building and deploying them. To obtain high-quality performance results, an optimal model generally depends on the appropriate architectural settings, such as the number of hidden layers and the number of neurons at each layer. A challenging and time-consuming task is to select and test various combinations of these settings manually. This paper presents an extensive empirical analysis of various deep learning algorithms trained recursively using permutated settings to establish benchmarks and find an optimal model. The paper analyzed the Stack Overflow dataset to predict the quality of posted questions. The extensive empirical analysis revealed that some famous deep learning algorithms such as CNN are the least effective algorithm in solving this problem compared to multilayer perceptron (MLP), which provides efficient computing and the best results in terms of prediction accuracy. The analysis also shows that manipulating the number of neurons alone at each layer in a network does not influence model optimization. This paper\u2019s findings will help to recognize the fact that future models should be built by considering a vast range of model architectural settings for an optimal solution.","doi":"10.1007\/s43674-023-00055-1","cleaned_title":"optimization of deep learning models: benchmark and analysis","cleaned_abstract":"model optimization in deep learning (dl) and neural networks is concerned about how and why the model can be successfully trained towards one or more objective functions. the evolutionary learning or training process continuously considers the dynamic parameters of the model. many researchers propose a deep learning-based solution by randomly selecting a single classifier model architecture. such approaches generally overlook the hidden and complex nature of the model\u2019s internal working, producing biased results. larger and deeper nn models bring many complexities and logistic challenges while building and deploying them. to obtain high-quality performance results, an optimal model generally depends on the appropriate architectural settings, such as the number of hidden layers and the number of neurons at each layer. a challenging and time-consuming task is to select and test various combinations of these settings manually. this paper presents an extensive empirical analysis of various deep learning algorithms trained recursively using permutated settings to establish benchmarks and find an optimal model. the paper analyzed the stack overflow dataset to predict the quality of posted questions. the extensive empirical analysis revealed that some famous deep learning algorithms such as cnn are the least effective algorithm in solving this problem compared to multilayer perceptron (mlp), which provides efficient computing and the best results in terms of prediction accuracy. the analysis also shows that manipulating the number of neurons alone at each layer in a network does not influence model optimization. this paper\u2019s findings will help to recognize the fact that future models should be built by considering a vast range of model architectural settings for an optimal solution.","key_phrases":["model optimization","deep learning","dl","neural networks","the model","one or more objective functions","the evolutionary learning or training process","the dynamic parameters","the model","many researchers","a deep learning-based solution","a single classifier model architecture","such approaches","the hidden and complex nature","the model\u2019s internal working","biased results","models","many complexities","logistic challenges","them","high-quality performance results","an optimal model","the appropriate architectural settings","the number","hidden layers","the number","neurons","each layer","a challenging and time-consuming task","various combinations","these settings","this paper","an extensive empirical analysis","various deep learning algorithms","permutated settings","benchmarks","an optimal model","the paper","the stack overflow","the quality","posted questions","the extensive empirical analysis","some famous deep learning algorithms","cnn","the least effective algorithm","this problem","multilayer perceptron","mlp","which","efficient computing","the best results","terms","prediction accuracy","the analysis","the number","neurons","each layer","a network","model optimization","this paper\u2019s findings","the fact","future models","a vast range","model architectural settings","an optimal solution","one","cnn"]},{"title":"Classification and detection of natural disasters using machine learning and deep learning techniques: A review","authors":["Kibitok Abraham","Moataz Abdelwahab","Mohammed Abo-Zahhad"],"abstract":"For efficient disaster management, it is essential to identify and categorize natural disasters. The classical approaches and current technological advancements for identifying, categorizing, and reducing the harmful effects of natural catastrophes are discussed in this review article. They include human observation and reporting, satellite images, seismology, radar, infrared imagery, and sonar. The article explores natural disasters\u2019 challenges and harmful effects and their mitigation measures. The article explains the benefits and drawbacks of published approaches and emphasizes how they may be used to identify many kinds of natural catastrophes, including earthquakes, floods, wildfires, and hurricanes. Discussions on current technological advancements, including machine and deep learning applications, that can potentially increase the precision and efficiency of natural disaster detection and classification are presented. Overall, the review article emphasizes the significance of continuing research and improving current techniques to increase communities\u2019 and countries\u2019 resilience and preparedness for natural disasters. Moreover, future directions and suggestions to stakeholders in disaster management are highlighted.","doi":"10.1007\/s12145-023-01205-2","cleaned_title":"classification and detection of natural disasters using machine learning and deep learning techniques: a review","cleaned_abstract":"for efficient disaster management, it is essential to identify and categorize natural disasters. the classical approaches and current technological advancements for identifying, categorizing, and reducing the harmful effects of natural catastrophes are discussed in this review article. they include human observation and reporting, satellite images, seismology, radar, infrared imagery, and sonar. the article explores natural disasters\u2019 challenges and harmful effects and their mitigation measures. the article explains the benefits and drawbacks of published approaches and emphasizes how they may be used to identify many kinds of natural catastrophes, including earthquakes, floods, wildfires, and hurricanes. discussions on current technological advancements, including machine and deep learning applications, that can potentially increase the precision and efficiency of natural disaster detection and classification are presented. overall, the review article emphasizes the significance of continuing research and improving current techniques to increase communities\u2019 and countries\u2019 resilience and preparedness for natural disasters. moreover, future directions and suggestions to stakeholders in disaster management are highlighted.","key_phrases":["efficient disaster management","it","natural disasters","the classical approaches","current technological advancements","the harmful effects","natural catastrophes","this review article","they","human observation","reporting","satellite images","seismology","radar","infrared imagery","sonar","the article","natural disasters\u2019 challenges","harmful effects","their mitigation measures","the article","the benefits","drawbacks","published approaches","they","many kinds","natural catastrophes","earthquakes","floods","wildfires","hurricanes","discussions","current technological advancements","machine","deep learning applications","that","the precision","efficiency","natural disaster detection","classification","the review article","the significance","continuing research","current techniques","communities","countries\u2019 resilience","preparedness","natural disasters","future directions","suggestions","stakeholders","disaster management"]},{"title":"A comprehensive review of deep learning approaches for group activity analysis","authors":["Gang Zhang","Yang Geng","Zhao G. Gong"],"abstract":"The study of group activity analysis has garnered significant attention. Group activity offers a unique perspective on the relationships between individuals, providing insights that individual and crowd activities may not reveal. This paper aims to contribute to the existing body of knowledge by providing a comprehensive review of the methods employed in utilizing deep learning for the analysis of group activity. The review encompasses an overview of various methodologies for group detection, segmentation, feature extraction, and description. Additionally, it delves into the classification, recognition, and prediction of group activities, including crowd trajectory prediction. The representation of crowd activity patterns and labels, along with an exploration of datasets for crowd activity analysis, is also included. Ultimately, the paper concludes with a discussion of potential future research directions in the field. By offering a comprehensive review of the advancements in group activity analysis through the lens of deep learning, this paper aims to provide researchers with a better understanding of the field's progress, thereby contributing to the continued development of this area of study.","doi":"10.1007\/s00371-024-03479-z","cleaned_title":"a comprehensive review of deep learning approaches for group activity analysis","cleaned_abstract":"the study of group activity analysis has garnered significant attention. group activity offers a unique perspective on the relationships between individuals, providing insights that individual and crowd activities may not reveal. this paper aims to contribute to the existing body of knowledge by providing a comprehensive review of the methods employed in utilizing deep learning for the analysis of group activity. the review encompasses an overview of various methodologies for group detection, segmentation, feature extraction, and description. additionally, it delves into the classification, recognition, and prediction of group activities, including crowd trajectory prediction. the representation of crowd activity patterns and labels, along with an exploration of datasets for crowd activity analysis, is also included. ultimately, the paper concludes with a discussion of potential future research directions in the field. by offering a comprehensive review of the advancements in group activity analysis through the lens of deep learning, this paper aims to provide researchers with a better understanding of the field's progress, thereby contributing to the continued development of this area of study.","key_phrases":["the study","group activity analysis","significant attention","group activity","a unique perspective","the relationships","individuals","insights","that","individual and crowd activities","this paper","the existing body","knowledge","a comprehensive review","the methods","deep learning","the analysis","group activity","the review","an overview","various methodologies","group detection","segmentation","feature extraction","description","it","the classification","recognition","prediction","group activities","crowd","trajectory prediction","the representation","crowd activity patterns","labels","an exploration","datasets","crowd activity analysis","the paper","a discussion","potential future research directions","the field","a comprehensive review","the advancements","group activity analysis","the lens","deep learning","this paper","researchers","a better understanding","the field's progress","the continued development","this area","study"]},{"title":"Fire Hawks Optimizer with hybrid deep learning driven fall detection on multimodal sensor data","authors":["K. Durga Bhavani","M. Ferni Ukrit"],"abstract":"Falls are the main factor contributing to nonfatal and fatal injuries among older persons. Fall detection on sensor data uses different sensors to identify when an individual has fallen. The data from these sensors can be examined to determine if a fall event has occurred. This technology is highly capable of improving the well-being and safety of persons, particularly older adults, by automatically identifying falls and alerting emergency services or caregivers. Deep learning-based fall detection on sensor data is a significant area of research and development that mainly focuses on utilizing deep learning methods for identifying and predicting falls based on data gathered from different sensors. This study presents a novel approach for fall detection utilizing the Fire Hawks Optimizer with hybrid deep learning technique on multimodal sensor data. The proposed methodology integrates deep learning algorithms with the Fire Hawks Optimizer metaheuristic to enhance fall detection accuracy. The technique preprocesses input data using standard scaling and employs a convolutional-recurrent Hopfield neural network model for fall detection and classification. Hyperparameter tuning using the Fire Hawks Optimizer further improves detection outcomes. Experimental evaluation conducted on the KFall dataset demonstrates the effectiveness of the proposed approach. Key metrics including accuracy, precision, recall, F-score, and Matthews correlation coefficient are utilized for evaluation. Results indicate superior performance of the method compared to existing fall detection approaches, with accuracies exceeding 99% on both training and testing sets. Visual representations such as confusion matrices, precision-recall curves, and ROC curves further validate the method's robustness. The proposed model offers significant advancements in fall detection accuracy, making it a promising solution for ensuring the safety and well-being of individuals, particularly older adults.","doi":"10.1007\/s11042-024-19970-7","cleaned_title":"fire hawks optimizer with hybrid deep learning driven fall detection on multimodal sensor data","cleaned_abstract":"falls are the main factor contributing to nonfatal and fatal injuries among older persons. fall detection on sensor data uses different sensors to identify when an individual has fallen. the data from these sensors can be examined to determine if a fall event has occurred. this technology is highly capable of improving the well-being and safety of persons, particularly older adults, by automatically identifying falls and alerting emergency services or caregivers. deep learning-based fall detection on sensor data is a significant area of research and development that mainly focuses on utilizing deep learning methods for identifying and predicting falls based on data gathered from different sensors. this study presents a novel approach for fall detection utilizing the fire hawks optimizer with hybrid deep learning technique on multimodal sensor data. the proposed methodology integrates deep learning algorithms with the fire hawks optimizer metaheuristic to enhance fall detection accuracy. the technique preprocesses input data using standard scaling and employs a convolutional-recurrent hopfield neural network model for fall detection and classification. hyperparameter tuning using the fire hawks optimizer further improves detection outcomes. experimental evaluation conducted on the kfall dataset demonstrates the effectiveness of the proposed approach. key metrics including accuracy, precision, recall, f-score, and matthews correlation coefficient are utilized for evaluation. results indicate superior performance of the method compared to existing fall detection approaches, with accuracies exceeding 99% on both training and testing sets. visual representations such as confusion matrices, precision-recall curves, and roc curves further validate the method's robustness. the proposed model offers significant advancements in fall detection accuracy, making it a promising solution for ensuring the safety and well-being of individuals, particularly older adults.","key_phrases":["falls","the main factor","nonfatal and fatal injuries","older persons","fall detection","sensor data","different sensors","an individual","the data","these sensors","a fall event","this technology","the well-being","safety","persons","particularly older adults","falls","emergency services","caregivers","deep learning-based fall detection","sensor data","a significant area","research","development","that","deep learning methods","falls","data","different sensors","this study","a novel approach","fall detection","the fire hawks optimizer","hybrid deep learning technique","multimodal sensor data","the proposed methodology","algorithms","the fire hawks optimizer metaheuristic","fall detection accuracy","the technique preprocesses","standard scaling","a convolutional-recurrent hopfield neural network model","fall detection","classification","hyperparameter","the fire hawks optimizer","detection outcomes","experimental evaluation","the kfall dataset","the effectiveness","the proposed approach","key metrics","accuracy","precision","recall","f-score","matthews correlation coefficient","evaluation","results","superior performance","the method","existing fall detection approaches","accuracies","99%","both training and testing sets","visual representations","confusion matrices","precision-recall curves","the method's robustness","the proposed model","significant advancements","fall detection accuracy","it","the safety","well-being","individuals","particularly older adults","hopfield neural","99%","roc"]},{"title":"Optimization of deep learning models: benchmark and analysis","authors":["Rasheed Ahmad","Izzat Alsmadi","Mohammad Al-Ramahi"],"abstract":"Model optimization in deep learning (DL) and neural networks is concerned about how and why the model can be successfully trained towards one or more objective functions. The evolutionary learning or training process continuously considers the dynamic parameters of the model. Many researchers propose a deep learning-based solution by randomly selecting a single classifier model architecture. Such approaches generally overlook the hidden and complex nature of the model\u2019s internal working, producing biased results. Larger and deeper NN models bring many complexities and logistic challenges while building and deploying them. To obtain high-quality performance results, an optimal model generally depends on the appropriate architectural settings, such as the number of hidden layers and the number of neurons at each layer. A challenging and time-consuming task is to select and test various combinations of these settings manually. This paper presents an extensive empirical analysis of various deep learning algorithms trained recursively using permutated settings to establish benchmarks and find an optimal model. The paper analyzed the Stack Overflow dataset to predict the quality of posted questions. The extensive empirical analysis revealed that some famous deep learning algorithms such as CNN are the least effective algorithm in solving this problem compared to multilayer perceptron (MLP), which provides efficient computing and the best results in terms of prediction accuracy. The analysis also shows that manipulating the number of neurons alone at each layer in a network does not influence model optimization. This paper\u2019s findings will help to recognize the fact that future models should be built by considering a vast range of model architectural settings for an optimal solution.","doi":"10.1007\/s43674-023-00055-1","cleaned_title":"optimization of deep learning models: benchmark and analysis","cleaned_abstract":"model optimization in deep learning (dl) and neural networks is concerned about how and why the model can be successfully trained towards one or more objective functions. the evolutionary learning or training process continuously considers the dynamic parameters of the model. many researchers propose a deep learning-based solution by randomly selecting a single classifier model architecture. such approaches generally overlook the hidden and complex nature of the model\u2019s internal working, producing biased results. larger and deeper nn models bring many complexities and logistic challenges while building and deploying them. to obtain high-quality performance results, an optimal model generally depends on the appropriate architectural settings, such as the number of hidden layers and the number of neurons at each layer. a challenging and time-consuming task is to select and test various combinations of these settings manually. this paper presents an extensive empirical analysis of various deep learning algorithms trained recursively using permutated settings to establish benchmarks and find an optimal model. the paper analyzed the stack overflow dataset to predict the quality of posted questions. the extensive empirical analysis revealed that some famous deep learning algorithms such as cnn are the least effective algorithm in solving this problem compared to multilayer perceptron (mlp), which provides efficient computing and the best results in terms of prediction accuracy. the analysis also shows that manipulating the number of neurons alone at each layer in a network does not influence model optimization. this paper\u2019s findings will help to recognize the fact that future models should be built by considering a vast range of model architectural settings for an optimal solution.","key_phrases":["model optimization","deep learning","dl","neural networks","the model","one or more objective functions","the evolutionary learning or training process","the dynamic parameters","the model","many researchers","a deep learning-based solution","a single classifier model architecture","such approaches","the hidden and complex nature","the model\u2019s internal working","biased results","models","many complexities","logistic challenges","them","high-quality performance results","an optimal model","the appropriate architectural settings","the number","hidden layers","the number","neurons","each layer","a challenging and time-consuming task","various combinations","these settings","this paper","an extensive empirical analysis","various deep learning algorithms","permutated settings","benchmarks","an optimal model","the paper","the stack overflow","the quality","posted questions","the extensive empirical analysis","some famous deep learning algorithms","cnn","the least effective algorithm","this problem","multilayer perceptron","mlp","which","efficient computing","the best results","terms","prediction accuracy","the analysis","the number","neurons","each layer","a network","model optimization","this paper\u2019s findings","the fact","future models","a vast range","model architectural settings","an optimal solution","one","cnn"]},{"title":"Literature survey on deep learning methods for liver segmentation from CT images: a comprehensive review","authors":["Kumar S. S.","Vinod Kumar R. S."],"abstract":"Segmentation of the liver from computed tomography images is an essential and critical task in medical image analysis, with significant implications for liver disease diagnosis and treatment. Deep learning techniques have emerged as a powerful tool in this domain, offering unprecedented accuracy and robustness. This literature survey paper provides a comprehensive overview of deep learning techniques for segmentation of liver from CT images, aiming to synthesize recent advancements, identify key contributions, and address challenges in this rapidly evolving field. The survey covers various deep learning architectures, including convolution neural networks, U-Net, attention mechanisms, generative adversarial neural networks and transformer models, highlighting their strengths and weaknesses. Evaluation metrics and benchmark datasets commonly used for performance assessment are discussed in this survey. Furthermore, the survey delves into the challenges and limitations of deep learning methods, including interpretability, model robustness, and ethical considerations. The survey concludes by summarizing key findings, highlighting advancements, and outlining future research directions, such as interpretable models, ethical considerations, and bridging the gap between research and clinical implementation. This literature survey serves as a valuable reference for researchers, healthcare professionals, and developers in their pursuit of accurate liver segmentation and advancing medical image analysis.","doi":"10.1007\/s11042-024-18388-5","cleaned_title":"literature survey on deep learning methods for liver segmentation from ct images: a comprehensive review","cleaned_abstract":"segmentation of the liver from computed tomography images is an essential and critical task in medical image analysis, with significant implications for liver disease diagnosis and treatment. deep learning techniques have emerged as a powerful tool in this domain, offering unprecedented accuracy and robustness. this literature survey paper provides a comprehensive overview of deep learning techniques for segmentation of liver from ct images, aiming to synthesize recent advancements, identify key contributions, and address challenges in this rapidly evolving field. the survey covers various deep learning architectures, including convolution neural networks, u-net, attention mechanisms, generative adversarial neural networks and transformer models, highlighting their strengths and weaknesses. evaluation metrics and benchmark datasets commonly used for performance assessment are discussed in this survey. furthermore, the survey delves into the challenges and limitations of deep learning methods, including interpretability, model robustness, and ethical considerations. the survey concludes by summarizing key findings, highlighting advancements, and outlining future research directions, such as interpretable models, ethical considerations, and bridging the gap between research and clinical implementation. this literature survey serves as a valuable reference for researchers, healthcare professionals, and developers in their pursuit of accurate liver segmentation and advancing medical image analysis.","key_phrases":["segmentation","the liver","computed tomography images","an essential and critical task","medical image analysis","significant implications","liver disease diagnosis","treatment","deep learning techniques","a powerful tool","this domain","unprecedented accuracy","robustness","this literature survey paper","a comprehensive overview","deep learning techniques","segmentation","liver","ct images","recent advancements","key contributions","address challenges","this rapidly evolving field","the survey","various deep learning architectures","convolution neural networks","u","net","attention mechanisms","generative adversarial neural networks","transformer models","their strengths","weaknesses","evaluation metrics","benchmark datasets","performance assessment","this survey","the survey","the challenges","limitations","deep learning methods","interpretability","model robustness","ethical considerations","the survey","key findings","advancements","future research directions","interpretable models","ethical considerations","the gap","research and clinical implementation","this literature survey","a valuable reference","researchers","healthcare professionals","developers","their pursuit","accurate liver segmentation","medical image analysis"]},{"title":"Evaluating the impact of reinforcement learning on automatic deep brain stimulation planning","authors":["Anja Pantovic","Caroline Essert"],"abstract":"PurposeTraditional techniques for automating the planning of brain electrode placement based on multi-objective optimization involving many parameters are subject to limitations, especially in terms of sensitivity to local optima, and tend to be replaced by machine learning approaches. This paper explores the feasibility of using deep reinforcement learning (DRL) in this context, starting with the single-electrode use-case of deep brain stimulation (DBS).MethodsWe propose a DRL approach based on deep Q-learning where the states represent the electrode trajectory and associated information, and actions are the possible motions. Deep neural networks allow to navigate the complex state space derived from MRI data. The chosen reward function emphasizes safety and accuracy in reaching the target structure. The results were compared with a reference (segmented electrode) and a conventional technique.ResultsThe DRL approach excelled in navigating the complex anatomy, consistently providing safer and more precise electrode placements than the reference. Compared to conventional techniques, it showed an improvement in accuracy of 2.3% in average proximity to obstacles and 19.4% in average orientation angle. Expectedly, computation times rose significantly, from 2 to 18\u00a0min.ConclusionOur investigation into DRL for DBS electrode trajectory planning has showcased its promising potential. Despite only delivering modest accuracy gains compared to traditional methods in the single-electrode case, its relevance for problems with high-dimensional state and action spaces and its resilience against local optima highlight its promising role for complex scenarios. This preliminary study constitutes a first step toward the more challenging problem of multiple-electrodes planning.","doi":"10.1007\/s11548-024-03078-2","cleaned_title":"evaluating the impact of reinforcement learning on automatic deep brain stimulation planning","cleaned_abstract":"purposetraditional techniques for automating the planning of brain electrode placement based on multi-objective optimization involving many parameters are subject to limitations, especially in terms of sensitivity to local optima, and tend to be replaced by machine learning approaches. this paper explores the feasibility of using deep reinforcement learning (drl) in this context, starting with the single-electrode use-case of deep brain stimulation (dbs).methodswe propose a drl approach based on deep q-learning where the states represent the electrode trajectory and associated information, and actions are the possible motions. deep neural networks allow to navigate the complex state space derived from mri data. the chosen reward function emphasizes safety and accuracy in reaching the target structure. the results were compared with a reference (segmented electrode) and a conventional technique.resultsthe drl approach excelled in navigating the complex anatomy, consistently providing safer and more precise electrode placements than the reference. compared to conventional techniques, it showed an improvement in accuracy of 2.3% in average proximity to obstacles and 19.4% in average orientation angle. expectedly, computation times rose significantly, from 2 to 18 min.conclusionour investigation into drl for dbs electrode trajectory planning has showcased its promising potential. despite only delivering modest accuracy gains compared to traditional methods in the single-electrode case, its relevance for problems with high-dimensional state and action spaces and its resilience against local optima highlight its promising role for complex scenarios. this preliminary study constitutes a first step toward the more challenging problem of multiple-electrodes planning.","key_phrases":["purposetraditional techniques","the planning","brain electrode placement","multi-objective optimization","many parameters","limitations","terms","sensitivity","local optima","machine learning approaches","this paper","the feasibility","deep reinforcement learning","drl","this context","the single-electrode use-case","deep brain stimulation","dbs).methodswe","a drl approach","deep q-learning","the states","the electrode trajectory","associated information","actions","the possible motions","deep neural networks","the complex state space","mri data","the chosen reward function","safety","accuracy","the target structure","the results","a reference","a conventional technique.resultsthe drl approach","the complex anatomy","safer and more precise electrode placements","the reference","conventional techniques","it","an improvement","accuracy","2.3%","average proximity","obstacles","19.4%","average orientation angle","computation times","2 to 18 min.conclusionour investigation","drl","dbs electrode trajectory planning","its promising potential","modest accuracy gains","traditional methods","the single-electrode case","its relevance","problems","high-dimensional state and action spaces","its resilience","local optima","its promising role","complex scenarios","this preliminary study","a first step","the more challenging problem","multiple-electrodes planning","the chosen reward function","2.3%","19.4%","2","first"]},{"title":"Deep learning for determining the difficulty of endodontic treatment: a pilot study","authors":["Hamed Karkehabadi","Elham Khoshbin","Nikoo Ghasemi","Amal Mahavi","Hossein Mohammad-Rahimi","Soroush Sadr"],"abstract":"BackgroundTo develop and validate a deep learning model for automated assessment of endodontic case difficulty from periapical radiographs.MethodsA dataset of 1,386 periapical radiographs was compiled from two clinical sites. Two dentists and two endodontists annotated the radiographs for difficulty using the \u201csimple assessment\u201d criteria from the American Association of Endodontists\u2019 case difficulty assessment form in the Endocase application. A classification task labeled cases as \u201ceasy\u201d or \u201chard\u201d, while regression predicted overall difficulty scores. Convolutional neural networks (i.e. VGG16, ResNet18, ResNet50, ResNext50, and Inception v2) were used, with a baseline model trained via transfer learning from ImageNet weights. Other models was pre-trained using self-supervised contrastive learning (i.e. BYOL, SimCLR, MoCo, and DINO) on 20,295 unlabeled dental radiographs to learn representation without manual labels. Both models were evaluated using 10-fold cross-validation, with performance compared to seven human examiners (three general dentists and four endodontists) on a hold-out test set.ResultsThe baseline VGG16 model attained 87.62% accuracy in classifying difficulty. Self-supervised pretraining did not improve performance. Regression predicted scores with \u00b1\u20093.21 score error. All models outperformed human raters, with poor inter-examiner reliability.ConclusionThis pilot study demonstrated the feasibility of automated endodontic difficulty assessment via deep learning models.","doi":"10.1186\/s12903-024-04235-4","cleaned_title":"deep learning for determining the difficulty of endodontic treatment: a pilot study","cleaned_abstract":"backgroundto develop and validate a deep learning model for automated assessment of endodontic case difficulty from periapical radiographs.methodsa dataset of 1,386 periapical radiographs was compiled from two clinical sites. two dentists and two endodontists annotated the radiographs for difficulty using the \u201csimple assessment\u201d criteria from the american association of endodontists\u2019 case difficulty assessment form in the endocase application. a classification task labeled cases as \u201ceasy\u201d or \u201chard\u201d, while regression predicted overall difficulty scores. convolutional neural networks (i.e. vgg16, resnet18, resnet50, resnext50, and inception v2) were used, with a baseline model trained via transfer learning from imagenet weights. other models was pre-trained using self-supervised contrastive learning (i.e. byol, simclr, moco, and dino) on 20,295 unlabeled dental radiographs to learn representation without manual labels. both models were evaluated using 10-fold cross-validation, with performance compared to seven human examiners (three general dentists and four endodontists) on a hold-out test set.resultsthe baseline vgg16 model attained 87.62% accuracy in classifying difficulty. self-supervised pretraining did not improve performance. regression predicted scores with \u00b1 3.21 score error. all models outperformed human raters, with poor inter-examiner reliability.conclusionthis pilot study demonstrated the feasibility of automated endodontic difficulty assessment via deep learning models.","key_phrases":["backgroundto","a deep learning model","automated assessment","endodontic case difficulty","periapical radiographs.methodsa dataset","1,386 periapical radiographs","two clinical sites","two dentists","two endodontists","the radiographs","difficulty","the \u201csimple assessment\u201d criteria","the american association","endodontists\u2019 case difficulty assessment form","the endocase application","a classification task","cases","regression","overall difficulty scores","convolutional neural networks","i.e. vgg16","resnet18","resnet50","resnext50","inception v2","a baseline model","imagenet weights","other models","self-supervised contrastive learning","i.e. byol","moco","dino","20,295 unlabeled dental radiographs","representation","manual labels","both models","10-fold cross","-","validation","performance","seven human examiners","three general dentists","four endodontists","a hold-out test set.resultsthe baseline vgg16 model","87.62% accuracy","difficulty","self-supervised pretraining","performance","regression","scores","\u00b1 3.21 score error","all models","human raters","poor inter-examiner reliability.conclusionthis pilot study","the feasibility","automated endodontic difficulty assessment","deep learning models","backgroundto","radiographs.methodsa","1,386","two","two","two","the american association of endodontists\u2019","resnet18","resnet50","resnext50","20,295","10-fold","seven","three","four","87.62%","3.21"]},{"title":"Application Level Resource Scheduling for Deep Learning Acceleration on MPSoC","authors":["Cong Gao","Sangeet Saha","Xuqi Zhu","Hongyuan Jing","Klaus D. McDonald-Maier","Xiaojun Zhai"],"abstract":"Deep Neutral Networks (DNNs) have been widely used in many applications, such as self-driving cars, natural language processing (NLP), image classification, visual object recognition, and so on. Field-programmable gate array (FPGA) based Multiprocessor System on a Chip (MPSoC) is recently considered one of the popular choices for deploying DNN models. However, the limited resource capacity of MPSoC imposes a challenge for such practical implementation. Recent studies revealed the trade-off between the \u201cresources consumed\" vs. the \u201cperformance achieved\". Taking a cue from these findings, we address the problem of efficient implementation of deep learning into the resource-constrained MPSoC in this paper, where each deep learning network is run with different service levels based on resource usage (where a higher service level implies higher performance with increased resource consumption). To this end, we propose a heuristic-based strategy, Application Wise Level Selector (AWLS), for selecting service levels to maximize the overall performance subject to a given resource bound. AWLS can achieve higher performance within a constrained resource budget under various simulation scenarios. Further, we verify the proposed strategy using an AMD-Xilinx Zynq UltraScale+ XCZU9EG SoC. Using a framework designed to deploy multi-DNN on multi-DPUs (Deep Learning Units), it is proved that an optimal solution is achieved from the algorithm, which obtains the highest performance (Frames Per Second) using the same resource budget.","doi":"10.1007\/s11265-023-01881-9","cleaned_title":"application level resource scheduling for deep learning acceleration on mpsoc","cleaned_abstract":"deep neutral networks (dnns) have been widely used in many applications, such as self-driving cars, natural language processing (nlp), image classification, visual object recognition, and so on. field-programmable gate array (fpga) based multiprocessor system on a chip (mpsoc) is recently considered one of the popular choices for deploying dnn models. however, the limited resource capacity of mpsoc imposes a challenge for such practical implementation. recent studies revealed the trade-off between the \u201cresources consumed\" vs. the \u201cperformance achieved\". taking a cue from these findings, we address the problem of efficient implementation of deep learning into the resource-constrained mpsoc in this paper, where each deep learning network is run with different service levels based on resource usage (where a higher service level implies higher performance with increased resource consumption). to this end, we propose a heuristic-based strategy, application wise level selector (awls), for selecting service levels to maximize the overall performance subject to a given resource bound. awls can achieve higher performance within a constrained resource budget under various simulation scenarios. further, we verify the proposed strategy using an amd-xilinx zynq ultrascale+ xczu9eg soc. using a framework designed to deploy multi-dnn on multi-dpus (deep learning units), it is proved that an optimal solution is achieved from the algorithm, which obtains the highest performance (frames per second) using the same resource budget.","key_phrases":["deep neutral networks","dnns","many applications","self-driving cars","natural language processing","nlp","image classification","visual object recognition","field-programmable gate array","fpga","a chip (mpsoc","the popular choices","dnn models","the limited resource capacity","mpsoc","a challenge","such practical implementation","recent studies","the trade-off","the \u201cresources","the \u201cperformance","a cue","these findings","we","the problem","efficient implementation","deep learning","the resource-constrained mpsoc","this paper","each deep learning network","different service levels","resource usage","a higher service level","higher performance","increased resource consumption","this end","we","a heuristic-based strategy","application wise level selector","(awls","service levels","the overall performance","a given resource","awls","higher performance","a constrained resource budget","various simulation scenarios","we","the proposed strategy","an amd-xilinx zynq","xczu9eg soc","a framework","-","dnn","multi-dpus (deep learning units","it","an optimal solution","the algorithm","which","the highest performance","frames","the same resource budget","xczu9eg","second"]},{"title":"Radiomic and deep learning analysis of dermoscopic images for skin lesion pattern decoding","authors":["Zheng Wang","Chong Wang","Li Peng","Kaibin Lin","Yang Xue","Xiao Chen","Linlin Bao","Chao Liu","Jianglin Zhang","Yang Xie"],"abstract":"This study aims to explore the efficacy of a hybrid deep learning and radiomics approach, supplemented with patient metadata, in the noninvasive dermoscopic imaging-based diagnosis of skin lesions. We analyzed dermoscopic images from the International Skin Imaging Collaboration (ISIC) dataset, spanning 2016\u20132020, encompassing a variety of skin lesions. Our approach integrates deep learning with a comprehensive radiomics analysis, utilizing a vast array of quantitative image features to precisely quantify skin lesion patterns. The dataset includes cases of three, four, and eight different skin lesion types. Our methodology was benchmarked against seven classification methods from the ISIC 2020 challenge and prior research using a binary decision framework. The proposed hybrid model demonstrated superior performance in distinguishing benign from malignant lesions, achieving area under the receiver operating characteristic curve (AUROC) scores of 99%, 95%, and 96%, and multiclass decoding AUROCs of 98.5%, 94.9%, and 96.4%, with sensitivities of 97.6%, 93.9%, and 96.0% and specificities of 98.4%, 96.7%, and 96.9% in the internal ISIC 2018 challenge, as well as in the external Jinan and Longhua datasets, respectively. Our findings suggest that the integration of radiomics and deep learning, utilizing dermoscopic images, effectively captures the heterogeneity and pattern expression of skin lesions.","doi":"10.1038\/s41598-024-70231-x","cleaned_title":"radiomic and deep learning analysis of dermoscopic images for skin lesion pattern decoding","cleaned_abstract":"this study aims to explore the efficacy of a hybrid deep learning and radiomics approach, supplemented with patient metadata, in the noninvasive dermoscopic imaging-based diagnosis of skin lesions. we analyzed dermoscopic images from the international skin imaging collaboration (isic) dataset, spanning 2016\u20132020, encompassing a variety of skin lesions. our approach integrates deep learning with a comprehensive radiomics analysis, utilizing a vast array of quantitative image features to precisely quantify skin lesion patterns. the dataset includes cases of three, four, and eight different skin lesion types. our methodology was benchmarked against seven classification methods from the isic 2020 challenge and prior research using a binary decision framework. the proposed hybrid model demonstrated superior performance in distinguishing benign from malignant lesions, achieving area under the receiver operating characteristic curve (auroc) scores of 99%, 95%, and 96%, and multiclass decoding aurocs of 98.5%, 94.9%, and 96.4%, with sensitivities of 97.6%, 93.9%, and 96.0% and specificities of 98.4%, 96.7%, and 96.9% in the internal isic 2018 challenge, as well as in the external jinan and longhua datasets, respectively. our findings suggest that the integration of radiomics and deep learning, utilizing dermoscopic images, effectively captures the heterogeneity and pattern expression of skin lesions.","key_phrases":["this study","the efficacy","a hybrid deep learning and radiomics approach","patient metadata","the noninvasive dermoscopic imaging-based diagnosis","skin lesions","we","dermoscopic images","the international skin imaging collaboration","isic) dataset","a variety","skin lesions","our approach","a comprehensive radiomics analysis","a vast array","quantitative image features","skin lesion patterns","the dataset","cases","eight different skin lesion types","our methodology","seven classification methods","the isic 2020 challenge","prior research","a binary decision framework","the proposed hybrid model","superior performance","malignant lesions","area","the receiver operating characteristic curve","auroc) scores","99%","95%","96%","multiclass decoding aurocs","98.5%","94.9%","96.4%","sensitivities","97.6%","93.9%","96.0%","specificities","98.4%","96.7%","96.9%","the internal isic 2018 challenge","the external jinan","longhua datasets","our findings","the integration","radiomics","deep learning","dermoscopic images","the heterogeneity and pattern expression","skin lesions","metadata","2016\u20132020","three","four","eight","seven","2020","99%","95%","96%","98.5%","94.9%","96.4%","97.6%","93.9%","96.0%","98.4%","96.7%","96.9%","2018","longhua datasets"]},{"title":"Fusing deep learning features for parameter identification of a stochastic airfoil system","authors":["Jing Feng","Xiaolong Wang","Qi Liu","Yong Xu","J\u00fcrgen Kurths"],"abstract":"This work proposes a data-driven parameter identification approach for a two-degree-of-freedom airfoil system with cubic nonlinearity and stochasticity, where the random turbulent flow is quantified by non-Gaussian L\u00e9vy colored noise. The joint identification of the parameters controlling the flow velocity, airfoil geometry and structural stiffness is shaped as a unified machine learning task that includes three stages. (1) The first stage extracts local deep learning features from measurement data. (2) Next, the local features are fused to construct fixed-length global features representing the whole sample trajectory. (3) The global features are mapped to the parameter estimates and the accuracy indicators for uncertainty quantification. The numerical studies show that the obtained parameter estimation neural network can identify the system parameters from a sample trajectory with partially observed state measurements, namely, system parameters can be fully identified if only one or two of the pitch and plunge degrees of freedom are available. The intermediate deep features extracted by the PENN are compact representations of the stochastic system, as they carry key information of the system parameters. Suitable rules for information fusion are further designed, adapting the PENN to identify the system parameters from multiple short trajectories or time-varying parameters from a sample trajectory. The results suggest that the proposed deep learning approach is a flexible and versatile computation device for information extraction and fusion from limited data of stochastic nonlinear systems.","doi":"10.1007\/s11071-024-10152-6","cleaned_title":"fusing deep learning features for parameter identification of a stochastic airfoil system","cleaned_abstract":"this work proposes a data-driven parameter identification approach for a two-degree-of-freedom airfoil system with cubic nonlinearity and stochasticity, where the random turbulent flow is quantified by non-gaussian l\u00e9vy colored noise. the joint identification of the parameters controlling the flow velocity, airfoil geometry and structural stiffness is shaped as a unified machine learning task that includes three stages. (1) the first stage extracts local deep learning features from measurement data. (2) next, the local features are fused to construct fixed-length global features representing the whole sample trajectory. (3) the global features are mapped to the parameter estimates and the accuracy indicators for uncertainty quantification. the numerical studies show that the obtained parameter estimation neural network can identify the system parameters from a sample trajectory with partially observed state measurements, namely, system parameters can be fully identified if only one or two of the pitch and plunge degrees of freedom are available. the intermediate deep features extracted by the penn are compact representations of the stochastic system, as they carry key information of the system parameters. suitable rules for information fusion are further designed, adapting the penn to identify the system parameters from multiple short trajectories or time-varying parameters from a sample trajectory. the results suggest that the proposed deep learning approach is a flexible and versatile computation device for information extraction and fusion from limited data of stochastic nonlinear systems.","key_phrases":["this work","a data-driven parameter identification approach","freedom","cubic nonlinearity","stochasticity","the random turbulent flow","non-gaussian l\u00e9vy colored noise","the joint identification","the parameters","the flow velocity","airfoil geometry","structural stiffness","a unified machine learning task","that","three stages","the first stage","local deep learning features","measurement data","the local features","fixed-length global features","the whole sample trajectory","the global features","the parameter estimates","the accuracy indicators","uncertainty quantification","the numerical studies","the obtained parameter estimation neural network","the system parameters","a sample trajectory","partially observed state measurements","system parameters","the pitch and plunge degrees","freedom","the intermediate deep features","the penn","compact representations","the stochastic system","they","key information","the system parameters","suitable rules","information fusion","the penn","the system parameters","multiple short trajectories","time-varying parameters","a sample trajectory","the results","the proposed deep learning approach","a flexible and versatile computation device","information extraction","fusion","limited data","stochastic nonlinear systems","two-degree","non-gaussian","three","1","first","2","3","only one","two"]},{"title":"Biologically informed deep learning for explainable epigenetic clocks","authors":["Aurel Prosz","Orsolya Pipek","Judit B\u00f6rcs\u00f6k","Gergely Palla","Zoltan Szallasi","Sandor Spisak","Istv\u00e1n Csabai"],"abstract":"Ageing is often characterised by progressive accumulation of damage, and it is one of the most important risk factors for chronic disease development. Epigenetic mechanisms including DNA methylation could functionally contribute to organismal aging, however the key functions and biological processes may govern ageing are still not understood. Although age predictors called epigenetic clocks can accurately estimate the biological age of an individual based on cellular DNA methylation, their models have limited ability to explain the prediction algorithm behind and underlying key biological processes controlling ageing. Here we present XAI-AGE, a biologically informed, explainable deep neural network model for accurate biological age prediction across multiple tissue types. We show that XAI-AGE outperforms the first-generation age predictors and achieves similar results to deep learning-based models, while opening up the possibility to infer biologically meaningful insights of the activity of pathways and other abstract biological processes directly from the model.","doi":"10.1038\/s41598-023-50495-5","cleaned_title":"biologically informed deep learning for explainable epigenetic clocks","cleaned_abstract":"ageing is often characterised by progressive accumulation of damage, and it is one of the most important risk factors for chronic disease development. epigenetic mechanisms including dna methylation could functionally contribute to organismal aging, however the key functions and biological processes may govern ageing are still not understood. although age predictors called epigenetic clocks can accurately estimate the biological age of an individual based on cellular dna methylation, their models have limited ability to explain the prediction algorithm behind and underlying key biological processes controlling ageing. here we present xai-age, a biologically informed, explainable deep neural network model for accurate biological age prediction across multiple tissue types. we show that xai-age outperforms the first-generation age predictors and achieves similar results to deep learning-based models, while opening up the possibility to infer biologically meaningful insights of the activity of pathways and other abstract biological processes directly from the model.","key_phrases":["ageing","progressive accumulation","damage","it","the most important risk factors","chronic disease development","epigenetic mechanisms","dna methylation","organismal aging","the key functions","biological processes","ageing","age predictors","epigenetic clocks","the biological age","an individual","cellular dna methylation","their models","limited ability","the prediction algorithm","key biological processes","ageing","we","xai-age","a biologically informed, explainable deep neural network model","accurate biological age prediction","multiple tissue types","we","xai-age","the first-generation age predictors","similar results","deep learning-based models","the possibility","biologically meaningful insights","the activity","pathways","other abstract biological processes","the model","first"]},{"title":"Remote Cardiac System Monitoring Using 6G-IoT Communication and Deep Learning","authors":["Abdulbasid S. Banga","Mohammed M. Alenazi","Nisreen Innab","Mansor Alohali","Fahad M. Alhomayani","Mohammad H. Algarni","Taoufik Saidani"],"abstract":"Remote patient monitoring has recently been popularised due to advanced technological innovations. The advent of the Sixth Generation Internet of Things (6G-IoT) communication technology, combined with deep learning algorithms, presents a groundbreaking opportunity for enhancing remote cardiac system monitoring. This paper proposes an innovative framework leveraging the ultra-reliable, low-latency communication capabilities of 6G-IoT to transmit real-time cardiac data from wearable devices directly to healthcare providers. Integrating deep learning models facilitates the accurate analysis and prediction of cardiac anomalies, significantly improving traditional monitoring systems. Our methodology involves the deployment of cutting-edge wearable sensors capable of capturing high-fidelity cardiac signals. These signals are transmitted via 6G-IoT networks, ensuring minimal delay and maximum reliability. Upon receiving the data, a densely connected deep neural network with an optimised swish activation function\u2014designed explicitly for cardiac anomaly detection is employed to analyse the data in real-time. These algorithms are trained on vast datasets to recognise patterns indicative of potential cardiac issues, allowing immediate intervention when necessary. The proposed system\u2019s efficacy is validated through extensive testing in simulated environments, demonstrating its ability to accurately detect and swiftly predict a wide range of cardiac conditions. Moreover, implementing 6G-IoT communication ensures the system's scalability and adaptability to future technological advancements.","doi":"10.1007\/s11277-024-11217-w","cleaned_title":"remote cardiac system monitoring using 6g-iot communication and deep learning","cleaned_abstract":"remote patient monitoring has recently been popularised due to advanced technological innovations. the advent of the sixth generation internet of things (6g-iot) communication technology, combined with deep learning algorithms, presents a groundbreaking opportunity for enhancing remote cardiac system monitoring. this paper proposes an innovative framework leveraging the ultra-reliable, low-latency communication capabilities of 6g-iot to transmit real-time cardiac data from wearable devices directly to healthcare providers. integrating deep learning models facilitates the accurate analysis and prediction of cardiac anomalies, significantly improving traditional monitoring systems. our methodology involves the deployment of cutting-edge wearable sensors capable of capturing high-fidelity cardiac signals. these signals are transmitted via 6g-iot networks, ensuring minimal delay and maximum reliability. upon receiving the data, a densely connected deep neural network with an optimised swish activation function\u2014designed explicitly for cardiac anomaly detection is employed to analyse the data in real-time. these algorithms are trained on vast datasets to recognise patterns indicative of potential cardiac issues, allowing immediate intervention when necessary. the proposed system\u2019s efficacy is validated through extensive testing in simulated environments, demonstrating its ability to accurately detect and swiftly predict a wide range of cardiac conditions. moreover, implementing 6g-iot communication ensures the system's scalability and adaptability to future technological advancements.","key_phrases":["remote patient monitoring","advanced technological innovations","the advent","the sixth generation internet","things","6g-iot) communication technology","deep learning algorithms","a groundbreaking opportunity","remote cardiac system monitoring","this paper","an innovative framework","the ultra-reliable, low-latency communication capabilities","6g-iot","real-time cardiac data","wearable devices","providers","deep learning models","the accurate analysis","prediction","cardiac anomalies","traditional monitoring systems","our methodology","the deployment","cutting-edge wearable sensors","high-fidelity cardiac signals","these signals","6g-iot networks","minimal delay","maximum reliability","the data","a densely connected deep neural network","an optimised swish activation function","cardiac anomaly detection","the data","real-time","these algorithms","vast datasets","patterns","potential cardiac issues","immediate intervention","the proposed system\u2019s efficacy","extensive testing","simulated environments","its ability","a wide range","cardiac conditions","6g-iot communication","the system's scalability","adaptability","future technological advancements","sixth","6g-iot","6g-iot","6g","6g-iot"]},{"title":"A deep learning approach for host-based cryptojacking malware detection","authors":["Olanrewaju Sanda","Michalis Pavlidis","Nikolaos Polatidis"],"abstract":"With the continued growth and popularity of blockchain-based cryptocurrencies there is a parallel growth in illegal mining to earn cryptocurrency. Since mining for cryptocurrencies requires high computational resource; malicious actors have resorted to using malicious file downloads and other methods to illegally use a victim\u2019s system to mine for cryptocurrency without them knowing. This process is known as host-based cryptojacking and is gradually becoming one of the most popular cyberthreats in recent years. There are some proposed traditional machine learning methods to detect host-based cryptojacking but only a few have proposed using deep-learning models for detection. This paper presents a novel approach, dubbed CryptoJackingModel. This approach is a deep-learning host-based cryptojacking detection model that will effectively detect evolving host-based cryptojacking techniques and reduce false positives and false negatives. The approach has an overall accuracy of 98% on a dataset of 129,380 samples and a low performance overhead making it highly scalable. This approach will be an improvement of current countermeasures for detecting, mitigating, and preventing cryptojacking.","doi":"10.1007\/s12530-023-09534-9","cleaned_title":"a deep learning approach for host-based cryptojacking malware detection","cleaned_abstract":"with the continued growth and popularity of blockchain-based cryptocurrencies there is a parallel growth in illegal mining to earn cryptocurrency. since mining for cryptocurrencies requires high computational resource; malicious actors have resorted to using malicious file downloads and other methods to illegally use a victim\u2019s system to mine for cryptocurrency without them knowing. this process is known as host-based cryptojacking and is gradually becoming one of the most popular cyberthreats in recent years. there are some proposed traditional machine learning methods to detect host-based cryptojacking but only a few have proposed using deep-learning models for detection. this paper presents a novel approach, dubbed cryptojackingmodel. this approach is a deep-learning host-based cryptojacking detection model that will effectively detect evolving host-based cryptojacking techniques and reduce false positives and false negatives. the approach has an overall accuracy of 98% on a dataset of 129,380 samples and a low performance overhead making it highly scalable. this approach will be an improvement of current countermeasures for detecting, mitigating, and preventing cryptojacking.","key_phrases":["the continued growth","popularity","blockchain-based cryptocurrencies","a parallel growth","illegal mining","cryptocurrency","mining","cryptocurrencies","high computational resource","malicious actors","malicious file downloads","other methods","a victim\u2019s system","mine","cryptocurrency","them","this process","host-based cryptojacking","the most popular cyberthreats","recent years","some proposed traditional machine learning methods","host-based cryptojacking","deep-learning models","detection","this paper","a novel approach","cryptojackingmodel","this approach","a deep-learning host-based cryptojacking detection model","that","evolving host-based cryptojacking techniques","false positives","false negatives","the approach","an overall accuracy","98%","a dataset","129,380 samples","a low performance","it","this approach","an improvement","current countermeasures","mitigating","cryptojacking","recent years","98%","129,380"]},{"title":"Pneumonia detection based on RSNA dataset and anchor-free deep learning detector","authors":["Linghua Wu","Jing Zhang","Yilin Wang","Rong Ding","Yueqin Cao","Guiqin Liu","Changsheng Liufu","Baowei Xie","Shanping Kang","Rui Liu","Wenle Li","Furen Guan"],"abstract":"Pneumonia is a highly lethal disease, and research on its treatment and early screening tools has received extensive attention from researchers. Due to the maturity and cost reduction of chest X-ray technology, and with the development of artificial intelligence technology, pneumonia identification based on deep learning and chest X-ray has attracted attention from all over the world. Although the feature extraction capability of deep learning is strong, existing deep learning object detection frameworks are based on pre-defined anchors, which require a lot of tuning and experience to guarantee their excellent results in the face of new applications or data. To avoid the influence of anchor settings in pneumonia detection, this paper proposes an anchor-free object detection framework and RSNA dataset based on pneumonia detection. First, a data enhancement scheme is used to preprocess the chest X-ray images; second, an anchor-free object detection framework is used for pneumonia detection, which contains a feature pyramid, two-branch detection head, and focal loss. The average precision of 51.5 obtained by Intersection over Union (IoU) calculation shows that the pneumonia detection results obtained in this paper can surpass the existing classical object detection framework, providing an idea for future research and exploration.","doi":"10.1038\/s41598-024-52156-7","cleaned_title":"pneumonia detection based on rsna dataset and anchor-free deep learning detector","cleaned_abstract":"pneumonia is a highly lethal disease, and research on its treatment and early screening tools has received extensive attention from researchers. due to the maturity and cost reduction of chest x-ray technology, and with the development of artificial intelligence technology, pneumonia identification based on deep learning and chest x-ray has attracted attention from all over the world. although the feature extraction capability of deep learning is strong, existing deep learning object detection frameworks are based on pre-defined anchors, which require a lot of tuning and experience to guarantee their excellent results in the face of new applications or data. to avoid the influence of anchor settings in pneumonia detection, this paper proposes an anchor-free object detection framework and rsna dataset based on pneumonia detection. first, a data enhancement scheme is used to preprocess the chest x-ray images; second, an anchor-free object detection framework is used for pneumonia detection, which contains a feature pyramid, two-branch detection head, and focal loss. the average precision of 51.5 obtained by intersection over union (iou) calculation shows that the pneumonia detection results obtained in this paper can surpass the existing classical object detection framework, providing an idea for future research and exploration.","key_phrases":["pneumonia","a highly lethal disease","research","its treatment","early screening tools","extensive attention","researchers","the maturity and cost reduction","chest x-ray technology","the development","artificial intelligence technology","pneumonia identification","deep learning","chest x","-","ray","attention","the world","the feature extraction capability","deep learning","existing deep learning object detection frameworks","pre-defined anchors","which","a lot","tuning","their excellent results","the face","new applications","data","the influence","anchor settings","pneumonia detection","this paper","an anchor-free object detection framework","rsna dataset","pneumonia detection","a data enhancement scheme","the chest x-ray images","an anchor-free object detection framework","pneumonia detection","which","a feature pyramid","two-branch detection head","focal loss","the average precision","intersection","union (iou) calculation","the pneumonia detection results","this paper","the existing classical object detection framework","an idea","future research","exploration","first","second","two","51.5"]},{"title":"Comparative Study for Optimized Deep Learning-Based Road Accidents Severity Prediction Models","authors":["Hussam Hijazi","Karim Sattar","Hassan M. Al-Ahmadi","Sami El-Ferik"],"abstract":"Road traffic accidents remain a major cause of fatalities and injuries worldwide. Effective classification of accident type and severity is crucial for prompt post-accident protocols and the development of comprehensive road safety policies. This study explores the application of deep learning techniques for predicting crash injury severity in the Eastern Province of Saudi Arabia. Five deep learning models were trained and evaluated, including various variants of feedforward multilayer perceptron, a back-propagated artificial neural network (ANN), an ANN with radial basis function (RPF), and tabular data learning network (TabNet). The models were optimized using Bayesian optimization (BO) and employed the synthetic minority oversampling technique (SMOTE) for oversampling the training dataset. While SMOTE enhanced balanced accuracy for ANN with RBF and TabNet, it compromised precision and increased recall. The results indicated that oversampling techniques did not consistently improve model performance. Additionally, significant features were identified using least absolute shrinkage and selection operator (LASSO) regularization, feature importance, and permutation importance. The results indicated that oversampling techniques did not consistently improve model performance. While SMOTE enhanced balanced accuracy for ANN with RBF and TabNet, it compromised precision and increased recall. The study's findings emphasize the consistent significance of the 'Number of Injuries Major' feature as a vital predictor in deep learning models, regardless of the selection techniques employed. These results shed light on the pivotal role played by the count of individuals with major injuries in influencing the severity of crash injuries, highlighting its potential relevance in shaping road safety policy development.","doi":"10.1007\/s13369-023-08510-4","cleaned_title":"comparative study for optimized deep learning-based road accidents severity prediction models","cleaned_abstract":"road traffic accidents remain a major cause of fatalities and injuries worldwide. effective classification of accident type and severity is crucial for prompt post-accident protocols and the development of comprehensive road safety policies. this study explores the application of deep learning techniques for predicting crash injury severity in the eastern province of saudi arabia. five deep learning models were trained and evaluated, including various variants of feedforward multilayer perceptron, a back-propagated artificial neural network (ann), an ann with radial basis function (rpf), and tabular data learning network (tabnet). the models were optimized using bayesian optimization (bo) and employed the synthetic minority oversampling technique (smote) for oversampling the training dataset. while smote enhanced balanced accuracy for ann with rbf and tabnet, it compromised precision and increased recall. the results indicated that oversampling techniques did not consistently improve model performance. additionally, significant features were identified using least absolute shrinkage and selection operator (lasso) regularization, feature importance, and permutation importance. the results indicated that oversampling techniques did not consistently improve model performance. while smote enhanced balanced accuracy for ann with rbf and tabnet, it compromised precision and increased recall. the study's findings emphasize the consistent significance of the 'number of injuries major' feature as a vital predictor in deep learning models, regardless of the selection techniques employed. these results shed light on the pivotal role played by the count of individuals with major injuries in influencing the severity of crash injuries, highlighting its potential relevance in shaping road safety policy development.","key_phrases":["road traffic accidents","a major cause","fatalities","injuries","effective classification","accident type","severity","prompt post-accident protocols","the development","comprehensive road safety policies","this study","the application","deep learning techniques","crash injury severity","the eastern province","saudi arabia","five deep learning models","various variants","feedforward multilayer perceptron","a back-propagated artificial neural network","ann","radial basis function","rpf","data learning network","tabnet","the models","bayesian optimization","bo","the synthetic minority oversampling technique","smote","the training dataset","smote","balanced accuracy","ann","rbf","tabnet","it","precision","increased recall","the results","techniques","model performance","significant features","lasso","feature importance","permutation importance","the results","techniques","model performance","smote","balanced accuracy","ann","rbf","tabnet","it","precision","increased recall","the study's findings","the consistent significance","the 'number","injuries major' feature","a vital predictor","deep learning models","the selection techniques","these results","light","the pivotal role","the count","individuals","major injuries","the severity","crash injuries","its potential relevance","road safety policy development","saudi arabia","five"]},{"title":"Blockchain-based multi-diagnosis deep learning application for various diseases classification","authors":["Hakima Rym Rahal","Sihem Slatnia","Okba Kazar","Ezedin Barka","Saad Harous"],"abstract":"Misdiagnosis is a critical issue in healthcare, which can lead to severe consequences for patients, including delayed or inappropriate treatment, unnecessary procedures, psychological distress, financial burden, and legal implications. To mitigate this issue, we propose using deep learning algorithms to improve diagnostic accuracy. However, building accurate deep learning models for medical diagnosis requires substantial amounts of high-quality data, which can be challenging for individual healthcare sectors or organizations to acquire. Therefore, combining data from multiple sources to create a diverse dataset for efficient training is needed. However, sharing medical data between different healthcare sectors can be problematic from a security standpoint due to sensitive information and privacy laws. To address these challenges, we propose using blockchain technology to provide a secure, decentralized, and privacy-respecting way to share locally trained deep learning models instead of the data itself. Our proposed method of model ensembling, which combines the weights of several local deep learning models to build a single global model, that enables accurate diagnosis of complex medical conditions across multiple locations while preserving patient privacy and data security. Our research demonstrates the effectiveness of this approach in accurately diagnosing three diseases (breast cancer, lung cancer, and diabetes) with high accuracy rates, surpassing the accuracy of local models and building a multi-diagnosis application.","doi":"10.1007\/s10207-023-00733-8","cleaned_title":"blockchain-based multi-diagnosis deep learning application for various diseases classification","cleaned_abstract":"misdiagnosis is a critical issue in healthcare, which can lead to severe consequences for patients, including delayed or inappropriate treatment, unnecessary procedures, psychological distress, financial burden, and legal implications. to mitigate this issue, we propose using deep learning algorithms to improve diagnostic accuracy. however, building accurate deep learning models for medical diagnosis requires substantial amounts of high-quality data, which can be challenging for individual healthcare sectors or organizations to acquire. therefore, combining data from multiple sources to create a diverse dataset for efficient training is needed. however, sharing medical data between different healthcare sectors can be problematic from a security standpoint due to sensitive information and privacy laws. to address these challenges, we propose using blockchain technology to provide a secure, decentralized, and privacy-respecting way to share locally trained deep learning models instead of the data itself. our proposed method of model ensembling, which combines the weights of several local deep learning models to build a single global model, that enables accurate diagnosis of complex medical conditions across multiple locations while preserving patient privacy and data security. our research demonstrates the effectiveness of this approach in accurately diagnosing three diseases (breast cancer, lung cancer, and diabetes) with high accuracy rates, surpassing the accuracy of local models and building a multi-diagnosis application.","key_phrases":["misdiagnosis","a critical issue","healthcare","which","severe consequences","patients","delayed or inappropriate treatment","unnecessary procedures","psychological distress","financial burden","legal implications","this issue","we","deep learning algorithms","diagnostic accuracy","accurate deep learning models","medical diagnosis","substantial amounts","high-quality data","which","individual healthcare sectors","organizations","data","multiple sources","a diverse dataset","efficient training","medical data","different healthcare sectors","a security standpoint","sensitive information","privacy laws","these challenges","we","blockchain technology","a secure, decentralized, and privacy-respecting way","locally trained deep learning models","the data","itself","our proposed method","model ensembling","which","the weights","several local deep learning models","a single global model","that","accurate diagnosis","complex medical conditions","multiple locations","patient privacy","data security","our research","the effectiveness","this approach","three diseases","breast cancer","lung cancer","high accuracy rates","the accuracy","local models","a multi-diagnosis application","three"]},{"title":"Deep learning-based question answering: a survey","authors":["Heba Abdel-Nabi","Arafat Awajan","Mostafa Z. Ali"],"abstract":"Question Answering is a crucial natural language processing task. This field of research has attracted a sudden amount of interest lately due mainly to the integration of the deep learning models in the Question Answering Systems which consequently power up many advancements and improvements. This survey aims to explore and shed light upon the recent and most powerful deep learning-based Question Answering Systems and classify them based on the deep learning model used, stating the details of the used word representation, datasets, and evaluation metrics. It aims to highlight and discuss the currently used models and give insights that direct future research to enhance this increasingly growing field.","doi":"10.1007\/s10115-022-01783-5","cleaned_title":"deep learning-based question answering: a survey","cleaned_abstract":"question answering is a crucial natural language processing task. this field of research has attracted a sudden amount of interest lately due mainly to the integration of the deep learning models in the question answering systems which consequently power up many advancements and improvements. this survey aims to explore and shed light upon the recent and most powerful deep learning-based question answering systems and classify them based on the deep learning model used, stating the details of the used word representation, datasets, and evaluation metrics. it aims to highlight and discuss the currently used models and give insights that direct future research to enhance this increasingly growing field.","key_phrases":["a crucial natural language processing task","this field","research","a sudden amount","interest","the integration","the deep learning models","the question","systems","which","many advancements","improvements","this survey","light","the recent and most powerful deep learning-based question","systems","them","the deep learning model","the details","the used word representation","datasets","evaluation metrics","it","the currently used models","insights","that","future research","this increasingly growing field"]},{"title":"Deep Learning Architecture for Computer Vision-based Structural Defect Detection","authors":["Ruoyu Yang","Shubhendu Kumar Singh","Mostafa Tavakkoli","M. Amin Karami","Rahul Rai"],"abstract":"Structural health monitoring (SHM) refers to the implementation of a damage detection strategy for structures. Fault occurrence in these structural systems during the operation is inevitable. Efficient, fast, and precise health monitoring methods are required to proactively perform the necessary repairs and maintenance on time before it is too late. The current structural health monitoring methods involve physically attached sensors or non-contact vision-based vibration measurements. However, these methods have significant drawbacks due to the low spatial resolution, weight influence on the lightweight structure, and time\/labor consumption. Recently, computer-vison-based deep learning methods like convolutional neural network (CNN) and fully convolutional neural network (FCN) have been applied for defect detection and localization, which address the aforementioned problems and obtain high accuracy. This paper proposes a novel hybrid deep learning architecture comprising CNN and temporal convolutional networks (CNN-TCN) for the computer vision-based defect detection task. Various beam samples, consisting of five different materials and various structural defects, were used to evaluate the proposed deep learning algorithms\u2019 performance. The proposed deep learning methods treat each pixel of the video frame like a sensor to extract valuable features for defect detection. Through empirical results, we demonstrate that this \u2019pixel-sensor\u2019 approach is more efficient and accurate and can achieve a better defect detection performance on different beam samples compared with the current state-of-the-art approaches, including CNN-long short-term memory (LSTM), CNN-bidirectional long short-term memory (BiLSTM), multi-scale CNN-LSTM, and CNN-gated recurrent unit(GRU) methods.","doi":"10.1007\/s10489-023-04654-w","cleaned_title":"deep learning architecture for computer vision-based structural defect detection","cleaned_abstract":"structural health monitoring (shm) refers to the implementation of a damage detection strategy for structures. fault occurrence in these structural systems during the operation is inevitable. efficient, fast, and precise health monitoring methods are required to proactively perform the necessary repairs and maintenance on time before it is too late. the current structural health monitoring methods involve physically attached sensors or non-contact vision-based vibration measurements. however, these methods have significant drawbacks due to the low spatial resolution, weight influence on the lightweight structure, and time\/labor consumption. recently, computer-vison-based deep learning methods like convolutional neural network (cnn) and fully convolutional neural network (fcn) have been applied for defect detection and localization, which address the aforementioned problems and obtain high accuracy. this paper proposes a novel hybrid deep learning architecture comprising cnn and temporal convolutional networks (cnn-tcn) for the computer vision-based defect detection task. various beam samples, consisting of five different materials and various structural defects, were used to evaluate the proposed deep learning algorithms\u2019 performance. the proposed deep learning methods treat each pixel of the video frame like a sensor to extract valuable features for defect detection. through empirical results, we demonstrate that this \u2019pixel-sensor\u2019 approach is more efficient and accurate and can achieve a better defect detection performance on different beam samples compared with the current state-of-the-art approaches, including cnn-long short-term memory (lstm), cnn-bidirectional long short-term memory (bilstm), multi-scale cnn-lstm, and cnn-gated recurrent unit(gru) methods.","key_phrases":["structural health monitoring","shm","the implementation","a damage detection strategy","structures","fault occurrence","these structural systems","the operation","precise health monitoring methods","the necessary repairs","maintenance","time","it","the current structural health monitoring methods","physically attached sensors","non-contact vision-based vibration measurements","these methods","significant drawbacks","the low spatial resolution","weight influence","the lightweight structure","time\/labor consumption","computer-vison-based deep learning methods","convolutional neural network","cnn","fully convolutional neural network","fcn","defect detection","localization","which","the aforementioned problems","high accuracy","this paper","a novel hybrid deep learning architecture","cnn","temporal convolutional networks","cnn-tcn","the computer vision-based defect detection task","various beam samples","five different materials","various structural defects","the proposed deep learning algorithms\u2019 performance","the proposed deep learning methods","each pixel","the video frame","a sensor","valuable features","defect detection","empirical results","we","this \u2019pixel-sensor\u2019 approach","a better defect detection performance","different beam samples","the-art","cnn-long short-term memory","lstm","cnn-bidirectional long short-term memory","bilstm","multi-scale cnn-lstm","cnn-gated recurrent unit(gru) methods","cnn","cnn","cnn-tcn","five","cnn","cnn","cnn","cnn"]},{"title":"Using machine learning and deep learning algorithms for downtime minimization in manufacturing systems: an early failure detection diagnostic service","authors":["Mohammad Shahin","F. Frank Chen","Ali Hosseinzadeh","Neda Zand"],"abstract":"Accurate detection of possible machine failure allows manufacturers to identify potential fault situations in processes to avoid downtimes caused by unexpected tool wear or unacceptable workpiece quality. This paper aims to report the study of more than 20 fault detection models using machine learning (ML), deep learning (DL), and deep hybrid learning (DHL). Predicting how the system could fail based on certain features or system settings (input variables) can help avoid future breakdowns and minimize downtime. The effectiveness of the proposed algorithms was experimented with a synthetic predictive maintenance dataset published by the School of Engineering of the University of Applied Sciences in Berlin, Germany. The fidelity of these algorithms was evaluated using performance measurement values such as accuracy, precision, recall, and the F-score. Final results demonstrated that deep forest and gradient boosting algorithms had shown very high levels of average accuracy (exceeded 90%). Additionally, the multinomial logistic regression and long short-term memory-based algorithms have shown satisfactory average accuracy (above 80%). Further analysis of models suggests that some models outperformed others. The research concluded that, through various ML, DL, and DHL algorithms, operational data analytics, and health monitoring system, engineers could optimize maintenance and reduce reliability risks.","doi":"10.1007\/s00170-023-12020-w","cleaned_title":"using machine learning and deep learning algorithms for downtime minimization in manufacturing systems: an early failure detection diagnostic service","cleaned_abstract":"accurate detection of possible machine failure allows manufacturers to identify potential fault situations in processes to avoid downtimes caused by unexpected tool wear or unacceptable workpiece quality. this paper aims to report the study of more than 20 fault detection models using machine learning (ml), deep learning (dl), and deep hybrid learning (dhl). predicting how the system could fail based on certain features or system settings (input variables) can help avoid future breakdowns and minimize downtime. the effectiveness of the proposed algorithms was experimented with a synthetic predictive maintenance dataset published by the school of engineering of the university of applied sciences in berlin, germany. the fidelity of these algorithms was evaluated using performance measurement values such as accuracy, precision, recall, and the f-score. final results demonstrated that deep forest and gradient boosting algorithms had shown very high levels of average accuracy (exceeded 90%). additionally, the multinomial logistic regression and long short-term memory-based algorithms have shown satisfactory average accuracy (above 80%). further analysis of models suggests that some models outperformed others. the research concluded that, through various ml, dl, and dhl algorithms, operational data analytics, and health monitoring system, engineers could optimize maintenance and reduce reliability risks.","key_phrases":["accurate detection","possible machine failure","manufacturers","potential fault situations","processes","downtimes","unexpected tool","this paper","the study","more than 20 fault detection models","machine learning","ml","deep learning","dl","deep hybrid learning","dhl","the system","certain features","system settings","input variables","future breakdowns","downtime","the effectiveness","the proposed algorithms","a synthetic predictive maintenance dataset","the school","engineering","the university","applied sciences","berlin","germany","the fidelity","these algorithms","performance measurement values","accuracy","precision","recall","the f-score","final results","deep forest and gradient boosting algorithms","very high levels","average accuracy","the multinomial logistic regression","long short-term memory-based algorithms","satisfactory average accuracy","80%","further analysis","models","some models","others","the research","various ml","dl","dhl algorithms","operational data analytics","health monitoring system","engineers","maintenance","reliability risks","more than 20","the university of applied sciences","berlin","germany","90%","80%"]},{"title":"Secure Communications with THz Reconfigurable Intelligent Surfaces and Deep Learning in 6G Systems","authors":["Ajmeera Kiran","Abhilash Sonker","Sachin Jadhav","Makarand Mohan Jadhav","Janjhyam Venkata Naga Ramesh","Elangovan Muniyandy"],"abstract":"In anticipation of the 6G era, this paper explores the integration of terahertz (THz) communications with Reconfigurable Intelligent Surfaces (RIS) and deep learning to establish a secure wireless network capable of ultra-high data rates. Addressing the non-convex challenge of maximizing secure energy efficiency, we introduce a novel deep learning framework that employs a variety of neural network architectures for optimizing RIS reflection and beamforming. Our simulations, set against scenarios with varying eavesdropper cooperation, confirm the efficacy of the proposed solution, achieving 97% of the optimal performance benchmarked against a genie-aided model. This research underlines a significant advancement in 6G network security, potentially influencing future standards and laying the groundwork for practical deployment, thereby marking a milestone in the convergence of THz technology, intelligent surfaces, and AI for future-proof secure communications.","doi":"10.1007\/s11277-024-11163-7","cleaned_title":"secure communications with thz reconfigurable intelligent surfaces and deep learning in 6g systems","cleaned_abstract":"in anticipation of the 6g era, this paper explores the integration of terahertz (thz) communications with reconfigurable intelligent surfaces (ris) and deep learning to establish a secure wireless network capable of ultra-high data rates. addressing the non-convex challenge of maximizing secure energy efficiency, we introduce a novel deep learning framework that employs a variety of neural network architectures for optimizing ris reflection and beamforming. our simulations, set against scenarios with varying eavesdropper cooperation, confirm the efficacy of the proposed solution, achieving 97% of the optimal performance benchmarked against a genie-aided model. this research underlines a significant advancement in 6g network security, potentially influencing future standards and laying the groundwork for practical deployment, thereby marking a milestone in the convergence of thz technology, intelligent surfaces, and ai for future-proof secure communications.","key_phrases":["anticipation","the 6g era","this paper","the integration","thz","reconfigurable intelligent surfaces","ris","deep learning","a secure wireless network","ultra-high data rates","the non-convex challenge","secure energy efficiency","we","a novel deep learning framework","that","a variety","neural network","ris reflection","beamforming","our simulations","scenarios","varying eavesdropper cooperation","the efficacy","the proposed solution","97%","the optimal performance","a genie-aided model","this research","a significant advancement","6g network security","future standards","the groundwork","practical deployment","a milestone","the convergence","thz technology","intelligent surfaces","future-proof secure communications","6","97%","genie","6"]},{"title":"HCCNet Fusion: a synergistic approach for accurate hepatocellular carcinoma staging using deep learning paradigm","authors":["Devi Rajeev","S. Remya","Anand Nayyar"],"abstract":"Hepatocellular carcinoma (HCC) stands as the second most prevalent cancer and a leading cause of cancer-related mortality globally, necessitating precise diagnostic and prognostic methodologies. The study introduces an innovative approach centered around the HCCNet Fusion model; a robust integration of advanced deep-learning techniques designed to elevate the accuracy of HCC stage recognition. Leveraging the synergies between the VGG16 architecture and U-Net and incorporating sophisticated data pre-processing methods such as Otsu\u2019s binary thresholding and marker-based watershed segmentation, this approach aims to strengthen the precision of HCC stage identification. Furthermore, transfer learning plays a pivotal role in HCCNet Fusion, enabling the models to integrate knowledge from diverse medical image settings through pre-trained weights from VGG16 and U-Net architectures. This strategic integration demonstrates the efficacy of advanced deep learning strategies in addressing intricate medical challenges, and outperforms conventional methods with a remarkable accuracy rate of 95%, underscoring the potential of cutting-edge deep learning techniques in medical diagnostics. The evaluation and validation of the proposed HCCNet Fusion model demonstrate its strong performance across many metrics like AUC ROC, Loss, Accuracy, Precision, Recall, and F1. Additionally, a comparative study was done against well-known methods, including CNN, Inception ResNetV2, VGG16, Inception V3, EfficientNet-B0, and ResNet50 and the results state that the proposed system not only advances HCC detection but also sets a pattern for leveraging state-of-the-art methodologies in addressing complex medical issues.","doi":"10.1007\/s11042-024-19446-8","cleaned_title":"hccnet fusion: a synergistic approach for accurate hepatocellular carcinoma staging using deep learning paradigm","cleaned_abstract":"hepatocellular carcinoma (hcc) stands as the second most prevalent cancer and a leading cause of cancer-related mortality globally, necessitating precise diagnostic and prognostic methodologies. the study introduces an innovative approach centered around the hccnet fusion model; a robust integration of advanced deep-learning techniques designed to elevate the accuracy of hcc stage recognition. leveraging the synergies between the vgg16 architecture and u-net and incorporating sophisticated data pre-processing methods such as otsu\u2019s binary thresholding and marker-based watershed segmentation, this approach aims to strengthen the precision of hcc stage identification. furthermore, transfer learning plays a pivotal role in hccnet fusion, enabling the models to integrate knowledge from diverse medical image settings through pre-trained weights from vgg16 and u-net architectures. this strategic integration demonstrates the efficacy of advanced deep learning strategies in addressing intricate medical challenges, and outperforms conventional methods with a remarkable accuracy rate of 95%, underscoring the potential of cutting-edge deep learning techniques in medical diagnostics. the evaluation and validation of the proposed hccnet fusion model demonstrate its strong performance across many metrics like auc roc, loss, accuracy, precision, recall, and f1. additionally, a comparative study was done against well-known methods, including cnn, inception resnetv2, vgg16, inception v3, efficientnet-b0, and resnet50 and the results state that the proposed system not only advances hcc detection but also sets a pattern for leveraging state-of-the-art methodologies in addressing complex medical issues.","key_phrases":["hepatocellular carcinoma","hcc","the second most prevalent cancer","a leading cause","cancer-related mortality","precise diagnostic and prognostic methodologies","the study","an innovative approach","the hccnet fusion model","a robust integration","advanced deep-learning techniques","the accuracy","hcc stage recognition","the synergies","the vgg16 architecture","u","-","net","sophisticated data pre-processing methods","otsu\u2019s binary thresholding and marker-based watershed segmentation","this approach","the precision","hcc stage identification","transfer learning","a pivotal role","hccnet fusion","the models","knowledge","diverse medical image settings","pre-trained weights","vgg16 and u-net architectures","this strategic integration","the efficacy","advanced deep learning strategies","intricate medical challenges","conventional methods","a remarkable accuracy rate","95%","the potential","cutting-edge deep learning techniques","medical diagnostics","the evaluation","validation","the proposed hccnet fusion model","its strong performance","many metrics","auc roc","loss","accuracy","precision","recall","f1","a comparative study","well-known methods","cnn","inception resnetv2","vgg16","inception v3","efficientnet-b0","resnet50","the results","the proposed system","hcc detection","a pattern","the-art","complex medical issues","second","95%","roc","cnn","resnetv2","v3","resnet50"]},{"title":"DVNE-DRL: dynamic virtual network embedding algorithm based on deep reinforcement learning","authors":["Xiancui Xiao"],"abstract":"Virtual network embedding (VNE), as the key challenge of network resource management technology, lies in the contradiction between online embedding decision and pursuing long-term average revenue goals. Most of the previous work ignored the dynamics in Virtual Network (VN) modeling, or could not automatically detect the complex and time-varying network state to provide a reasonable network embedding scheme. In view of this, we model a network embedding framework where the topology and resource allocation change dynamically with the number of network users and workload, and then introduce a deep reinforcement learning method to solve the VNE problem. Further, a dynamic virtual network embedding algorithm based on Deep Reinforcement Learning (DRL), named DVNE-DRL, is proposed. In DVNE-DRL, VNE is modeled as a Markov Decision Process (MDP), and then deep learning is introduced to perceive the current network state through historical data and embedded knowledge, while utilizing reinforcement learning decision-making capabilities to implement the network embedding process. In addition, we improve the method of feature extraction and matrix optimization, and consider the characteristics of virtual network and physical network together to alleviate the problem of redundancy and slow convergence. The simulation results show that compared with the existing advanced algorithms, the acceptance rate and average revenue of DVNE-DRL are increased by about 25% and 35%, respectively.","doi":"10.1038\/s41598-023-47195-5","cleaned_title":"dvne-drl: dynamic virtual network embedding algorithm based on deep reinforcement learning","cleaned_abstract":"virtual network embedding (vne), as the key challenge of network resource management technology, lies in the contradiction between online embedding decision and pursuing long-term average revenue goals. most of the previous work ignored the dynamics in virtual network (vn) modeling, or could not automatically detect the complex and time-varying network state to provide a reasonable network embedding scheme. in view of this, we model a network embedding framework where the topology and resource allocation change dynamically with the number of network users and workload, and then introduce a deep reinforcement learning method to solve the vne problem. further, a dynamic virtual network embedding algorithm based on deep reinforcement learning (drl), named dvne-drl, is proposed. in dvne-drl, vne is modeled as a markov decision process (mdp), and then deep learning is introduced to perceive the current network state through historical data and embedded knowledge, while utilizing reinforcement learning decision-making capabilities to implement the network embedding process. in addition, we improve the method of feature extraction and matrix optimization, and consider the characteristics of virtual network and physical network together to alleviate the problem of redundancy and slow convergence. the simulation results show that compared with the existing advanced algorithms, the acceptance rate and average revenue of dvne-drl are increased by about 25% and 35%, respectively.","key_phrases":["virtual network","the key challenge","network resource management technology","the contradiction","online embedding decision","long-term average revenue goals","the previous work","the dynamics","virtual network","modeling","the complex and time-varying network state","a reasonable network","scheme","view","this","we","a network","framework","the number","network users","workload","a deep reinforcement learning method","the vne problem","a dynamic virtual network","algorithm","deep reinforcement learning","drl","dvne-drl","dvne-drl","vne","a markov decision process","mdp","deep learning","the current network state","historical data","embedded knowledge","reinforcement learning decision-making capabilities","the network embedding process","addition","we","the method","feature extraction and matrix optimization","the characteristics","virtual network","physical network","the problem","redundancy","slow convergence","the simulation results","the existing advanced algorithms","the acceptance rate","average revenue","dvne-drl","about 25%","35%","about 25% and 35%"]},{"title":"Deep learning-based 1-D magnetotelluric inversion: performance comparison of architectures","authors":["Mehdi Rahmani Jevinani","Banafsheh Habibian Dehkordi","Ian J. Ferguson","Mohammad Hossein Rohban"],"abstract":"The study compares the three deep learning approaches and assesses their relative performance solving the 1-D magnetotellurics (MT) inverse problem. MT data from a 1-D geothermal-type structure are used as an example to examine Variational Autoencoder (VAE), Residual Network (Res-Net), and U-Net architectures, adapted for 1-D MT inversion. Root Mean Square Error (RMSE) and Pearson correlation coefficient are applied as misfit measure and similarity criterion, and box plot tools are used to parameterize individual model parameters. The results show that the U-Net provides the most successful recovery of the 1-D resistivity models, even though all three approaches can produce accurate inversions of MT data. To investigate applicability of results to real data sets, the models performance are examined for the case of data containing noise. Three deep learning algorithms are robust with respect to data noise, although the U-Net is relatively superior. The study results provide a platform for more complex magnetotelluric inverse problems and ones involving real data sets.","doi":"10.1007\/s12145-024-01233-6","cleaned_title":"deep learning-based 1-d magnetotelluric inversion: performance comparison of architectures","cleaned_abstract":"the study compares the three deep learning approaches and assesses their relative performance solving the 1-d magnetotellurics (mt) inverse problem. mt data from a 1-d geothermal-type structure are used as an example to examine variational autoencoder (vae), residual network (res-net), and u-net architectures, adapted for 1-d mt inversion. root mean square error (rmse) and pearson correlation coefficient are applied as misfit measure and similarity criterion, and box plot tools are used to parameterize individual model parameters. the results show that the u-net provides the most successful recovery of the 1-d resistivity models, even though all three approaches can produce accurate inversions of mt data. to investigate applicability of results to real data sets, the models performance are examined for the case of data containing noise. three deep learning algorithms are robust with respect to data noise, although the u-net is relatively superior. the study results provide a platform for more complex magnetotelluric inverse problems and ones involving real data sets.","key_phrases":["the study","the three deep learning approaches","their relative performance","the 1-d magnetotellurics (mt) inverse problem","mt data","a 1-d geothermal-type structure","an example","variational autoencoder","vae","residual network","res-net","u-net architectures","1-d mt inversion","root mean square error","rmse","pearson correlation coefficient","misfit measure","similarity criterion","box plot tools","individual model parameters","the results","the u","-","net","the most successful recovery","the 1-d resistivity models","all three approaches","accurate inversions","mt data","applicability","results","real data sets","the models performance","the case","data","noise","three deep learning algorithms","respect","data noise","the u","-","net","the study results","a platform","more complex magnetotelluric inverse problems","ones","real data sets","three","1-d","1","1-d mt inversion","root","1","three","three"]},{"title":"Automatic segmentation of inconstant fractured fragments for tibia\/fibula from CT images using deep learning","authors":["Hyeonjoo Kim","Young Dae Jeon","Ki Bong Park","Hayeong Cha","Moo-Sub Kim","Juyeon You","Se-Won Lee","Seung-Han Shin","Yang-Guk Chung","Sung Bin Kang","Won Seuk Jang","Do-Kun Yoon"],"abstract":"Orthopaedic surgeons need to correctly identify bone fragments using 2D\/3D CT images before trauma surgery. Advances in deep learning technology provide good insights into trauma surgery over manual diagnosis. This study demonstrates the application of the DeepLab v3+\u2009-based deep learning model for the automatic segmentation of fragments of the fractured tibia and fibula from CT images and the results of the evaluation of the performance of the automatic segmentation. The deep learning model, which was trained using over 11 million images, showed good performance with a global accuracy of 98.92%, a weighted intersection over the union of 0.9841, and a mean boundary F1 score of 0.8921. Moreover, deep learning performed 5\u20138 times faster than the experts\u2019 recognition performed manually, which is comparatively inefficient, with almost the same significance. This study will play an important role in preoperative surgical planning for trauma surgery with convenience and speed.","doi":"10.1038\/s41598-023-47706-4","cleaned_title":"automatic segmentation of inconstant fractured fragments for tibia\/fibula from ct images using deep learning","cleaned_abstract":"orthopaedic surgeons need to correctly identify bone fragments using 2d\/3d ct images before trauma surgery. advances in deep learning technology provide good insights into trauma surgery over manual diagnosis. this study demonstrates the application of the deeplab v3+ -based deep learning model for the automatic segmentation of fragments of the fractured tibia and fibula from ct images and the results of the evaluation of the performance of the automatic segmentation. the deep learning model, which was trained using over 11 million images, showed good performance with a global accuracy of 98.92%, a weighted intersection over the union of 0.9841, and a mean boundary f1 score of 0.8921. moreover, deep learning performed 5\u20138 times faster than the experts\u2019 recognition performed manually, which is comparatively inefficient, with almost the same significance. this study will play an important role in preoperative surgical planning for trauma surgery with convenience and speed.","key_phrases":["orthopaedic surgeons","bone fragments","2d\/3d ct images","trauma surgery","advances","deep learning technology","good insights","trauma surgery","manual diagnosis","this study","the application","the deeplab v3","-based deep learning model","the automatic segmentation","fragments","the fractured tibia","ct images","the results","the evaluation","the performance","the automatic segmentation","the deep learning model","which","over 11 million images","good performance","a global accuracy","98.92%","a weighted intersection","the union","a mean boundary f1 score","deep learning","the experts\u2019 recognition","which","almost the same significance","this study","an important role","preoperative surgical planning","trauma surgery","convenience","speed","2d\/3d","over 11 million","98.92%","0.9841","0.8921","5\u20138"]},{"title":"Smart city urban planning using an evolutionary deep learning model","authors":["Mansoor Alghamdi"],"abstract":"Following the evolution of big data collection, storage, and manipulation techniques, deep learning has drawn the attention of numerous recent studies proposing solutions for smart cities. These solutions were focusing especially on energy consumption, pollution levels, public services, and traffic management issues. Predicting urban evolution and planning is another recent concern for smart cities. In this context, this paper introduces a hybrid model that incorporates evolutionary optimization algorithms, such as Teaching\u2013learning-based optimization (TLBO), into the functioning process of neural deep learning models, such as recurrent neural network (RNN) networks. According to the achieved simulations, deep learning enhanced by evolutionary optimizers can be an effective and promising method for predicting urban evolution of future smart cities.","doi":"10.1007\/s00500-023-08219-4","cleaned_title":"smart city urban planning using an evolutionary deep learning model","cleaned_abstract":"following the evolution of big data collection, storage, and manipulation techniques, deep learning has drawn the attention of numerous recent studies proposing solutions for smart cities. these solutions were focusing especially on energy consumption, pollution levels, public services, and traffic management issues. predicting urban evolution and planning is another recent concern for smart cities. in this context, this paper introduces a hybrid model that incorporates evolutionary optimization algorithms, such as teaching\u2013learning-based optimization (tlbo), into the functioning process of neural deep learning models, such as recurrent neural network (rnn) networks. according to the achieved simulations, deep learning enhanced by evolutionary optimizers can be an effective and promising method for predicting urban evolution of future smart cities.","key_phrases":["the evolution","big data collection","storage","manipulation techniques","deep learning","the attention","numerous recent studies","solutions","smart cities","these solutions","energy consumption","pollution levels","public services","traffic management issues","urban evolution","planning","another recent concern","smart cities","this context","this paper","a hybrid model","that","evolutionary optimization algorithms","teaching","learning-based optimization","tlbo","the functioning process","neural deep learning models","recurrent neural network (rnn) networks","the achieved simulations","deep learning","evolutionary optimizers","an effective and promising method","urban evolution","future smart cities"]},{"title":"Recent advances in deep learning models: a systematic literature review","authors":["Ruchika Malhotra","Priya Singh"],"abstract":"In recent years, deep learning has evolved as a rapidly growing and stimulating field of machine learning and has redefined state-of-the-art performances in a variety of applications. There are multiple deep learning models that have distinct architectures and capabilities. Up to the present, a large number of novel variants of these baseline deep learning models is proposed to address the shortcomings of the existing baseline models. This paper provides a comprehensive review of one hundred seven novel variants of six baseline deep learning models viz. Convolutional Neural Network, Recurrent Neural Network, Long Short Term Memory, Generative Adversarial Network, Autoencoder and Transformer Neural Network. The current review thoroughly examines the novel variants of each of the six baseline models to identify the advancements adopted by them to address one or more limitations of the respective baseline model. It is achieved by critically reviewing the novel variants based on their improved approach. It further provides the merits and demerits of incorporating the advancements in novel variants compared to the baseline deep learning model. Additionally, it reports the domain, datasets and performance measures exploited by the novel variants to make an overall judgment in terms of the improvements. This is because the performance of the deep learning models are subject to the application domain, type of datasets and may also vary on different performance measures. The critical findings of the review would facilitate the researchers and practitioners with the most recent progressions and advancements in the baseline deep learning models and guide them in selecting an appropriate novel variant of the baseline to solve deep learning based tasks in a similar setting.","doi":"10.1007\/s11042-023-15295-z","cleaned_title":"recent advances in deep learning models: a systematic literature review","cleaned_abstract":"in recent years, deep learning has evolved as a rapidly growing and stimulating field of machine learning and has redefined state-of-the-art performances in a variety of applications. there are multiple deep learning models that have distinct architectures and capabilities. up to the present, a large number of novel variants of these baseline deep learning models is proposed to address the shortcomings of the existing baseline models. this paper provides a comprehensive review of one hundred seven novel variants of six baseline deep learning models viz. convolutional neural network, recurrent neural network, long short term memory, generative adversarial network, autoencoder and transformer neural network. the current review thoroughly examines the novel variants of each of the six baseline models to identify the advancements adopted by them to address one or more limitations of the respective baseline model. it is achieved by critically reviewing the novel variants based on their improved approach. it further provides the merits and demerits of incorporating the advancements in novel variants compared to the baseline deep learning model. additionally, it reports the domain, datasets and performance measures exploited by the novel variants to make an overall judgment in terms of the improvements. this is because the performance of the deep learning models are subject to the application domain, type of datasets and may also vary on different performance measures. the critical findings of the review would facilitate the researchers and practitioners with the most recent progressions and advancements in the baseline deep learning models and guide them in selecting an appropriate novel variant of the baseline to solve deep learning based tasks in a similar setting.","key_phrases":["recent years","deep learning","a rapidly growing and stimulating field","machine learning","the-art","a variety","applications","multiple deep learning models","that","distinct architectures","capabilities","the present","a large number","novel variants","these baseline deep learning models","the shortcomings","the existing baseline models","this paper","a comprehensive review","one hundred seven novel variants","six baseline deep learning models","convolutional neural network","recurrent neural network","long short term memory","generative adversarial network","autoencoder","transformer neural network","the current review","the novel variants","each","the six baseline models","the advancements","them","one or more limitations","the respective baseline model","it","the novel variants","their improved approach","it","the merits","demerits","the advancements","novel variants","the baseline deep learning model","it","the domain","datasets","performance measures","the novel variants","an overall judgment","terms","the improvements","this","the performance","the deep learning models","the application domain","type","datasets","different performance measures","the critical findings","the review","the researchers","practitioners","the most recent progressions","advancements","the baseline deep learning models","them","an appropriate novel variant","the baseline","deep learning based tasks","a similar setting","recent years","one hundred seven","six","six","one"]},{"title":"Deep Learning Based Alzheimer Disease Diagnosis: A Comprehensive Review","authors":["S. Suganyadevi","A. Shiny Pershiya","K. Balasamy","V. Seethalakshmi","Saroj Bala","Kumud Arora"],"abstract":"Dementia encompasses a range of cognitive disorders, with Alzheimer\u2019s Disease being the utmost widespread and devastating. AD gradually erodes memory and daily functioning through the progressive deterioration of brain cells. It poses a significant global health challenge, necessitating early identification and intervention. Detecting AD at its onset holds immense potential to predict future health outcomes for individuals. By harnessing the power of artificial intelligence and leveraging MRI scans, we could utilize advanced technology to not only classify AD patients but also predict the likelihood of them developing this life-altering condition. This paper delves into the latest advancements in Deep Learning techniques and their functions in image analysis in medical field. Its primary goals are to elucidate the intricacies of medical image processing and to elucidate and implement key findings and recommendations from recent research.","doi":"10.1007\/s42979-024-02743-2","cleaned_title":"deep learning based alzheimer disease diagnosis: a comprehensive review","cleaned_abstract":"dementia encompasses a range of cognitive disorders, with alzheimer\u2019s disease being the utmost widespread and devastating. ad gradually erodes memory and daily functioning through the progressive deterioration of brain cells. it poses a significant global health challenge, necessitating early identification and intervention. detecting ad at its onset holds immense potential to predict future health outcomes for individuals. by harnessing the power of artificial intelligence and leveraging mri scans, we could utilize advanced technology to not only classify ad patients but also predict the likelihood of them developing this life-altering condition. this paper delves into the latest advancements in deep learning techniques and their functions in image analysis in medical field. its primary goals are to elucidate the intricacies of medical image processing and to elucidate and implement key findings and recommendations from recent research.","key_phrases":["dementia","a range","cognitive disorders","alzheimer\u2019s disease","ad","memory","the progressive deterioration","brain cells","it","a significant global health challenge","early identification","intervention","ad","its onset","immense potential","future health outcomes","individuals","the power","artificial intelligence","mri scans","we","advanced technology","ad patients","the likelihood","them","this life-altering condition","this paper","the latest advancements","deep learning techniques","their functions","image analysis","medical field","its primary goals","the intricacies","medical image processing","key findings","recommendations","recent research","daily"]},{"title":"Can deep learning replace histopathological examinations in the differential diagnosis of cervical lymphadenopathy?","authors":["Sermin Can","\u00d6mer T\u00fcrk","Muhammed Ayral","G\u00fcnay Kozan","Hamza Ar\u0131","Mehmet Akda\u011f","M\u00fczeyyen Y\u0131ld\u0131r\u0131m Baylan"],"abstract":"IntroductionWe aimed to develop a diagnostic deep learning model using contrast-enhanced CT images and to investigate whether cervical lymphadenopathies can be diagnosed with these deep learning methods without radiologist interpretations and histopathological examinations.Material methodA total of 400 patients who underwent surgery for lymphadenopathy in the neck between 2010 and 2022 were retrospectively analyzed. They were examined in four groups of 100 patients: the granulomatous diseases group, the lymphoma group, the squamous cell tumor group, and the reactive hyperplasia group. The diagnoses of the patients were confirmed histopathologically. Two CT images from all the patients in each group were used in the study. The CT images were classified using ResNet50, NASNetMobile, and DenseNet121 architecture input.ResultsThe classification accuracies obtained with ResNet50, DenseNet121, and NASNetMobile were 92.5%, 90.62, and 87.5, respectively.ConclusionDeep learning is a useful diagnostic tool in diagnosing cervical lymphadenopathy. In the near future, many diseases could be diagnosed with deep learning models without radiologist interpretations and invasive examinations such as histopathological examinations. However, further studies with much larger case series are needed to develop accurate deep-learning models.","doi":"10.1007\/s00405-023-08181-9","cleaned_title":"can deep learning replace histopathological examinations in the differential diagnosis of cervical lymphadenopathy?","cleaned_abstract":"introductionwe aimed to develop a diagnostic deep learning model using contrast-enhanced ct images and to investigate whether cervical lymphadenopathies can be diagnosed with these deep learning methods without radiologist interpretations and histopathological examinations.material methoda total of 400 patients who underwent surgery for lymphadenopathy in the neck between 2010 and 2022 were retrospectively analyzed. they were examined in four groups of 100 patients: the granulomatous diseases group, the lymphoma group, the squamous cell tumor group, and the reactive hyperplasia group. the diagnoses of the patients were confirmed histopathologically. two ct images from all the patients in each group were used in the study. the ct images were classified using resnet50, nasnetmobile, and densenet121 architecture input.resultsthe classification accuracies obtained with resnet50, densenet121, and nasnetmobile were 92.5%, 90.62, and 87.5, respectively.conclusiondeep learning is a useful diagnostic tool in diagnosing cervical lymphadenopathy. in the near future, many diseases could be diagnosed with deep learning models without radiologist interpretations and invasive examinations such as histopathological examinations. however, further studies with much larger case series are needed to develop accurate deep-learning models.","key_phrases":["introductionwe","a diagnostic deep learning model","contrast-enhanced ct images","cervical lymphadenopathies","these deep learning methods","radiologist interpretations","histopathological examinations.material methoda total","400 patients","who","surgery","lymphadenopathy","the neck","they","four groups","100 patients","the granulomatous diseases group","the lymphoma group","the squamous cell tumor group","the reactive hyperplasia group","the diagnoses","the patients","two ct images","all the patients","each group","the study","the ct images","resnet50","nasnetmobile","architecture","input.resultsthe classification accuracies","resnet50","densenet121","92.5%","respectively.conclusiondeep learning","a useful diagnostic tool","cervical lymphadenopathy","the near future","many diseases","deep learning models","radiologist interpretations","invasive examinations","histopathological examinations","further studies","much larger case series","accurate deep-learning models","400","between 2010 and 2022","four","100","two","resnet50","resnet50","92.5%","90.62","87.5"]},{"title":"Deep Learning Models for Diagnosis of Schizophrenia Using EEG Signals: Emerging Trends, Challenges, and Prospects","authors":["Rakesh Ranjan","Bikash Chandra Sahana","Ashish Kumar Bhandari"],"abstract":"Schizophrenia (ScZ) is a chronic neuropsychiatric disorder characterized by disruptions in cognitive, perceptual, social, emotional, and behavioral functions. In the traditional approach, the diagnosis of ScZ primarily relies on the subject\u2019s response and the psychiatrist\u2019s experience, making it highly subjective, prejudiced, and time-consuming. In recent medical research, incorporating deep learning (DL) into the diagnostic process improves performance by reducing inter-observer variation and providing qualitative and quantitative support for clinical decisions. Compared with other modalities, such as magnetic resonance images (MRI) or computed tomography (CT) scans, electroencephalogram (EEG) signals give better insights into the underlying neural mechanisms and brain biomarkers of ScZ. Deep learning models show promising results but the utilization of EEG signals as an effective biomarker for ScZ is still under research. Numerous deep learning models have recently been developed for automated ScZ diagnosis with EEG signals exclusively, yet a comprehensive assessment of these approaches still does not exist in the literature. To fill this gap, we comprehensively review the current advancements in deep learning-based schizophrenia diagnosis using EEG signals. This review is intended to provide systematic details of prominent components: deep learning models, ScZ EEG datasets, data preprocessing approaches, input data formulations for DL, chronological DL methodology advancement in ScZ diagnosis, and design trends of DL architecture. Finally, few challenges in both clinical and technical aspects that create hindrances in achieving the full potential of DL models in EEG-based ScZ diagnosis are expounded along with future outlooks.","doi":"10.1007\/s11831-023-10047-6","cleaned_title":"deep learning models for diagnosis of schizophrenia using eeg signals: emerging trends, challenges, and prospects","cleaned_abstract":"schizophrenia (scz) is a chronic neuropsychiatric disorder characterized by disruptions in cognitive, perceptual, social, emotional, and behavioral functions. in the traditional approach, the diagnosis of scz primarily relies on the subject\u2019s response and the psychiatrist\u2019s experience, making it highly subjective, prejudiced, and time-consuming. in recent medical research, incorporating deep learning (dl) into the diagnostic process improves performance by reducing inter-observer variation and providing qualitative and quantitative support for clinical decisions. compared with other modalities, such as magnetic resonance images (mri) or computed tomography (ct) scans, electroencephalogram (eeg) signals give better insights into the underlying neural mechanisms and brain biomarkers of scz. deep learning models show promising results but the utilization of eeg signals as an effective biomarker for scz is still under research. numerous deep learning models have recently been developed for automated scz diagnosis with eeg signals exclusively, yet a comprehensive assessment of these approaches still does not exist in the literature. to fill this gap, we comprehensively review the current advancements in deep learning-based schizophrenia diagnosis using eeg signals. this review is intended to provide systematic details of prominent components: deep learning models, scz eeg datasets, data preprocessing approaches, input data formulations for dl, chronological dl methodology advancement in scz diagnosis, and design trends of dl architecture. finally, few challenges in both clinical and technical aspects that create hindrances in achieving the full potential of dl models in eeg-based scz diagnosis are expounded along with future outlooks.","key_phrases":["schizophrenia","scz","a chronic neuropsychiatric disorder","disruptions","cognitive, perceptual, social, emotional, and behavioral functions","the traditional approach","the diagnosis","scz","the subject\u2019s response","the psychiatrist\u2019s experience","it","recent medical research","deep learning","dl","the diagnostic process","performance","inter-observer variation","qualitative and quantitative support","clinical decisions","other modalities","magnetic resonance images","mri","tomography (ct) scans","electroencephalogram (eeg) signals","better insights","the underlying neural mechanisms","brain biomarkers","scz","deep learning models","promising results","the utilization","eeg signals","an effective biomarker","scz","research","numerous deep learning models","automated scz diagnosis","eeg signals","a comprehensive assessment","these approaches","the literature","this gap","we","the current advancements","deep learning-based schizophrenia diagnosis","eeg signals","this review","systematic details","prominent components","deep learning models","scz eeg datasets","data","approaches","input data formulations","dl, chronological dl methodology advancement","scz diagnosis","design trends","dl architecture","few challenges","both clinical and technical aspects","that","hindrances","the full potential","dl models","eeg-based scz diagnosis","future outlooks"]},{"title":"GMPP-NN: a deep learning architecture for graph molecular property prediction","authors":["Outhman Abbassi","Soumia Ziti","Meryam Belhiah","Souad Najoua Lagmiri","Yassine Zaoui Seghroucheni"],"abstract":"The pharmacy industry is highly focused on drug discovery and development for the identification and optimization of potential drug candidates. One of the key aspects of this process is the prediction of various molecular properties that justify their potential effectiveness in treating specific diseases. Recently, graph neural networks have gained significant attention, primarily due to their strong suitability for predicting complex relationships that exist between atoms and other molecular structures. GNNs require significant depth to capture global features and to allow the network to iteratively aggregate and propagate information across the entire graph structure. In this research study, we present a deep learning architecture known as a graph molecular property prediction neural network. which combines MPNN feature extraction with a multilayer perceptron classifier. The deep learning architecture was evaluated on four benchmark datasets, and its performance was compared to the smiles transformer, fingerprint to vector, deeper graph convolutional networks, geometry-enhanced molecular, and atom-bond transformer-based message-passing neural network. The results showed that the architecture outperformed the other models using the receiver operating characteristic area under the curve metric. These findings offer an exciting opportunity to enhance and improve molecular property prediction in drug discovery and development.","doi":"10.1007\/s42452-024-05944-9","cleaned_title":"gmpp-nn: a deep learning architecture for graph molecular property prediction","cleaned_abstract":"the pharmacy industry is highly focused on drug discovery and development for the identification and optimization of potential drug candidates. one of the key aspects of this process is the prediction of various molecular properties that justify their potential effectiveness in treating specific diseases. recently, graph neural networks have gained significant attention, primarily due to their strong suitability for predicting complex relationships that exist between atoms and other molecular structures. gnns require significant depth to capture global features and to allow the network to iteratively aggregate and propagate information across the entire graph structure. in this research study, we present a deep learning architecture known as a graph molecular property prediction neural network. which combines mpnn feature extraction with a multilayer perceptron classifier. the deep learning architecture was evaluated on four benchmark datasets, and its performance was compared to the smiles transformer, fingerprint to vector, deeper graph convolutional networks, geometry-enhanced molecular, and atom-bond transformer-based message-passing neural network. the results showed that the architecture outperformed the other models using the receiver operating characteristic area under the curve metric. these findings offer an exciting opportunity to enhance and improve molecular property prediction in drug discovery and development.","key_phrases":["the pharmacy industry","drug discovery","development","the identification","optimization","potential drug candidates","the key aspects","this process","the prediction","various molecular properties","that","their potential effectiveness","specific diseases","graph neural networks","significant attention","their strong suitability","complex relationships","that","atoms","other molecular structures","gnns","significant depth","global features","the network","information","the entire graph structure","this research study","we","a deep learning architecture","a graph molecular property prediction neural network","which","mpnn feature extraction","a multilayer perceptron classifier","the deep learning architecture","four benchmark datasets","its performance","the smiles transformer","vector","deeper graph convolutional networks","atom-bond transformer-based message-passing neural network","the results","the architecture","the other models","the receiver operating characteristic area","the curve metric","these findings","an exciting opportunity","molecular property prediction","drug discovery","development","one","four"]},{"title":"High-dimensional stochastic control models for newsvendor problems and deep learning resolution","authors":["Jingtang Ma","Shan Yang"],"abstract":"This paper studies continuous-time models for newsvendor problems with dynamic replenishment, financial hedging and Stackelberg competition. These factors are considered simultaneously and the high-dimensional stochastic control models are established. High-dimensional Hamilton-Jacobi-Bellman (HJB) equations are derived for the value functions. To circumvent the curse of dimensionality, a deep learning algorithm is proposed to solve the HJB equations. A projection is introduced in the algorithm to avoid the gradient explosion during the training phase. The deep learning algorithm is implemented for HJB equations derived from the newsvendor models with dimensions up to six. Numerical outcomes validate the algorithm\u2019s accuracy and demonstrate that the high-dimensional stochastic control models can successfully mitigate the risk.","doi":"10.1007\/s10479-024-05872-2","cleaned_title":"high-dimensional stochastic control models for newsvendor problems and deep learning resolution","cleaned_abstract":"this paper studies continuous-time models for newsvendor problems with dynamic replenishment, financial hedging and stackelberg competition. these factors are considered simultaneously and the high-dimensional stochastic control models are established. high-dimensional hamilton-jacobi-bellman (hjb) equations are derived for the value functions. to circumvent the curse of dimensionality, a deep learning algorithm is proposed to solve the hjb equations. a projection is introduced in the algorithm to avoid the gradient explosion during the training phase. the deep learning algorithm is implemented for hjb equations derived from the newsvendor models with dimensions up to six. numerical outcomes validate the algorithm\u2019s accuracy and demonstrate that the high-dimensional stochastic control models can successfully mitigate the risk.","key_phrases":["this paper","newsvendor problems","dynamic replenishment","financial hedging","stackelberg competition","these factors","the high-dimensional stochastic control models","hjb","the value functions","the curse","dimensionality","a deep learning algorithm","the hjb equations","a projection","the algorithm","the gradient explosion","the training phase","the deep learning algorithm","hjb equations","the newsvendor models","dimensions","numerical outcomes","the algorithm\u2019s accuracy","the high-dimensional stochastic control models","the risk","hamilton-jacobi-bellman","six"]},{"title":"Integrated deep learning approach for automatic coronary artery segmentation and classification on computed tomographic coronary angiography","authors":["Chitra Devi Muthusamy","Ramaswami  Murugesh"],"abstract":"The field of coronary artery disease (CAD) has seen a rapid development in coronary computed tomography angiography (CCTA). However, manual coronary artery tree segmentation and reconstruction take time and effort. Deep learning algorithms have been created effectively to analyze large amounts of data for medical image analysis. The primary goal of this research\u00a0is to create an automated CAD diagnostic model and a deep learning tool for automatic coronary artery reconstruction\u00a0using a large, single-center retrospective CCTA cohort. We propose an integrated deep learning-based intelligent system for human heart blood vessel position within heart coronary CT angiography images using a multi-class ensemble classification mechanism in this research. The Modified DenseNet201 will segment the cardiac blood vessels in the proposed work. Then, the low-level features are successfully extracted using the ResNet-152 model. Finally, the improved deep residual shrinkage network (IDRSN) model classifies heart blood vessels into\u00a0four distinct classes: normal, block, narrow, and blood flow-reduced. The CCTA image dataset is used for the experiment analysis. The integrated deep learning-based blood vessel segmentation and classification system (Modified DenseNet201-IDRSN) was developed using the Python tool, and the system\u2019s efficiency was determined using different performance metrics. The experimental results of this research demonstrate that the proposed system is effective and efficient concerning related studies. Compared to the U-Net segmentation model, the segmentation result\u00a0in the proposed\u00a0research\u00a0is smoother and nearest to the segmentation result of the human expert. The proposed integrated deep learning intelligent system improves the efficiency of disease diagnosis, reduces dependence on medical personnel, reduces manual interaction in diagnosis, and offers auxiliary strategies for subsequent medical diagnosis systems based on cardiac coronary angiography.","doi":"10.1007\/s13721-024-00473-2","cleaned_title":"integrated deep learning approach for automatic coronary artery segmentation and classification on computed tomographic coronary angiography","cleaned_abstract":"the field of coronary artery disease (cad) has seen a rapid development in coronary computed tomography angiography (ccta). however, manual coronary artery tree segmentation and reconstruction take time and effort. deep learning algorithms have been created effectively to analyze large amounts of data for medical image analysis. the primary goal of this research is to create an automated cad diagnostic model and a deep learning tool for automatic coronary artery reconstruction using a large, single-center retrospective ccta cohort. we propose an integrated deep learning-based intelligent system for human heart blood vessel position within heart coronary ct angiography images using a multi-class ensemble classification mechanism in this research. the modified densenet201 will segment the cardiac blood vessels in the proposed work. then, the low-level features are successfully extracted using the resnet-152 model. finally, the improved deep residual shrinkage network (idrsn) model classifies heart blood vessels into four distinct classes: normal, block, narrow, and blood flow-reduced. the ccta image dataset is used for the experiment analysis. the integrated deep learning-based blood vessel segmentation and classification system (modified densenet201-idrsn) was developed using the python tool, and the system\u2019s efficiency was determined using different performance metrics. the experimental results of this research demonstrate that the proposed system is effective and efficient concerning related studies. compared to the u-net segmentation model, the segmentation result in the proposed research is smoother and nearest to the segmentation result of the human expert. the proposed integrated deep learning intelligent system improves the efficiency of disease diagnosis, reduces dependence on medical personnel, reduces manual interaction in diagnosis, and offers auxiliary strategies for subsequent medical diagnosis systems based on cardiac coronary angiography.","key_phrases":["the field","coronary artery disease","cad","a rapid development","coronary computed tomography angiography","ccta","manual coronary artery tree segmentation","reconstruction","time","effort","deep learning algorithms","large amounts","data","medical image analysis","the primary goal","this research","an automated cad diagnostic model","a deep learning tool","automatic coronary artery reconstruction","a large, single-center retrospective ccta cohort","we","an integrated deep learning-based intelligent system","human heart blood vessel position","heart coronary ct angiography images","a multi-class ensemble classification mechanism","this research","the modified densenet201","the cardiac blood vessels","the proposed work","the low-level features","the resnet-152 model","the improved deep residual shrinkage network","four distinct classes","the ccta image dataset","the experiment analysis","the integrated deep learning-based blood vessel segmentation and classification system","modified densenet201-idrsn","the python tool","the system\u2019s efficiency","different performance metrics","the experimental results","this research demonstrate","the proposed system","related studies","the u-net segmentation model","the segmentation","the proposed research","the segmentation result","the human expert","the proposed integrated deep learning intelligent system","the efficiency","disease diagnosis","dependence","medical personnel","manual interaction","diagnosis","auxiliary strategies","subsequent medical diagnosis systems","cardiac coronary angiography","resnet-152","four"]},{"title":"Deep learning with autoencoders and LSTM for ENSO forecasting","authors":["Chibuike Chiedozie Ibebuchi","Michael B. Richman"],"abstract":"El Ni\u00f1o Southern Oscillation (ENSO) is the prominent recurrent climatic pattern in the tropical Pacific Ocean with global impacts on regional climates. This study utilizes deep learning to predict the Ni\u00f1o 3.4 index by encoding non-linear sea surface temperature patterns in the tropical Pacific using an autoencoder neural network. The resulting encoded patterns identify crucial centers of action in the Pacific that serve as predictors of the ENSO mode. These patterns are utilized as predictors for forecasting the Ni\u00f1o 3.4 index with a lead time of at least 6 months using the Long Short-Term Memory (LSTM) deep learning model. The analysis uncovers multiple non-linear dipole patterns in the tropical Pacific, with anomalies that are both regionalized and latitudinally oriented that should support a single inter-tropical convergence zone for modeling efforts. Leveraging these encoded patterns as predictors, the LSTM - trained on monthly data from 1950 to 2007 and tested from 2008 to 2022 - shows fidelity in predicting the Ni\u00f1o 3.4 index. The encoded patterns captured the annual cycle of ENSO with a 0.94 correlation between the actual and predicted Ni\u00f1o 3.4 index for lag 12 and 0.91 for lags 6 and 18. Additionally, the 6-month lag predictions excel in detecting extreme ENSO events, achieving an 85% hit rate, outperforming the 70% hit rate at lag 12 and 55% hit rate at lag 18. The prediction accuracy peaks from November to March, with correlations ranging from 0.94 to 0.96. The average correlations in the boreal spring were as large as 0.84, indicating the method has the capability to decrease the spring predictability barrier.","doi":"10.1007\/s00382-024-07180-8","cleaned_title":"deep learning with autoencoders and lstm for enso forecasting","cleaned_abstract":"el ni\u00f1o southern oscillation (enso) is the prominent recurrent climatic pattern in the tropical pacific ocean with global impacts on regional climates. this study utilizes deep learning to predict the ni\u00f1o 3.4 index by encoding non-linear sea surface temperature patterns in the tropical pacific using an autoencoder neural network. the resulting encoded patterns identify crucial centers of action in the pacific that serve as predictors of the enso mode. these patterns are utilized as predictors for forecasting the ni\u00f1o 3.4 index with a lead time of at least 6 months using the long short-term memory (lstm) deep learning model. the analysis uncovers multiple non-linear dipole patterns in the tropical pacific, with anomalies that are both regionalized and latitudinally oriented that should support a single inter-tropical convergence zone for modeling efforts. leveraging these encoded patterns as predictors, the lstm - trained on monthly data from 1950 to 2007 and tested from 2008 to 2022 - shows fidelity in predicting the ni\u00f1o 3.4 index. the encoded patterns captured the annual cycle of enso with a 0.94 correlation between the actual and predicted ni\u00f1o 3.4 index for lag 12 and 0.91 for lags 6 and 18. additionally, the 6-month lag predictions excel in detecting extreme enso events, achieving an 85% hit rate, outperforming the 70% hit rate at lag 12 and 55% hit rate at lag 18. the prediction accuracy peaks from november to march, with correlations ranging from 0.94 to 0.96. the average correlations in the boreal spring were as large as 0.84, indicating the method has the capability to decrease the spring predictability barrier.","key_phrases":["el ni\u00f1o southern oscillation","enso","the prominent recurrent climatic pattern","the tropical pacific ocean","global impacts","regional climates","this study","deep learning","the ni\u00f1o 3.4 index","non-linear sea surface temperature patterns","the tropical pacific","an autoencoder neural network","the resulting encoded patterns","crucial centers","action","the pacific","that","predictors","the enso mode","these patterns","predictors","the ni\u00f1o 3.4 index","a lead time","at least 6 months","the long short-term memory","lstm","deep learning model","the analysis uncovers","non-linear dipole patterns","the tropical pacific","anomalies","that","that","a single inter-tropical convergence zone","modeling efforts","these encoded patterns","predictors","monthly data","fidelity","the ni\u00f1o 3.4 index","the encoded patterns","the annual cycle","enso","a 0.94 correlation","3.4 index","lag","lags","the 6-month lag predictions","extreme enso events","an 85% hit rate","the 70% hit rate","lag","hit rate","lag","the prediction accuracy peaks","november","march","correlations","the average correlations","the boreal spring","the method","the capability","the spring predictability barrier","el ni\u00f1o southern oscillation","pacific","3.4","non-linear","3.4","at least 6 months","monthly","1950","2007","2008","2022","3.4","annual","0.94","3.4","12","0.91","6","6-month","85%","70%","55%","november to march","0.94","0.96","as large as 0.84"]},{"title":"Survey on deep learning in multimodal medical imaging for cancer detection","authors":["Yan Tian","Zhaocheng Xu","Yujun Ma","Weiping Ding","Ruili Wang","Zhihong Gao","Guohua Cheng","Linyang He","Xuran Zhao"],"abstract":"The task of multimodal cancer detection is to determine the locations and categories of lesions by using different imaging techniques, which is one of the key research methods for cancer diagnosis. Recently, deep learning-based object detection has made significant developments due to its strength in semantic feature extraction and nonlinear function fitting. However, multimodal cancer detection remains challenging due to morphological differences in lesions, interpatient variability, difficulty in annotation, and imaging artifacts. In this survey, we mainly investigate over 150 papers in recent years with respect to multimodal cancer detection using deep learning, with a focus on datasets and solutions to various challenges such as data annotation, variance between classes, small-scale lesions, and occlusion. We also provide an overview of the advantages and drawbacks of each approach. Finally, we discuss the current scope of work and provide directions for the future development of multimodal cancer detection.","doi":"10.1007\/s00521-023-09214-4","cleaned_title":"survey on deep learning in multimodal medical imaging for cancer detection","cleaned_abstract":"the task of multimodal cancer detection is to determine the locations and categories of lesions by using different imaging techniques, which is one of the key research methods for cancer diagnosis. recently, deep learning-based object detection has made significant developments due to its strength in semantic feature extraction and nonlinear function fitting. however, multimodal cancer detection remains challenging due to morphological differences in lesions, interpatient variability, difficulty in annotation, and imaging artifacts. in this survey, we mainly investigate over 150 papers in recent years with respect to multimodal cancer detection using deep learning, with a focus on datasets and solutions to various challenges such as data annotation, variance between classes, small-scale lesions, and occlusion. we also provide an overview of the advantages and drawbacks of each approach. finally, we discuss the current scope of work and provide directions for the future development of multimodal cancer detection.","key_phrases":["the task","multimodal cancer detection","the locations","categories","lesions","different imaging techniques","which","the key research methods","cancer diagnosis","deep learning-based object detection","significant developments","its strength","semantic feature extraction and nonlinear function","multimodal cancer detection","morphological differences","lesions","interpatient variability","difficulty","annotation","imaging artifacts","this survey","we","over 150 papers","recent years","respect","multimodal cancer detection","deep learning","a focus","datasets","solutions","various challenges","data annotation","variance","classes","small-scale lesions","occlusion","we","an overview","the advantages","drawbacks","each approach","we","the current scope","work","directions","the future development","multimodal cancer detection","150","recent years"]},{"title":"A systematic review and meta-analysis of artificial neural network, machine learning, deep learning, and ensemble learning approaches in field of geotechnical engineering","authors":["Elaheh Yaghoubi","Elnaz Yaghoubi","Ahmed Khamees","Amir Hossein Vakili"],"abstract":"Artificial neural networks (ANN), machine learning (ML), deep learning (DL), and ensemble learning (EL) are four outstanding approaches that enable algorithms to extract information from data and make predictions or decisions autonomously without the need for direct instructions. ANN, ML, DL, and EL models have found extensive application in predicting geotechnical and geoenvironmental parameters. This research aims to provide a comprehensive assessment of the applications of ANN, ML, DL, and EL in addressing forecasting within the field related to geotechnical engineering, including soil mechanics, foundation engineering, rock mechanics, environmental geotechnics, and transportation geotechnics. Previous studies have not collectively examined all four algorithms\u2014ANN, ML, DL, and EL\u2014and have not explored their advantages and disadvantages in the field of geotechnical engineering. This research aims to categorize and address this gap in the existing literature systematically. An extensive dataset of relevant research studies was gathered from the Web of Science and subjected to an analysis based on their approach, primary focus and objectives, year of publication, geographical distribution, and results. Additionally, this study included a co-occurrence keyword analysis that covered ANN, ML, DL, and EL techniques, systematic reviews, geotechnical engineering, and review articles that the data, sourced from the Scopus database through the Elsevier Journal, were then visualized using VOS Viewer for further examination. The results demonstrated that ANN is widely utilized despite the proven potential of ML, DL, and EL methods in geotechnical engineering due to the need for real-world laboratory data that civil and geotechnical engineers often encounter. However, when it comes to predicting behavior in geotechnical scenarios, EL techniques outperform all three other methods. Additionally, the techniques discussed here assist geotechnical engineering in understanding the benefits and disadvantages of ANN, ML, DL, and EL within the geo techniques area. This understanding enables geotechnical practitioners to select the most suitable techniques for creating a certainty and resilient ecosystem.","doi":"10.1007\/s00521-024-09893-7","cleaned_title":"a systematic review and meta-analysis of artificial neural network, machine learning, deep learning, and ensemble learning approaches in field of geotechnical engineering","cleaned_abstract":"artificial neural networks (ann), machine learning (ml), deep learning (dl), and ensemble learning (el) are four outstanding approaches that enable algorithms to extract information from data and make predictions or decisions autonomously without the need for direct instructions. ann, ml, dl, and el models have found extensive application in predicting geotechnical and geoenvironmental parameters. this research aims to provide a comprehensive assessment of the applications of ann, ml, dl, and el in addressing forecasting within the field related to geotechnical engineering, including soil mechanics, foundation engineering, rock mechanics, environmental geotechnics, and transportation geotechnics. previous studies have not collectively examined all four algorithms\u2014ann, ml, dl, and el\u2014and have not explored their advantages and disadvantages in the field of geotechnical engineering. this research aims to categorize and address this gap in the existing literature systematically. an extensive dataset of relevant research studies was gathered from the web of science and subjected to an analysis based on their approach, primary focus and objectives, year of publication, geographical distribution, and results. additionally, this study included a co-occurrence keyword analysis that covered ann, ml, dl, and el techniques, systematic reviews, geotechnical engineering, and review articles that the data, sourced from the scopus database through the elsevier journal, were then visualized using vos viewer for further examination. the results demonstrated that ann is widely utilized despite the proven potential of ml, dl, and el methods in geotechnical engineering due to the need for real-world laboratory data that civil and geotechnical engineers often encounter. however, when it comes to predicting behavior in geotechnical scenarios, el techniques outperform all three other methods. additionally, the techniques discussed here assist geotechnical engineering in understanding the benefits and disadvantages of ann, ml, dl, and el within the geo techniques area. this understanding enables geotechnical practitioners to select the most suitable techniques for creating a certainty and resilient ecosystem.","key_phrases":["artificial neural networks","ann","machine learning","ml","deep learning","dl","ensemble learning","el","four outstanding approaches","that","algorithms","information","data","predictions","decisions","the need","direct instructions","ann","el models","extensive application","geotechnical and geoenvironmental parameters","this research","a comprehensive assessment","the applications","ann","ml","dl","el","forecasting","the field","geotechnical engineering","soil mechanics","foundation engineering","rock mechanics","environmental geotechnics","transportation geotechnics","previous studies","all four algorithms","ann","ml","dl","el","their advantages","disadvantages","the field","geotechnical engineering","this research","this gap","the existing literature","an extensive dataset","relevant research studies","the web","science","an analysis","their approach","primary focus","objectives","year","publication","geographical distribution","results","this study","a co-occurrence keyword analysis","that","ann","ml","dl","el techniques","systematic reviews","geotechnical engineering","articles","the data","the scopus database","the elsevier journal","vos viewer","further examination","the results","ann","the proven potential","ml","dl","el methods","geotechnical engineering","the need","real-world laboratory data","that","civil and geotechnical engineers","it","behavior","geotechnical scenarios","el techniques","all three other methods","the techniques","geotechnical engineering","the benefits","disadvantages","ann","ml","dl","el","the geo techniques area","this understanding","geotechnical practitioners","the most suitable techniques","a certainty and resilient ecosystem","four","ann","el models","four","el techniques","el techniques","three"]},{"title":"HCR-Net: a deep learning based script independent handwritten character recognition network","authors":["Vinod Kumar Chauhan","Sukhdeep Singh","Anuj Sharma"],"abstract":"Handwritten character recognition (HCR) remains a challenging pattern recognition problem despite decades of research, and lacks research on script independent recognition techniques. This is mainly because of similar character structures, different handwriting styles, diverse scripts, handcrafted feature extraction techniques, unavailability of data and code, and the development of script-specific deep learning techniques. To address these limitations, we have proposed a script independent deep learning network for HCR research, called HCR-Net, that sets a new research direction for the field. HCR-Net is based on a novel transfer learning approach for HCR, which partly utilizes feature extraction layers of a pre-trained network. Due to transfer learning and image augmentation, HCR-Net provides faster and computationally efficient training, better performance and generalizations, and can work with small datasets. HCR-Net is extensively evaluated on 40 publicly available datasets of Bangla, Punjabi, Hindi, English, Swedish, Urdu, Farsi, Tibetan, Kannada, Malayalam, Telugu, Marathi, Nepali and Arabic languages, and established 26 new benchmark results while performed close to the best results in the rest cases. HCR-Net showed performance improvements up to 11% against the existing results and achieved a fast convergence rate showing up to 99% of final performance in the very first epoch. HCR-Net significantly outperformed the state-of-the-art transfer learning techniques and also reduced the number of trainable parameters by 34% as compared with the corresponding pre-trained network. To facilitate reproducibility and further advancements of HCR research, the complete code is publicly released at https:\/\/github.com\/jmdvinodjmd\/HCR-Net.","doi":"10.1007\/s11042-024-18655-5","cleaned_title":"hcr-net: a deep learning based script independent handwritten character recognition network","cleaned_abstract":"handwritten character recognition (hcr) remains a challenging pattern recognition problem despite decades of research, and lacks research on script independent recognition techniques. this is mainly because of similar character structures, different handwriting styles, diverse scripts, handcrafted feature extraction techniques, unavailability of data and code, and the development of script-specific deep learning techniques. to address these limitations, we have proposed a script independent deep learning network for hcr research, called hcr-net, that sets a new research direction for the field. hcr-net is based on a novel transfer learning approach for hcr, which partly utilizes feature extraction layers of a pre-trained network. due to transfer learning and image augmentation, hcr-net provides faster and computationally efficient training, better performance and generalizations, and can work with small datasets. hcr-net is extensively evaluated on 40 publicly available datasets of bangla, punjabi, hindi, english, swedish, urdu, farsi, tibetan, kannada, malayalam, telugu, marathi, nepali and arabic languages, and established 26 new benchmark results while performed close to the best results in the rest cases. hcr-net showed performance improvements up to 11% against the existing results and achieved a fast convergence rate showing up to 99% of final performance in the very first epoch. hcr-net significantly outperformed the state-of-the-art transfer learning techniques and also reduced the number of trainable parameters by 34% as compared with the corresponding pre-trained network. to facilitate reproducibility and further advancements of hcr research, the complete code is publicly released at https:\/\/github.com\/jmdvinodjmd\/hcr-net.","key_phrases":["handwritten character recognition","hcr","a challenging pattern recognition problem","decades","research","research","script independent recognition techniques","this","similar character structures","different handwriting styles","diverse scripts","handcrafted feature extraction techniques","unavailability","data","code","the development","script-specific deep learning techniques","these limitations","we","a script independent deep learning network","hcr research","hcr","-","net","that","a new research direction","the field","hcr","net","a novel transfer learning approach","hcr","which","feature extraction layers","a pre-trained network","learning","image augmentation","hcr-net","faster and computationally efficient training","better performance","generalizations","small datasets","hcr","net","40 publicly available datasets","bangla","punjabi","hindi","english","urdu","farsi","tibetan","kannada","malayalam","telugu","marathi","nepali","arabic languages","26 new benchmark results","the best results","the rest cases","net","performance improvements","the existing results","a fast convergence rate","up to 99%","final performance","the very first epoch","hcr","net","the-art","techniques","the number","trainable parameters","34%","the corresponding pre-trained network","reproducibility","further advancements","hcr research","the complete code","https:\/\/github.com\/jmdvinodjmd\/hcr-net","decades","40","english","swedish","tibetan","kannada","malayalam","arabic","26","11%","up to 99%","first","34%"]},{"title":"A review of deep learning approaches in clinical and healthcare systems based on medical image analysis","authors":["Hadeer A. Helaly","Mahmoud Badawy","Amira Y. Haikal"],"abstract":"Healthcare is a high-priority sector where people expect the highest levels of care and service, regardless of cost. That makes it distinct from other sectors. Due to the promising results of deep learning in other practical applications, many deep learning algorithms have been proposed for use in healthcare and to solve traditional artificial intelligence issues. The main objective of this study is to review and analyze current deep learning algorithms in healthcare systems. In addition, it highlights the contributions and limitations of recent research papers. It combines deep learning methods with the interpretability of human healthcare by providing insights into deep learning applications in healthcare solutions. It first provides an overview of several deep learning models and their most recent developments. It then briefly examines how these models are applied in several medical practices. Finally, it summarizes current trends and issues in the design and training of deep neural networks besides the future direction in this field.","doi":"10.1007\/s11042-023-16605-1","cleaned_title":"a review of deep learning approaches in clinical and healthcare systems based on medical image analysis","cleaned_abstract":"healthcare is a high-priority sector where people expect the highest levels of care and service, regardless of cost. that makes it distinct from other sectors. due to the promising results of deep learning in other practical applications, many deep learning algorithms have been proposed for use in healthcare and to solve traditional artificial intelligence issues. the main objective of this study is to review and analyze current deep learning algorithms in healthcare systems. in addition, it highlights the contributions and limitations of recent research papers. it combines deep learning methods with the interpretability of human healthcare by providing insights into deep learning applications in healthcare solutions. it first provides an overview of several deep learning models and their most recent developments. it then briefly examines how these models are applied in several medical practices. finally, it summarizes current trends and issues in the design and training of deep neural networks besides the future direction in this field.","key_phrases":["healthcare","a high-priority sector","people","the highest levels","care","service","cost","that","it","other sectors","the promising results","deep learning","other practical applications","many deep learning algorithms","use","healthcare","traditional artificial intelligence issues","the main objective","this study","current deep learning algorithms","healthcare systems","addition","it","the contributions","limitations","recent research papers","it","deep learning methods","the interpretability","human healthcare","insights","deep learning applications","healthcare solutions","it","an overview","several deep learning models","their most recent developments","it","these models","several medical practices","it","current trends","issues","the design","training","deep neural networks","the future direction","this field","healthcare","first"]},{"title":"Automated multifocus pollen detection using deep learning","authors":["Ram\u00f3n Gallardo","Carlos J. Garc\u00eda-Orellana","Horacio M. Gonz\u00e1lez-Velasco","Antonio Garc\u00eda-Manso","Rafael Tormo-Molina","Miguel Mac\u00edas-Mac\u00edas","Eugenio Abeng\u00f3zar"],"abstract":"Pollen-induced allergies affect a significant part of the population in developed countries. Current palynological analysis in Europe is a slow and laborious process which provides pollen information in a weekly-cycle basis. In this paper, we describe a system that allows to locate and classify, in a single step, the pollen grains present in standard glass microscope slides. Besides, processing the samples in the z-axis allows us to increase the probability of detecting grains compared to solutions based on one image per sample. Our system has been trained to recognise 11 pollen types, achieving 97.6\u00a0% success rate locating grains, of which 96.3\u00a0% are also correctly identified (0.956 macro\u2013F1 score), and with a 2.4\u00a0% grains lost. Our results indicate that deep learning provides a robust framework to address automated identification of various pollen types, facilitating their daily measurement.","doi":"10.1007\/s11042-024-18450-2","cleaned_title":"automated multifocus pollen detection using deep learning","cleaned_abstract":"pollen-induced allergies affect a significant part of the population in developed countries. current palynological analysis in europe is a slow and laborious process which provides pollen information in a weekly-cycle basis. in this paper, we describe a system that allows to locate and classify, in a single step, the pollen grains present in standard glass microscope slides. besides, processing the samples in the z-axis allows us to increase the probability of detecting grains compared to solutions based on one image per sample. our system has been trained to recognise 11 pollen types, achieving 97.6 % success rate locating grains, of which 96.3 % are also correctly identified (0.956 macro\u2013f1 score), and with a 2.4 % grains lost. our results indicate that deep learning provides a robust framework to address automated identification of various pollen types, facilitating their daily measurement.","key_phrases":["pollen-induced allergies","a significant part","the population","developed countries","current palynological analysis","europe","a slow and laborious process","which","pollen information","a weekly-cycle basis","this paper","we","a system","that","a single step","the pollen grains","standard glass microscope","the samples","the z-axis","us","the probability","grains","solutions","one image","sample","our system","11 pollen types","97.6 % success rate locating grains","which","96.3 %","0.956 macro","f1 score","a 2.4 % grains","our results","deep learning","a robust framework","automated identification","various pollen types","their daily measurement","europe","weekly","one","11","97.6 %","96.3 %","0.956","2.4 %","daily"]},{"title":"Cataract-1K Dataset for Deep-Learning-Assisted Analysis of Cataract Surgery Videos","authors":["Negin Ghamsarian","Yosuf El-Shabrawi","Sahar Nasirihaghighi","Doris Putzgruber-Adamitsch","Martin Zinkernagel","Sebastian Wolf","Klaus Schoeffmann","Raphael Sznitman"],"abstract":"In recent years, the landscape of computer-assisted interventions and post-operative surgical video analysis has been dramatically reshaped by deep-learning techniques, resulting in significant advancements in surgeons\u2019 skills, operation room management, and overall surgical outcomes. However, the progression of deep-learning-powered surgical technologies is profoundly reliant on large-scale datasets and annotations. In particular, surgical scene understanding and phase recognition stand as pivotal pillars within the realm of computer-assisted surgery and post-operative assessment of cataract surgery videos. In this context, we present the largest cataract surgery video dataset that addresses diverse requisites for constructing computerized surgical workflow analysis and detecting post-operative irregularities in cataract surgery. We validate the quality of annotations by benchmarking the performance of several state-of-the-art neural network architectures for phase recognition and surgical scene segmentation. Besides, we initiate the research on domain adaptation for instrument segmentation in cataract surgery by evaluating cross-domain instrument segmentation performance in cataract surgery videos. The dataset and annotations are publicly available in Synapse.","doi":"10.1038\/s41597-024-03193-4","cleaned_title":"cataract-1k dataset for deep-learning-assisted analysis of cataract surgery videos","cleaned_abstract":"in recent years, the landscape of computer-assisted interventions and post-operative surgical video analysis has been dramatically reshaped by deep-learning techniques, resulting in significant advancements in surgeons\u2019 skills, operation room management, and overall surgical outcomes. however, the progression of deep-learning-powered surgical technologies is profoundly reliant on large-scale datasets and annotations. in particular, surgical scene understanding and phase recognition stand as pivotal pillars within the realm of computer-assisted surgery and post-operative assessment of cataract surgery videos. in this context, we present the largest cataract surgery video dataset that addresses diverse requisites for constructing computerized surgical workflow analysis and detecting post-operative irregularities in cataract surgery. we validate the quality of annotations by benchmarking the performance of several state-of-the-art neural network architectures for phase recognition and surgical scene segmentation. besides, we initiate the research on domain adaptation for instrument segmentation in cataract surgery by evaluating cross-domain instrument segmentation performance in cataract surgery videos. the dataset and annotations are publicly available in synapse.","key_phrases":["recent years","the landscape","computer-assisted interventions","post-operative surgical video analysis","deep-learning techniques","significant advancements","surgeons\u2019 skills","operation room management","overall surgical outcomes","the progression","deep-learning-powered surgical technologies","large-scale datasets","annotations","surgical scene understanding","phase recognition","pivotal pillars","the realm","computer-assisted surgery","post-operative assessment","cataract surgery videos","this context","we","the largest cataract surgery video dataset","that","diverse requisites","computerized surgical workflow analysis","post-operative irregularities","cataract surgery","we","the quality","annotations","the performance","the-art","phase recognition","surgical scene segmentation","we","the research","domain adaptation","instrument segmentation","cataract surgery","cross-domain instrument segmentation performance","cataract surgery videos","the dataset","annotations","synapse","recent years"]},{"title":"A systematic review and meta-analysis of artificial neural network, machine learning, deep learning, and ensemble learning approaches in field of geotechnical engineering","authors":["Elaheh Yaghoubi","Elnaz Yaghoubi","Ahmed Khamees","Amir Hossein Vakili"],"abstract":"Artificial neural networks (ANN), machine learning (ML), deep learning (DL), and ensemble learning (EL) are four outstanding approaches that enable algorithms to extract information from data and make predictions or decisions autonomously without the need for direct instructions. ANN, ML, DL, and EL models have found extensive application in predicting geotechnical and geoenvironmental parameters. This research aims to provide a comprehensive assessment of the applications of ANN, ML, DL, and EL in addressing forecasting within the field related to geotechnical engineering, including soil mechanics, foundation engineering, rock mechanics, environmental geotechnics, and transportation geotechnics. Previous studies have not collectively examined all four algorithms\u2014ANN, ML, DL, and EL\u2014and have not explored their advantages and disadvantages in the field of geotechnical engineering. This research aims to categorize and address this gap in the existing literature systematically. An extensive dataset of relevant research studies was gathered from the Web of Science and subjected to an analysis based on their approach, primary focus and objectives, year of publication, geographical distribution, and results. Additionally, this study included a co-occurrence keyword analysis that covered ANN, ML, DL, and EL techniques, systematic reviews, geotechnical engineering, and review articles that the data, sourced from the Scopus database through the Elsevier Journal, were then visualized using VOS Viewer for further examination. The results demonstrated that ANN is widely utilized despite the proven potential of ML, DL, and EL methods in geotechnical engineering due to the need for real-world laboratory data that civil and geotechnical engineers often encounter. However, when it comes to predicting behavior in geotechnical scenarios, EL techniques outperform all three other methods. Additionally, the techniques discussed here assist geotechnical engineering in understanding the benefits and disadvantages of ANN, ML, DL, and EL within the geo techniques area. This understanding enables geotechnical practitioners to select the most suitable techniques for creating a certainty and resilient ecosystem.","doi":"10.1007\/s00521-024-09893-7","cleaned_title":"a systematic review and meta-analysis of artificial neural network, machine learning, deep learning, and ensemble learning approaches in field of geotechnical engineering","cleaned_abstract":"artificial neural networks (ann), machine learning (ml), deep learning (dl), and ensemble learning (el) are four outstanding approaches that enable algorithms to extract information from data and make predictions or decisions autonomously without the need for direct instructions. ann, ml, dl, and el models have found extensive application in predicting geotechnical and geoenvironmental parameters. this research aims to provide a comprehensive assessment of the applications of ann, ml, dl, and el in addressing forecasting within the field related to geotechnical engineering, including soil mechanics, foundation engineering, rock mechanics, environmental geotechnics, and transportation geotechnics. previous studies have not collectively examined all four algorithms\u2014ann, ml, dl, and el\u2014and have not explored their advantages and disadvantages in the field of geotechnical engineering. this research aims to categorize and address this gap in the existing literature systematically. an extensive dataset of relevant research studies was gathered from the web of science and subjected to an analysis based on their approach, primary focus and objectives, year of publication, geographical distribution, and results. additionally, this study included a co-occurrence keyword analysis that covered ann, ml, dl, and el techniques, systematic reviews, geotechnical engineering, and review articles that the data, sourced from the scopus database through the elsevier journal, were then visualized using vos viewer for further examination. the results demonstrated that ann is widely utilized despite the proven potential of ml, dl, and el methods in geotechnical engineering due to the need for real-world laboratory data that civil and geotechnical engineers often encounter. however, when it comes to predicting behavior in geotechnical scenarios, el techniques outperform all three other methods. additionally, the techniques discussed here assist geotechnical engineering in understanding the benefits and disadvantages of ann, ml, dl, and el within the geo techniques area. this understanding enables geotechnical practitioners to select the most suitable techniques for creating a certainty and resilient ecosystem.","key_phrases":["artificial neural networks","ann","machine learning","ml","deep learning","dl","ensemble learning","el","four outstanding approaches","that","algorithms","information","data","predictions","decisions","the need","direct instructions","ann","el models","extensive application","geotechnical and geoenvironmental parameters","this research","a comprehensive assessment","the applications","ann","ml","dl","el","forecasting","the field","geotechnical engineering","soil mechanics","foundation engineering","rock mechanics","environmental geotechnics","transportation geotechnics","previous studies","all four algorithms","ann","ml","dl","el","their advantages","disadvantages","the field","geotechnical engineering","this research","this gap","the existing literature","an extensive dataset","relevant research studies","the web","science","an analysis","their approach","primary focus","objectives","year","publication","geographical distribution","results","this study","a co-occurrence keyword analysis","that","ann","ml","dl","el techniques","systematic reviews","geotechnical engineering","articles","the data","the scopus database","the elsevier journal","vos viewer","further examination","the results","ann","the proven potential","ml","dl","el methods","geotechnical engineering","the need","real-world laboratory data","that","civil and geotechnical engineers","it","behavior","geotechnical scenarios","el techniques","all three other methods","the techniques","geotechnical engineering","the benefits","disadvantages","ann","ml","dl","el","the geo techniques area","this understanding","geotechnical practitioners","the most suitable techniques","a certainty and resilient ecosystem","four","ann","el models","four","el techniques","el techniques","three"]},{"title":"Integrated deep learning approach for automatic coronary artery segmentation and classification on computed tomographic coronary angiography","authors":["Chitra Devi Muthusamy","Ramaswami  Murugesh"],"abstract":"The field of coronary artery disease (CAD) has seen a rapid development in coronary computed tomography angiography (CCTA). However, manual coronary artery tree segmentation and reconstruction take time and effort. Deep learning algorithms have been created effectively to analyze large amounts of data for medical image analysis. The primary goal of this research\u00a0is to create an automated CAD diagnostic model and a deep learning tool for automatic coronary artery reconstruction\u00a0using a large, single-center retrospective CCTA cohort. We propose an integrated deep learning-based intelligent system for human heart blood vessel position within heart coronary CT angiography images using a multi-class ensemble classification mechanism in this research. The Modified DenseNet201 will segment the cardiac blood vessels in the proposed work. Then, the low-level features are successfully extracted using the ResNet-152 model. Finally, the improved deep residual shrinkage network (IDRSN) model classifies heart blood vessels into\u00a0four distinct classes: normal, block, narrow, and blood flow-reduced. The CCTA image dataset is used for the experiment analysis. The integrated deep learning-based blood vessel segmentation and classification system (Modified DenseNet201-IDRSN) was developed using the Python tool, and the system\u2019s efficiency was determined using different performance metrics. The experimental results of this research demonstrate that the proposed system is effective and efficient concerning related studies. Compared to the U-Net segmentation model, the segmentation result\u00a0in the proposed\u00a0research\u00a0is smoother and nearest to the segmentation result of the human expert. The proposed integrated deep learning intelligent system improves the efficiency of disease diagnosis, reduces dependence on medical personnel, reduces manual interaction in diagnosis, and offers auxiliary strategies for subsequent medical diagnosis systems based on cardiac coronary angiography.","doi":"10.1007\/s13721-024-00473-2","cleaned_title":"integrated deep learning approach for automatic coronary artery segmentation and classification on computed tomographic coronary angiography","cleaned_abstract":"the field of coronary artery disease (cad) has seen a rapid development in coronary computed tomography angiography (ccta). however, manual coronary artery tree segmentation and reconstruction take time and effort. deep learning algorithms have been created effectively to analyze large amounts of data for medical image analysis. the primary goal of this research is to create an automated cad diagnostic model and a deep learning tool for automatic coronary artery reconstruction using a large, single-center retrospective ccta cohort. we propose an integrated deep learning-based intelligent system for human heart blood vessel position within heart coronary ct angiography images using a multi-class ensemble classification mechanism in this research. the modified densenet201 will segment the cardiac blood vessels in the proposed work. then, the low-level features are successfully extracted using the resnet-152 model. finally, the improved deep residual shrinkage network (idrsn) model classifies heart blood vessels into four distinct classes: normal, block, narrow, and blood flow-reduced. the ccta image dataset is used for the experiment analysis. the integrated deep learning-based blood vessel segmentation and classification system (modified densenet201-idrsn) was developed using the python tool, and the system\u2019s efficiency was determined using different performance metrics. the experimental results of this research demonstrate that the proposed system is effective and efficient concerning related studies. compared to the u-net segmentation model, the segmentation result in the proposed research is smoother and nearest to the segmentation result of the human expert. the proposed integrated deep learning intelligent system improves the efficiency of disease diagnosis, reduces dependence on medical personnel, reduces manual interaction in diagnosis, and offers auxiliary strategies for subsequent medical diagnosis systems based on cardiac coronary angiography.","key_phrases":["the field","coronary artery disease","cad","a rapid development","coronary computed tomography angiography","ccta","manual coronary artery tree segmentation","reconstruction","time","effort","deep learning algorithms","large amounts","data","medical image analysis","the primary goal","this research","an automated cad diagnostic model","a deep learning tool","automatic coronary artery reconstruction","a large, single-center retrospective ccta cohort","we","an integrated deep learning-based intelligent system","human heart blood vessel position","heart coronary ct angiography images","a multi-class ensemble classification mechanism","this research","the modified densenet201","the cardiac blood vessels","the proposed work","the low-level features","the resnet-152 model","the improved deep residual shrinkage network","four distinct classes","the ccta image dataset","the experiment analysis","the integrated deep learning-based blood vessel segmentation and classification system","modified densenet201-idrsn","the python tool","the system\u2019s efficiency","different performance metrics","the experimental results","this research demonstrate","the proposed system","related studies","the u-net segmentation model","the segmentation","the proposed research","the segmentation result","the human expert","the proposed integrated deep learning intelligent system","the efficiency","disease diagnosis","dependence","medical personnel","manual interaction","diagnosis","auxiliary strategies","subsequent medical diagnosis systems","cardiac coronary angiography","resnet-152","four"]},{"title":"Can deep learning replace histopathological examinations in the differential diagnosis of cervical lymphadenopathy?","authors":["Sermin Can","\u00d6mer T\u00fcrk","Muhammed Ayral","G\u00fcnay Kozan","Hamza Ar\u0131","Mehmet Akda\u011f","M\u00fczeyyen Y\u0131ld\u0131r\u0131m Baylan"],"abstract":"IntroductionWe aimed to develop a diagnostic deep learning model using contrast-enhanced CT images and to investigate whether cervical lymphadenopathies can be diagnosed with these deep learning methods without radiologist interpretations and histopathological examinations.Material methodA total of 400 patients who underwent surgery for lymphadenopathy in the neck between 2010 and 2022 were retrospectively analyzed. They were examined in four groups of 100 patients: the granulomatous diseases group, the lymphoma group, the squamous cell tumor group, and the reactive hyperplasia group. The diagnoses of the patients were confirmed histopathologically. Two CT images from all the patients in each group were used in the study. The CT images were classified using ResNet50, NASNetMobile, and DenseNet121 architecture input.ResultsThe classification accuracies obtained with ResNet50, DenseNet121, and NASNetMobile were 92.5%, 90.62, and 87.5, respectively.ConclusionDeep learning is a useful diagnostic tool in diagnosing cervical lymphadenopathy. In the near future, many diseases could be diagnosed with deep learning models without radiologist interpretations and invasive examinations such as histopathological examinations. However, further studies with much larger case series are needed to develop accurate deep-learning models.","doi":"10.1007\/s00405-023-08181-9","cleaned_title":"can deep learning replace histopathological examinations in the differential diagnosis of cervical lymphadenopathy?","cleaned_abstract":"introductionwe aimed to develop a diagnostic deep learning model using contrast-enhanced ct images and to investigate whether cervical lymphadenopathies can be diagnosed with these deep learning methods without radiologist interpretations and histopathological examinations.material methoda total of 400 patients who underwent surgery for lymphadenopathy in the neck between 2010 and 2022 were retrospectively analyzed. they were examined in four groups of 100 patients: the granulomatous diseases group, the lymphoma group, the squamous cell tumor group, and the reactive hyperplasia group. the diagnoses of the patients were confirmed histopathologically. two ct images from all the patients in each group were used in the study. the ct images were classified using resnet50, nasnetmobile, and densenet121 architecture input.resultsthe classification accuracies obtained with resnet50, densenet121, and nasnetmobile were 92.5%, 90.62, and 87.5, respectively.conclusiondeep learning is a useful diagnostic tool in diagnosing cervical lymphadenopathy. in the near future, many diseases could be diagnosed with deep learning models without radiologist interpretations and invasive examinations such as histopathological examinations. however, further studies with much larger case series are needed to develop accurate deep-learning models.","key_phrases":["introductionwe","a diagnostic deep learning model","contrast-enhanced ct images","cervical lymphadenopathies","these deep learning methods","radiologist interpretations","histopathological examinations.material methoda total","400 patients","who","surgery","lymphadenopathy","the neck","they","four groups","100 patients","the granulomatous diseases group","the lymphoma group","the squamous cell tumor group","the reactive hyperplasia group","the diagnoses","the patients","two ct images","all the patients","each group","the study","the ct images","resnet50","nasnetmobile","architecture","input.resultsthe classification accuracies","resnet50","densenet121","92.5%","respectively.conclusiondeep learning","a useful diagnostic tool","cervical lymphadenopathy","the near future","many diseases","deep learning models","radiologist interpretations","invasive examinations","histopathological examinations","further studies","much larger case series","accurate deep-learning models","400","between 2010 and 2022","four","100","two","resnet50","resnet50","92.5%","90.62","87.5"]},{"title":"Survey on deep learning in multimodal medical imaging for cancer detection","authors":["Yan Tian","Zhaocheng Xu","Yujun Ma","Weiping Ding","Ruili Wang","Zhihong Gao","Guohua Cheng","Linyang He","Xuran Zhao"],"abstract":"The task of multimodal cancer detection is to determine the locations and categories of lesions by using different imaging techniques, which is one of the key research methods for cancer diagnosis. Recently, deep learning-based object detection has made significant developments due to its strength in semantic feature extraction and nonlinear function fitting. However, multimodal cancer detection remains challenging due to morphological differences in lesions, interpatient variability, difficulty in annotation, and imaging artifacts. In this survey, we mainly investigate over 150 papers in recent years with respect to multimodal cancer detection using deep learning, with a focus on datasets and solutions to various challenges such as data annotation, variance between classes, small-scale lesions, and occlusion. We also provide an overview of the advantages and drawbacks of each approach. Finally, we discuss the current scope of work and provide directions for the future development of multimodal cancer detection.","doi":"10.1007\/s00521-023-09214-4","cleaned_title":"survey on deep learning in multimodal medical imaging for cancer detection","cleaned_abstract":"the task of multimodal cancer detection is to determine the locations and categories of lesions by using different imaging techniques, which is one of the key research methods for cancer diagnosis. recently, deep learning-based object detection has made significant developments due to its strength in semantic feature extraction and nonlinear function fitting. however, multimodal cancer detection remains challenging due to morphological differences in lesions, interpatient variability, difficulty in annotation, and imaging artifacts. in this survey, we mainly investigate over 150 papers in recent years with respect to multimodal cancer detection using deep learning, with a focus on datasets and solutions to various challenges such as data annotation, variance between classes, small-scale lesions, and occlusion. we also provide an overview of the advantages and drawbacks of each approach. finally, we discuss the current scope of work and provide directions for the future development of multimodal cancer detection.","key_phrases":["the task","multimodal cancer detection","the locations","categories","lesions","different imaging techniques","which","the key research methods","cancer diagnosis","deep learning-based object detection","significant developments","its strength","semantic feature extraction and nonlinear function","multimodal cancer detection","morphological differences","lesions","interpatient variability","difficulty","annotation","imaging artifacts","this survey","we","over 150 papers","recent years","respect","multimodal cancer detection","deep learning","a focus","datasets","solutions","various challenges","data annotation","variance","classes","small-scale lesions","occlusion","we","an overview","the advantages","drawbacks","each approach","we","the current scope","work","directions","the future development","multimodal cancer detection","150","recent years"]},{"title":"A real-time traffic sign detection in intelligent transportation system using YOLOv8-based deep learning approach","authors":["Mingdeng Tang"],"abstract":"Intelligent transportation systems rely heavily on accurate traffic sign detection (TSD) to enhance road safety and traffic management. Various methods have been explored in the literature for this purpose, with deep learning methods consistently demonstrating superior accuracy. However, existing research highlights the persistent challenge of achieving high accuracy rates while maintaining non-destructive and real-time requirements. In this study, we propose a deep learning model based on the YOLOv8 architecture to address this challenge. The model is trained and evaluated using a custom dataset, and extensive experiments and performance analysis demonstrate its ability to achieve precise results, thus offering a promising solution to the current research challenge in deep learning-based TSD.","doi":"10.1007\/s11760-024-03300-3","cleaned_title":"a real-time traffic sign detection in intelligent transportation system using yolov8-based deep learning approach","cleaned_abstract":"intelligent transportation systems rely heavily on accurate traffic sign detection (tsd) to enhance road safety and traffic management. various methods have been explored in the literature for this purpose, with deep learning methods consistently demonstrating superior accuracy. however, existing research highlights the persistent challenge of achieving high accuracy rates while maintaining non-destructive and real-time requirements. in this study, we propose a deep learning model based on the yolov8 architecture to address this challenge. the model is trained and evaluated using a custom dataset, and extensive experiments and performance analysis demonstrate its ability to achieve precise results, thus offering a promising solution to the current research challenge in deep learning-based tsd.","key_phrases":["intelligent transportation systems","accurate traffic sign detection","(tsd","road safety and traffic management","various methods","the literature","this purpose","deep learning methods","superior accuracy","the persistent challenge","high accuracy rates","-destructive and real-time requirements","this study","we","a deep learning model","the yolov8 architecture","this challenge","the model","a custom dataset","extensive experiments","performance analysis","its ability","precise results","a promising solution","the current research challenge","deep learning-based tsd","yolov8"]},{"title":"Cataract-1K Dataset for Deep-Learning-Assisted Analysis of Cataract Surgery Videos","authors":["Negin Ghamsarian","Yosuf El-Shabrawi","Sahar Nasirihaghighi","Doris Putzgruber-Adamitsch","Martin Zinkernagel","Sebastian Wolf","Klaus Schoeffmann","Raphael Sznitman"],"abstract":"In recent years, the landscape of computer-assisted interventions and post-operative surgical video analysis has been dramatically reshaped by deep-learning techniques, resulting in significant advancements in surgeons\u2019 skills, operation room management, and overall surgical outcomes. However, the progression of deep-learning-powered surgical technologies is profoundly reliant on large-scale datasets and annotations. In particular, surgical scene understanding and phase recognition stand as pivotal pillars within the realm of computer-assisted surgery and post-operative assessment of cataract surgery videos. In this context, we present the largest cataract surgery video dataset that addresses diverse requisites for constructing computerized surgical workflow analysis and detecting post-operative irregularities in cataract surgery. We validate the quality of annotations by benchmarking the performance of several state-of-the-art neural network architectures for phase recognition and surgical scene segmentation. Besides, we initiate the research on domain adaptation for instrument segmentation in cataract surgery by evaluating cross-domain instrument segmentation performance in cataract surgery videos. The dataset and annotations are publicly available in Synapse.","doi":"10.1038\/s41597-024-03193-4","cleaned_title":"cataract-1k dataset for deep-learning-assisted analysis of cataract surgery videos","cleaned_abstract":"in recent years, the landscape of computer-assisted interventions and post-operative surgical video analysis has been dramatically reshaped by deep-learning techniques, resulting in significant advancements in surgeons\u2019 skills, operation room management, and overall surgical outcomes. however, the progression of deep-learning-powered surgical technologies is profoundly reliant on large-scale datasets and annotations. in particular, surgical scene understanding and phase recognition stand as pivotal pillars within the realm of computer-assisted surgery and post-operative assessment of cataract surgery videos. in this context, we present the largest cataract surgery video dataset that addresses diverse requisites for constructing computerized surgical workflow analysis and detecting post-operative irregularities in cataract surgery. we validate the quality of annotations by benchmarking the performance of several state-of-the-art neural network architectures for phase recognition and surgical scene segmentation. besides, we initiate the research on domain adaptation for instrument segmentation in cataract surgery by evaluating cross-domain instrument segmentation performance in cataract surgery videos. the dataset and annotations are publicly available in synapse.","key_phrases":["recent years","the landscape","computer-assisted interventions","post-operative surgical video analysis","deep-learning techniques","significant advancements","surgeons\u2019 skills","operation room management","overall surgical outcomes","the progression","deep-learning-powered surgical technologies","large-scale datasets","annotations","surgical scene understanding","phase recognition","pivotal pillars","the realm","computer-assisted surgery","post-operative assessment","cataract surgery videos","this context","we","the largest cataract surgery video dataset","that","diverse requisites","computerized surgical workflow analysis","post-operative irregularities","cataract surgery","we","the quality","annotations","the performance","the-art","phase recognition","surgical scene segmentation","we","the research","domain adaptation","instrument segmentation","cataract surgery","cross-domain instrument segmentation performance","cataract surgery videos","the dataset","annotations","synapse","recent years"]},{"title":"Analyzing and identifying predictable time range for stress prediction based on chaos theory and deep learning","authors":["Ningyun Li","Huijun Zhang","Ling Feng","Yang Ding","Haichuan Li"],"abstract":"ProposeStress is a common problem globally. Prediction of stress in advance could help people take effective measures to manage stress before bad consequences occur. Considering the chaotic features of human psychological states, in this study, we integrate deep learning and chaos theory to address the stress prediction problem.MethodsBased on chaos theory, we embed one\u2019s seemingly disordered stress sequence into a high dimensional phase space so as to reveal the underlying dynamics and patterns of the stress system, and meanwhile are able to identify the stress predictable time range. We then conduct deep learning with a two-layer (dimension and temporal) attention mechanism to simulate the nonlinear state of the embedded stress sequence for stress prediction.ResultsWe validate the effectiveness of the proposed method on the public available Tesserae dataset. The experimental results show that the proposed method outperforms the pure deep learning method and Chaos method in both 2-label and 3-label stress prediction.ConclusionIntegrating deep learning and chaos theory for stress prediction is effective, and can improve the prediction accuracy over 2% and 8% more than those of the deep learning and the Chaos method respectively. Implications and further possible improvements are also discussed at the end of the paper.","doi":"10.1007\/s13755-024-00280-z","cleaned_title":"analyzing and identifying predictable time range for stress prediction based on chaos theory and deep learning","cleaned_abstract":"proposestress is a common problem globally. prediction of stress in advance could help people take effective measures to manage stress before bad consequences occur. considering the chaotic features of human psychological states, in this study, we integrate deep learning and chaos theory to address the stress prediction problem.methodsbased on chaos theory, we embed one\u2019s seemingly disordered stress sequence into a high dimensional phase space so as to reveal the underlying dynamics and patterns of the stress system, and meanwhile are able to identify the stress predictable time range. we then conduct deep learning with a two-layer (dimension and temporal) attention mechanism to simulate the nonlinear state of the embedded stress sequence for stress prediction.resultswe validate the effectiveness of the proposed method on the public available tesserae dataset. the experimental results show that the proposed method outperforms the pure deep learning method and chaos method in both 2-label and 3-label stress prediction.conclusionintegrating deep learning and chaos theory for stress prediction is effective, and can improve the prediction accuracy over 2% and 8% more than those of the deep learning and the chaos method respectively. implications and further possible improvements are also discussed at the end of the paper.","key_phrases":["proposestress","a common problem","prediction","stress","advance","people","effective measures","stress","bad consequences","the chaotic features","human psychological states","this study","we","deep learning","chaos theory","the stress prediction","chaos theory","we","one","stress sequence","a high dimensional phase space","the underlying dynamics","patterns","the stress system","the stress predictable time range","we","deep learning","dimension","the nonlinear state","the embedded stress sequence","stress","prediction.resultswe validate","the effectiveness","the proposed method","the public available tesserae dataset","the experimental results","the proposed method","the pure deep learning method","chaos method","both 2-label and 3-label stress","deep learning","chaos theory","stress prediction","the prediction accuracy","2%","those","the deep learning","the chaos","implications","further possible improvements","the end","the paper","two","2","3","over 2% and 8%"]},{"title":"Arithmetic Circuits, Structured Matrices and (not so) Deep Learning","authors":["Atri Rudra"],"abstract":"This survey presents a necessarily incomplete (and biased) overview of results at the intersection of arithmetic circuit complexity, structured matrices and deep learning. Recently there has been some research activity in replacing unstructured weight matrices in neural networks by structured ones (with the aim of reducing the size of the corresponding deep learning models). Most of this work has been experimental and in this survey, we formalize the research question and show how a recent work that combines arithmetic circuit complexity, structured matrices and deep learning essentially answers this question. This survey is targeted at complexity theorists who might enjoy reading about how tools developed in arithmetic circuit complexity helped design (to the best of our knowledge) a new family of structured matrices, which in turn seem well-suited for applications in deep learning. However, we hope that folks primarily interested in deep learning would also appreciate the connections to complexity theory.","doi":"10.1007\/s00224-022-10112-w","cleaned_title":"arithmetic circuits, structured matrices and (not so) deep learning","cleaned_abstract":"this survey presents a necessarily incomplete (and biased) overview of results at the intersection of arithmetic circuit complexity, structured matrices and deep learning. recently there has been some research activity in replacing unstructured weight matrices in neural networks by structured ones (with the aim of reducing the size of the corresponding deep learning models). most of this work has been experimental and in this survey, we formalize the research question and show how a recent work that combines arithmetic circuit complexity, structured matrices and deep learning essentially answers this question. this survey is targeted at complexity theorists who might enjoy reading about how tools developed in arithmetic circuit complexity helped design (to the best of our knowledge) a new family of structured matrices, which in turn seem well-suited for applications in deep learning. however, we hope that folks primarily interested in deep learning would also appreciate the connections to complexity theory.","key_phrases":["this survey","a necessarily incomplete (and biased) overview","results","the intersection","arithmetic circuit complexity","structured matrices","deep learning","some research activity","unstructured weight matrices","neural networks","structured ones","the aim","the size","the corresponding deep learning models","this work","this survey","we","the research question","a recent work","that","arithmetic circuit complexity","structured matrices","deep learning","this question","this survey","complexity theorists","who","tools","arithmetic circuit complexity","our knowledge","a new family","structured matrices","which","turn","applications","deep learning","we","folks","deep learning","the connections","complexity theory"]},{"title":"Geometric deep learning for molecular property predictions with chemical accuracy across chemical space","authors":["Maarten R. Dobbelaere","Istv\u00e1n Lengyel","Christian V. Stevens","Kevin M. Van Geem"],"abstract":"Chemical engineers heavily rely on precise knowledge of physicochemical properties to model chemical processes. Despite the growing popularity of deep learning, it is only rarely applied for property prediction due to data scarcity and limited accuracy for compounds in industrially-relevant areas of the chemical space. Herein, we present a geometric deep learning framework for predicting gas- and liquid-phase properties based on novel quantum chemical datasets comprising 124,000 molecules. Our findings reveal that the necessity for quantum-chemical information in deep learning models varies significantly depending on the modeled physicochemical property. Specifically, our top-performing geometric model meets the most stringent criteria for \u201cchemically accurate\u201d thermochemistry predictions. We also show that by carefully selecting the appropriate model featurization and evaluating prediction uncertainties, the reliability of the predictions can be strongly enhanced. These insights represent a crucial step towards establishing deep learning as the standard property prediction workflow in both industry and academia.Scientific contributionWe propose a flexible property prediction tool that can handle two-dimensional and three-dimensional molecular information. A thermochemistry prediction methodology that achieves high-level quantum chemistry accuracy for a broad application range is presented. Trained deep learning models and large novel molecular databases of real-world molecules are provided to offer a directly usable and fast property prediction solution to practitioners.","doi":"10.1186\/s13321-024-00895-0","cleaned_title":"geometric deep learning for molecular property predictions with chemical accuracy across chemical space","cleaned_abstract":"chemical engineers heavily rely on precise knowledge of physicochemical properties to model chemical processes. despite the growing popularity of deep learning, it is only rarely applied for property prediction due to data scarcity and limited accuracy for compounds in industrially-relevant areas of the chemical space. herein, we present a geometric deep learning framework for predicting gas- and liquid-phase properties based on novel quantum chemical datasets comprising 124,000 molecules. our findings reveal that the necessity for quantum-chemical information in deep learning models varies significantly depending on the modeled physicochemical property. specifically, our top-performing geometric model meets the most stringent criteria for \u201cchemically accurate\u201d thermochemistry predictions. we also show that by carefully selecting the appropriate model featurization and evaluating prediction uncertainties, the reliability of the predictions can be strongly enhanced. these insights represent a crucial step towards establishing deep learning as the standard property prediction workflow in both industry and academia.scientific contributionwe propose a flexible property prediction tool that can handle two-dimensional and three-dimensional molecular information. a thermochemistry prediction methodology that achieves high-level quantum chemistry accuracy for a broad application range is presented. trained deep learning models and large novel molecular databases of real-world molecules are provided to offer a directly usable and fast property prediction solution to practitioners.","key_phrases":["chemical engineers","precise knowledge","physicochemical properties","chemical processes","the growing popularity","deep learning","it","property prediction","data scarcity","limited accuracy","compounds","industrially-relevant areas","the chemical space","we","a geometric deep learning framework","gas- and liquid-phase properties","novel quantum chemical datasets","124,000 molecules","our findings","the necessity","quantum-chemical information","deep learning models","the modeled physicochemical property","our top-performing geometric model","the most stringent criteria","\u201cchemically accurate\u201d thermochemistry predictions","we","the appropriate model featurization","prediction uncertainties","the reliability","the predictions","these insights","a crucial step","deep learning","the standard property prediction","both industry","academia.scientific contributionwe","a flexible property prediction tool","that","two-dimensional and three-dimensional molecular information","a thermochemistry prediction methodology","that","high-level quantum chemistry accuracy","a broad application range","trained deep learning models","large novel molecular databases","real-world molecules","a directly usable and fast property prediction solution","practitioners","124,000","two","three"]},{"title":"DVNE-DRL: dynamic virtual network embedding algorithm based on deep reinforcement learning","authors":["Xiancui Xiao"],"abstract":"Virtual network embedding (VNE), as the key challenge of network resource management technology, lies in the contradiction between online embedding decision and pursuing long-term average revenue goals. Most of the previous work ignored the dynamics in Virtual Network (VN) modeling, or could not automatically detect the complex and time-varying network state to provide a reasonable network embedding scheme. In view of this, we model a network embedding framework where the topology and resource allocation change dynamically with the number of network users and workload, and then introduce a deep reinforcement learning method to solve the VNE problem. Further, a dynamic virtual network embedding algorithm based on Deep Reinforcement Learning (DRL), named DVNE-DRL, is proposed. In DVNE-DRL, VNE is modeled as a Markov Decision Process (MDP), and then deep learning is introduced to perceive the current network state through historical data and embedded knowledge, while utilizing reinforcement learning decision-making capabilities to implement the network embedding process. In addition, we improve the method of feature extraction and matrix optimization, and consider the characteristics of virtual network and physical network together to alleviate the problem of redundancy and slow convergence. The simulation results show that compared with the existing advanced algorithms, the acceptance rate and average revenue of DVNE-DRL are increased by about 25% and 35%, respectively.","doi":"10.1038\/s41598-023-47195-5","cleaned_title":"dvne-drl: dynamic virtual network embedding algorithm based on deep reinforcement learning","cleaned_abstract":"virtual network embedding (vne), as the key challenge of network resource management technology, lies in the contradiction between online embedding decision and pursuing long-term average revenue goals. most of the previous work ignored the dynamics in virtual network (vn) modeling, or could not automatically detect the complex and time-varying network state to provide a reasonable network embedding scheme. in view of this, we model a network embedding framework where the topology and resource allocation change dynamically with the number of network users and workload, and then introduce a deep reinforcement learning method to solve the vne problem. further, a dynamic virtual network embedding algorithm based on deep reinforcement learning (drl), named dvne-drl, is proposed. in dvne-drl, vne is modeled as a markov decision process (mdp), and then deep learning is introduced to perceive the current network state through historical data and embedded knowledge, while utilizing reinforcement learning decision-making capabilities to implement the network embedding process. in addition, we improve the method of feature extraction and matrix optimization, and consider the characteristics of virtual network and physical network together to alleviate the problem of redundancy and slow convergence. the simulation results show that compared with the existing advanced algorithms, the acceptance rate and average revenue of dvne-drl are increased by about 25% and 35%, respectively.","key_phrases":["virtual network","the key challenge","network resource management technology","the contradiction","online embedding decision","long-term average revenue goals","the previous work","the dynamics","virtual network","modeling","the complex and time-varying network state","a reasonable network","scheme","view","this","we","a network","framework","the number","network users","workload","a deep reinforcement learning method","the vne problem","a dynamic virtual network","algorithm","deep reinforcement learning","drl","dvne-drl","dvne-drl","vne","a markov decision process","mdp","deep learning","the current network state","historical data","embedded knowledge","reinforcement learning decision-making capabilities","the network embedding process","addition","we","the method","feature extraction and matrix optimization","the characteristics","virtual network","physical network","the problem","redundancy","slow convergence","the simulation results","the existing advanced algorithms","the acceptance rate","average revenue","dvne-drl","about 25%","35%","about 25% and 35%"]},{"title":"Deep learning-enabled segmentation of ambiguous bioimages with deepflash2","authors":["Matthias Griebel","Dennis Segebarth","Nikolai Stein","Nina Schukraft","Philip Tovote","Robert Blum","Christoph M. Flath"],"abstract":"Bioimages frequently exhibit low signal-to-noise ratios due to experimental conditions, specimen characteristics, and imaging trade-offs. Reliable segmentation of such ambiguous images is difficult and laborious. Here we introduce deepflash2, a deep learning-enabled segmentation tool for bioimage analysis. The tool addresses typical challenges that may arise during the training, evaluation, and application of deep learning models on ambiguous data. The tool\u2019s training and evaluation pipeline uses multiple expert annotations and deep model ensembles to achieve accurate results. The application pipeline supports various use-cases for expert annotations and includes a quality assurance mechanism in the form of uncertainty measures. Benchmarked against other tools, deepflash2 offers both high predictive accuracy and efficient computational resource usage. The tool is built upon established deep learning libraries and enables sharing of trained model ensembles with the research community. deepflash2 aims to simplify the integration of deep learning into bioimage analysis projects while improving accuracy and reliability.","doi":"10.1038\/s41467-023-36960-9","cleaned_title":"deep learning-enabled segmentation of ambiguous bioimages with deepflash2","cleaned_abstract":"bioimages frequently exhibit low signal-to-noise ratios due to experimental conditions, specimen characteristics, and imaging trade-offs. reliable segmentation of such ambiguous images is difficult and laborious. here we introduce deepflash2, a deep learning-enabled segmentation tool for bioimage analysis. the tool addresses typical challenges that may arise during the training, evaluation, and application of deep learning models on ambiguous data. the tool\u2019s training and evaluation pipeline uses multiple expert annotations and deep model ensembles to achieve accurate results. the application pipeline supports various use-cases for expert annotations and includes a quality assurance mechanism in the form of uncertainty measures. benchmarked against other tools, deepflash2 offers both high predictive accuracy and efficient computational resource usage. the tool is built upon established deep learning libraries and enables sharing of trained model ensembles with the research community. deepflash2 aims to simplify the integration of deep learning into bioimage analysis projects while improving accuracy and reliability.","key_phrases":["bioimages","noise","experimental conditions","specimen characteristics","imaging trade-offs","reliable segmentation","such ambiguous images","we","deepflash2","a deep learning-enabled segmentation tool","bioimage analysis","the tool","typical challenges","that","the training, evaluation","application","deep learning models","ambiguous data","the tool\u2019s training and evaluation pipeline","multiple expert annotations","deep model ensembles","accurate results","the application pipeline","various use-cases","expert annotations","a quality assurance mechanism","the form","uncertainty measures","other tools","deepflash2","both high predictive accuracy","efficient computational resource usage","the tool","established deep learning libraries","sharing","trained model ensembles","the research community","deepflash2","the integration","deep learning","bioimage analysis projects","accuracy","reliability","deepflash2","deepflash2","deepflash2"]},{"title":"Automatic segmentation of inconstant fractured fragments for tibia\/fibula from CT images using deep learning","authors":["Hyeonjoo Kim","Young Dae Jeon","Ki Bong Park","Hayeong Cha","Moo-Sub Kim","Juyeon You","Se-Won Lee","Seung-Han Shin","Yang-Guk Chung","Sung Bin Kang","Won Seuk Jang","Do-Kun Yoon"],"abstract":"Orthopaedic surgeons need to correctly identify bone fragments using 2D\/3D CT images before trauma surgery. Advances in deep learning technology provide good insights into trauma surgery over manual diagnosis. This study demonstrates the application of the DeepLab v3+\u2009-based deep learning model for the automatic segmentation of fragments of the fractured tibia and fibula from CT images and the results of the evaluation of the performance of the automatic segmentation. The deep learning model, which was trained using over 11 million images, showed good performance with a global accuracy of 98.92%, a weighted intersection over the union of 0.9841, and a mean boundary F1 score of 0.8921. Moreover, deep learning performed 5\u20138 times faster than the experts\u2019 recognition performed manually, which is comparatively inefficient, with almost the same significance. This study will play an important role in preoperative surgical planning for trauma surgery with convenience and speed.","doi":"10.1038\/s41598-023-47706-4","cleaned_title":"automatic segmentation of inconstant fractured fragments for tibia\/fibula from ct images using deep learning","cleaned_abstract":"orthopaedic surgeons need to correctly identify bone fragments using 2d\/3d ct images before trauma surgery. advances in deep learning technology provide good insights into trauma surgery over manual diagnosis. this study demonstrates the application of the deeplab v3+ -based deep learning model for the automatic segmentation of fragments of the fractured tibia and fibula from ct images and the results of the evaluation of the performance of the automatic segmentation. the deep learning model, which was trained using over 11 million images, showed good performance with a global accuracy of 98.92%, a weighted intersection over the union of 0.9841, and a mean boundary f1 score of 0.8921. moreover, deep learning performed 5\u20138 times faster than the experts\u2019 recognition performed manually, which is comparatively inefficient, with almost the same significance. this study will play an important role in preoperative surgical planning for trauma surgery with convenience and speed.","key_phrases":["orthopaedic surgeons","bone fragments","2d\/3d ct images","trauma surgery","advances","deep learning technology","good insights","trauma surgery","manual diagnosis","this study","the application","the deeplab v3","-based deep learning model","the automatic segmentation","fragments","the fractured tibia","ct images","the results","the evaluation","the performance","the automatic segmentation","the deep learning model","which","over 11 million images","good performance","a global accuracy","98.92%","a weighted intersection","the union","a mean boundary f1 score","deep learning","the experts\u2019 recognition","which","almost the same significance","this study","an important role","preoperative surgical planning","trauma surgery","convenience","speed","2d\/3d","over 11 million","98.92%","0.9841","0.8921","5\u20138"]},{"title":"Smart city urban planning using an evolutionary deep learning model","authors":["Mansoor Alghamdi"],"abstract":"Following the evolution of big data collection, storage, and manipulation techniques, deep learning has drawn the attention of numerous recent studies proposing solutions for smart cities. These solutions were focusing especially on energy consumption, pollution levels, public services, and traffic management issues. Predicting urban evolution and planning is another recent concern for smart cities. In this context, this paper introduces a hybrid model that incorporates evolutionary optimization algorithms, such as Teaching\u2013learning-based optimization (TLBO), into the functioning process of neural deep learning models, such as recurrent neural network (RNN) networks. According to the achieved simulations, deep learning enhanced by evolutionary optimizers can be an effective and promising method for predicting urban evolution of future smart cities.","doi":"10.1007\/s00500-023-08219-4","cleaned_title":"smart city urban planning using an evolutionary deep learning model","cleaned_abstract":"following the evolution of big data collection, storage, and manipulation techniques, deep learning has drawn the attention of numerous recent studies proposing solutions for smart cities. these solutions were focusing especially on energy consumption, pollution levels, public services, and traffic management issues. predicting urban evolution and planning is another recent concern for smart cities. in this context, this paper introduces a hybrid model that incorporates evolutionary optimization algorithms, such as teaching\u2013learning-based optimization (tlbo), into the functioning process of neural deep learning models, such as recurrent neural network (rnn) networks. according to the achieved simulations, deep learning enhanced by evolutionary optimizers can be an effective and promising method for predicting urban evolution of future smart cities.","key_phrases":["the evolution","big data collection","storage","manipulation techniques","deep learning","the attention","numerous recent studies","solutions","smart cities","these solutions","energy consumption","pollution levels","public services","traffic management issues","urban evolution","planning","another recent concern","smart cities","this context","this paper","a hybrid model","that","evolutionary optimization algorithms","teaching","learning-based optimization","tlbo","the functioning process","neural deep learning models","recurrent neural network (rnn) networks","the achieved simulations","deep learning","evolutionary optimizers","an effective and promising method","urban evolution","future smart cities"]},{"title":"Deep Multi-task Learning for Animal Chest Circumference Estimation from Monocular Images","authors":["Hongtao Zhang","Dongbing Gu"],"abstract":"The applications of deep learning algorithms with images to various scenarios have attracted significant research attention. However, application scenarios in animal breeding managements are still limited. In this paper we propose a new deep learning framework to estimate the chest circumference of domestic animals from images. This parameter is a key metric for breeding and monitoring the quality of animal in animal husbandry. We design a set of feature extraction methods based on a multi-task learning framework to address the challenging issues in the main estimation task. The multiple tasks in our proposed framework include object segmentation, keypoint estimation, and depth estimation of cow from monocular images. The domain-specific features extracted from these tasks improve upon our main estimation task. In addition, we also attempt to reduce unnecessary computations during the framework design to reduce the cost of subsequent practical implementation of the developed system. Our proposed framework is tested on our own collected dataset to evaluate its performance.","doi":"10.1007\/s12559-024-10250-y","cleaned_title":"deep multi-task learning for animal chest circumference estimation from monocular images","cleaned_abstract":"the applications of deep learning algorithms with images to various scenarios have attracted significant research attention. however, application scenarios in animal breeding managements are still limited. in this paper we propose a new deep learning framework to estimate the chest circumference of domestic animals from images. this parameter is a key metric for breeding and monitoring the quality of animal in animal husbandry. we design a set of feature extraction methods based on a multi-task learning framework to address the challenging issues in the main estimation task. the multiple tasks in our proposed framework include object segmentation, keypoint estimation, and depth estimation of cow from monocular images. the domain-specific features extracted from these tasks improve upon our main estimation task. in addition, we also attempt to reduce unnecessary computations during the framework design to reduce the cost of subsequent practical implementation of the developed system. our proposed framework is tested on our own collected dataset to evaluate its performance.","key_phrases":["the applications","deep learning algorithms","images","various scenarios","significant research attention","application scenarios","animal breeding managements","this paper","we","a new deep learning framework","the chest circumference","domestic animals","images","this parameter","a key metric","the quality","animal","animal husbandry","we","a set","feature extraction methods","a multi-task learning framework","the challenging issues","the main estimation task","the multiple tasks","our proposed framework","object segmentation","keypoint estimation","depth estimation","cow","monocular images","the domain-specific features","these tasks","our main estimation task","addition","we","unnecessary computations","the framework design","the cost","subsequent practical implementation","the developed system","our proposed framework","dataset","its performance"]},{"title":"Recent advances in deep learning models: a systematic literature review","authors":["Ruchika Malhotra","Priya Singh"],"abstract":"In recent years, deep learning has evolved as a rapidly growing and stimulating field of machine learning and has redefined state-of-the-art performances in a variety of applications. There are multiple deep learning models that have distinct architectures and capabilities. Up to the present, a large number of novel variants of these baseline deep learning models is proposed to address the shortcomings of the existing baseline models. This paper provides a comprehensive review of one hundred seven novel variants of six baseline deep learning models viz. Convolutional Neural Network, Recurrent Neural Network, Long Short Term Memory, Generative Adversarial Network, Autoencoder and Transformer Neural Network. The current review thoroughly examines the novel variants of each of the six baseline models to identify the advancements adopted by them to address one or more limitations of the respective baseline model. It is achieved by critically reviewing the novel variants based on their improved approach. It further provides the merits and demerits of incorporating the advancements in novel variants compared to the baseline deep learning model. Additionally, it reports the domain, datasets and performance measures exploited by the novel variants to make an overall judgment in terms of the improvements. This is because the performance of the deep learning models are subject to the application domain, type of datasets and may also vary on different performance measures. The critical findings of the review would facilitate the researchers and practitioners with the most recent progressions and advancements in the baseline deep learning models and guide them in selecting an appropriate novel variant of the baseline to solve deep learning based tasks in a similar setting.","doi":"10.1007\/s11042-023-15295-z","cleaned_title":"recent advances in deep learning models: a systematic literature review","cleaned_abstract":"in recent years, deep learning has evolved as a rapidly growing and stimulating field of machine learning and has redefined state-of-the-art performances in a variety of applications. there are multiple deep learning models that have distinct architectures and capabilities. up to the present, a large number of novel variants of these baseline deep learning models is proposed to address the shortcomings of the existing baseline models. this paper provides a comprehensive review of one hundred seven novel variants of six baseline deep learning models viz. convolutional neural network, recurrent neural network, long short term memory, generative adversarial network, autoencoder and transformer neural network. the current review thoroughly examines the novel variants of each of the six baseline models to identify the advancements adopted by them to address one or more limitations of the respective baseline model. it is achieved by critically reviewing the novel variants based on their improved approach. it further provides the merits and demerits of incorporating the advancements in novel variants compared to the baseline deep learning model. additionally, it reports the domain, datasets and performance measures exploited by the novel variants to make an overall judgment in terms of the improvements. this is because the performance of the deep learning models are subject to the application domain, type of datasets and may also vary on different performance measures. the critical findings of the review would facilitate the researchers and practitioners with the most recent progressions and advancements in the baseline deep learning models and guide them in selecting an appropriate novel variant of the baseline to solve deep learning based tasks in a similar setting.","key_phrases":["recent years","deep learning","a rapidly growing and stimulating field","machine learning","the-art","a variety","applications","multiple deep learning models","that","distinct architectures","capabilities","the present","a large number","novel variants","these baseline deep learning models","the shortcomings","the existing baseline models","this paper","a comprehensive review","one hundred seven novel variants","six baseline deep learning models","convolutional neural network","recurrent neural network","long short term memory","generative adversarial network","autoencoder","transformer neural network","the current review","the novel variants","each","the six baseline models","the advancements","them","one or more limitations","the respective baseline model","it","the novel variants","their improved approach","it","the merits","demerits","the advancements","novel variants","the baseline deep learning model","it","the domain","datasets","performance measures","the novel variants","an overall judgment","terms","the improvements","this","the performance","the deep learning models","the application domain","type","datasets","different performance measures","the critical findings","the review","the researchers","practitioners","the most recent progressions","advancements","the baseline deep learning models","them","an appropriate novel variant","the baseline","deep learning based tasks","a similar setting","recent years","one hundred seven","six","six","one"]},{"title":"Deep Learning Based Alzheimer Disease Diagnosis: A Comprehensive Review","authors":["S. Suganyadevi","A. Shiny Pershiya","K. Balasamy","V. Seethalakshmi","Saroj Bala","Kumud Arora"],"abstract":"Dementia encompasses a range of cognitive disorders, with Alzheimer\u2019s Disease being the utmost widespread and devastating. AD gradually erodes memory and daily functioning through the progressive deterioration of brain cells. It poses a significant global health challenge, necessitating early identification and intervention. Detecting AD at its onset holds immense potential to predict future health outcomes for individuals. By harnessing the power of artificial intelligence and leveraging MRI scans, we could utilize advanced technology to not only classify AD patients but also predict the likelihood of them developing this life-altering condition. This paper delves into the latest advancements in Deep Learning techniques and their functions in image analysis in medical field. Its primary goals are to elucidate the intricacies of medical image processing and to elucidate and implement key findings and recommendations from recent research.","doi":"10.1007\/s42979-024-02743-2","cleaned_title":"deep learning based alzheimer disease diagnosis: a comprehensive review","cleaned_abstract":"dementia encompasses a range of cognitive disorders, with alzheimer\u2019s disease being the utmost widespread and devastating. ad gradually erodes memory and daily functioning through the progressive deterioration of brain cells. it poses a significant global health challenge, necessitating early identification and intervention. detecting ad at its onset holds immense potential to predict future health outcomes for individuals. by harnessing the power of artificial intelligence and leveraging mri scans, we could utilize advanced technology to not only classify ad patients but also predict the likelihood of them developing this life-altering condition. this paper delves into the latest advancements in deep learning techniques and their functions in image analysis in medical field. its primary goals are to elucidate the intricacies of medical image processing and to elucidate and implement key findings and recommendations from recent research.","key_phrases":["dementia","a range","cognitive disorders","alzheimer\u2019s disease","ad","memory","the progressive deterioration","brain cells","it","a significant global health challenge","early identification","intervention","ad","its onset","immense potential","future health outcomes","individuals","the power","artificial intelligence","mri scans","we","advanced technology","ad patients","the likelihood","them","this life-altering condition","this paper","the latest advancements","deep learning techniques","their functions","image analysis","medical field","its primary goals","the intricacies","medical image processing","key findings","recommendations","recent research","daily"]},{"title":"Deep Learning Models for Diagnosis of Schizophrenia Using EEG Signals: Emerging Trends, Challenges, and Prospects","authors":["Rakesh Ranjan","Bikash Chandra Sahana","Ashish Kumar Bhandari"],"abstract":"Schizophrenia (ScZ) is a chronic neuropsychiatric disorder characterized by disruptions in cognitive, perceptual, social, emotional, and behavioral functions. In the traditional approach, the diagnosis of ScZ primarily relies on the subject\u2019s response and the psychiatrist\u2019s experience, making it highly subjective, prejudiced, and time-consuming. In recent medical research, incorporating deep learning (DL) into the diagnostic process improves performance by reducing inter-observer variation and providing qualitative and quantitative support for clinical decisions. Compared with other modalities, such as magnetic resonance images (MRI) or computed tomography (CT) scans, electroencephalogram (EEG) signals give better insights into the underlying neural mechanisms and brain biomarkers of ScZ. Deep learning models show promising results but the utilization of EEG signals as an effective biomarker for ScZ is still under research. Numerous deep learning models have recently been developed for automated ScZ diagnosis with EEG signals exclusively, yet a comprehensive assessment of these approaches still does not exist in the literature. To fill this gap, we comprehensively review the current advancements in deep learning-based schizophrenia diagnosis using EEG signals. This review is intended to provide systematic details of prominent components: deep learning models, ScZ EEG datasets, data preprocessing approaches, input data formulations for DL, chronological DL methodology advancement in ScZ diagnosis, and design trends of DL architecture. Finally, few challenges in both clinical and technical aspects that create hindrances in achieving the full potential of DL models in EEG-based ScZ diagnosis are expounded along with future outlooks.","doi":"10.1007\/s11831-023-10047-6","cleaned_title":"deep learning models for diagnosis of schizophrenia using eeg signals: emerging trends, challenges, and prospects","cleaned_abstract":"schizophrenia (scz) is a chronic neuropsychiatric disorder characterized by disruptions in cognitive, perceptual, social, emotional, and behavioral functions. in the traditional approach, the diagnosis of scz primarily relies on the subject\u2019s response and the psychiatrist\u2019s experience, making it highly subjective, prejudiced, and time-consuming. in recent medical research, incorporating deep learning (dl) into the diagnostic process improves performance by reducing inter-observer variation and providing qualitative and quantitative support for clinical decisions. compared with other modalities, such as magnetic resonance images (mri) or computed tomography (ct) scans, electroencephalogram (eeg) signals give better insights into the underlying neural mechanisms and brain biomarkers of scz. deep learning models show promising results but the utilization of eeg signals as an effective biomarker for scz is still under research. numerous deep learning models have recently been developed for automated scz diagnosis with eeg signals exclusively, yet a comprehensive assessment of these approaches still does not exist in the literature. to fill this gap, we comprehensively review the current advancements in deep learning-based schizophrenia diagnosis using eeg signals. this review is intended to provide systematic details of prominent components: deep learning models, scz eeg datasets, data preprocessing approaches, input data formulations for dl, chronological dl methodology advancement in scz diagnosis, and design trends of dl architecture. finally, few challenges in both clinical and technical aspects that create hindrances in achieving the full potential of dl models in eeg-based scz diagnosis are expounded along with future outlooks.","key_phrases":["schizophrenia","scz","a chronic neuropsychiatric disorder","disruptions","cognitive, perceptual, social, emotional, and behavioral functions","the traditional approach","the diagnosis","scz","the subject\u2019s response","the psychiatrist\u2019s experience","it","recent medical research","deep learning","dl","the diagnostic process","performance","inter-observer variation","qualitative and quantitative support","clinical decisions","other modalities","magnetic resonance images","mri","tomography (ct) scans","electroencephalogram (eeg) signals","better insights","the underlying neural mechanisms","brain biomarkers","scz","deep learning models","promising results","the utilization","eeg signals","an effective biomarker","scz","research","numerous deep learning models","automated scz diagnosis","eeg signals","a comprehensive assessment","these approaches","the literature","this gap","we","the current advancements","deep learning-based schizophrenia diagnosis","eeg signals","this review","systematic details","prominent components","deep learning models","scz eeg datasets","data","approaches","input data formulations","dl, chronological dl methodology advancement","scz diagnosis","design trends","dl architecture","few challenges","both clinical and technical aspects","that","hindrances","the full potential","dl models","eeg-based scz diagnosis","future outlooks"]},{"title":"GMPP-NN: a deep learning architecture for graph molecular property prediction","authors":["Outhman Abbassi","Soumia Ziti","Meryam Belhiah","Souad Najoua Lagmiri","Yassine Zaoui Seghroucheni"],"abstract":"The pharmacy industry is highly focused on drug discovery and development for the identification and optimization of potential drug candidates. One of the key aspects of this process is the prediction of various molecular properties that justify their potential effectiveness in treating specific diseases. Recently, graph neural networks have gained significant attention, primarily due to their strong suitability for predicting complex relationships that exist between atoms and other molecular structures. GNNs require significant depth to capture global features and to allow the network to iteratively aggregate and propagate information across the entire graph structure. In this research study, we present a deep learning architecture known as a graph molecular property prediction neural network. which combines MPNN feature extraction with a multilayer perceptron classifier. The deep learning architecture was evaluated on four benchmark datasets, and its performance was compared to the smiles transformer, fingerprint to vector, deeper graph convolutional networks, geometry-enhanced molecular, and atom-bond transformer-based message-passing neural network. The results showed that the architecture outperformed the other models using the receiver operating characteristic area under the curve metric. These findings offer an exciting opportunity to enhance and improve molecular property prediction in drug discovery and development.","doi":"10.1007\/s42452-024-05944-9","cleaned_title":"gmpp-nn: a deep learning architecture for graph molecular property prediction","cleaned_abstract":"the pharmacy industry is highly focused on drug discovery and development for the identification and optimization of potential drug candidates. one of the key aspects of this process is the prediction of various molecular properties that justify their potential effectiveness in treating specific diseases. recently, graph neural networks have gained significant attention, primarily due to their strong suitability for predicting complex relationships that exist between atoms and other molecular structures. gnns require significant depth to capture global features and to allow the network to iteratively aggregate and propagate information across the entire graph structure. in this research study, we present a deep learning architecture known as a graph molecular property prediction neural network. which combines mpnn feature extraction with a multilayer perceptron classifier. the deep learning architecture was evaluated on four benchmark datasets, and its performance was compared to the smiles transformer, fingerprint to vector, deeper graph convolutional networks, geometry-enhanced molecular, and atom-bond transformer-based message-passing neural network. the results showed that the architecture outperformed the other models using the receiver operating characteristic area under the curve metric. these findings offer an exciting opportunity to enhance and improve molecular property prediction in drug discovery and development.","key_phrases":["the pharmacy industry","drug discovery","development","the identification","optimization","potential drug candidates","the key aspects","this process","the prediction","various molecular properties","that","their potential effectiveness","specific diseases","graph neural networks","significant attention","their strong suitability","complex relationships","that","atoms","other molecular structures","gnns","significant depth","global features","the network","information","the entire graph structure","this research study","we","a deep learning architecture","a graph molecular property prediction neural network","which","mpnn feature extraction","a multilayer perceptron classifier","the deep learning architecture","four benchmark datasets","its performance","the smiles transformer","vector","deeper graph convolutional networks","atom-bond transformer-based message-passing neural network","the results","the architecture","the other models","the receiver operating characteristic area","the curve metric","these findings","an exciting opportunity","molecular property prediction","drug discovery","development","one","four"]},{"title":"White-box inference attack: compromising the security of deep learning-based COVID-19 diagnosis systems","authors":["Burhan Ul Haque Sheikh","Aasim Zafar"],"abstract":"The COVID-19 pandemic has necessitated the exploration of innovative diagnostic approaches, including the utilization of machine learning (ML) and deep learning (DL) technologies. However, recent findings shed light on the susceptibility of deep learning-based models to adversarial attacks, leading to erroneous predictions. This study investigates the vulnerability of a deep COVID-19 diagnosis model to the Fast Gradient Sign Method (FGSM) adversarial attack. Leveraging transfer learning of EfficientNet-B2 on a publicly available dataset, a deep learning-based COVID-19 diagnosis model is developed, achieving an impressive average accuracy of 94.56% on clean test data. However, when subjected to an untargeted FGSM attack with varying epsilon values, the model\u2019s accuracy is severely compromised, plummeting to 21.72% at epsilon 0.008. Notably, the attack successfully misclassifies adversarial COVID-19 images as normal with 100% confidence. This study underscores the critical need for further research and development to address these vulnerabilities and ensure the reliability and accuracy of deep learning models in the diagnosis of COVID-19 patients.","doi":"10.1007\/s41870-023-01538-7","cleaned_title":"white-box inference attack: compromising the security of deep learning-based covid-19 diagnosis systems","cleaned_abstract":"the covid-19 pandemic has necessitated the exploration of innovative diagnostic approaches, including the utilization of machine learning (ml) and deep learning (dl) technologies. however, recent findings shed light on the susceptibility of deep learning-based models to adversarial attacks, leading to erroneous predictions. this study investigates the vulnerability of a deep covid-19 diagnosis model to the fast gradient sign method (fgsm) adversarial attack. leveraging transfer learning of efficientnet-b2 on a publicly available dataset, a deep learning-based covid-19 diagnosis model is developed, achieving an impressive average accuracy of 94.56% on clean test data. however, when subjected to an untargeted fgsm attack with varying epsilon values, the model\u2019s accuracy is severely compromised, plummeting to 21.72% at epsilon 0.008. notably, the attack successfully misclassifies adversarial covid-19 images as normal with 100% confidence. this study underscores the critical need for further research and development to address these vulnerabilities and ensure the reliability and accuracy of deep learning models in the diagnosis of covid-19 patients.","key_phrases":["the exploration","innovative diagnostic approaches","the utilization","machine learning","ml","deep learning","(dl) technologies","recent findings","light","the susceptibility","deep learning-based models","adversarial attacks","erroneous predictions","this study","the vulnerability","a deep covid-19 diagnosis model","the fast gradient sign method","fgsm) adversarial attack","transfer learning","efficientnet-b2","a publicly available dataset","a deep learning-based covid-19 diagnosis model","an impressive average accuracy","94.56%","clean test data","an untargeted fgsm attack","varying epsilon values","the model\u2019s accuracy","21.72%","epsilon","the attack","adversarial covid-19 images","100% confidence","this study","the critical need","further research","development","these vulnerabilities","the reliability","accuracy","deep learning models","the diagnosis","covid-19 patients","covid-19","covid-19","covid-19","94.56%","21.72%","0.008","covid-19","100%","covid-19"]},{"title":"Detection of vulnerabilities in blockchain smart contracts using deep learning","authors":["Namya Aankur Gupta","Mansi Bansal","Seema Sharma","Deepti Mehrotra","Misha Kakkar"],"abstract":"Blockchain helps to give a sense of security as there is only one history of transactions visible to all the involved parties. Smart contracts enable users to manage significant asset amounts of finances on the blockchain without the involvement of any intermediaries. The conditions and checks that have been written in smart contract and executed to the application cannot be changed again. However, these unique features pose some other risks to the smart contract. Smart contracts have several flaws in its programmable language and methods of execution, despite being a developing technology. To build smart contracts and implement numerous complicated business logics, high-level languages are used by the developers to code smart contracts. Thus, blockchain smart contract is the most important element of any decentralized application, posing the risk for it to be attacked. So, the presence of vulnerabilities are to be taken care of on a priority basis. It is important for detection of vulnerabilities in a smart contract and only then implement and connect it with applications to ensure security of funds. The motive of the paper is to discuss how deep learning may be utilized to deliver bug-free secure smart contracts. Objective of the paper is to detect three kinds of vulnerabilities- reentrancy, timestamp and infinite loop. A deep learning model has been created for detection of smart contract vulnerabilities using graph neural networks. The performance of this model has been compared to the present automated tools and other independent methods. It has been shown that this model has greater accuracy than other methods while comparing the prediction of smart contract vulnerabilities in existing models.","doi":"10.1007\/s11276-024-03755-9","cleaned_title":"detection of vulnerabilities in blockchain smart contracts using deep learning","cleaned_abstract":"blockchain helps to give a sense of security as there is only one history of transactions visible to all the involved parties. smart contracts enable users to manage significant asset amounts of finances on the blockchain without the involvement of any intermediaries. the conditions and checks that have been written in smart contract and executed to the application cannot be changed again. however, these unique features pose some other risks to the smart contract. smart contracts have several flaws in its programmable language and methods of execution, despite being a developing technology. to build smart contracts and implement numerous complicated business logics, high-level languages are used by the developers to code smart contracts. thus, blockchain smart contract is the most important element of any decentralized application, posing the risk for it to be attacked. so, the presence of vulnerabilities are to be taken care of on a priority basis. it is important for detection of vulnerabilities in a smart contract and only then implement and connect it with applications to ensure security of funds. the motive of the paper is to discuss how deep learning may be utilized to deliver bug-free secure smart contracts. objective of the paper is to detect three kinds of vulnerabilities- reentrancy, timestamp and infinite loop. a deep learning model has been created for detection of smart contract vulnerabilities using graph neural networks. the performance of this model has been compared to the present automated tools and other independent methods. it has been shown that this model has greater accuracy than other methods while comparing the prediction of smart contract vulnerabilities in existing models.","key_phrases":["blockchain","a sense","security","only one history","transactions","all the involved parties","smart contracts","users","significant asset amounts","finances","the blockchain","the involvement","any intermediaries","the conditions","checks","that","smart contract","the application","these unique features","some other risks","the smart contract","smart contracts","several flaws","its programmable language","methods","execution","a developing technology","smart contracts","numerous complicated business logics","high-level languages","the developers","smart contracts","blockchain smart contract","the most important element","any decentralized application","the risk","it","the presence","vulnerabilities","care","a priority basis","it","detection","vulnerabilities","a smart contract","it","applications","security","funds","the motive","the paper","how deep learning","bug-free secure smart contracts","objective","the paper","three kinds","vulnerabilities- reentrancy","timestamp","infinite loop","a deep learning model","detection","smart contract vulnerabilities","graph neural networks","the performance","this model","the present automated tools","other independent methods","it","this model","greater accuracy","other methods","the prediction","smart contract vulnerabilities","existing models","three"]},{"title":"Integrating Deep Learning and Reinforcement Learning for Enhanced Financial Risk Forecasting in Supply Chain Management","authors":["Yuanfei Cui","Fengtong Yao"],"abstract":"In today\u2019s dynamic business landscape, the integration of supply chain management and financial risk forecasting is imperative for sustained success. This research paper introduces a groundbreaking approach that seamlessly merges deep autoencoder (DAE) models with reinforcement learning (RL) techniques to enhance financial risk forecasting within the realm of supply chain management. The primary objective of this research is to optimize financial decision-making processes by extracting key feature representations from financial data and leveraging RL for decision optimization. To achieve this, the paper presents the PSO-SDAE model, a novel and sophisticated approach to financial risk forecasting. By incorporating advanced noise reduction features and optimization algorithms, the PSO-SDAE model significantly enhances the accuracy and reliability of financial risk predictions. Notably, the PSO-SDAE model goes beyond traditional forecasting methods by addressing the need for real-time decision-making in the rapidly evolving landscape of financial risk management. This is achieved through the utilization of a distributed RL algorithm, which expedites the processing of supply chain data while maintaining both efficiency and accuracy. The results of our study showcase the exceptional precision of the PSO-SDAE model in predicting financial risks, underscoring its efficacy for proactive risk management within supply chain operations. Moreover, the augmented processing speed of the model enables real-time analysis and decision-making \u2014 a critical capability in today\u2019s fast-paced business environment.","doi":"10.1007\/s13132-024-01946-5","cleaned_title":"integrating deep learning and reinforcement learning for enhanced financial risk forecasting in supply chain management","cleaned_abstract":"in today\u2019s dynamic business landscape, the integration of supply chain management and financial risk forecasting is imperative for sustained success. this research paper introduces a groundbreaking approach that seamlessly merges deep autoencoder (dae) models with reinforcement learning (rl) techniques to enhance financial risk forecasting within the realm of supply chain management. the primary objective of this research is to optimize financial decision-making processes by extracting key feature representations from financial data and leveraging rl for decision optimization. to achieve this, the paper presents the pso-sdae model, a novel and sophisticated approach to financial risk forecasting. by incorporating advanced noise reduction features and optimization algorithms, the pso-sdae model significantly enhances the accuracy and reliability of financial risk predictions. notably, the pso-sdae model goes beyond traditional forecasting methods by addressing the need for real-time decision-making in the rapidly evolving landscape of financial risk management. this is achieved through the utilization of a distributed rl algorithm, which expedites the processing of supply chain data while maintaining both efficiency and accuracy. the results of our study showcase the exceptional precision of the pso-sdae model in predicting financial risks, underscoring its efficacy for proactive risk management within supply chain operations. moreover, the augmented processing speed of the model enables real-time analysis and decision-making \u2014 a critical capability in today\u2019s fast-paced business environment.","key_phrases":["today\u2019s dynamic business landscape","the integration","supply chain management","financial risk forecasting","sustained success","this research paper","a groundbreaking approach","that","deep autoencoder (dae) models","reinforcement learning","(rl) techniques","financial risk forecasting","the realm","supply chain management","the primary objective","this research","financial decision-making processes","key feature representations","financial data","rl","decision optimization","this","the paper","the pso-sdae model","a novel and sophisticated approach","financial risk forecasting","advanced noise reduction features","optimization algorithms","the pso-sdae model","the accuracy","reliability","financial risk predictions","the pso-sdae model","traditional forecasting methods","the need","real-time decision-making","the rapidly evolving landscape","financial risk management","this","the utilization","a distributed rl algorithm","which","the processing","supply chain data","both efficiency","accuracy","the results","our study showcase","the exceptional precision","the pso-sdae model","financial risks","its efficacy","proactive risk management","supply chain operations","the augmented processing speed","the model","real-time analysis","decision-making","a critical capability","today\u2019s fast-paced business environment","today","dae","today"]},{"title":"ZS-DML: Zero-Shot Deep Metric Learning approach for plant leaf disease classification","authors":["Davood Zabihzadeh","Mina Masoudifar"],"abstract":"Automatic plant disease detection plays an important role in food security. Deep learning methods are able to detect precisely various types of plant diseases but at the expense of using huge amounts of resources (processors and data). Therefore, employing few-shot or zero-shot learning methods is unavoidable. Deep Metric Learning (DML) is a widely used technique for few\/zero shot learning. Existing DML methods extract features from the last hidden layer of a pre-trained deep network, which increases the dependence of the specific features on the observed classes. In this paper, the general discriminative feature learning method is used to learn general features of plant leaves. Moreover, a proxy-based loss is utilized that learns the embedding without sampling phase while having a higher convergence rate. The network is trained on the Plant Village dataset where the images are split into 32 and 6 classes as source and target, respectively. The knowledge learned from the source domain is transferred to the target in a zero-shot setting. A few samples of the target domain are presented to the network as a gallery. The network is then evaluated on the target domain. The experimental results show that by presenting few or even only one sample of new classes to the network without fine-tuning step, our method can achieve a classification accuracy of 99%\/80.64% for few\/one image(s) per class.","doi":"10.1007\/s11042-023-17136-5","cleaned_title":"zs-dml: zero-shot deep metric learning approach for plant leaf disease classification","cleaned_abstract":"automatic plant disease detection plays an important role in food security. deep learning methods are able to detect precisely various types of plant diseases but at the expense of using huge amounts of resources (processors and data). therefore, employing few-shot or zero-shot learning methods is unavoidable. deep metric learning (dml) is a widely used technique for few\/zero shot learning. existing dml methods extract features from the last hidden layer of a pre-trained deep network, which increases the dependence of the specific features on the observed classes. in this paper, the general discriminative feature learning method is used to learn general features of plant leaves. moreover, a proxy-based loss is utilized that learns the embedding without sampling phase while having a higher convergence rate. the network is trained on the plant village dataset where the images are split into 32 and 6 classes as source and target, respectively. the knowledge learned from the source domain is transferred to the target in a zero-shot setting. a few samples of the target domain are presented to the network as a gallery. the network is then evaluated on the target domain. the experimental results show that by presenting few or even only one sample of new classes to the network without fine-tuning step, our method can achieve a classification accuracy of 99%\/80.64% for few\/one image(s) per class.","key_phrases":["automatic plant disease detection","an important role","food security","deep learning methods","precisely various types","plant diseases","the expense","huge amounts","resources","processors","data","few-shot or zero-shot learning methods","deep metric learning","dml","a widely used technique","few\/zero shot learning","existing dml methods","features","the last hidden layer","a pre-trained deep network","which","the dependence","the specific features","the observed classes","this paper","the general discriminative feature learning method","general features","plant leaves","a proxy-based loss","that","the embedding","sampling phase","a higher convergence rate","the network","the plant village dataset","the images","32 and 6 classes","source","target","the knowledge","the source domain","the target","a zero-shot setting","a few samples","the target domain","the network","a gallery","the network","the target domain","the experimental results","few or even only one sample","new classes","the network","fine-tuning step","our method","a classification accuracy","99%\/80.64%","few\/one image(s","class","zero","32","6","zero","only one","99%\/80.64%","one"]},{"title":"Deep Learning Approaches for Automatic Quality Assurance of Magnetic Resonance Images Using ACR Phantom","authors":["Tarraf Torfeh","Souha Aouadi","SA Yoganathan","Satheesh Paloor","Rabih Hammoud","Noora Al-Hammadi"],"abstract":"BackgroundIn recent years, there has been a growing trend towards utilizing Artificial Intelligence (AI) and machine learning techniques in medical imaging, including for the purpose of automating quality assurance. In this research, we aimed to develop and evaluate various deep learning-based approaches for automatic quality assurance of Magnetic Resonance (MR) images using the American College of Radiology (ACR) standards.MethodsThe study involved the development, optimization, and testing of custom convolutional neural network (CNN) models. Additionally, popular pre-trained models such as VGG16, VGG19, ResNet50, InceptionV3, EfficientNetB0, and EfficientNetB5 were trained and tested. The use of pre-trained models, particularly those trained on the ImageNet dataset, for transfer learning was also explored. Two-class classification models were employed for assessing spatial resolution and geometric distortion, while an approach classifying the image into 10 classes representing the number of visible spokes was used for the low contrast.ResultsOur results showed that deep learning-based methods can be effectively used for MR image quality assurance and can improve the performance of these models. The low contrast test was one of the most challenging tests within the ACR phantom.ConclusionsOverall, for geometric distortion and spatial resolution, all of the deep learning models tested produced prediction accuracy of 80% or higher. The study also revealed that training the models from scratch performed slightly better compared to transfer learning. For the low contrast, our investigation emphasized the adaptability and potential of deep learning models. The custom CNN models excelled in predicting the number of visible spokes, achieving commendable accuracy, recall, precision, and F1 scores.","doi":"10.1186\/s12880-023-01157-5","cleaned_title":"deep learning approaches for automatic quality assurance of magnetic resonance images using acr phantom","cleaned_abstract":"backgroundin recent years, there has been a growing trend towards utilizing artificial intelligence (ai) and machine learning techniques in medical imaging, including for the purpose of automating quality assurance. in this research, we aimed to develop and evaluate various deep learning-based approaches for automatic quality assurance of magnetic resonance (mr) images using the american college of radiology (acr) standards.methodsthe study involved the development, optimization, and testing of custom convolutional neural network (cnn) models. additionally, popular pre-trained models such as vgg16, vgg19, resnet50, inceptionv3, efficientnetb0, and efficientnetb5 were trained and tested. the use of pre-trained models, particularly those trained on the imagenet dataset, for transfer learning was also explored. two-class classification models were employed for assessing spatial resolution and geometric distortion, while an approach classifying the image into 10 classes representing the number of visible spokes was used for the low contrast.resultsour results showed that deep learning-based methods can be effectively used for mr image quality assurance and can improve the performance of these models. the low contrast test was one of the most challenging tests within the acr phantom.conclusionsoverall, for geometric distortion and spatial resolution, all of the deep learning models tested produced prediction accuracy of 80% or higher. the study also revealed that training the models from scratch performed slightly better compared to transfer learning. for the low contrast, our investigation emphasized the adaptability and potential of deep learning models. the custom cnn models excelled in predicting the number of visible spokes, achieving commendable accuracy, recall, precision, and f1 scores.","key_phrases":["a growing trend","artificial intelligence","ai","techniques","medical imaging","the purpose","quality assurance","this research","we","various deep learning-based approaches","automatic quality assurance","magnetic resonance","(mr) images","the american college","radiology","standards.methodsthe study","the development","optimization","testing","custom convolutional neural network (cnn) models","popular pre-trained models","vgg16","vgg19","resnet50","inceptionv3","efficientnetb0","efficientnetb5","the use","pre-trained models","particularly those","the imagenet dataset","transfer learning","two-class classification models","spatial resolution","geometric distortion","an approach","the image","10 classes","the number","visible spokes","the low contrast.resultsour results","deep learning-based methods","mr image quality assurance","the performance","these models","the low contrast test","the most challenging tests","the acr phantom.conclusionsoverall","geometric distortion","spatial resolution","all","the deep learning models","produced prediction accuracy","80%","the study","the models","scratch","learning","the low contrast","our investigation","the adaptability","potential","deep learning models","the custom cnn models","the number","visible spokes","commendable accuracy","recall","precision","f1 scores","backgroundin recent years","american","cnn","resnet50","inceptionv3","efficientnetb0","efficientnetb5","two","10","80%","cnn"]},{"title":"Depression clinical detection model based on social media: a federated deep learning approach","authors":["Yang Liu"],"abstract":"Depression can significantly impact people\u2019s mental health, and recent research shows that social media can provide decision-making support for health-care professionals and serve as supplementary information for understanding patients\u2019 health status. Deep learning models are also able to assess an individual\u2019s likelihood of experiencing depression. However, data availability on social media is often limited due to privacy concerns, even though deep learning models benefit from having more data to analyze. To address this issue, this study proposes a methodological framework system for clinical decision support that uses federated deep learning (FDL) to identify individuals experiencing depression and provide intervention decisions for clinicians. The proposed framework involves evaluation of datasets from three social media platforms, and the experimental results demonstrate that our method achieves state-of-the-art results. The study aims to provide a clinical decision support system with evolvable features that can deliver precise solutions and assist health-care professionals in medical diagnosis. The proposed framework that incorporates social media data and deep learning models can provide valuable insights into patients\u2019 health status, support personalized treatment decisions, and adapt to changing health-care needs.","doi":"10.1007\/s11227-023-05754-7","cleaned_title":"depression clinical detection model based on social media: a federated deep learning approach","cleaned_abstract":"depression can significantly impact people\u2019s mental health, and recent research shows that social media can provide decision-making support for health-care professionals and serve as supplementary information for understanding patients\u2019 health status. deep learning models are also able to assess an individual\u2019s likelihood of experiencing depression. however, data availability on social media is often limited due to privacy concerns, even though deep learning models benefit from having more data to analyze. to address this issue, this study proposes a methodological framework system for clinical decision support that uses federated deep learning (fdl) to identify individuals experiencing depression and provide intervention decisions for clinicians. the proposed framework involves evaluation of datasets from three social media platforms, and the experimental results demonstrate that our method achieves state-of-the-art results. the study aims to provide a clinical decision support system with evolvable features that can deliver precise solutions and assist health-care professionals in medical diagnosis. the proposed framework that incorporates social media data and deep learning models can provide valuable insights into patients\u2019 health status, support personalized treatment decisions, and adapt to changing health-care needs.","key_phrases":["depression","people\u2019s mental health","recent research","social media","decision-making support","health-care professionals","supplementary information","patients\u2019 health status","deep learning models","an individual\u2019s likelihood","depression","data availability","social media","privacy concerns","deep learning models","more data","this issue","this study","a methodological framework system","clinical decision support","that","deep learning","individuals","depression","intervention decisions","clinicians","the proposed framework","evaluation","datasets","three social media platforms","the experimental results","our method","the-art","the study","a clinical decision support system","evolvable features","that","precise solutions","health-care professionals","medical diagnosis","the proposed framework","that","social media data","deep learning models","valuable insights","patients\u2019 health status","personalized treatment decisions","health-care needs","clinicians","three"]},{"title":"Detecting schizophrenia with 3D structural brain MRI using deep learning","authors":["Junhao Zhang","Vishwanatha M. Rao","Ye Tian","Yanting Yang","Nicolas Acosta","Zihan Wan","Pin-Yu Lee","Chloe Zhang","Lawrence S. Kegeles","Scott A. Small","Jia Guo"],"abstract":"Schizophrenia is a chronic neuropsychiatric disorder that causes distinct structural alterations within the brain. We hypothesize that deep learning applied to a structural neuroimaging dataset could detect disease-related alteration and improve classification and diagnostic accuracy. We tested this hypothesis using a single, widely available, and conventional T1-weighted MRI scan, from which we extracted the 3D whole-brain structure using standard post-processing methods. A deep learning model was then developed, optimized, and evaluated on three open datasets with T1-weighted MRI scans of patients with schizophrenia. Our proposed model outperformed the benchmark model, which was also trained with structural MR images using a 3D CNN architecture. Our model is capable of almost perfectly (area under the ROC curve\u2009=\u20090.987) distinguishing schizophrenia patients from healthy controls on unseen structural MRI scans. Regional analysis localized subcortical regions and ventricles as the most predictive brain regions. Subcortical structures serve a pivotal role in cognitive, affective, and social functions in humans, and structural abnormalities of these regions have been associated with schizophrenia. Our finding corroborates that schizophrenia is associated with widespread alterations in subcortical brain structure and the subcortical structural information provides prominent features in diagnostic classification. Together, these results further demonstrate the potential of deep learning to improve schizophrenia diagnosis and identify its structural neuroimaging signatures from a single, standard T1-weighted brain MRI.","doi":"10.1038\/s41598-023-41359-z","cleaned_title":"detecting schizophrenia with 3d structural brain mri using deep learning","cleaned_abstract":"schizophrenia is a chronic neuropsychiatric disorder that causes distinct structural alterations within the brain. we hypothesize that deep learning applied to a structural neuroimaging dataset could detect disease-related alteration and improve classification and diagnostic accuracy. we tested this hypothesis using a single, widely available, and conventional t1-weighted mri scan, from which we extracted the 3d whole-brain structure using standard post-processing methods. a deep learning model was then developed, optimized, and evaluated on three open datasets with t1-weighted mri scans of patients with schizophrenia. our proposed model outperformed the benchmark model, which was also trained with structural mr images using a 3d cnn architecture. our model is capable of almost perfectly (area under the roc curve = 0.987) distinguishing schizophrenia patients from healthy controls on unseen structural mri scans. regional analysis localized subcortical regions and ventricles as the most predictive brain regions. subcortical structures serve a pivotal role in cognitive, affective, and social functions in humans, and structural abnormalities of these regions have been associated with schizophrenia. our finding corroborates that schizophrenia is associated with widespread alterations in subcortical brain structure and the subcortical structural information provides prominent features in diagnostic classification. together, these results further demonstrate the potential of deep learning to improve schizophrenia diagnosis and identify its structural neuroimaging signatures from a single, standard t1-weighted brain mri.","key_phrases":["schizophrenia","a chronic neuropsychiatric disorder","that","distinct structural alterations","the brain","we","deep learning","a structural neuroimaging dataset","disease-related alteration","classification and diagnostic accuracy","we","this hypothesis","conventional t1-weighted mri scan","which","we","the 3d whole-brain structure","standard post-processing methods","a deep learning model","three open datasets","t1-weighted mri scans","patients","schizophrenia","our proposed model","the benchmark model","which","structural mr images","a 3d cnn architecture","our model","almost perfectly (area","the roc curve","schizophrenia patients","healthy controls","unseen structural mri scans","regional analysis","subcortical regions","ventricles","the most predictive brain regions","subcortical structures","a pivotal role","cognitive, affective, and social functions","humans","structural abnormalities","these regions","schizophrenia","our finding","schizophrenia","widespread alterations","subcortical brain structure","the subcortical structural information","prominent features","diagnostic classification","these results","the potential","deep learning","schizophrenia diagnosis","its structural neuroimaging signatures","a single, standard t1-weighted brain mri","schizophrenia","3d","three","3d","cnn","roc","0.987","schizophrenia"]},{"title":"Application of simultaneous uncertainty quantification and segmentation for oropharyngeal cancer use-case with Bayesian deep learning","authors":["Jaakko Sahlsten","Joel Jaskari","Kareem A. Wahid","Sara Ahmed","Enrico Glerean","Renjie He","Benjamin H. Kann","Antti M\u00e4kitie","Clifton D. Fuller","Mohamed A. Naser","Kimmo Kaski"],"abstract":"BackgroundRadiotherapy is a core treatment modality for oropharyngeal cancer (OPC), where the primary gross tumor volume (GTVp) is manually segmented with high interobserver variability. This calls for reliable and trustworthy automated tools in clinician workflow. Therefore, accurate uncertainty quantification and its downstream utilization is critical.MethodsHere we propose uncertainty-aware deep learning for OPC GTVp segmentation, and illustrate the utility of uncertainty in multiple applications. We examine two Bayesian deep learning (BDL) models and eight uncertainty measures, and utilize a large multi-institute dataset of 292 PET\/CT scans to systematically analyze our approach.ResultsWe show that our uncertainty-based approach accurately predicts the quality of the deep learning segmentation in 86.6% of cases, identifies low performance cases for semi-automated correction, and visualizes regions of the scans where the segmentations likely fail.ConclusionsOur BDL-based analysis provides a first-step towards more widespread implementation of uncertainty quantification in OPC GTVp segmentation.","doi":"10.1038\/s43856-024-00528-5","cleaned_title":"application of simultaneous uncertainty quantification and segmentation for oropharyngeal cancer use-case with bayesian deep learning","cleaned_abstract":"backgroundradiotherapy is a core treatment modality for oropharyngeal cancer (opc), where the primary gross tumor volume (gtvp) is manually segmented with high interobserver variability. this calls for reliable and trustworthy automated tools in clinician workflow. therefore, accurate uncertainty quantification and its downstream utilization is critical.methodshere we propose uncertainty-aware deep learning for opc gtvp segmentation, and illustrate the utility of uncertainty in multiple applications. we examine two bayesian deep learning (bdl) models and eight uncertainty measures, and utilize a large multi-institute dataset of 292 pet\/ct scans to systematically analyze our approach.resultswe show that our uncertainty-based approach accurately predicts the quality of the deep learning segmentation in 86.6% of cases, identifies low performance cases for semi-automated correction, and visualizes regions of the scans where the segmentations likely fail.conclusionsour bdl-based analysis provides a first-step towards more widespread implementation of uncertainty quantification in opc gtvp segmentation.","key_phrases":["backgroundradiotherapy","a core treatment modality","oropharyngeal cancer","the primary gross tumor volume","gtvp","high interobserver variability","this","reliable and trustworthy automated tools","clinician workflow","accurate uncertainty quantification","its downstream utilization","we","uncertainty-aware deep learning","opc","gtvp segmentation","the utility","uncertainty","multiple applications","we","two bayesian deep learning","bdl) models","eight uncertainty measures","a large multi-institute dataset","292 pet\/ct","our approach.resultswe show","our uncertainty-based approach","the quality","the deep learning segmentation","86.6%","cases","low performance cases","semi-automated correction","regions","the scans","the segmentations","bdl-based analysis","a first-step","more widespread implementation","uncertainty quantification","gtvp segmentation","bayesian","eight","292","86.6%","fail.conclusionsour bdl","first"]},{"title":"Insights from EEG analysis of evoked memory recalls using deep learning for emotion charting","authors":["Muhammad Najam Dar","Muhammad Usman Akram","Ahmad Rauf Subhani","Sajid Gul Khawaja","Constantino Carlos Reyes-Aldasoro","Sarah Gul"],"abstract":"Affect recognition in a real-world, less constrained environment is the principal prerequisite of the industrial-level usefulness of this technology. Monitoring the psychological profile using smart, wearable electroencephalogram (EEG) sensors during daily activities without external stimuli, such as memory-induced emotions, is a challenging research gap in emotion recognition. This paper proposed a deep learning framework for improved memory-induced emotion recognition leveraging a combination of 1D-CNN and LSTM as feature extractors integrated with an Extreme Learning Machine (ELM) classifier. The proposed deep learning architecture, combined with the EEG preprocessing, such as the removal of the average baseline signal from each sample and extraction of EEG rhythms (delta, theta, alpha, beta, and gamma), aims to capture repetitive and continuous patterns for memory-induced emotion recognition, underexplored with deep learning techniques. This work has analyzed EEG signals using a wearable, ultra-mobile sports cap while recalling autobiographical emotional memories evoked by affect-denoting words, with self-annotation on the scale of valence and arousal. With extensive experimentation using the same dataset, the proposed framework empirically outperforms existing techniques for the emerging area of memory-induced emotion recognition with an accuracy of 65.6%. The EEG rhythms analysis, such as delta, theta, alpha, beta, and gamma, achieved 65.5%, 52.1%, 65.1%, 64.6%, and 65.0% accuracies for classification with four quadrants of valence and arousal. These results underscore the significant advancement achieved by our proposed method for the real-world environment of memory-induced emotion recognition.","doi":"10.1038\/s41598-024-61832-7","cleaned_title":"insights from eeg analysis of evoked memory recalls using deep learning for emotion charting","cleaned_abstract":"affect recognition in a real-world, less constrained environment is the principal prerequisite of the industrial-level usefulness of this technology. monitoring the psychological profile using smart, wearable electroencephalogram (eeg) sensors during daily activities without external stimuli, such as memory-induced emotions, is a challenging research gap in emotion recognition. this paper proposed a deep learning framework for improved memory-induced emotion recognition leveraging a combination of 1d-cnn and lstm as feature extractors integrated with an extreme learning machine (elm) classifier. the proposed deep learning architecture, combined with the eeg preprocessing, such as the removal of the average baseline signal from each sample and extraction of eeg rhythms (delta, theta, alpha, beta, and gamma), aims to capture repetitive and continuous patterns for memory-induced emotion recognition, underexplored with deep learning techniques. this work has analyzed eeg signals using a wearable, ultra-mobile sports cap while recalling autobiographical emotional memories evoked by affect-denoting words, with self-annotation on the scale of valence and arousal. with extensive experimentation using the same dataset, the proposed framework empirically outperforms existing techniques for the emerging area of memory-induced emotion recognition with an accuracy of 65.6%. the eeg rhythms analysis, such as delta, theta, alpha, beta, and gamma, achieved 65.5%, 52.1%, 65.1%, 64.6%, and 65.0% accuracies for classification with four quadrants of valence and arousal. these results underscore the significant advancement achieved by our proposed method for the real-world environment of memory-induced emotion recognition.","key_phrases":["recognition","a real-world, less constrained environment","the principal prerequisite","the industrial-level usefulness","this technology","the psychological profile","smart, wearable electroencephalogram (eeg) sensors","daily activities","external stimuli","memory-induced emotions","a challenging research gap","emotion recognition","this paper","a deep learning framework","improved memory-induced emotion recognition","a combination","1d-cnn","lstm","feature extractors","an extreme learning machine","(elm) classifier","the proposed deep learning architecture","the eeg preprocessing","the removal","the average baseline signal","each sample","extraction","eeg rhythms","delta","theta","alpha","beta","gamma","repetitive and continuous patterns","memory-induced emotion recognition","deep learning techniques","this work","eeg signals","a wearable, ultra-mobile sports cap","autobiographical emotional memories","affect-denoting words","self-annotation","the scale","valence","arousal","extensive experimentation","the same dataset","the proposed framework","existing techniques","the emerging area","memory-induced emotion recognition","an accuracy","65.6%","the eeg rhythms analysis","delta","theta","alpha","beta","gamma","65.5%","52.1%","65.1%","64.6%","65.0% accuracies","classification","four quadrants","valence","arousal","these results","the significant advancement","our proposed method","the real-world environment","memory-induced emotion recognition","daily","1d","65.6%","65.5%","52.1%","65.1%","64.6%","65.0%","four"]},{"title":"Pneumonia detection based on RSNA dataset and anchor-free deep learning detector","authors":["Linghua Wu","Jing Zhang","Yilin Wang","Rong Ding","Yueqin Cao","Guiqin Liu","Changsheng Liufu","Baowei Xie","Shanping Kang","Rui Liu","Wenle Li","Furen Guan"],"abstract":"Pneumonia is a highly lethal disease, and research on its treatment and early screening tools has received extensive attention from researchers. Due to the maturity and cost reduction of chest X-ray technology, and with the development of artificial intelligence technology, pneumonia identification based on deep learning and chest X-ray has attracted attention from all over the world. Although the feature extraction capability of deep learning is strong, existing deep learning object detection frameworks are based on pre-defined anchors, which require a lot of tuning and experience to guarantee their excellent results in the face of new applications or data. To avoid the influence of anchor settings in pneumonia detection, this paper proposes an anchor-free object detection framework and RSNA dataset based on pneumonia detection. First, a data enhancement scheme is used to preprocess the chest X-ray images; second, an anchor-free object detection framework is used for pneumonia detection, which contains a feature pyramid, two-branch detection head, and focal loss. The average precision of 51.5 obtained by Intersection over Union (IoU) calculation shows that the pneumonia detection results obtained in this paper can surpass the existing classical object detection framework, providing an idea for future research and exploration.","doi":"10.1038\/s41598-024-52156-7","cleaned_title":"pneumonia detection based on rsna dataset and anchor-free deep learning detector","cleaned_abstract":"pneumonia is a highly lethal disease, and research on its treatment and early screening tools has received extensive attention from researchers. due to the maturity and cost reduction of chest x-ray technology, and with the development of artificial intelligence technology, pneumonia identification based on deep learning and chest x-ray has attracted attention from all over the world. although the feature extraction capability of deep learning is strong, existing deep learning object detection frameworks are based on pre-defined anchors, which require a lot of tuning and experience to guarantee their excellent results in the face of new applications or data. to avoid the influence of anchor settings in pneumonia detection, this paper proposes an anchor-free object detection framework and rsna dataset based on pneumonia detection. first, a data enhancement scheme is used to preprocess the chest x-ray images; second, an anchor-free object detection framework is used for pneumonia detection, which contains a feature pyramid, two-branch detection head, and focal loss. the average precision of 51.5 obtained by intersection over union (iou) calculation shows that the pneumonia detection results obtained in this paper can surpass the existing classical object detection framework, providing an idea for future research and exploration.","key_phrases":["pneumonia","a highly lethal disease","research","its treatment","early screening tools","extensive attention","researchers","the maturity and cost reduction","chest x-ray technology","the development","artificial intelligence technology","pneumonia identification","deep learning","chest x","-","ray","attention","the world","the feature extraction capability","deep learning","existing deep learning object detection frameworks","pre-defined anchors","which","a lot","tuning","their excellent results","the face","new applications","data","the influence","anchor settings","pneumonia detection","this paper","an anchor-free object detection framework","rsna dataset","pneumonia detection","a data enhancement scheme","the chest x-ray images","an anchor-free object detection framework","pneumonia detection","which","a feature pyramid","two-branch detection head","focal loss","the average precision","intersection","union (iou) calculation","the pneumonia detection results","this paper","the existing classical object detection framework","an idea","future research","exploration","first","second","two","51.5"]},{"title":"Comparative Study for Optimized Deep Learning-Based Road Accidents Severity Prediction Models","authors":["Hussam Hijazi","Karim Sattar","Hassan M. Al-Ahmadi","Sami El-Ferik"],"abstract":"Road traffic accidents remain a major cause of fatalities and injuries worldwide. Effective classification of accident type and severity is crucial for prompt post-accident protocols and the development of comprehensive road safety policies. This study explores the application of deep learning techniques for predicting crash injury severity in the Eastern Province of Saudi Arabia. Five deep learning models were trained and evaluated, including various variants of feedforward multilayer perceptron, a back-propagated artificial neural network (ANN), an ANN with radial basis function (RPF), and tabular data learning network (TabNet). The models were optimized using Bayesian optimization (BO) and employed the synthetic minority oversampling technique (SMOTE) for oversampling the training dataset. While SMOTE enhanced balanced accuracy for ANN with RBF and TabNet, it compromised precision and increased recall. The results indicated that oversampling techniques did not consistently improve model performance. Additionally, significant features were identified using least absolute shrinkage and selection operator (LASSO) regularization, feature importance, and permutation importance. The results indicated that oversampling techniques did not consistently improve model performance. While SMOTE enhanced balanced accuracy for ANN with RBF and TabNet, it compromised precision and increased recall. The study's findings emphasize the consistent significance of the 'Number of Injuries Major' feature as a vital predictor in deep learning models, regardless of the selection techniques employed. These results shed light on the pivotal role played by the count of individuals with major injuries in influencing the severity of crash injuries, highlighting its potential relevance in shaping road safety policy development.","doi":"10.1007\/s13369-023-08510-4","cleaned_title":"comparative study for optimized deep learning-based road accidents severity prediction models","cleaned_abstract":"road traffic accidents remain a major cause of fatalities and injuries worldwide. effective classification of accident type and severity is crucial for prompt post-accident protocols and the development of comprehensive road safety policies. this study explores the application of deep learning techniques for predicting crash injury severity in the eastern province of saudi arabia. five deep learning models were trained and evaluated, including various variants of feedforward multilayer perceptron, a back-propagated artificial neural network (ann), an ann with radial basis function (rpf), and tabular data learning network (tabnet). the models were optimized using bayesian optimization (bo) and employed the synthetic minority oversampling technique (smote) for oversampling the training dataset. while smote enhanced balanced accuracy for ann with rbf and tabnet, it compromised precision and increased recall. the results indicated that oversampling techniques did not consistently improve model performance. additionally, significant features were identified using least absolute shrinkage and selection operator (lasso) regularization, feature importance, and permutation importance. the results indicated that oversampling techniques did not consistently improve model performance. while smote enhanced balanced accuracy for ann with rbf and tabnet, it compromised precision and increased recall. the study's findings emphasize the consistent significance of the 'number of injuries major' feature as a vital predictor in deep learning models, regardless of the selection techniques employed. these results shed light on the pivotal role played by the count of individuals with major injuries in influencing the severity of crash injuries, highlighting its potential relevance in shaping road safety policy development.","key_phrases":["road traffic accidents","a major cause","fatalities","injuries","effective classification","accident type","severity","prompt post-accident protocols","the development","comprehensive road safety policies","this study","the application","deep learning techniques","crash injury severity","the eastern province","saudi arabia","five deep learning models","various variants","feedforward multilayer perceptron","a back-propagated artificial neural network","ann","radial basis function","rpf","data learning network","tabnet","the models","bayesian optimization","bo","the synthetic minority oversampling technique","smote","the training dataset","smote","balanced accuracy","ann","rbf","tabnet","it","precision","increased recall","the results","techniques","model performance","significant features","lasso","feature importance","permutation importance","the results","techniques","model performance","smote","balanced accuracy","ann","rbf","tabnet","it","precision","increased recall","the study's findings","the consistent significance","the 'number","injuries major' feature","a vital predictor","deep learning models","the selection techniques","these results","light","the pivotal role","the count","individuals","major injuries","the severity","crash injuries","its potential relevance","road safety policy development","saudi arabia","five"]},{"title":"A Resource-Efficient Deep Learning Approach to Visual-Based Cattle Geographic Origin Prediction","authors":["Camellia Ray","Sambit Bakshi","Pankaj Kumar Sa","Ganapati Panda"],"abstract":"Customized healthcare for cattle health monitoring is essential, which aims to optimize individual animal health, thereby enhancing productivity, minimizing illness-related risks, and improving overall welfare. Tailoring healthcare practices to individual requirements guarantees that individual animals receive proper attention and intervention, resulting in better health outcomes and sustainable cattle farming practices. In this regard, the manuscript proposes a visual cues-based region prediction methodology to design a customized cattle healthcare system. The proposed automated AI healthcare system uses resource-efficient deep learning-inspired architecture for computer vision applications like performing region-wise classification. The classification mechanism can be used further to identify a cattle and the regions it belongs. Extensive experimentation has been conducted on a redesigned image dataset to identify the best-suited deep-learning framework to perform region classification for livestock, such as cattle. MobileNetV2 outperforms the considered state-of-the-art frameworks by achieving an accuracy of 93% in identifying the regions of the cattle.","doi":"10.1007\/s11036-024-02350-8","cleaned_title":"a resource-efficient deep learning approach to visual-based cattle geographic origin prediction","cleaned_abstract":"customized healthcare for cattle health monitoring is essential, which aims to optimize individual animal health, thereby enhancing productivity, minimizing illness-related risks, and improving overall welfare. tailoring healthcare practices to individual requirements guarantees that individual animals receive proper attention and intervention, resulting in better health outcomes and sustainable cattle farming practices. in this regard, the manuscript proposes a visual cues-based region prediction methodology to design a customized cattle healthcare system. the proposed automated ai healthcare system uses resource-efficient deep learning-inspired architecture for computer vision applications like performing region-wise classification. the classification mechanism can be used further to identify a cattle and the regions it belongs. extensive experimentation has been conducted on a redesigned image dataset to identify the best-suited deep-learning framework to perform region classification for livestock, such as cattle. mobilenetv2 outperforms the considered state-of-the-art frameworks by achieving an accuracy of 93% in identifying the regions of the cattle.","key_phrases":["customized healthcare","cattle health monitoring","which","individual animal health","productivity","illness-related risks","overall welfare","healthcare practices","individual requirements","individual animals","proper attention","intervention","better health outcomes","sustainable cattle farming practices","this regard","the manuscript","a visual cues-based region prediction methodology","a customized cattle healthcare system","the proposed automated ai healthcare system","resource-efficient deep learning-inspired architecture","computer vision applications","region-wise classification","the classification mechanism","a cattle","the regions","it","extensive experimentation","a redesigned image dataset","the best-suited deep-learning framework","region classification","livestock","cattle","mobilenetv2","the-art","an accuracy","93%","the regions","the cattle","tailoring healthcare","mobilenetv2","93%"]},{"title":"Hybrid deep learning framework for weather forecast with rainfall prediction using weather bigdata analytics","authors":["C. Lalitha","D. Ravindran"],"abstract":"The volume and complexity of weather data, along with missing values and high correlation between collected variables, make it challenging to develop efficient deep learning frameworks that can handle data with more features. This leads to a lack of accurate and predictable weather forecasts. To develop a hybrid deep learning framework for weather forecast with rainfall prediction using weather big data analytics to ensure high detection rates. A modified planet optimization (MPO) algorithm is used for data preprocessing to remove unwanted artifacts. An improved Tuna optimization (ITO) algorithm is presented to select optimal features to avoid data dimensionality issues. A hybrid memory-augmented artificial neural network (MA-ANN) classifier is developed to improve weather early forecast detection rates. The proposed framework is validated against standard benchmark datasets such as weather underground and climate forecast system reanalysis (CFSR). The simulation results are compared with other existing state-of-the-art frameworks based on error measures (RMSE, MAPE, BIAS, R) and quality measures (accuracy, sensitivity, specificity, precision, F1-measure).The MA-ANN classifier accuracy obtained 97.65% for wunderground.com Delhi and 98.88% for Tamilnadu. The hybrid deep learning framework with rainfall prediction using weather big data analytics has shown promising results for accurate and predictable weather forecasts. The proposed framework outperforms other existing state-of-the-art frameworks, and the MA-ANN classifier has improved weather early forecast detection rates. The study demonstrates the potential of utilizing big data techniques in weather forecasting and highlights the importance of developing efficient deep learning frameworks to handle complex and high-dimensional weather data.","doi":"10.1007\/s11042-023-17801-9","cleaned_title":"hybrid deep learning framework for weather forecast with rainfall prediction using weather bigdata analytics","cleaned_abstract":"the volume and complexity of weather data, along with missing values and high correlation between collected variables, make it challenging to develop efficient deep learning frameworks that can handle data with more features. this leads to a lack of accurate and predictable weather forecasts. to develop a hybrid deep learning framework for weather forecast with rainfall prediction using weather big data analytics to ensure high detection rates. a modified planet optimization (mpo) algorithm is used for data preprocessing to remove unwanted artifacts. an improved tuna optimization (ito) algorithm is presented to select optimal features to avoid data dimensionality issues. a hybrid memory-augmented artificial neural network (ma-ann) classifier is developed to improve weather early forecast detection rates. the proposed framework is validated against standard benchmark datasets such as weather underground and climate forecast system reanalysis (cfsr). the simulation results are compared with other existing state-of-the-art frameworks based on error measures (rmse, mape, bias, r) and quality measures (accuracy, sensitivity, specificity, precision, f1-measure).the ma-ann classifier accuracy obtained 97.65% for wunderground.com delhi and 98.88% for tamilnadu. the hybrid deep learning framework with rainfall prediction using weather big data analytics has shown promising results for accurate and predictable weather forecasts. the proposed framework outperforms other existing state-of-the-art frameworks, and the ma-ann classifier has improved weather early forecast detection rates. the study demonstrates the potential of utilizing big data techniques in weather forecasting and highlights the importance of developing efficient deep learning frameworks to handle complex and high-dimensional weather data.","key_phrases":["the volume","complexity","weather data","missing values","high correlation","collected variables","it","efficient deep learning frameworks","that","data","more features","this","a lack","accurate and predictable weather forecasts","a hybrid deep learning framework","weather forecast","rainfall prediction","weather big data analytics","high detection rates","a modified planet optimization","mpo","algorithm","data","unwanted artifacts","an improved tuna optimization","ito","algorithm","optimal features","data dimensionality issues","a hybrid memory-augmented artificial neural network","ma-ann) classifier","weather early forecast detection rates","the proposed framework","standard benchmark datasets","weather","climate forecast system reanalysis","cfsr","the simulation results","the-art","error measures","accuracy","sensitivity","specificity","precision","f1-measure).the ma-ann classifier accuracy","97.65%","98.88%","tamilnadu","the hybrid deep learning framework","rainfall prediction","weather big data analytics","promising results","accurate and predictable weather forecasts","the proposed framework","the-art","the ma-ann classifier","weather early forecast detection rates","the study","the potential","big data techniques","weather forecasting","the importance","efficient deep learning frameworks","complex and high-dimensional weather data","rmse","97.65%","wunderground.com delhi","98.88%"]},{"title":"Deep learning-based question answering: a survey","authors":["Heba Abdel-Nabi","Arafat Awajan","Mostafa Z. Ali"],"abstract":"Question Answering is a crucial natural language processing task. This field of research has attracted a sudden amount of interest lately due mainly to the integration of the deep learning models in the Question Answering Systems which consequently power up many advancements and improvements. This survey aims to explore and shed light upon the recent and most powerful deep learning-based Question Answering Systems and classify them based on the deep learning model used, stating the details of the used word representation, datasets, and evaluation metrics. It aims to highlight and discuss the currently used models and give insights that direct future research to enhance this increasingly growing field.","doi":"10.1007\/s10115-022-01783-5","cleaned_title":"deep learning-based question answering: a survey","cleaned_abstract":"question answering is a crucial natural language processing task. this field of research has attracted a sudden amount of interest lately due mainly to the integration of the deep learning models in the question answering systems which consequently power up many advancements and improvements. this survey aims to explore and shed light upon the recent and most powerful deep learning-based question answering systems and classify them based on the deep learning model used, stating the details of the used word representation, datasets, and evaluation metrics. it aims to highlight and discuss the currently used models and give insights that direct future research to enhance this increasingly growing field.","key_phrases":["a crucial natural language processing task","this field","research","a sudden amount","interest","the integration","the deep learning models","the question","systems","which","many advancements","improvements","this survey","light","the recent and most powerful deep learning-based question","systems","them","the deep learning model","the details","the used word representation","datasets","evaluation metrics","it","the currently used models","insights","that","future research","this increasingly growing field"]},{"title":"Deep Learning Architecture for Computer Vision-based Structural Defect Detection","authors":["Ruoyu Yang","Shubhendu Kumar Singh","Mostafa Tavakkoli","M. Amin Karami","Rahul Rai"],"abstract":"Structural health monitoring (SHM) refers to the implementation of a damage detection strategy for structures. Fault occurrence in these structural systems during the operation is inevitable. Efficient, fast, and precise health monitoring methods are required to proactively perform the necessary repairs and maintenance on time before it is too late. The current structural health monitoring methods involve physically attached sensors or non-contact vision-based vibration measurements. However, these methods have significant drawbacks due to the low spatial resolution, weight influence on the lightweight structure, and time\/labor consumption. Recently, computer-vison-based deep learning methods like convolutional neural network (CNN) and fully convolutional neural network (FCN) have been applied for defect detection and localization, which address the aforementioned problems and obtain high accuracy. This paper proposes a novel hybrid deep learning architecture comprising CNN and temporal convolutional networks (CNN-TCN) for the computer vision-based defect detection task. Various beam samples, consisting of five different materials and various structural defects, were used to evaluate the proposed deep learning algorithms\u2019 performance. The proposed deep learning methods treat each pixel of the video frame like a sensor to extract valuable features for defect detection. Through empirical results, we demonstrate that this \u2019pixel-sensor\u2019 approach is more efficient and accurate and can achieve a better defect detection performance on different beam samples compared with the current state-of-the-art approaches, including CNN-long short-term memory (LSTM), CNN-bidirectional long short-term memory (BiLSTM), multi-scale CNN-LSTM, and CNN-gated recurrent unit(GRU) methods.","doi":"10.1007\/s10489-023-04654-w","cleaned_title":"deep learning architecture for computer vision-based structural defect detection","cleaned_abstract":"structural health monitoring (shm) refers to the implementation of a damage detection strategy for structures. fault occurrence in these structural systems during the operation is inevitable. efficient, fast, and precise health monitoring methods are required to proactively perform the necessary repairs and maintenance on time before it is too late. the current structural health monitoring methods involve physically attached sensors or non-contact vision-based vibration measurements. however, these methods have significant drawbacks due to the low spatial resolution, weight influence on the lightweight structure, and time\/labor consumption. recently, computer-vison-based deep learning methods like convolutional neural network (cnn) and fully convolutional neural network (fcn) have been applied for defect detection and localization, which address the aforementioned problems and obtain high accuracy. this paper proposes a novel hybrid deep learning architecture comprising cnn and temporal convolutional networks (cnn-tcn) for the computer vision-based defect detection task. various beam samples, consisting of five different materials and various structural defects, were used to evaluate the proposed deep learning algorithms\u2019 performance. the proposed deep learning methods treat each pixel of the video frame like a sensor to extract valuable features for defect detection. through empirical results, we demonstrate that this \u2019pixel-sensor\u2019 approach is more efficient and accurate and can achieve a better defect detection performance on different beam samples compared with the current state-of-the-art approaches, including cnn-long short-term memory (lstm), cnn-bidirectional long short-term memory (bilstm), multi-scale cnn-lstm, and cnn-gated recurrent unit(gru) methods.","key_phrases":["structural health monitoring","shm","the implementation","a damage detection strategy","structures","fault occurrence","these structural systems","the operation","precise health monitoring methods","the necessary repairs","maintenance","time","it","the current structural health monitoring methods","physically attached sensors","non-contact vision-based vibration measurements","these methods","significant drawbacks","the low spatial resolution","weight influence","the lightweight structure","time\/labor consumption","computer-vison-based deep learning methods","convolutional neural network","cnn","fully convolutional neural network","fcn","defect detection","localization","which","the aforementioned problems","high accuracy","this paper","a novel hybrid deep learning architecture","cnn","temporal convolutional networks","cnn-tcn","the computer vision-based defect detection task","various beam samples","five different materials","various structural defects","the proposed deep learning algorithms\u2019 performance","the proposed deep learning methods","each pixel","the video frame","a sensor","valuable features","defect detection","empirical results","we","this \u2019pixel-sensor\u2019 approach","a better defect detection performance","different beam samples","the-art","cnn-long short-term memory","lstm","cnn-bidirectional long short-term memory","bilstm","multi-scale cnn-lstm","cnn-gated recurrent unit(gru) methods","cnn","cnn","cnn-tcn","five","cnn","cnn","cnn","cnn"]},{"title":"DCLGM: Fusion Recommendation Model Based on LightGBM and Deep Learning","authors":["Bin Zhao","Bin Li","Jiqun Zhang","Wei Cao","Yilong Gao"],"abstract":"The recommendation system can mine valuable information according to user preferences, so it is widely used in various industries. However, the performance of recommendation systems is generally affected by the problem of data sparsity, and LightGBM can alleviate the impact caused by data sparsity to a certain extent. To this end, this paper proposes a fusion recommendation model based on the LightGBM and deep learning\u2014CLGM model. The model is composed of LighGBM, cross network and deep neural network. First, the features in the dataset are fused and extracted through LightGBM, and the feature with the highest classification accuracy is selected as the input of the neural network layer; Then, using the cross network and the deep neural network, the linear cross combination feature relationship and nonlinear correlation relationship between high-order features are respectively obtained; finally, the results obtained by the pre-order network are linearly weighted and combined to obtain the final recommendation result. In this paper, AUC and Logloss are used as evaluation indicators to verify the model on the public dataset Criteo and dataset Avazu. The simulation experiment results show that, compared with the four typical recommendation models, the recommendation effect of this model is better.","doi":"10.1007\/s11063-024-11504-4","cleaned_title":"dclgm: fusion recommendation model based on lightgbm and deep learning","cleaned_abstract":"the recommendation system can mine valuable information according to user preferences, so it is widely used in various industries. however, the performance of recommendation systems is generally affected by the problem of data sparsity, and lightgbm can alleviate the impact caused by data sparsity to a certain extent. to this end, this paper proposes a fusion recommendation model based on the lightgbm and deep learning\u2014clgm model. the model is composed of lighgbm, cross network and deep neural network. first, the features in the dataset are fused and extracted through lightgbm, and the feature with the highest classification accuracy is selected as the input of the neural network layer; then, using the cross network and the deep neural network, the linear cross combination feature relationship and nonlinear correlation relationship between high-order features are respectively obtained; finally, the results obtained by the pre-order network are linearly weighted and combined to obtain the final recommendation result. in this paper, auc and logloss are used as evaluation indicators to verify the model on the public dataset criteo and dataset avazu. the simulation experiment results show that, compared with the four typical recommendation models, the recommendation effect of this model is better.","key_phrases":["the recommendation system","valuable information","user preferences","it","various industries","the performance","recommendation systems","the problem","data sparsity","lightgbm","the impact","data sparsity","a certain extent","this end","this paper","a fusion recommendation model","the lightgbm","deep learning","clgm model","the model","lighgbm","cross network","deep neural network","the features","the dataset","lightgbm","the feature","the highest classification accuracy","the input","the neural network layer","the cross network","the deep neural network","the linear cross combination feature relationship","nonlinear correlation relationship","high-order features","the results","the pre-order network","the final recommendation result","this paper","auc","logloss","evaluation indicators","the model","the public dataset criteo","dataset avazu","the simulation experiment results","the four typical recommendation models","the recommendation effect","this model","first","four"]},{"title":"Automated assembly quality inspection by deep learning with 2D and 3D synthetic CAD data","authors":["Xiaomeng Zhu","P\u00e4r M\u00e5rtensson","Lars Hanson","M\u00e5rten Bj\u00f6rkman","Atsuto Maki"],"abstract":"In the manufacturing industry, automatic quality inspections can lead to improved product quality and productivity. Deep learning-based computer vision technologies, with their superior performance in many applications, can be a possible solution for automatic quality inspections. However, collecting a large amount of annotated training data for deep learning is expensive and time-consuming, especially for processes involving various products and human activities such as assembly. To address this challenge, we propose a method for automated assembly quality inspection using synthetic data generated from computer-aided design (CAD) models. The method involves two steps: automatic data generation and model implementation. In the first step, we generate synthetic data in two formats: two-dimensional (2D) images and three-dimensional (3D) point clouds. In the second step, we apply different state-of-the-art deep learning approaches to the data for quality inspection, including unsupervised domain adaptation, i.e., a method of adapting models across different data distributions, and transfer learning, which transfers knowledge between related tasks. We evaluate the methods in a case study of pedal car front-wheel assembly quality inspection to identify the possible optimal approach for assembly quality inspection. Our results show that the method using Transfer Learning on 2D synthetic images achieves superior performance compared with others. Specifically, it attained 95% accuracy through fine-tuning with only five annotated real images per class. With promising results, our method may be suggested for other similar quality inspection use cases. By utilizing synthetic CAD data, our method reduces the need for manual data collection and annotation. Furthermore, our method performs well on test data with different backgrounds, making it suitable for different manufacturing environments.","doi":"10.1007\/s10845-024-02375-6","cleaned_title":"automated assembly quality inspection by deep learning with 2d and 3d synthetic cad data","cleaned_abstract":"in the manufacturing industry, automatic quality inspections can lead to improved product quality and productivity. deep learning-based computer vision technologies, with their superior performance in many applications, can be a possible solution for automatic quality inspections. however, collecting a large amount of annotated training data for deep learning is expensive and time-consuming, especially for processes involving various products and human activities such as assembly. to address this challenge, we propose a method for automated assembly quality inspection using synthetic data generated from computer-aided design (cad) models. the method involves two steps: automatic data generation and model implementation. in the first step, we generate synthetic data in two formats: two-dimensional (2d) images and three-dimensional (3d) point clouds. in the second step, we apply different state-of-the-art deep learning approaches to the data for quality inspection, including unsupervised domain adaptation, i.e., a method of adapting models across different data distributions, and transfer learning, which transfers knowledge between related tasks. we evaluate the methods in a case study of pedal car front-wheel assembly quality inspection to identify the possible optimal approach for assembly quality inspection. our results show that the method using transfer learning on 2d synthetic images achieves superior performance compared with others. specifically, it attained 95% accuracy through fine-tuning with only five annotated real images per class. with promising results, our method may be suggested for other similar quality inspection use cases. by utilizing synthetic cad data, our method reduces the need for manual data collection and annotation. furthermore, our method performs well on test data with different backgrounds, making it suitable for different manufacturing environments.","key_phrases":["the manufacturing industry","automatic quality inspections","improved product quality","productivity","deep learning-based computer vision technologies","their superior performance","many applications","a possible solution","automatic quality inspections","a large amount","annotated training data","deep learning","processes","various products","human activities","assembly","this challenge","we","a method","automated assembly quality inspection","synthetic data","computer-aided design (cad) models","the method","two steps","automatic data generation and model implementation","the first step","we","synthetic data","two formats","two-dimensional (2d) images","three-dimensional (3d","the second step","we","the-art","the data","quality inspection","unsupervised domain adaptation","i.e., a method","adapting models","different data distributions","transfer learning","which","knowledge","related tasks","we","the methods","a case study","pedal car front-wheel assembly quality inspection","the possible optimal approach","assembly quality inspection","our results","the method","transfer learning","2d synthetic images","superior performance","others","it","95% accuracy","fine-tuning","only five annotated real images","class","promising results","our method","other similar quality inspection use cases","synthetic cad data","our method","the need","manual data collection","annotation","our method","test data","different backgrounds","it","different manufacturing environments","two","first","two","two","2d","three","3d","second","2d","95%","only five"]},{"title":"Domain knowledge enhanced deep learning for electrocardiogram arrhythmia classification","authors":["Jie Sun\u00a0\n            (\u5b59\u6d01)"],"abstract":"Deep learning provides an effective way for automatic classification of cardiac arrhythmias, but in clinical decision-making, pure data-driven methods working as black-boxes may lead to unsatisfactory results. A promising solution is combining domain knowledge with deep learning. This paper develops a flexible and extensible framework for integrating domain knowledge with a deep neural network. The model consists of a deep neural network to capture the statistical pattern between input data and the ground-truth label, and a knowledge module to guarantee consistency with the domain knowledge. These two components are trained interactively to bring the best of both worlds. The experiments show that the domain knowledge is valuable in refining the neural network prediction and thus improves accuracy.","doi":"10.1631\/FITEE.2100519","cleaned_title":"domain knowledge enhanced deep learning for electrocardiogram arrhythmia classification","cleaned_abstract":"deep learning provides an effective way for automatic classification of cardiac arrhythmias, but in clinical decision-making, pure data-driven methods working as black-boxes may lead to unsatisfactory results. a promising solution is combining domain knowledge with deep learning. this paper develops a flexible and extensible framework for integrating domain knowledge with a deep neural network. the model consists of a deep neural network to capture the statistical pattern between input data and the ground-truth label, and a knowledge module to guarantee consistency with the domain knowledge. these two components are trained interactively to bring the best of both worlds. the experiments show that the domain knowledge is valuable in refining the neural network prediction and thus improves accuracy.","key_phrases":["deep learning","an effective way","automatic classification","cardiac arrhythmias","clinical decision-making",", pure data-driven methods","black-boxes","unsatisfactory results","a promising solution","domain knowledge","deep learning","this paper","a flexible and extensible framework","domain knowledge","a deep neural network","the model","a deep neural network","the statistical pattern","input data","the ground-truth label","a knowledge module","consistency","the domain knowledge","these two components","both worlds","the experiments","the domain knowledge","the neural network prediction","accuracy","two"]},{"title":"Development of Deep Learning Color Recognition Model for Color Measurement Processes","authors":["Sanghun Lee","Ki-Sub Kim","Jeong Won Kang"],"abstract":"We present a deep learning color recognition model for the color measurement process in the paint industry. Currently, spectrophotometers are primarily used for color measurements owing to their accuracy. The measurement method involves manually injecting the sample into a spectrophotometer. Our proposed method uses a webcam with a deep learning model on the stand of a spectrophotometer. Deep learning models are widely used for image and color detection. In this study, the \u201cyou only look once (YOLO)\u201d algorithm is applied for real-time detection of color samples. Upon training various sample images using YOLO, the model could detect the sample area in real time using a webcam. An open source computer vision (OpenCV) library was used for the color recognition model, and the detected RGB color value was converted to the international commission on illumination color space (CIELAB) value, which is primarily used in the color measuring process. However, because of the mirror-like reflection of light from a surface with specular reflection, it is difficult to implement the color value using a camera. To address this problem, we compare several specular removal methods and propose the most suitable model for the color recognition model of color samples. The accuracy of the proposed model was verified by comparing the colors of various samples. Our proposed approach can easily detect samples and color values, which can contribute significantly to automatically calculating the exact amount of coloring required for the target color.","doi":"10.1007\/s42835-024-01791-1","cleaned_title":"development of deep learning color recognition model for color measurement processes","cleaned_abstract":"we present a deep learning color recognition model for the color measurement process in the paint industry. currently, spectrophotometers are primarily used for color measurements owing to their accuracy. the measurement method involves manually injecting the sample into a spectrophotometer. our proposed method uses a webcam with a deep learning model on the stand of a spectrophotometer. deep learning models are widely used for image and color detection. in this study, the \u201cyou only look once (yolo)\u201d algorithm is applied for real-time detection of color samples. upon training various sample images using yolo, the model could detect the sample area in real time using a webcam. an open source computer vision (opencv) library was used for the color recognition model, and the detected rgb color value was converted to the international commission on illumination color space (cielab) value, which is primarily used in the color measuring process. however, because of the mirror-like reflection of light from a surface with specular reflection, it is difficult to implement the color value using a camera. to address this problem, we compare several specular removal methods and propose the most suitable model for the color recognition model of color samples. the accuracy of the proposed model was verified by comparing the colors of various samples. our proposed approach can easily detect samples and color values, which can contribute significantly to automatically calculating the exact amount of coloring required for the target color.","key_phrases":["we","a deep learning color recognition model","the color measurement process","the paint industry","spectrophotometers","color measurements","their accuracy","the measurement method","the sample","a spectrophotometer","our proposed method","a webcam","a deep learning model","the stand","a spectrophotometer","deep learning models","image","color detection","this study","you","algorithm","real-time detection","color samples","various sample images","yolo","the model","the sample area","real time","a webcam","an open source computer vision","(opencv) library","the color recognition model","the detected rgb color value","the international commission","illumination color space","(cielab) value","which","the color measuring process","the mirror-like reflection","light","a surface","specular reflection","it","the color value","a camera","this problem","we","several specular removal methods","the most suitable model","the color recognition model","color samples","the accuracy","the proposed model","the colors","various samples","our proposed approach","samples","color values","which","the exact amount","the target color","deep"]},{"title":"Deep learning nomogram for predicting neoadjuvant chemotherapy response in locally advanced gastric cancer patients","authors":["Jingjing Zhang","Qiang Zhang","Bo Zhao","Gaofeng Shi"],"abstract":"PurposeDeveloped and validated a deep learning radiomics nomogram using multi-phase contrast-enhanced computed tomography (CECT) images to predict neoadjuvant chemotherapy (NAC) response in locally advanced gastric cancer (LAGC) patients.MethodsThis multi-center study retrospectively included 322 patients diagnosed with gastric cancer from January 2013 to June 2023 at two hospitals. Handcrafted radiomics technique and the EfficientNet V2 neural network were applied to arterial, portal venous, and delayed phase CT images to extract two-dimensional handcrafted and deep learning features. A nomogram model was built by integrating the handcrafted signature, the deep learning signature, with clinical features. Discriminative ability was assessed using the receiver operating characteristics (ROC) curve and the precision-recall (P-R) curve. Model fitting was evaluated using calibration curves, and clinical utility was assessed through decision curve analysis (DCA).ResultsThe nomogram exhibited excellent performance. The area under the ROC curve (AUC) was 0.848 [95% confidence interval (CI), 0.793\u20130.893)], 0.802 (95% CI 0.688\u20130.889), and 0.751 (95% CI 0.652\u20130.833) for the training, internal validation, and external validation sets, respectively. The AUCs of the P-R curves were 0.838 (95% CI 0.756\u20130.895), 0.541 (95% CI 0.329\u20130.740), and 0.556 (95% CI 0.376\u20130.722) for the corresponding sets. The nomogram outperformed the clinical model and handcrafted signature across all sets (all P\u2009<\u20090.05). The nomogram model demonstrated good calibration and provided greater net benefit within the relevant threshold range compared to other models.ConclusionThis study created a deep learning nomogram using CECT images and clinical data to predict NAC response in LAGC patients undergoing surgical resection, offering personalized treatment insights.Graphical abstract","doi":"10.1007\/s00261-024-04331-7","cleaned_title":"deep learning nomogram for predicting neoadjuvant chemotherapy response in locally advanced gastric cancer patients","cleaned_abstract":"purposedeveloped and validated a deep learning radiomics nomogram using multi-phase contrast-enhanced computed tomography (cect) images to predict neoadjuvant chemotherapy (nac) response in locally advanced gastric cancer (lagc) patients.methodsthis multi-center study retrospectively included 322 patients diagnosed with gastric cancer from january 2013 to june 2023 at two hospitals. handcrafted radiomics technique and the efficientnet v2 neural network were applied to arterial, portal venous, and delayed phase ct images to extract two-dimensional handcrafted and deep learning features. a nomogram model was built by integrating the handcrafted signature, the deep learning signature, with clinical features. discriminative ability was assessed using the receiver operating characteristics (roc) curve and the precision-recall (p-r) curve. model fitting was evaluated using calibration curves, and clinical utility was assessed through decision curve analysis (dca).resultsthe nomogram exhibited excellent performance. the area under the roc curve (auc) was 0.848 [95% confidence interval (ci), 0.793\u20130.893)], 0.802 (95% ci 0.688\u20130.889), and 0.751 (95% ci 0.652\u20130.833) for the training, internal validation, and external validation sets, respectively. the aucs of the p-r curves were 0.838 (95% ci 0.756\u20130.895), 0.541 (95% ci 0.329\u20130.740), and 0.556 (95% ci 0.376\u20130.722) for the corresponding sets. the nomogram outperformed the clinical model and handcrafted signature across all sets (all p < 0.05). the nomogram model demonstrated good calibration and provided greater net benefit within the relevant threshold range compared to other models.conclusionthis study created a deep learning nomogram using cect images and clinical data to predict nac response in lagc patients undergoing surgical resection, offering personalized treatment insights.graphical abstract","key_phrases":["a deep learning radiomics nomogram","multi-phase contrast-enhanced computed tomography","cect","neoadjuvant chemotherapy (nac) response","locally advanced gastric cancer","lagc","patients.methodsthis multi-center study","322 patients","gastric cancer","january","june","two hospitals","handcrafted radiomics technique","the efficientnet v2 neural network","arterial, portal venous, and delayed phase ct images","two-dimensional handcrafted and deep learning features","a nomogram model","the handcrafted signature","the deep learning signature","clinical features","discriminative ability","the receiver operating characteristics","(roc) curve","the precision-recall (p-r) curve","model fitting","calibration curves","clinical utility","decision curve analysis","dca).resultsthe nomogram","excellent performance","the area","the roc curve","auc","[95% confidence interval","ci","0.793\u20130.893","0.802 (95%","ci 0.688\u20130.889","(95%","ci","the training","internal validation","external validation sets","the aucs","the p-r curves","0.756\u20130.895","(95%","ci 0.329\u20130.740","ci 0.376\u20130.722","the corresponding sets","the nomogram","the clinical model","signature","all sets","all p","the nomogram model","good calibration","greater net benefit","the relevant threshold range","other models.conclusionthis study","a deep learning nomogram","cect images","clinical data","nac response","lagc patients","surgical resection","personalized treatment insights.graphical abstract","322","january 2013 to june 2023","two","two","roc","roc","0.848","95%","0.802","95%","0.688\u20130.889","0.751","95%","0.652\u20130.833","0.838","95%","0.541","95%","0.329\u20130.740","0.556","95%"]},{"title":"Towards a universal mechanism for successful deep learning","authors":["Yuval Meir","Yarden Tzach","Shiri Hodassman","Ofek Tevet","Ido Kanter"],"abstract":"Recently, the underlying mechanism for successful deep learning (DL) was presented based on a quantitative method that measures the quality of a single filter in each layer of a DL model, particularly VGG-16 trained on CIFAR-10. This method exemplifies that each filter identifies small clusters of possible output labels, with additional noise selected as labels outside the clusters. This feature is progressively sharpened with each layer, resulting in an enhanced signal-to-noise ratio (SNR), which leads to an increase in the accuracy of the DL network. In this study, this mechanism is verified for VGG-16 and EfficientNet-B0 trained on the CIFAR-100 and ImageNet datasets, and the main results are as follows. First, the accuracy and SNR progressively increase with the layers. Second, for a given deep architecture, the maximal error rate increases approximately linearly with the number of output labels. Third, similar trends were obtained for dataset labels in the range [3, 1000], thus supporting the universality of this mechanism. Understanding the performance of a single filter and its dominating features paves the way to highly dilute the deep architecture without affecting its overall accuracy, and this can be achieved by applying the filter\u2019s cluster connections (AFCC).","doi":"10.1038\/s41598-024-56609-x","cleaned_title":"towards a universal mechanism for successful deep learning","cleaned_abstract":"recently, the underlying mechanism for successful deep learning (dl) was presented based on a quantitative method that measures the quality of a single filter in each layer of a dl model, particularly vgg-16 trained on cifar-10. this method exemplifies that each filter identifies small clusters of possible output labels, with additional noise selected as labels outside the clusters. this feature is progressively sharpened with each layer, resulting in an enhanced signal-to-noise ratio (snr), which leads to an increase in the accuracy of the dl network. in this study, this mechanism is verified for vgg-16 and efficientnet-b0 trained on the cifar-100 and imagenet datasets, and the main results are as follows. first, the accuracy and snr progressively increase with the layers. second, for a given deep architecture, the maximal error rate increases approximately linearly with the number of output labels. third, similar trends were obtained for dataset labels in the range [3, 1000], thus supporting the universality of this mechanism. understanding the performance of a single filter and its dominating features paves the way to highly dilute the deep architecture without affecting its overall accuracy, and this can be achieved by applying the filter\u2019s cluster connections (afcc).","key_phrases":["the underlying mechanism","successful deep learning","dl","a quantitative method","that","the quality","a single filter","each layer","a dl model","particularly vgg-16","cifar-10","this method","each filter","small clusters","possible output labels","additional noise","labels","the clusters","this feature","each layer","noise","snr","which","an increase","the accuracy","the dl network","this study","this mechanism","vgg-16","efficientnet-b0","the cifar-100","imagenet","datasets","the main results","the accuracy","the layers","a given deep architecture","the maximal error rate","the number","output labels","similar trends","dataset labels","the range","the universality","this mechanism","the performance","a single filter","its dominating features","the way","the deep architecture","its overall accuracy","this","the filter\u2019s cluster connections","afcc","cifar-10","vgg-16","first","second","third","3","1000"]},{"title":"White-box inference attack: compromising the security of deep learning-based COVID-19 diagnosis systems","authors":["Burhan Ul Haque Sheikh","Aasim Zafar"],"abstract":"The COVID-19 pandemic has necessitated the exploration of innovative diagnostic approaches, including the utilization of machine learning (ML) and deep learning (DL) technologies. However, recent findings shed light on the susceptibility of deep learning-based models to adversarial attacks, leading to erroneous predictions. This study investigates the vulnerability of a deep COVID-19 diagnosis model to the Fast Gradient Sign Method (FGSM) adversarial attack. Leveraging transfer learning of EfficientNet-B2 on a publicly available dataset, a deep learning-based COVID-19 diagnosis model is developed, achieving an impressive average accuracy of 94.56% on clean test data. However, when subjected to an untargeted FGSM attack with varying epsilon values, the model\u2019s accuracy is severely compromised, plummeting to 21.72% at epsilon 0.008. Notably, the attack successfully misclassifies adversarial COVID-19 images as normal with 100% confidence. This study underscores the critical need for further research and development to address these vulnerabilities and ensure the reliability and accuracy of deep learning models in the diagnosis of COVID-19 patients.","doi":"10.1007\/s41870-023-01538-7","cleaned_title":"white-box inference attack: compromising the security of deep learning-based covid-19 diagnosis systems","cleaned_abstract":"the covid-19 pandemic has necessitated the exploration of innovative diagnostic approaches, including the utilization of machine learning (ml) and deep learning (dl) technologies. however, recent findings shed light on the susceptibility of deep learning-based models to adversarial attacks, leading to erroneous predictions. this study investigates the vulnerability of a deep covid-19 diagnosis model to the fast gradient sign method (fgsm) adversarial attack. leveraging transfer learning of efficientnet-b2 on a publicly available dataset, a deep learning-based covid-19 diagnosis model is developed, achieving an impressive average accuracy of 94.56% on clean test data. however, when subjected to an untargeted fgsm attack with varying epsilon values, the model\u2019s accuracy is severely compromised, plummeting to 21.72% at epsilon 0.008. notably, the attack successfully misclassifies adversarial covid-19 images as normal with 100% confidence. this study underscores the critical need for further research and development to address these vulnerabilities and ensure the reliability and accuracy of deep learning models in the diagnosis of covid-19 patients.","key_phrases":["the exploration","innovative diagnostic approaches","the utilization","machine learning","ml","deep learning","(dl) technologies","recent findings","light","the susceptibility","deep learning-based models","adversarial attacks","erroneous predictions","this study","the vulnerability","a deep covid-19 diagnosis model","the fast gradient sign method","fgsm) adversarial attack","transfer learning","efficientnet-b2","a publicly available dataset","a deep learning-based covid-19 diagnosis model","an impressive average accuracy","94.56%","clean test data","an untargeted fgsm attack","varying epsilon values","the model\u2019s accuracy","21.72%","epsilon","the attack","adversarial covid-19 images","100% confidence","this study","the critical need","further research","development","these vulnerabilities","the reliability","accuracy","deep learning models","the diagnosis","covid-19 patients","covid-19","covid-19","covid-19","94.56%","21.72%","0.008","covid-19","100%","covid-19"]},{"title":"Detection of vulnerabilities in blockchain smart contracts using deep learning","authors":["Namya Aankur Gupta","Mansi Bansal","Seema Sharma","Deepti Mehrotra","Misha Kakkar"],"abstract":"Blockchain helps to give a sense of security as there is only one history of transactions visible to all the involved parties. Smart contracts enable users to manage significant asset amounts of finances on the blockchain without the involvement of any intermediaries. The conditions and checks that have been written in smart contract and executed to the application cannot be changed again. However, these unique features pose some other risks to the smart contract. Smart contracts have several flaws in its programmable language and methods of execution, despite being a developing technology. To build smart contracts and implement numerous complicated business logics, high-level languages are used by the developers to code smart contracts. Thus, blockchain smart contract is the most important element of any decentralized application, posing the risk for it to be attacked. So, the presence of vulnerabilities are to be taken care of on a priority basis. It is important for detection of vulnerabilities in a smart contract and only then implement and connect it with applications to ensure security of funds. The motive of the paper is to discuss how deep learning may be utilized to deliver bug-free secure smart contracts. Objective of the paper is to detect three kinds of vulnerabilities- reentrancy, timestamp and infinite loop. A deep learning model has been created for detection of smart contract vulnerabilities using graph neural networks. The performance of this model has been compared to the present automated tools and other independent methods. It has been shown that this model has greater accuracy than other methods while comparing the prediction of smart contract vulnerabilities in existing models.","doi":"10.1007\/s11276-024-03755-9","cleaned_title":"detection of vulnerabilities in blockchain smart contracts using deep learning","cleaned_abstract":"blockchain helps to give a sense of security as there is only one history of transactions visible to all the involved parties. smart contracts enable users to manage significant asset amounts of finances on the blockchain without the involvement of any intermediaries. the conditions and checks that have been written in smart contract and executed to the application cannot be changed again. however, these unique features pose some other risks to the smart contract. smart contracts have several flaws in its programmable language and methods of execution, despite being a developing technology. to build smart contracts and implement numerous complicated business logics, high-level languages are used by the developers to code smart contracts. thus, blockchain smart contract is the most important element of any decentralized application, posing the risk for it to be attacked. so, the presence of vulnerabilities are to be taken care of on a priority basis. it is important for detection of vulnerabilities in a smart contract and only then implement and connect it with applications to ensure security of funds. the motive of the paper is to discuss how deep learning may be utilized to deliver bug-free secure smart contracts. objective of the paper is to detect three kinds of vulnerabilities- reentrancy, timestamp and infinite loop. a deep learning model has been created for detection of smart contract vulnerabilities using graph neural networks. the performance of this model has been compared to the present automated tools and other independent methods. it has been shown that this model has greater accuracy than other methods while comparing the prediction of smart contract vulnerabilities in existing models.","key_phrases":["blockchain","a sense","security","only one history","transactions","all the involved parties","smart contracts","users","significant asset amounts","finances","the blockchain","the involvement","any intermediaries","the conditions","checks","that","smart contract","the application","these unique features","some other risks","the smart contract","smart contracts","several flaws","its programmable language","methods","execution","a developing technology","smart contracts","numerous complicated business logics","high-level languages","the developers","smart contracts","blockchain smart contract","the most important element","any decentralized application","the risk","it","the presence","vulnerabilities","care","a priority basis","it","detection","vulnerabilities","a smart contract","it","applications","security","funds","the motive","the paper","how deep learning","bug-free secure smart contracts","objective","the paper","three kinds","vulnerabilities- reentrancy","timestamp","infinite loop","a deep learning model","detection","smart contract vulnerabilities","graph neural networks","the performance","this model","the present automated tools","other independent methods","it","this model","greater accuracy","other methods","the prediction","smart contract vulnerabilities","existing models","three"]},{"title":"Integrating Deep Learning and Reinforcement Learning for Enhanced Financial Risk Forecasting in Supply Chain Management","authors":["Yuanfei Cui","Fengtong Yao"],"abstract":"In today\u2019s dynamic business landscape, the integration of supply chain management and financial risk forecasting is imperative for sustained success. This research paper introduces a groundbreaking approach that seamlessly merges deep autoencoder (DAE) models with reinforcement learning (RL) techniques to enhance financial risk forecasting within the realm of supply chain management. The primary objective of this research is to optimize financial decision-making processes by extracting key feature representations from financial data and leveraging RL for decision optimization. To achieve this, the paper presents the PSO-SDAE model, a novel and sophisticated approach to financial risk forecasting. By incorporating advanced noise reduction features and optimization algorithms, the PSO-SDAE model significantly enhances the accuracy and reliability of financial risk predictions. Notably, the PSO-SDAE model goes beyond traditional forecasting methods by addressing the need for real-time decision-making in the rapidly evolving landscape of financial risk management. This is achieved through the utilization of a distributed RL algorithm, which expedites the processing of supply chain data while maintaining both efficiency and accuracy. The results of our study showcase the exceptional precision of the PSO-SDAE model in predicting financial risks, underscoring its efficacy for proactive risk management within supply chain operations. Moreover, the augmented processing speed of the model enables real-time analysis and decision-making \u2014 a critical capability in today\u2019s fast-paced business environment.","doi":"10.1007\/s13132-024-01946-5","cleaned_title":"integrating deep learning and reinforcement learning for enhanced financial risk forecasting in supply chain management","cleaned_abstract":"in today\u2019s dynamic business landscape, the integration of supply chain management and financial risk forecasting is imperative for sustained success. this research paper introduces a groundbreaking approach that seamlessly merges deep autoencoder (dae) models with reinforcement learning (rl) techniques to enhance financial risk forecasting within the realm of supply chain management. the primary objective of this research is to optimize financial decision-making processes by extracting key feature representations from financial data and leveraging rl for decision optimization. to achieve this, the paper presents the pso-sdae model, a novel and sophisticated approach to financial risk forecasting. by incorporating advanced noise reduction features and optimization algorithms, the pso-sdae model significantly enhances the accuracy and reliability of financial risk predictions. notably, the pso-sdae model goes beyond traditional forecasting methods by addressing the need for real-time decision-making in the rapidly evolving landscape of financial risk management. this is achieved through the utilization of a distributed rl algorithm, which expedites the processing of supply chain data while maintaining both efficiency and accuracy. the results of our study showcase the exceptional precision of the pso-sdae model in predicting financial risks, underscoring its efficacy for proactive risk management within supply chain operations. moreover, the augmented processing speed of the model enables real-time analysis and decision-making \u2014 a critical capability in today\u2019s fast-paced business environment.","key_phrases":["today\u2019s dynamic business landscape","the integration","supply chain management","financial risk forecasting","sustained success","this research paper","a groundbreaking approach","that","deep autoencoder (dae) models","reinforcement learning","(rl) techniques","financial risk forecasting","the realm","supply chain management","the primary objective","this research","financial decision-making processes","key feature representations","financial data","rl","decision optimization","this","the paper","the pso-sdae model","a novel and sophisticated approach","financial risk forecasting","advanced noise reduction features","optimization algorithms","the pso-sdae model","the accuracy","reliability","financial risk predictions","the pso-sdae model","traditional forecasting methods","the need","real-time decision-making","the rapidly evolving landscape","financial risk management","this","the utilization","a distributed rl algorithm","which","the processing","supply chain data","both efficiency","accuracy","the results","our study showcase","the exceptional precision","the pso-sdae model","financial risks","its efficacy","proactive risk management","supply chain operations","the augmented processing speed","the model","real-time analysis","decision-making","a critical capability","today\u2019s fast-paced business environment","today","dae","today"]},{"title":"ZS-DML: Zero-Shot Deep Metric Learning approach for plant leaf disease classification","authors":["Davood Zabihzadeh","Mina Masoudifar"],"abstract":"Automatic plant disease detection plays an important role in food security. Deep learning methods are able to detect precisely various types of plant diseases but at the expense of using huge amounts of resources (processors and data). Therefore, employing few-shot or zero-shot learning methods is unavoidable. Deep Metric Learning (DML) is a widely used technique for few\/zero shot learning. Existing DML methods extract features from the last hidden layer of a pre-trained deep network, which increases the dependence of the specific features on the observed classes. In this paper, the general discriminative feature learning method is used to learn general features of plant leaves. Moreover, a proxy-based loss is utilized that learns the embedding without sampling phase while having a higher convergence rate. The network is trained on the Plant Village dataset where the images are split into 32 and 6 classes as source and target, respectively. The knowledge learned from the source domain is transferred to the target in a zero-shot setting. A few samples of the target domain are presented to the network as a gallery. The network is then evaluated on the target domain. The experimental results show that by presenting few or even only one sample of new classes to the network without fine-tuning step, our method can achieve a classification accuracy of 99%\/80.64% for few\/one image(s) per class.","doi":"10.1007\/s11042-023-17136-5","cleaned_title":"zs-dml: zero-shot deep metric learning approach for plant leaf disease classification","cleaned_abstract":"automatic plant disease detection plays an important role in food security. deep learning methods are able to detect precisely various types of plant diseases but at the expense of using huge amounts of resources (processors and data). therefore, employing few-shot or zero-shot learning methods is unavoidable. deep metric learning (dml) is a widely used technique for few\/zero shot learning. existing dml methods extract features from the last hidden layer of a pre-trained deep network, which increases the dependence of the specific features on the observed classes. in this paper, the general discriminative feature learning method is used to learn general features of plant leaves. moreover, a proxy-based loss is utilized that learns the embedding without sampling phase while having a higher convergence rate. the network is trained on the plant village dataset where the images are split into 32 and 6 classes as source and target, respectively. the knowledge learned from the source domain is transferred to the target in a zero-shot setting. a few samples of the target domain are presented to the network as a gallery. the network is then evaluated on the target domain. the experimental results show that by presenting few or even only one sample of new classes to the network without fine-tuning step, our method can achieve a classification accuracy of 99%\/80.64% for few\/one image(s) per class.","key_phrases":["automatic plant disease detection","an important role","food security","deep learning methods","precisely various types","plant diseases","the expense","huge amounts","resources","processors","data","few-shot or zero-shot learning methods","deep metric learning","dml","a widely used technique","few\/zero shot learning","existing dml methods","features","the last hidden layer","a pre-trained deep network","which","the dependence","the specific features","the observed classes","this paper","the general discriminative feature learning method","general features","plant leaves","a proxy-based loss","that","the embedding","sampling phase","a higher convergence rate","the network","the plant village dataset","the images","32 and 6 classes","source","target","the knowledge","the source domain","the target","a zero-shot setting","a few samples","the target domain","the network","a gallery","the network","the target domain","the experimental results","few or even only one sample","new classes","the network","fine-tuning step","our method","a classification accuracy","99%\/80.64%","few\/one image(s","class","zero","32","6","zero","only one","99%\/80.64%","one"]},{"title":"Deep Learning Approaches for Automatic Quality Assurance of Magnetic Resonance Images Using ACR Phantom","authors":["Tarraf Torfeh","Souha Aouadi","SA Yoganathan","Satheesh Paloor","Rabih Hammoud","Noora Al-Hammadi"],"abstract":"BackgroundIn recent years, there has been a growing trend towards utilizing Artificial Intelligence (AI) and machine learning techniques in medical imaging, including for the purpose of automating quality assurance. In this research, we aimed to develop and evaluate various deep learning-based approaches for automatic quality assurance of Magnetic Resonance (MR) images using the American College of Radiology (ACR) standards.MethodsThe study involved the development, optimization, and testing of custom convolutional neural network (CNN) models. Additionally, popular pre-trained models such as VGG16, VGG19, ResNet50, InceptionV3, EfficientNetB0, and EfficientNetB5 were trained and tested. The use of pre-trained models, particularly those trained on the ImageNet dataset, for transfer learning was also explored. Two-class classification models were employed for assessing spatial resolution and geometric distortion, while an approach classifying the image into 10 classes representing the number of visible spokes was used for the low contrast.ResultsOur results showed that deep learning-based methods can be effectively used for MR image quality assurance and can improve the performance of these models. The low contrast test was one of the most challenging tests within the ACR phantom.ConclusionsOverall, for geometric distortion and spatial resolution, all of the deep learning models tested produced prediction accuracy of 80% or higher. The study also revealed that training the models from scratch performed slightly better compared to transfer learning. For the low contrast, our investigation emphasized the adaptability and potential of deep learning models. The custom CNN models excelled in predicting the number of visible spokes, achieving commendable accuracy, recall, precision, and F1 scores.","doi":"10.1186\/s12880-023-01157-5","cleaned_title":"deep learning approaches for automatic quality assurance of magnetic resonance images using acr phantom","cleaned_abstract":"backgroundin recent years, there has been a growing trend towards utilizing artificial intelligence (ai) and machine learning techniques in medical imaging, including for the purpose of automating quality assurance. in this research, we aimed to develop and evaluate various deep learning-based approaches for automatic quality assurance of magnetic resonance (mr) images using the american college of radiology (acr) standards.methodsthe study involved the development, optimization, and testing of custom convolutional neural network (cnn) models. additionally, popular pre-trained models such as vgg16, vgg19, resnet50, inceptionv3, efficientnetb0, and efficientnetb5 were trained and tested. the use of pre-trained models, particularly those trained on the imagenet dataset, for transfer learning was also explored. two-class classification models were employed for assessing spatial resolution and geometric distortion, while an approach classifying the image into 10 classes representing the number of visible spokes was used for the low contrast.resultsour results showed that deep learning-based methods can be effectively used for mr image quality assurance and can improve the performance of these models. the low contrast test was one of the most challenging tests within the acr phantom.conclusionsoverall, for geometric distortion and spatial resolution, all of the deep learning models tested produced prediction accuracy of 80% or higher. the study also revealed that training the models from scratch performed slightly better compared to transfer learning. for the low contrast, our investigation emphasized the adaptability and potential of deep learning models. the custom cnn models excelled in predicting the number of visible spokes, achieving commendable accuracy, recall, precision, and f1 scores.","key_phrases":["a growing trend","artificial intelligence","ai","techniques","medical imaging","the purpose","quality assurance","this research","we","various deep learning-based approaches","automatic quality assurance","magnetic resonance","(mr) images","the american college","radiology","standards.methodsthe study","the development","optimization","testing","custom convolutional neural network (cnn) models","popular pre-trained models","vgg16","vgg19","resnet50","inceptionv3","efficientnetb0","efficientnetb5","the use","pre-trained models","particularly those","the imagenet dataset","transfer learning","two-class classification models","spatial resolution","geometric distortion","an approach","the image","10 classes","the number","visible spokes","the low contrast.resultsour results","deep learning-based methods","mr image quality assurance","the performance","these models","the low contrast test","the most challenging tests","the acr phantom.conclusionsoverall","geometric distortion","spatial resolution","all","the deep learning models","produced prediction accuracy","80%","the study","the models","scratch","learning","the low contrast","our investigation","the adaptability","potential","deep learning models","the custom cnn models","the number","visible spokes","commendable accuracy","recall","precision","f1 scores","backgroundin recent years","american","cnn","resnet50","inceptionv3","efficientnetb0","efficientnetb5","two","10","80%","cnn"]},{"title":"Depression clinical detection model based on social media: a federated deep learning approach","authors":["Yang Liu"],"abstract":"Depression can significantly impact people\u2019s mental health, and recent research shows that social media can provide decision-making support for health-care professionals and serve as supplementary information for understanding patients\u2019 health status. Deep learning models are also able to assess an individual\u2019s likelihood of experiencing depression. However, data availability on social media is often limited due to privacy concerns, even though deep learning models benefit from having more data to analyze. To address this issue, this study proposes a methodological framework system for clinical decision support that uses federated deep learning (FDL) to identify individuals experiencing depression and provide intervention decisions for clinicians. The proposed framework involves evaluation of datasets from three social media platforms, and the experimental results demonstrate that our method achieves state-of-the-art results. The study aims to provide a clinical decision support system with evolvable features that can deliver precise solutions and assist health-care professionals in medical diagnosis. The proposed framework that incorporates social media data and deep learning models can provide valuable insights into patients\u2019 health status, support personalized treatment decisions, and adapt to changing health-care needs.","doi":"10.1007\/s11227-023-05754-7","cleaned_title":"depression clinical detection model based on social media: a federated deep learning approach","cleaned_abstract":"depression can significantly impact people\u2019s mental health, and recent research shows that social media can provide decision-making support for health-care professionals and serve as supplementary information for understanding patients\u2019 health status. deep learning models are also able to assess an individual\u2019s likelihood of experiencing depression. however, data availability on social media is often limited due to privacy concerns, even though deep learning models benefit from having more data to analyze. to address this issue, this study proposes a methodological framework system for clinical decision support that uses federated deep learning (fdl) to identify individuals experiencing depression and provide intervention decisions for clinicians. the proposed framework involves evaluation of datasets from three social media platforms, and the experimental results demonstrate that our method achieves state-of-the-art results. the study aims to provide a clinical decision support system with evolvable features that can deliver precise solutions and assist health-care professionals in medical diagnosis. the proposed framework that incorporates social media data and deep learning models can provide valuable insights into patients\u2019 health status, support personalized treatment decisions, and adapt to changing health-care needs.","key_phrases":["depression","people\u2019s mental health","recent research","social media","decision-making support","health-care professionals","supplementary information","patients\u2019 health status","deep learning models","an individual\u2019s likelihood","depression","data availability","social media","privacy concerns","deep learning models","more data","this issue","this study","a methodological framework system","clinical decision support","that","deep learning","individuals","depression","intervention decisions","clinicians","the proposed framework","evaluation","datasets","three social media platforms","the experimental results","our method","the-art","the study","a clinical decision support system","evolvable features","that","precise solutions","health-care professionals","medical diagnosis","the proposed framework","that","social media data","deep learning models","valuable insights","patients\u2019 health status","personalized treatment decisions","health-care needs","clinicians","three"]},{"title":"Fast reconstruction of EEG signal compression sensing based on deep learning","authors":["XiuLi Du","KuanYang Liang","YaNa Lv","ShaoMing Qiu"],"abstract":"When traditional EEG signals are collected based on the Nyquist theorem, long-time recordings of EEG signals will produce a large amount of data. At the same time, limited bandwidth, end-to-end delay, and memory space will bring great pressure on the effective transmission of data. The birth of compressed sensing alleviates this transmission pressure. However, using an iterative compressed sensing reconstruction algorithm for EEG signal reconstruction faces complex calculation problems and slow data processing speed, limiting the application of compressed sensing in EEG signal rapid monitoring systems. As such, this paper presents a non-iterative and fast algorithm for reconstructing EEG signals using compressed sensing and deep learning techniques. This algorithm uses the improved residual network model, extracts the feature information of the EEG signal by one-dimensional dilated convolution, directly learns the nonlinear mapping relationship between the measured value and the original signal, and can quickly and accurately reconstruct the EEG signal. The method proposed in this paper has been verified by simulation on the open BCI contest dataset. Overall, it is proved that the proposed method has higher reconstruction accuracy and faster reconstruction speed than the traditional CS reconstruction algorithm and the existing deep learning reconstruction algorithm. In addition, it can realize the rapid reconstruction of EEG signals.","doi":"10.1038\/s41598-024-55334-9","cleaned_title":"fast reconstruction of eeg signal compression sensing based on deep learning","cleaned_abstract":"when traditional eeg signals are collected based on the nyquist theorem, long-time recordings of eeg signals will produce a large amount of data. at the same time, limited bandwidth, end-to-end delay, and memory space will bring great pressure on the effective transmission of data. the birth of compressed sensing alleviates this transmission pressure. however, using an iterative compressed sensing reconstruction algorithm for eeg signal reconstruction faces complex calculation problems and slow data processing speed, limiting the application of compressed sensing in eeg signal rapid monitoring systems. as such, this paper presents a non-iterative and fast algorithm for reconstructing eeg signals using compressed sensing and deep learning techniques. this algorithm uses the improved residual network model, extracts the feature information of the eeg signal by one-dimensional dilated convolution, directly learns the nonlinear mapping relationship between the measured value and the original signal, and can quickly and accurately reconstruct the eeg signal. the method proposed in this paper has been verified by simulation on the open bci contest dataset. overall, it is proved that the proposed method has higher reconstruction accuracy and faster reconstruction speed than the traditional cs reconstruction algorithm and the existing deep learning reconstruction algorithm. in addition, it can realize the rapid reconstruction of eeg signals.","key_phrases":["traditional eeg signals","the nyquist theorem","long-time recordings","eeg signals","a large amount","data","the same time","end","memory space","great pressure","the effective transmission","data","the birth","this transmission pressure","an iterative compressed sensing reconstruction algorithm","eeg signal reconstruction","complex calculation problems","slow data processing speed","the application","compressed sensing","eeg signal rapid monitoring systems","this paper","a non-iterative and fast algorithm","eeg signals","compressed sensing","deep learning techniques","this algorithm","the improved residual network model","the feature information","the eeg signal","one-dimensional dilated convolution","the nonlinear mapping relationship","the measured value","the original signal","the eeg signal","the method","this paper","simulation","the open bci contest dataset","it","the proposed method","higher reconstruction accuracy","faster reconstruction speed","the traditional cs reconstruction algorithm","the existing deep learning reconstruction algorithm","addition","it","the rapid reconstruction","eeg signals","one","bci contest"]},{"title":"Detecting schizophrenia with 3D structural brain MRI using deep learning","authors":["Junhao Zhang","Vishwanatha M. Rao","Ye Tian","Yanting Yang","Nicolas Acosta","Zihan Wan","Pin-Yu Lee","Chloe Zhang","Lawrence S. Kegeles","Scott A. Small","Jia Guo"],"abstract":"Schizophrenia is a chronic neuropsychiatric disorder that causes distinct structural alterations within the brain. We hypothesize that deep learning applied to a structural neuroimaging dataset could detect disease-related alteration and improve classification and diagnostic accuracy. We tested this hypothesis using a single, widely available, and conventional T1-weighted MRI scan, from which we extracted the 3D whole-brain structure using standard post-processing methods. A deep learning model was then developed, optimized, and evaluated on three open datasets with T1-weighted MRI scans of patients with schizophrenia. Our proposed model outperformed the benchmark model, which was also trained with structural MR images using a 3D CNN architecture. Our model is capable of almost perfectly (area under the ROC curve\u2009=\u20090.987) distinguishing schizophrenia patients from healthy controls on unseen structural MRI scans. Regional analysis localized subcortical regions and ventricles as the most predictive brain regions. Subcortical structures serve a pivotal role in cognitive, affective, and social functions in humans, and structural abnormalities of these regions have been associated with schizophrenia. Our finding corroborates that schizophrenia is associated with widespread alterations in subcortical brain structure and the subcortical structural information provides prominent features in diagnostic classification. Together, these results further demonstrate the potential of deep learning to improve schizophrenia diagnosis and identify its structural neuroimaging signatures from a single, standard T1-weighted brain MRI.","doi":"10.1038\/s41598-023-41359-z","cleaned_title":"detecting schizophrenia with 3d structural brain mri using deep learning","cleaned_abstract":"schizophrenia is a chronic neuropsychiatric disorder that causes distinct structural alterations within the brain. we hypothesize that deep learning applied to a structural neuroimaging dataset could detect disease-related alteration and improve classification and diagnostic accuracy. we tested this hypothesis using a single, widely available, and conventional t1-weighted mri scan, from which we extracted the 3d whole-brain structure using standard post-processing methods. a deep learning model was then developed, optimized, and evaluated on three open datasets with t1-weighted mri scans of patients with schizophrenia. our proposed model outperformed the benchmark model, which was also trained with structural mr images using a 3d cnn architecture. our model is capable of almost perfectly (area under the roc curve = 0.987) distinguishing schizophrenia patients from healthy controls on unseen structural mri scans. regional analysis localized subcortical regions and ventricles as the most predictive brain regions. subcortical structures serve a pivotal role in cognitive, affective, and social functions in humans, and structural abnormalities of these regions have been associated with schizophrenia. our finding corroborates that schizophrenia is associated with widespread alterations in subcortical brain structure and the subcortical structural information provides prominent features in diagnostic classification. together, these results further demonstrate the potential of deep learning to improve schizophrenia diagnosis and identify its structural neuroimaging signatures from a single, standard t1-weighted brain mri.","key_phrases":["schizophrenia","a chronic neuropsychiatric disorder","that","distinct structural alterations","the brain","we","deep learning","a structural neuroimaging dataset","disease-related alteration","classification and diagnostic accuracy","we","this hypothesis","conventional t1-weighted mri scan","which","we","the 3d whole-brain structure","standard post-processing methods","a deep learning model","three open datasets","t1-weighted mri scans","patients","schizophrenia","our proposed model","the benchmark model","which","structural mr images","a 3d cnn architecture","our model","almost perfectly (area","the roc curve","schizophrenia patients","healthy controls","unseen structural mri scans","regional analysis","subcortical regions","ventricles","the most predictive brain regions","subcortical structures","a pivotal role","cognitive, affective, and social functions","humans","structural abnormalities","these regions","schizophrenia","our finding","schizophrenia","widespread alterations","subcortical brain structure","the subcortical structural information","prominent features","diagnostic classification","these results","the potential","deep learning","schizophrenia diagnosis","its structural neuroimaging signatures","a single, standard t1-weighted brain mri","schizophrenia","3d","three","3d","cnn","roc","0.987","schizophrenia"]},{"title":"An end-to-end intrusion detection system with IoT dataset using deep learning with unsupervised feature extraction","authors":["Yesi Novaria Kunang","Siti Nurmaini","Deris Stiawan","Bhakti Yudho Suprapto"],"abstract":"The rapid growth of the Internet of things (IoT) platform has implications on security vulnerabilities that need to be resolved. This requires an intrusion detection system (IDS) to secure attacks on the platforms. In line with this, numerous machine and deep learning algorithms have been adopted to detect cyber-attacks. Real-time IoT devices transmit massive amounts of heterogeneous data, which affects the network. Traffic networks generate redundant and large amounts of data that must be reduced before processing. This study proposed a hybrid deep learning model for an IDS on the IoT platform. We used unsupervised approaches to extract data dimensions and features, then a neural network for classification. Several approaches were used to determine the effectiveness of the deep learning-based IoT IDS with two scenarios of feature extraction. The first case used autoencoder variants such as deep autoencoder (DAE), deep LSTM autoencoder (LSTM-DAE), and deep convolutional autoencoder. The second case used stacked models for feature extraction, including stacked autoencoder and deep belief network. The feature extraction output from the five models was fine-tuned to the fully connected layer using the BoT-IoT dataset. The results showed a good detection performance of almost 100% and a false positive rate (FPR) of nearly 0%. On the CSE-CIC-IDS2018 dataset, the proposed deep learning model was evaluated using a transfer learning approach with the highest detection rate of 99.17% and the lowest FPR of 0.18%. The model developed from the feature extraction process recognized attacks significantly better than the previous approach.","doi":"10.1007\/s10207-023-00807-7","cleaned_title":"an end-to-end intrusion detection system with iot dataset using deep learning with unsupervised feature extraction","cleaned_abstract":"the rapid growth of the internet of things (iot) platform has implications on security vulnerabilities that need to be resolved. this requires an intrusion detection system (ids) to secure attacks on the platforms. in line with this, numerous machine and deep learning algorithms have been adopted to detect cyber-attacks. real-time iot devices transmit massive amounts of heterogeneous data, which affects the network. traffic networks generate redundant and large amounts of data that must be reduced before processing. this study proposed a hybrid deep learning model for an ids on the iot platform. we used unsupervised approaches to extract data dimensions and features, then a neural network for classification. several approaches were used to determine the effectiveness of the deep learning-based iot ids with two scenarios of feature extraction. the first case used autoencoder variants such as deep autoencoder (dae), deep lstm autoencoder (lstm-dae), and deep convolutional autoencoder. the second case used stacked models for feature extraction, including stacked autoencoder and deep belief network. the feature extraction output from the five models was fine-tuned to the fully connected layer using the bot-iot dataset. the results showed a good detection performance of almost 100% and a false positive rate (fpr) of nearly 0%. on the cse-cic-ids2018 dataset, the proposed deep learning model was evaluated using a transfer learning approach with the highest detection rate of 99.17% and the lowest fpr of 0.18%. the model developed from the feature extraction process recognized attacks significantly better than the previous approach.","key_phrases":["the rapid growth","the internet","things","(iot) platform","implications","security vulnerabilities","that","this","an intrusion detection system","ids","attacks","the platforms","line","this","numerous machine","deep learning algorithms","cyber-attacks","real-time iot devices","massive amounts","heterogeneous data","which","the network","traffic networks","redundant and large amounts","data","that","processing","this study","a hybrid deep learning model","an ids","the iot platform","we","unsupervised approaches","data dimensions","features",", then a neural network","classification","several approaches","the effectiveness","the deep learning-based iot ids","two scenarios","feature extraction","the first case","autoencoder variants","deep autoencoder","dae","deep lstm autoencoder","lstm-dae","deep convolutional autoencoder","the second case","stacked models","feature extraction","autoencoder","deep belief network","the feature extraction output","the five models","the fully connected layer","the bot-iot dataset","the results","a good detection performance","almost 100%","a false positive rate","fpr","nearly 0%","the cse-cic-ids2018 dataset","the proposed deep learning model","a transfer learning approach","the highest detection rate","99.17%","the lowest fpr","0.18%","the model","the feature extraction process","attacks","the previous approach","two","first","dae","second","five","almost 100%","nearly 0%","99.17%","0.18%"]},{"title":"Application of simultaneous uncertainty quantification and segmentation for oropharyngeal cancer use-case with Bayesian deep learning","authors":["Jaakko Sahlsten","Joel Jaskari","Kareem A. Wahid","Sara Ahmed","Enrico Glerean","Renjie He","Benjamin H. Kann","Antti M\u00e4kitie","Clifton D. Fuller","Mohamed A. Naser","Kimmo Kaski"],"abstract":"BackgroundRadiotherapy is a core treatment modality for oropharyngeal cancer (OPC), where the primary gross tumor volume (GTVp) is manually segmented with high interobserver variability. This calls for reliable and trustworthy automated tools in clinician workflow. Therefore, accurate uncertainty quantification and its downstream utilization is critical.MethodsHere we propose uncertainty-aware deep learning for OPC GTVp segmentation, and illustrate the utility of uncertainty in multiple applications. We examine two Bayesian deep learning (BDL) models and eight uncertainty measures, and utilize a large multi-institute dataset of 292 PET\/CT scans to systematically analyze our approach.ResultsWe show that our uncertainty-based approach accurately predicts the quality of the deep learning segmentation in 86.6% of cases, identifies low performance cases for semi-automated correction, and visualizes regions of the scans where the segmentations likely fail.ConclusionsOur BDL-based analysis provides a first-step towards more widespread implementation of uncertainty quantification in OPC GTVp segmentation.","doi":"10.1038\/s43856-024-00528-5","cleaned_title":"application of simultaneous uncertainty quantification and segmentation for oropharyngeal cancer use-case with bayesian deep learning","cleaned_abstract":"backgroundradiotherapy is a core treatment modality for oropharyngeal cancer (opc), where the primary gross tumor volume (gtvp) is manually segmented with high interobserver variability. this calls for reliable and trustworthy automated tools in clinician workflow. therefore, accurate uncertainty quantification and its downstream utilization is critical.methodshere we propose uncertainty-aware deep learning for opc gtvp segmentation, and illustrate the utility of uncertainty in multiple applications. we examine two bayesian deep learning (bdl) models and eight uncertainty measures, and utilize a large multi-institute dataset of 292 pet\/ct scans to systematically analyze our approach.resultswe show that our uncertainty-based approach accurately predicts the quality of the deep learning segmentation in 86.6% of cases, identifies low performance cases for semi-automated correction, and visualizes regions of the scans where the segmentations likely fail.conclusionsour bdl-based analysis provides a first-step towards more widespread implementation of uncertainty quantification in opc gtvp segmentation.","key_phrases":["backgroundradiotherapy","a core treatment modality","oropharyngeal cancer","the primary gross tumor volume","gtvp","high interobserver variability","this","reliable and trustworthy automated tools","clinician workflow","accurate uncertainty quantification","its downstream utilization","we","uncertainty-aware deep learning","opc","gtvp segmentation","the utility","uncertainty","multiple applications","we","two bayesian deep learning","bdl) models","eight uncertainty measures","a large multi-institute dataset","292 pet\/ct","our approach.resultswe show","our uncertainty-based approach","the quality","the deep learning segmentation","86.6%","cases","low performance cases","semi-automated correction","regions","the scans","the segmentations","bdl-based analysis","a first-step","more widespread implementation","uncertainty quantification","gtvp segmentation","bayesian","eight","292","86.6%","fail.conclusionsour bdl","first"]},{"title":"Insights from EEG analysis of evoked memory recalls using deep learning for emotion charting","authors":["Muhammad Najam Dar","Muhammad Usman Akram","Ahmad Rauf Subhani","Sajid Gul Khawaja","Constantino Carlos Reyes-Aldasoro","Sarah Gul"],"abstract":"Affect recognition in a real-world, less constrained environment is the principal prerequisite of the industrial-level usefulness of this technology. Monitoring the psychological profile using smart, wearable electroencephalogram (EEG) sensors during daily activities without external stimuli, such as memory-induced emotions, is a challenging research gap in emotion recognition. This paper proposed a deep learning framework for improved memory-induced emotion recognition leveraging a combination of 1D-CNN and LSTM as feature extractors integrated with an Extreme Learning Machine (ELM) classifier. The proposed deep learning architecture, combined with the EEG preprocessing, such as the removal of the average baseline signal from each sample and extraction of EEG rhythms (delta, theta, alpha, beta, and gamma), aims to capture repetitive and continuous patterns for memory-induced emotion recognition, underexplored with deep learning techniques. This work has analyzed EEG signals using a wearable, ultra-mobile sports cap while recalling autobiographical emotional memories evoked by affect-denoting words, with self-annotation on the scale of valence and arousal. With extensive experimentation using the same dataset, the proposed framework empirically outperforms existing techniques for the emerging area of memory-induced emotion recognition with an accuracy of 65.6%. The EEG rhythms analysis, such as delta, theta, alpha, beta, and gamma, achieved 65.5%, 52.1%, 65.1%, 64.6%, and 65.0% accuracies for classification with four quadrants of valence and arousal. These results underscore the significant advancement achieved by our proposed method for the real-world environment of memory-induced emotion recognition.","doi":"10.1038\/s41598-024-61832-7","cleaned_title":"insights from eeg analysis of evoked memory recalls using deep learning for emotion charting","cleaned_abstract":"affect recognition in a real-world, less constrained environment is the principal prerequisite of the industrial-level usefulness of this technology. monitoring the psychological profile using smart, wearable electroencephalogram (eeg) sensors during daily activities without external stimuli, such as memory-induced emotions, is a challenging research gap in emotion recognition. this paper proposed a deep learning framework for improved memory-induced emotion recognition leveraging a combination of 1d-cnn and lstm as feature extractors integrated with an extreme learning machine (elm) classifier. the proposed deep learning architecture, combined with the eeg preprocessing, such as the removal of the average baseline signal from each sample and extraction of eeg rhythms (delta, theta, alpha, beta, and gamma), aims to capture repetitive and continuous patterns for memory-induced emotion recognition, underexplored with deep learning techniques. this work has analyzed eeg signals using a wearable, ultra-mobile sports cap while recalling autobiographical emotional memories evoked by affect-denoting words, with self-annotation on the scale of valence and arousal. with extensive experimentation using the same dataset, the proposed framework empirically outperforms existing techniques for the emerging area of memory-induced emotion recognition with an accuracy of 65.6%. the eeg rhythms analysis, such as delta, theta, alpha, beta, and gamma, achieved 65.5%, 52.1%, 65.1%, 64.6%, and 65.0% accuracies for classification with four quadrants of valence and arousal. these results underscore the significant advancement achieved by our proposed method for the real-world environment of memory-induced emotion recognition.","key_phrases":["recognition","a real-world, less constrained environment","the principal prerequisite","the industrial-level usefulness","this technology","the psychological profile","smart, wearable electroencephalogram (eeg) sensors","daily activities","external stimuli","memory-induced emotions","a challenging research gap","emotion recognition","this paper","a deep learning framework","improved memory-induced emotion recognition","a combination","1d-cnn","lstm","feature extractors","an extreme learning machine","(elm) classifier","the proposed deep learning architecture","the eeg preprocessing","the removal","the average baseline signal","each sample","extraction","eeg rhythms","delta","theta","alpha","beta","gamma","repetitive and continuous patterns","memory-induced emotion recognition","deep learning techniques","this work","eeg signals","a wearable, ultra-mobile sports cap","autobiographical emotional memories","affect-denoting words","self-annotation","the scale","valence","arousal","extensive experimentation","the same dataset","the proposed framework","existing techniques","the emerging area","memory-induced emotion recognition","an accuracy","65.6%","the eeg rhythms analysis","delta","theta","alpha","beta","gamma","65.5%","52.1%","65.1%","64.6%","65.0% accuracies","classification","four quadrants","valence","arousal","these results","the significant advancement","our proposed method","the real-world environment","memory-induced emotion recognition","daily","1d","65.6%","65.5%","52.1%","65.1%","64.6%","65.0%","four"]},{"title":"Application of machine learning and deep learning for cancer vaccine (rapid review)","authors":["Mohaddeseh Nasiri Hooshmand","Elham Maserat"],"abstract":"Cancer is a common and dangerous disease based on the World Health Organization. Much research has been done on new and effective cancer treatments, including cancer vaccines and the prediction of neoantigens using machine learning. The purpose of this study is to review articles that use machine learning to design cancer vaccines. This study is a rapid review study using search strategies and related keywords in Google Scholar, PubMed, and science direct databases from 2010 to 2021 in 2021 and revised in August 2023. 1250 articles were searched and 13 articles were selected for this review. We investigated them and then due to the importance and popularity of using machine learning in cancer vaccines recently, we compared them based on their machine learning technique. it is shown that neural networks with Python are used to predict neoantigens in 4 articles and with MATLAB in 2 articles, one article was about using the Fontom, one article with PERL, and one article with R; Other studies were about data mining with flowsom algorithm, multiple linear regression, logistics, and oncopepVCA, and the rest of articles do not provide information about machine learning implementation tools. Providing neural networks with Python is useful in the prediction of neoantigens due to the precision and examination of complex data sets. They use to predict HLA and peptide binding affinity, vaccines outcome, personalized cancer vaccines based on new data, the immune response, processing RNA and DNA sequences, and immunological analysis.","doi":"10.1007\/s11042-023-17589-8","cleaned_title":"application of machine learning and deep learning for cancer vaccine (rapid review)","cleaned_abstract":"cancer is a common and dangerous disease based on the world health organization. much research has been done on new and effective cancer treatments, including cancer vaccines and the prediction of neoantigens using machine learning. the purpose of this study is to review articles that use machine learning to design cancer vaccines. this study is a rapid review study using search strategies and related keywords in google scholar, pubmed, and science direct databases from 2010 to 2021 in 2021 and revised in august 2023. 1250 articles were searched and 13 articles were selected for this review. we investigated them and then due to the importance and popularity of using machine learning in cancer vaccines recently, we compared them based on their machine learning technique. it is shown that neural networks with python are used to predict neoantigens in 4 articles and with matlab in 2 articles, one article was about using the fontom, one article with perl, and one article with r; other studies were about data mining with flowsom algorithm, multiple linear regression, logistics, and oncopepvca, and the rest of articles do not provide information about machine learning implementation tools. providing neural networks with python is useful in the prediction of neoantigens due to the precision and examination of complex data sets. they use to predict hla and peptide binding affinity, vaccines outcome, personalized cancer vaccines based on new data, the immune response, processing rna and dna sequences, and immunological analysis.","key_phrases":["cancer","a common and dangerous disease","the world health organization","much research","new and effective cancer treatments","cancer vaccines","the prediction","neoantigens","machine learning","the purpose","this study","articles","that","cancer vaccines","this study","a rapid review study","search strategies","related keywords","google scholar","direct databases","august","1250 articles","13 articles","this review","we","them","the importance","popularity","machine learning","cancer vaccines","we","them","their machine learning technique","it","neural networks","python","neoantigens","4 articles","matlab","2 articles","one article","the fontom","one article","perl","one article","r","other studies","data mining","flowsom algorithm","multiple linear regression","logistics","oncopepvca","the rest","articles","information","implementation tools","neural networks","python","the prediction","neoantigens","the precision","examination","complex data sets","they","hla and peptide binding affinity","vaccines","new data","the immune response","rna","dna sequences","immunological analysis","the world health organization","google","2010","2021","2021","august 2023.","13","4","2","one","one","one"]},{"title":"A Resource-Efficient Deep Learning Approach to Visual-Based Cattle Geographic Origin Prediction","authors":["Camellia Ray","Sambit Bakshi","Pankaj Kumar Sa","Ganapati Panda"],"abstract":"Customized healthcare for cattle health monitoring is essential, which aims to optimize individual animal health, thereby enhancing productivity, minimizing illness-related risks, and improving overall welfare. Tailoring healthcare practices to individual requirements guarantees that individual animals receive proper attention and intervention, resulting in better health outcomes and sustainable cattle farming practices. In this regard, the manuscript proposes a visual cues-based region prediction methodology to design a customized cattle healthcare system. The proposed automated AI healthcare system uses resource-efficient deep learning-inspired architecture for computer vision applications like performing region-wise classification. The classification mechanism can be used further to identify a cattle and the regions it belongs. Extensive experimentation has been conducted on a redesigned image dataset to identify the best-suited deep-learning framework to perform region classification for livestock, such as cattle. MobileNetV2 outperforms the considered state-of-the-art frameworks by achieving an accuracy of 93% in identifying the regions of the cattle.","doi":"10.1007\/s11036-024-02350-8","cleaned_title":"a resource-efficient deep learning approach to visual-based cattle geographic origin prediction","cleaned_abstract":"customized healthcare for cattle health monitoring is essential, which aims to optimize individual animal health, thereby enhancing productivity, minimizing illness-related risks, and improving overall welfare. tailoring healthcare practices to individual requirements guarantees that individual animals receive proper attention and intervention, resulting in better health outcomes and sustainable cattle farming practices. in this regard, the manuscript proposes a visual cues-based region prediction methodology to design a customized cattle healthcare system. the proposed automated ai healthcare system uses resource-efficient deep learning-inspired architecture for computer vision applications like performing region-wise classification. the classification mechanism can be used further to identify a cattle and the regions it belongs. extensive experimentation has been conducted on a redesigned image dataset to identify the best-suited deep-learning framework to perform region classification for livestock, such as cattle. mobilenetv2 outperforms the considered state-of-the-art frameworks by achieving an accuracy of 93% in identifying the regions of the cattle.","key_phrases":["customized healthcare","cattle health monitoring","which","individual animal health","productivity","illness-related risks","overall welfare","healthcare practices","individual requirements","individual animals","proper attention","intervention","better health outcomes","sustainable cattle farming practices","this regard","the manuscript","a visual cues-based region prediction methodology","a customized cattle healthcare system","the proposed automated ai healthcare system","resource-efficient deep learning-inspired architecture","computer vision applications","region-wise classification","the classification mechanism","a cattle","the regions","it","extensive experimentation","a redesigned image dataset","the best-suited deep-learning framework","region classification","livestock","cattle","mobilenetv2","the-art","an accuracy","93%","the regions","the cattle","tailoring healthcare","mobilenetv2","93%"]},{"title":"Hybrid deep learning framework for weather forecast with rainfall prediction using weather bigdata analytics","authors":["C. Lalitha","D. Ravindran"],"abstract":"The volume and complexity of weather data, along with missing values and high correlation between collected variables, make it challenging to develop efficient deep learning frameworks that can handle data with more features. This leads to a lack of accurate and predictable weather forecasts. To develop a hybrid deep learning framework for weather forecast with rainfall prediction using weather big data analytics to ensure high detection rates. A modified planet optimization (MPO) algorithm is used for data preprocessing to remove unwanted artifacts. An improved Tuna optimization (ITO) algorithm is presented to select optimal features to avoid data dimensionality issues. A hybrid memory-augmented artificial neural network (MA-ANN) classifier is developed to improve weather early forecast detection rates. The proposed framework is validated against standard benchmark datasets such as weather underground and climate forecast system reanalysis (CFSR). The simulation results are compared with other existing state-of-the-art frameworks based on error measures (RMSE, MAPE, BIAS, R) and quality measures (accuracy, sensitivity, specificity, precision, F1-measure).The MA-ANN classifier accuracy obtained 97.65% for wunderground.com Delhi and 98.88% for Tamilnadu. The hybrid deep learning framework with rainfall prediction using weather big data analytics has shown promising results for accurate and predictable weather forecasts. The proposed framework outperforms other existing state-of-the-art frameworks, and the MA-ANN classifier has improved weather early forecast detection rates. The study demonstrates the potential of utilizing big data techniques in weather forecasting and highlights the importance of developing efficient deep learning frameworks to handle complex and high-dimensional weather data.","doi":"10.1007\/s11042-023-17801-9","cleaned_title":"hybrid deep learning framework for weather forecast with rainfall prediction using weather bigdata analytics","cleaned_abstract":"the volume and complexity of weather data, along with missing values and high correlation between collected variables, make it challenging to develop efficient deep learning frameworks that can handle data with more features. this leads to a lack of accurate and predictable weather forecasts. to develop a hybrid deep learning framework for weather forecast with rainfall prediction using weather big data analytics to ensure high detection rates. a modified planet optimization (mpo) algorithm is used for data preprocessing to remove unwanted artifacts. an improved tuna optimization (ito) algorithm is presented to select optimal features to avoid data dimensionality issues. a hybrid memory-augmented artificial neural network (ma-ann) classifier is developed to improve weather early forecast detection rates. the proposed framework is validated against standard benchmark datasets such as weather underground and climate forecast system reanalysis (cfsr). the simulation results are compared with other existing state-of-the-art frameworks based on error measures (rmse, mape, bias, r) and quality measures (accuracy, sensitivity, specificity, precision, f1-measure).the ma-ann classifier accuracy obtained 97.65% for wunderground.com delhi and 98.88% for tamilnadu. the hybrid deep learning framework with rainfall prediction using weather big data analytics has shown promising results for accurate and predictable weather forecasts. the proposed framework outperforms other existing state-of-the-art frameworks, and the ma-ann classifier has improved weather early forecast detection rates. the study demonstrates the potential of utilizing big data techniques in weather forecasting and highlights the importance of developing efficient deep learning frameworks to handle complex and high-dimensional weather data.","key_phrases":["the volume","complexity","weather data","missing values","high correlation","collected variables","it","efficient deep learning frameworks","that","data","more features","this","a lack","accurate and predictable weather forecasts","a hybrid deep learning framework","weather forecast","rainfall prediction","weather big data analytics","high detection rates","a modified planet optimization","mpo","algorithm","data","unwanted artifacts","an improved tuna optimization","ito","algorithm","optimal features","data dimensionality issues","a hybrid memory-augmented artificial neural network","ma-ann) classifier","weather early forecast detection rates","the proposed framework","standard benchmark datasets","weather","climate forecast system reanalysis","cfsr","the simulation results","the-art","error measures","accuracy","sensitivity","specificity","precision","f1-measure).the ma-ann classifier accuracy","97.65%","98.88%","tamilnadu","the hybrid deep learning framework","rainfall prediction","weather big data analytics","promising results","accurate and predictable weather forecasts","the proposed framework","the-art","the ma-ann classifier","weather early forecast detection rates","the study","the potential","big data techniques","weather forecasting","the importance","efficient deep learning frameworks","complex and high-dimensional weather data","rmse","97.65%","wunderground.com delhi","98.88%"]},{"title":"Image Classification Algorithm Based on Proposal Region Clustering Learning-Unsupervised Deep Learning","authors":["Lei Li","Xiao-li Yin"],"abstract":"Although deep learning has achieved certain results in image classification, images are susceptible to factors such as lighting conditions, shooting angles, complex backgrounds, rotation transformations or scale scaling, and image data sets in some areas are difficult to obtain. They make the deep learning framework unable to give full play to its generalization ability and nonlinear modeling ability in image classification. Therefore, this paper first proposes a proposal region clustering learning algorithm, which clusters the proposal regions in each image so that each cluster corresponds to the category of the image. Then, different clusters can be regarded as different multi-instance learning packets, and each packet uses the multi-instance learning method to learn the unsupervised image classification detector. It can effectively improve the generalization and modeling capabilities of deep learning models. In addition, this paper proposes an unsupervised deep learning method, which designs an unsupervised deep learning network structure and loss function according to the characteristics of the classified image, and combines densely connected blocks to extract features from the source image. It retains the multi-scale features of the middle layer of the classified image, and effectively solves the problem of insufficient image feature extraction information caused by the lack of image data. It also guarantees the accuracy of subsequent image classification. The experimental results show that the image classification method proposed in this paper not only solves the problem of insufficient image data sets and the interference of various complex factors, but also can accurately classify various image data sets. The accuracy of the image classification method proposed in this paper is 1.38\u201319% higher than other mainstream deep learning methods.","doi":"10.1007\/s42835-022-01227-8","cleaned_title":"image classification algorithm based on proposal region clustering learning-unsupervised deep learning","cleaned_abstract":"although deep learning has achieved certain results in image classification, images are susceptible to factors such as lighting conditions, shooting angles, complex backgrounds, rotation transformations or scale scaling, and image data sets in some areas are difficult to obtain. they make the deep learning framework unable to give full play to its generalization ability and nonlinear modeling ability in image classification. therefore, this paper first proposes a proposal region clustering learning algorithm, which clusters the proposal regions in each image so that each cluster corresponds to the category of the image. then, different clusters can be regarded as different multi-instance learning packets, and each packet uses the multi-instance learning method to learn the unsupervised image classification detector. it can effectively improve the generalization and modeling capabilities of deep learning models. in addition, this paper proposes an unsupervised deep learning method, which designs an unsupervised deep learning network structure and loss function according to the characteristics of the classified image, and combines densely connected blocks to extract features from the source image. it retains the multi-scale features of the middle layer of the classified image, and effectively solves the problem of insufficient image feature extraction information caused by the lack of image data. it also guarantees the accuracy of subsequent image classification. the experimental results show that the image classification method proposed in this paper not only solves the problem of insufficient image data sets and the interference of various complex factors, but also can accurately classify various image data sets. the accuracy of the image classification method proposed in this paper is 1.38\u201319% higher than other mainstream deep learning methods.","key_phrases":["deep learning","certain results","image classification","images","factors","lighting conditions","angles","complex backgrounds","rotation transformations","scale scaling","image data sets","some areas","they","the deep learning framework","full play","its generalization ability","modeling ability","image classification","this paper","a proposal region","learning algorithm","which","the proposal regions","each image","each cluster","the category","the image","different clusters","different multi-instance learning packets","each packet","the multi-instance learning method","the unsupervised image classification detector","it","the generalization","modeling","capabilities","deep learning models","addition","this paper","an unsupervised deep learning method","which","an unsupervised deep learning network structure and loss function","the characteristics","the classified image","densely connected blocks","features","the source image","it","the multi-scale features","the middle layer","the classified image","the problem","insufficient image feature extraction information","the lack","image data","it","the accuracy","subsequent image classification","the experimental results","the image classification method","this paper","the problem","insufficient image data sets","the interference","various complex factors","various image data sets","the accuracy","the image classification method","this paper","other mainstream deep learning methods","first","1.38\u201319%"]},{"title":"SkinMultiNet: Advancements in Skin Cancer Prediction Using Deep Learning with Web Interface","authors":["Md Nur Hosain Likhon","Sahab Uddin Rana","Sadeka Akter","Md. Shorup Ahmed","Khadiza Akter Tanha","Md. Mahbubur Rahman","Md Emran Hussain Nayeem"],"abstract":"Cancer remains the leading cause of death worldwide, significantly impacting individuals and healthcare systems alike. In recent decades, skin cancer has surged in prevalence compared to other major cancer types. Various factors such as texture, color, morphological characteristics, and structure are employed in categorizing different forms of skin cancer. However, traditional methods of identification often prove time-consuming and costly. Skin cancer classification predominantly relies on machine learning, with the primary method being convolutional neural networks (CNNs). Our \u2018SkinMultiNet\u2019 framework, presented in this study and based on transfer learning principles, integrates the InceptionV3 and Xception CNN models for predicting skin cancer using image data. While other machine learning models such as ResNet50, NasNet, and MobileNet were explored, the 'SkinMultiNet' framework demonstrated the most promising outcomes. Utilizing a publicly available dataset comprising 6086 skin images, we trained, tested, and evaluated our models extensively. Proposed system employed a train generator to feed image data into our deep learning CNN models, followed by implementing a learning rate reducer on the datasets within the model. Through rigorous testing and validation procedures, our models successfully processed a substantial volume of skin image data. In contrast to conventional approaches, our proposed architecture offers the potential for more reliable diagnoses, achieving an optimal accuracy rate of 94% in skin cancer prediction. This advancement holds promise for early detection and improved patient outcomes following therapy.","doi":"10.1007\/s44174-024-00205-0","cleaned_title":"skinmultinet: advancements in skin cancer prediction using deep learning with web interface","cleaned_abstract":"cancer remains the leading cause of death worldwide, significantly impacting individuals and healthcare systems alike. in recent decades, skin cancer has surged in prevalence compared to other major cancer types. various factors such as texture, color, morphological characteristics, and structure are employed in categorizing different forms of skin cancer. however, traditional methods of identification often prove time-consuming and costly. skin cancer classification predominantly relies on machine learning, with the primary method being convolutional neural networks (cnns). our \u2018skinmultinet\u2019 framework, presented in this study and based on transfer learning principles, integrates the inceptionv3 and xception cnn models for predicting skin cancer using image data. while other machine learning models such as resnet50, nasnet, and mobilenet were explored, the 'skinmultinet' framework demonstrated the most promising outcomes. utilizing a publicly available dataset comprising 6086 skin images, we trained, tested, and evaluated our models extensively. proposed system employed a train generator to feed image data into our deep learning cnn models, followed by implementing a learning rate reducer on the datasets within the model. through rigorous testing and validation procedures, our models successfully processed a substantial volume of skin image data. in contrast to conventional approaches, our proposed architecture offers the potential for more reliable diagnoses, achieving an optimal accuracy rate of 94% in skin cancer prediction. this advancement holds promise for early detection and improved patient outcomes following therapy.","key_phrases":["cancer","the leading cause","death","significantly impacting individuals","healthcare systems","recent decades","skin cancer","prevalence","other major cancer types","various factors","texture","color","morphological characteristics","structure","different forms","skin cancer","traditional methods","identification","skin cancer classification","machine learning","the primary method","convolutional neural networks","cnns","our \u2018skinmultinet\u2019 framework","this study","transfer learning principles","the inceptionv3 and xception cnn models","skin cancer","image data","other machine learning models","resnet50","nasnet","mobilenet","the 'skinmultinet' framework","the most promising outcomes","a publicly available dataset","6086 skin images","we","our models","proposed system","a train generator","image data","our deep learning cnn models","a learning rate reducer","the datasets","the model","rigorous testing and validation procedures","our models","a substantial volume","skin image data","contrast","conventional approaches","our proposed architecture","the potential","more reliable diagnoses","an optimal accuracy rate","94%","skin cancer prediction","this advancement","promise","early detection","improved patient outcomes","therapy","recent decades","inceptionv3","cnn","resnet50","6086","cnn","94%"]},{"title":"Deep learning-based prediction of post-pancreaticoduodenectomy pancreatic fistula","authors":["Woohyung Lee","Hyo Jung Park","Hack-Jin Lee","Ki Byung Song","Dae Wook Hwang","Jae Hoon Lee","Kyongmook Lim","Yousun Ko","Hyoung Jung Kim","Kyung Won Kim","Song Cheol Kim"],"abstract":"Postoperative pancreatic fistula is a life-threatening complication with an unmet need for accurate prediction. This study was aimed to develop preoperative artificial intelligence-based prediction models. Patients who underwent pancreaticoduodenectomy were enrolled and stratified into model development and validation sets by surgery between 2016 and 2017 or in 2018, respectively. Machine learning models based on clinical and body composition data, and deep learning models based on computed tomographic data, were developed, combined by ensemble voting, and final models were selected comparison with earlier model. Among the 1333 participants (training, n\u2009=\u2009881; test, n\u2009=\u2009452), postoperative pancreatic fistula occurred in 421 (47.8%) and 134 (31.8%) and clinically relevant postoperative pancreatic fistula occurred in 59 (6.7%) and 27 (6.0%) participants in the training and test datasets, respectively. In the test dataset, the area under the receiver operating curve [AUC (95% confidence interval)] of the selected preoperative model for predicting all and clinically relevant postoperative pancreatic fistula was 0.75 (0.71\u20130.80) and 0.68 (0.58\u20130.78). The ensemble model showed better predictive performance than the individual ML and DL models.","doi":"10.1038\/s41598-024-51777-2","cleaned_title":"deep learning-based prediction of post-pancreaticoduodenectomy pancreatic fistula","cleaned_abstract":"postoperative pancreatic fistula is a life-threatening complication with an unmet need for accurate prediction. this study was aimed to develop preoperative artificial intelligence-based prediction models. patients who underwent pancreaticoduodenectomy were enrolled and stratified into model development and validation sets by surgery between 2016 and 2017 or in 2018, respectively. machine learning models based on clinical and body composition data, and deep learning models based on computed tomographic data, were developed, combined by ensemble voting, and final models were selected comparison with earlier model. among the 1333 participants (training, n = 881; test, n = 452), postoperative pancreatic fistula occurred in 421 (47.8%) and 134 (31.8%) and clinically relevant postoperative pancreatic fistula occurred in 59 (6.7%) and 27 (6.0%) participants in the training and test datasets, respectively. in the test dataset, the area under the receiver operating curve [auc (95% confidence interval)] of the selected preoperative model for predicting all and clinically relevant postoperative pancreatic fistula was 0.75 (0.71\u20130.80) and 0.68 (0.58\u20130.78). the ensemble model showed better predictive performance than the individual ml and dl models.","key_phrases":["postoperative pancreatic fistula","a life-threatening complication","an unmet need","accurate prediction","this study","preoperative artificial intelligence-based prediction models","patients","who","pancreaticoduodenectomy","model development","validation sets","surgery","machine learning models","clinical and body composition data","deep learning models","computed tomographic data","ensemble voting","final models","comparison","earlier model","the 1333 participants","training","test","postoperative pancreatic fistula","47.8%","31.8%","clinically relevant postoperative pancreatic fistula","6.7%","the training","test datasets","the test dataset","the area","the receiver operating curve","[auc","95% confidence interval","the selected preoperative model","all and clinically relevant postoperative pancreatic fistula","the ensemble model","better predictive performance","the individual ml","dl models","between 2016 and 2017","2018","1333","881","452","421","47.8%","134","31.8%","59","6.7%","27","6.0%","95%","0.75","0.71\u20130.80","0.68"]},{"title":"Debunking multi-lingual social media posts using deep learning","authors":["Bina Kotiyal","Heman Pathak","Nipur Singh"],"abstract":"Fake news on social media has become a growing concern due to its potential impact on shaping public opinion. The proposed Debunking Multi-Lingual Social Media Posts using Deep Learning (DSMPD) approach offers a promising solution to detect fake news. The DSMPD approach involves creating a dataset of English and Hindi social media posts using web scraping and Natural Language Processing (NLP) techniques. This dataset is then used to train, test, and validate a deep learning-based model that extracts various features, including Embedding from Language Models (ELMo), word and n-gram counts, Term Frequency-Inverse Document Frequency (TF-IDF), sentiments, polarity, and Named Entity Recognition (NER). Based on these features, the model classifies news items into five categories: real, could be real, could be fabricated, fabricated, or dangerously fabricated. To evaluate the performance of the classifiers, the researchers used two datasets comprising over 45,000 articles. Machine learning (ML) algorithms and Deep learning (DL) model are compared to choose the best option for classification and prediction.","doi":"10.1007\/s41870-023-01288-6","cleaned_title":"debunking multi-lingual social media posts using deep learning","cleaned_abstract":"fake news on social media has become a growing concern due to its potential impact on shaping public opinion. the proposed debunking multi-lingual social media posts using deep learning (dsmpd) approach offers a promising solution to detect fake news. the dsmpd approach involves creating a dataset of english and hindi social media posts using web scraping and natural language processing (nlp) techniques. this dataset is then used to train, test, and validate a deep learning-based model that extracts various features, including embedding from language models (elmo), word and n-gram counts, term frequency-inverse document frequency (tf-idf), sentiments, polarity, and named entity recognition (ner). based on these features, the model classifies news items into five categories: real, could be real, could be fabricated, fabricated, or dangerously fabricated. to evaluate the performance of the classifiers, the researchers used two datasets comprising over 45,000 articles. machine learning (ml) algorithms and deep learning (dl) model are compared to choose the best option for classification and prediction.","key_phrases":["fake news","social media","a growing concern","its potential impact","public opinion","the proposed debunking multi-lingual social media posts","(dsmpd","a promising solution","fake news","the dsmpd approach","a dataset","english and hindi social media posts","web scraping","natural language processing (nlp) techniques","this dataset","a deep learning-based model","that","various features","language models","word and n-gram counts","term frequency-inverse document frequency","tf-idf","sentiments","polarity","entity recognition","ner","these features","the model classifies","five categories","the performance","the classifiers","the researchers","two datasets","over 45,000 articles","ml","the best option","classification","prediction","english","n-gram","ner","five","two","over 45,000"]},{"title":"Designing observables for measurements with deep learning","authors":["Owen Long","Benjamin Nachman"],"abstract":"Many analyses in particle and nuclear physics use simulations to infer fundamental, effective, or phenomenological parameters of the underlying physics models. When the inference is performed with unfolded cross sections, the observables are designed using physics intuition and heuristics. We propose to design targeted observables with machine learning. Unfolded, differential cross sections in a neural network output contain the most information about parameters of interest and can be well-measured by construction. The networks are trained using a custom loss function that rewards outputs that are sensitive to the parameter(s) of interest while simultaneously penalizing outputs that are different between particle-level and detector-level (to minimize detector distortions). We demonstrate this idea in simulation using two physics models for inclusive measurements in deep inelastic scattering. We find that the new approach is more sensitive than classical observables at distinguishing the two models and also has a reduced unfolding uncertainty due to the reduced detector distortions.","doi":"10.1140\/epjc\/s10052-024-13135-4","cleaned_title":"designing observables for measurements with deep learning","cleaned_abstract":"many analyses in particle and nuclear physics use simulations to infer fundamental, effective, or phenomenological parameters of the underlying physics models. when the inference is performed with unfolded cross sections, the observables are designed using physics intuition and heuristics. we propose to design targeted observables with machine learning. unfolded, differential cross sections in a neural network output contain the most information about parameters of interest and can be well-measured by construction. the networks are trained using a custom loss function that rewards outputs that are sensitive to the parameter(s) of interest while simultaneously penalizing outputs that are different between particle-level and detector-level (to minimize detector distortions). we demonstrate this idea in simulation using two physics models for inclusive measurements in deep inelastic scattering. we find that the new approach is more sensitive than classical observables at distinguishing the two models and also has a reduced unfolding uncertainty due to the reduced detector distortions.","key_phrases":["many analyses","particle","nuclear physics","simulations","fundamental, effective, or phenomenological parameters","the underlying physics models","the inference","unfolded cross sections","the observables","physics intuition","heuristics","we","targeted observables","machine learning","unfolded, differential cross sections","a neural network output","the most information","parameters","interest","construction","the networks","a custom loss function","that","outputs","that","the parameter(s","interest","outputs","that","particle-level","detector-level","detector distortions","we","this idea","simulation","two physics models","inclusive measurements","deep inelastic scattering","we","the new approach","classical observables","the two models","a reduced unfolding uncertainty","the reduced detector distortions","two","two"]},{"title":"Autonomous multi-drone racing method based on deep reinforcement learning","authors":["Yu Kang","Jian Di","Ming Li","Yunbo Zhao","Yuhui Wang"],"abstract":"Racing drones have attracted increasing attention due to their remarkable high speed and excellent maneuverability. However, autonomous multi-drone racing is quite difficult since it requires quick and agile flight in intricate surroundings and rich drone interaction. To address these issues, we propose a novel autonomous multi-drone racing method based on deep reinforcement learning. A new set of reward functions is proposed to make racing drones learn the racing skills of human experts. Unlike previous methods that required global information about tracks and track boundary constraints, the proposed method requires only limited localized track information within the range of its own onboard sensors. Further, the dynamic response characteristics of racing drones are incorporated into the training environment, so that the proposed method is more in line with the requirements of real drone racing scenarios. In addition, our method has a low computational cost and can meet the requirements of real-time racing. Finally, the effectiveness and superiority of the proposed method are verified by extensive comparison with the state-of-the-art methods in a series of simulations and real-world experiments.","doi":"10.1007\/s11432-023-4029-9","cleaned_title":"autonomous multi-drone racing method based on deep reinforcement learning","cleaned_abstract":"racing drones have attracted increasing attention due to their remarkable high speed and excellent maneuverability. however, autonomous multi-drone racing is quite difficult since it requires quick and agile flight in intricate surroundings and rich drone interaction. to address these issues, we propose a novel autonomous multi-drone racing method based on deep reinforcement learning. a new set of reward functions is proposed to make racing drones learn the racing skills of human experts. unlike previous methods that required global information about tracks and track boundary constraints, the proposed method requires only limited localized track information within the range of its own onboard sensors. further, the dynamic response characteristics of racing drones are incorporated into the training environment, so that the proposed method is more in line with the requirements of real drone racing scenarios. in addition, our method has a low computational cost and can meet the requirements of real-time racing. finally, the effectiveness and superiority of the proposed method are verified by extensive comparison with the state-of-the-art methods in a series of simulations and real-world experiments.","key_phrases":["racing drones","increasing attention","their remarkable high speed","excellent maneuverability","autonomous multi-drone racing","it","quick and agile flight","intricate surroundings","rich drone interaction","these issues","we","a novel autonomous multi-drone racing method","deep reinforcement learning","a new set","reward functions","racing drones","the racing skills","human experts","previous methods","that","global information","tracks","boundary constraints","the proposed method","only limited localized track information","the range","its own onboard sensors","the dynamic response characteristics","racing drones","the training environment","the proposed method","line","the requirements","real drone racing scenarios","addition","our method","a low computational cost","the requirements","real-time racing","the effectiveness","superiority","the proposed method","extensive comparison","the-art","a series","simulations","real-world experiments"]},{"title":"Enhancing parasitic organism detection in microscopy images through deep learning and fine-tuned optimizer","authors":["Yogesh Kumar","Pertik Garg","Manu Raj Moudgil","Rupinder Singh","Marcin Wo\u017aniak","Jana Shafi","Muhammad Fazal Ijaz"],"abstract":"Parasitic organisms pose a major global health threat, mainly in regions that lack advanced medical facilities. Early and accurate detection of parasitic organisms is vital to saving lives. Deep learning models have uplifted the medical sector by providing promising results in diagnosing, detecting, and classifying diseases. This paper explores the role of deep learning techniques in detecting and classifying various parasitic organisms. The research works on a dataset consisting of 34,298 samples of parasites such as Toxoplasma Gondii, Trypanosome, Plasmodium, Leishmania, Babesia, and Trichomonad along with host cells like red blood cells and white blood cells. These images are initially converted from RGB to grayscale followed by the computation of morphological features such as perimeter, height, area, and width. Later, Otsu thresholding and watershed techniques are applied to differentiate foreground from background and create markers on the images for the identification of regions of interest. Deep transfer learning models such as VGG19, InceptionV3, ResNet50V2, ResNet152V2, EfficientNetB3, EfficientNetB0, MobileNetV2, Xception, DenseNet169, and a hybrid model, InceptionResNetV2, are employed. The parameters of these models are fine-tuned using three optimizers: SGD, RMSprop, and Adam. Experimental results reveal that when RMSprop is applied, VGG19, InceptionV3, and EfficientNetB0 achieve the highest accuracy of 99.1% with a loss of 0.09. Similarly, using the SGD optimizer, InceptionV3 performs exceptionally well, achieving the highest accuracy of 99.91% with a loss of 0.98. Finally, applying the Adam optimizer, InceptionResNetV2 excels, achieving the highest accuracy of 99.96% with a loss of 0.13, outperforming other optimizers. The findings of this research signify that using deep learning models coupled with image processing methods generates a highly accurate and efficient way to detect and classify parasitic organisms.","doi":"10.1038\/s41598-024-56323-8","cleaned_title":"enhancing parasitic organism detection in microscopy images through deep learning and fine-tuned optimizer","cleaned_abstract":"parasitic organisms pose a major global health threat, mainly in regions that lack advanced medical facilities. early and accurate detection of parasitic organisms is vital to saving lives. deep learning models have uplifted the medical sector by providing promising results in diagnosing, detecting, and classifying diseases. this paper explores the role of deep learning techniques in detecting and classifying various parasitic organisms. the research works on a dataset consisting of 34,298 samples of parasites such as toxoplasma gondii, trypanosome, plasmodium, leishmania, babesia, and trichomonad along with host cells like red blood cells and white blood cells. these images are initially converted from rgb to grayscale followed by the computation of morphological features such as perimeter, height, area, and width. later, otsu thresholding and watershed techniques are applied to differentiate foreground from background and create markers on the images for the identification of regions of interest. deep transfer learning models such as vgg19, inceptionv3, resnet50v2, resnet152v2, efficientnetb3, efficientnetb0, mobilenetv2, xception, densenet169, and a hybrid model, inceptionresnetv2, are employed. the parameters of these models are fine-tuned using three optimizers: sgd, rmsprop, and adam. experimental results reveal that when rmsprop is applied, vgg19, inceptionv3, and efficientnetb0 achieve the highest accuracy of 99.1% with a loss of 0.09. similarly, using the sgd optimizer, inceptionv3 performs exceptionally well, achieving the highest accuracy of 99.91% with a loss of 0.98. finally, applying the adam optimizer, inceptionresnetv2 excels, achieving the highest accuracy of 99.96% with a loss of 0.13, outperforming other optimizers. the findings of this research signify that using deep learning models coupled with image processing methods generates a highly accurate and efficient way to detect and classify parasitic organisms.","key_phrases":["parasitic organisms","a major global health threat","regions","that","advanced medical facilities","early and accurate detection","parasitic organisms","lives","deep learning models","the medical sector","promising results","diseases","this paper","the role","deep learning techniques","various parasitic organisms","the research","a dataset","34,298 samples","parasites","toxoplasma gondii","trypanosome","plasmodium","leishmania","babesia","trichomonad","host cells","red blood cells","white blood cells","these images","rgb","grayscale","the computation","morphological features","perimeter","height","area","width","otsu thresholding and watershed techniques","foreground","background","markers","the images","the identification","regions","interest","deep transfer learning models","vgg19","inceptionv3","resnet50v2","resnet152v2","efficientnetb3","efficientnetb0","mobilenetv2","xception","densenet169","a hybrid model","the parameters","these models","three optimizers","sgd","rmsprop","adam","experimental results","rmsprop","efficientnetb0","the highest accuracy","99.1%","a loss","the sgd optimizer","inceptionv3","the highest accuracy","99.91%","a loss","the adam optimizer","inceptionresnetv2 excels","the highest accuracy","99.96%","a loss","other optimizers","the findings","this research","deep learning models","image processing methods","a highly accurate and efficient way","parasitic organisms","34,298","leishmania","babesia","rgb","inceptionv3","efficientnetb3","efficientnetb0","mobilenetv2","inceptionresnetv2","three","inceptionv3","efficientnetb0","99.1%","0.09","inceptionv3","99.91%","0.98","inceptionresnetv2","99.96%","0.13"]},{"title":"A survey on automatic generation of medical imaging reports based on deep learning","authors":["Ting Pang","Peigao Li","Lijie Zhao"],"abstract":"Recent advances in deep learning have shown great potential for the automatic generation of medical imaging reports. Deep learning techniques, inspired by image captioning, have made significant progress in the field of diagnostic report generation. This paper provides a comprehensive overview of recent research efforts in deep learning-based medical imaging report generation and proposes future directions in this field. First, we summarize and analyze the data set, architecture, application, and evaluation of deep learning-based medical imaging report generation. Specially, we survey the deep learning architectures used in diagnostic report generation, including hierarchical RNN-based frameworks, attention-based frameworks, and reinforcement learning-based frameworks. In addition, we identify potential challenges and suggest future research directions to support clinical applications and decision-making using medical imaging report generation systems.","doi":"10.1186\/s12938-023-01113-y","cleaned_title":"a survey on automatic generation of medical imaging reports based on deep learning","cleaned_abstract":"recent advances in deep learning have shown great potential for the automatic generation of medical imaging reports. deep learning techniques, inspired by image captioning, have made significant progress in the field of diagnostic report generation. this paper provides a comprehensive overview of recent research efforts in deep learning-based medical imaging report generation and proposes future directions in this field. first, we summarize and analyze the data set, architecture, application, and evaluation of deep learning-based medical imaging report generation. specially, we survey the deep learning architectures used in diagnostic report generation, including hierarchical rnn-based frameworks, attention-based frameworks, and reinforcement learning-based frameworks. in addition, we identify potential challenges and suggest future research directions to support clinical applications and decision-making using medical imaging report generation systems.","key_phrases":["recent advances","deep learning","great potential","the automatic generation","medical imaging reports","deep learning techniques","image captioning","significant progress","the field","diagnostic report generation","this paper","a comprehensive overview","recent research efforts","deep learning-based medical imaging report generation","future directions","this field","we","the data set","architecture","application","evaluation","deep learning-based medical imaging report generation","we","the deep learning architectures","diagnostic report generation","hierarchical rnn-based frameworks","attention-based frameworks","reinforcement learning-based frameworks","addition","we","potential challenges","future research directions","clinical applications","decision-making","medical imaging report generation systems","first"]},{"title":"Development of Deep Learning Color Recognition Model for Color Measurement Processes","authors":["Sanghun Lee","Ki-Sub Kim","Jeong Won Kang"],"abstract":"We present a deep learning color recognition model for the color measurement process in the paint industry. Currently, spectrophotometers are primarily used for color measurements owing to their accuracy. The measurement method involves manually injecting the sample into a spectrophotometer. Our proposed method uses a webcam with a deep learning model on the stand of a spectrophotometer. Deep learning models are widely used for image and color detection. In this study, the \u201cyou only look once (YOLO)\u201d algorithm is applied for real-time detection of color samples. Upon training various sample images using YOLO, the model could detect the sample area in real time using a webcam. An open source computer vision (OpenCV) library was used for the color recognition model, and the detected RGB color value was converted to the international commission on illumination color space (CIELAB) value, which is primarily used in the color measuring process. However, because of the mirror-like reflection of light from a surface with specular reflection, it is difficult to implement the color value using a camera. To address this problem, we compare several specular removal methods and propose the most suitable model for the color recognition model of color samples. The accuracy of the proposed model was verified by comparing the colors of various samples. Our proposed approach can easily detect samples and color values, which can contribute significantly to automatically calculating the exact amount of coloring required for the target color.","doi":"10.1007\/s42835-024-01791-1","cleaned_title":"development of deep learning color recognition model for color measurement processes","cleaned_abstract":"we present a deep learning color recognition model for the color measurement process in the paint industry. currently, spectrophotometers are primarily used for color measurements owing to their accuracy. the measurement method involves manually injecting the sample into a spectrophotometer. our proposed method uses a webcam with a deep learning model on the stand of a spectrophotometer. deep learning models are widely used for image and color detection. in this study, the \u201cyou only look once (yolo)\u201d algorithm is applied for real-time detection of color samples. upon training various sample images using yolo, the model could detect the sample area in real time using a webcam. an open source computer vision (opencv) library was used for the color recognition model, and the detected rgb color value was converted to the international commission on illumination color space (cielab) value, which is primarily used in the color measuring process. however, because of the mirror-like reflection of light from a surface with specular reflection, it is difficult to implement the color value using a camera. to address this problem, we compare several specular removal methods and propose the most suitable model for the color recognition model of color samples. the accuracy of the proposed model was verified by comparing the colors of various samples. our proposed approach can easily detect samples and color values, which can contribute significantly to automatically calculating the exact amount of coloring required for the target color.","key_phrases":["we","a deep learning color recognition model","the color measurement process","the paint industry","spectrophotometers","color measurements","their accuracy","the measurement method","the sample","a spectrophotometer","our proposed method","a webcam","a deep learning model","the stand","a spectrophotometer","deep learning models","image","color detection","this study","you","algorithm","real-time detection","color samples","various sample images","yolo","the model","the sample area","real time","a webcam","an open source computer vision","(opencv) library","the color recognition model","the detected rgb color value","the international commission","illumination color space","(cielab) value","which","the color measuring process","the mirror-like reflection","light","a surface","specular reflection","it","the color value","a camera","this problem","we","several specular removal methods","the most suitable model","the color recognition model","color samples","the accuracy","the proposed model","the colors","various samples","our proposed approach","samples","color values","which","the exact amount","the target color","deep"]},{"title":"Rapid diagnosis of celiac disease based on plasma Raman spectroscopy combined with deep learning","authors":["Tian Shi","Jiahe Li","Na Li","Cheng Chen","Chen Chen","Chenjie Chang","Shenglong Xue","Weidong Liu","Ainur Maimaiti Reyim","Feng Gao","Xiaoyi Lv"],"abstract":"Celiac Disease (CD) is a primary malabsorption syndrome resulting from the interplay of genetic, immune, and dietary factors. CD negatively impacts daily activities and may lead to conditions such as osteoporosis, malignancies in the small intestine, ulcerative jejunitis, and enteritis, ultimately causing severe malnutrition. Therefore, an effective and rapid differentiation between healthy individuals and those with celiac disease is crucial for early diagnosis and treatment. This study utilizes Raman spectroscopy combined with deep learning models to achieve a non-invasive, rapid, and accurate diagnostic method for celiac disease and healthy controls. A total of 59 plasma samples, comprising 29 celiac disease cases and 30 healthy controls, were collected for experimental purposes. Convolutional Neural Network (CNN), Multi-Scale Convolutional Neural Network (MCNN), Residual Network (ResNet), and Deep Residual Shrinkage Network (DRSN) classification models were employed. The accuracy rates for these models were found to be 86.67%, 90.76%, 86.67% and 95.00%, respectively. Comparative validation results revealed that the DRSN model exhibited the best performance, with an AUC value and accuracy of 97.60% and 95%, respectively. This confirms the superiority of Raman spectroscopy combined with deep learning in the diagnosis of celiac disease.","doi":"10.1038\/s41598-024-64621-4","cleaned_title":"rapid diagnosis of celiac disease based on plasma raman spectroscopy combined with deep learning","cleaned_abstract":"celiac disease (cd) is a primary malabsorption syndrome resulting from the interplay of genetic, immune, and dietary factors. cd negatively impacts daily activities and may lead to conditions such as osteoporosis, malignancies in the small intestine, ulcerative jejunitis, and enteritis, ultimately causing severe malnutrition. therefore, an effective and rapid differentiation between healthy individuals and those with celiac disease is crucial for early diagnosis and treatment. this study utilizes raman spectroscopy combined with deep learning models to achieve a non-invasive, rapid, and accurate diagnostic method for celiac disease and healthy controls. a total of 59 plasma samples, comprising 29 celiac disease cases and 30 healthy controls, were collected for experimental purposes. convolutional neural network (cnn), multi-scale convolutional neural network (mcnn), residual network (resnet), and deep residual shrinkage network (drsn) classification models were employed. the accuracy rates for these models were found to be 86.67%, 90.76%, 86.67% and 95.00%, respectively. comparative validation results revealed that the drsn model exhibited the best performance, with an auc value and accuracy of 97.60% and 95%, respectively. this confirms the superiority of raman spectroscopy combined with deep learning in the diagnosis of celiac disease.","key_phrases":["celiac disease","cd","a primary malabsorption syndrome","the interplay","genetic, immune, and dietary factors","cd","daily activities","conditions","osteoporosis","malignancies","the small intestine","ulcerative jejunitis","enteritis","severe malnutrition","an effective and rapid differentiation","healthy individuals","those","celiac disease","early diagnosis","treatment","this study","raman spectroscopy","deep learning models","a non-invasive, rapid, and accurate diagnostic method","celiac disease","healthy controls","a total","59 plasma samples","29 celiac disease cases","30 healthy controls","experimental purposes","convolutional neural network","cnn","multi-scale convolutional neural network","mcnn","residual network","resnet","deep residual shrinkage network","drsn) classification models","the accuracy rates","these models","86.67%","90.76%","86.67%","95.00%","comparative validation results","the drsn model","the best performance","an auc value","accuracy","97.60%","95%","this","the superiority","raman spectroscopy","deep learning","the diagnosis","celiac disease","daily","59","29","30","cnn","86.67%","90.76%","86.67%","95.00%","97.60% and","95%"]},{"title":"White-box inference attack: compromising the security of deep learning-based COVID-19 diagnosis systems","authors":["Burhan Ul Haque Sheikh","Aasim Zafar"],"abstract":"The COVID-19 pandemic has necessitated the exploration of innovative diagnostic approaches, including the utilization of machine learning (ML) and deep learning (DL) technologies. However, recent findings shed light on the susceptibility of deep learning-based models to adversarial attacks, leading to erroneous predictions. This study investigates the vulnerability of a deep COVID-19 diagnosis model to the Fast Gradient Sign Method (FGSM) adversarial attack. Leveraging transfer learning of EfficientNet-B2 on a publicly available dataset, a deep learning-based COVID-19 diagnosis model is developed, achieving an impressive average accuracy of 94.56% on clean test data. However, when subjected to an untargeted FGSM attack with varying epsilon values, the model\u2019s accuracy is severely compromised, plummeting to 21.72% at epsilon 0.008. Notably, the attack successfully misclassifies adversarial COVID-19 images as normal with 100% confidence. This study underscores the critical need for further research and development to address these vulnerabilities and ensure the reliability and accuracy of deep learning models in the diagnosis of COVID-19 patients.","doi":"10.1007\/s41870-023-01538-7","cleaned_title":"white-box inference attack: compromising the security of deep learning-based covid-19 diagnosis systems","cleaned_abstract":"the covid-19 pandemic has necessitated the exploration of innovative diagnostic approaches, including the utilization of machine learning (ml) and deep learning (dl) technologies. however, recent findings shed light on the susceptibility of deep learning-based models to adversarial attacks, leading to erroneous predictions. this study investigates the vulnerability of a deep covid-19 diagnosis model to the fast gradient sign method (fgsm) adversarial attack. leveraging transfer learning of efficientnet-b2 on a publicly available dataset, a deep learning-based covid-19 diagnosis model is developed, achieving an impressive average accuracy of 94.56% on clean test data. however, when subjected to an untargeted fgsm attack with varying epsilon values, the model\u2019s accuracy is severely compromised, plummeting to 21.72% at epsilon 0.008. notably, the attack successfully misclassifies adversarial covid-19 images as normal with 100% confidence. this study underscores the critical need for further research and development to address these vulnerabilities and ensure the reliability and accuracy of deep learning models in the diagnosis of covid-19 patients.","key_phrases":["the exploration","innovative diagnostic approaches","the utilization","machine learning","ml","deep learning","(dl) technologies","recent findings","light","the susceptibility","deep learning-based models","adversarial attacks","erroneous predictions","this study","the vulnerability","a deep covid-19 diagnosis model","the fast gradient sign method","fgsm) adversarial attack","transfer learning","efficientnet-b2","a publicly available dataset","a deep learning-based covid-19 diagnosis model","an impressive average accuracy","94.56%","clean test data","an untargeted fgsm attack","varying epsilon values","the model\u2019s accuracy","21.72%","epsilon","the attack","adversarial covid-19 images","100% confidence","this study","the critical need","further research","development","these vulnerabilities","the reliability","accuracy","deep learning models","the diagnosis","covid-19 patients","covid-19","covid-19","covid-19","94.56%","21.72%","0.008","covid-19","100%","covid-19"]},{"title":"Prediction of intraoperative hypotension using deep learning models based on non-invasive monitoring devices","authors":["Heejoon Jeong","Donghee Kim","Dong Won Kim","Seungho Baek","Hyung-Chul Lee","Yusung Kim","Hyun Joo Ahn"],"abstract":"PurposeIntraoperative hypotension is associated with adverse outcomes. Predicting and proactively managing hypotension can reduce its incidence. Previously, hypotension prediction algorithms using artificial intelligence were developed for invasive arterial blood pressure monitors. This study tested whether routine non-invasive monitors could also predict intraoperative hypotension using deep learning algorithms.MethodsAn open-source database of non-cardiac surgery patients (https:\/\/vitadb.net\/dataset) was used to develop the deep learning algorithm. The algorithm was validated using external data obtained from a tertiary Korean hospital. Intraoperative hypotension was defined as a systolic blood pressure less than 90 mmHg. The input data included five monitors: non-invasive blood pressure, electrocardiography, photoplethysmography, capnography, and bispectral index. The primary outcome was the performance of the deep learning model as assessed by the area under the receiver operating characteristic curve (AUROC).ResultsData from 4754 and 421 patients were used for algorithm development and external validation, respectively. The fully connected model of Multi-head Attention architecture and the Globally Attentive Locally Recurrent model with Focal Loss function were able to predict intraoperative hypotension 5\u00a0min before its occurrence. The AUROC of the algorithm was 0.917 (95% confidence interval [CI], 0.915\u20130.918) for the original data and 0.833 (95% CI, 0.830\u20130.836) for the external validation data. Attention map, which quantified the contributions of each monitor, showed that our algorithm utilized data from each monitor with weights ranging from 8 to 22% for determining hypotension.ConclusionsA deep learning model utilizing multi-channel non-invasive monitors could predict intraoperative hypotension with high accuracy. Future prospective studies are needed to determine whether this model can assist clinicians in preventing hypotension in patients undergoing surgery with non-invasive monitoring.","doi":"10.1007\/s10877-024-01206-6","cleaned_title":"prediction of intraoperative hypotension using deep learning models based on non-invasive monitoring devices","cleaned_abstract":"purposeintraoperative hypotension is associated with adverse outcomes. predicting and proactively managing hypotension can reduce its incidence. previously, hypotension prediction algorithms using artificial intelligence were developed for invasive arterial blood pressure monitors. this study tested whether routine non-invasive monitors could also predict intraoperative hypotension using deep learning algorithms.methodsan open-source database of non-cardiac surgery patients (https:\/\/vitadb.net\/dataset) was used to develop the deep learning algorithm. the algorithm was validated using external data obtained from a tertiary korean hospital. intraoperative hypotension was defined as a systolic blood pressure less than 90 mmhg. the input data included five monitors: non-invasive blood pressure, electrocardiography, photoplethysmography, capnography, and bispectral index. the primary outcome was the performance of the deep learning model as assessed by the area under the receiver operating characteristic curve (auroc).resultsdata from 4754 and 421 patients were used for algorithm development and external validation, respectively. the fully connected model of multi-head attention architecture and the globally attentive locally recurrent model with focal loss function were able to predict intraoperative hypotension 5 min before its occurrence. the auroc of the algorithm was 0.917 (95% confidence interval [ci], 0.915\u20130.918) for the original data and 0.833 (95% ci, 0.830\u20130.836) for the external validation data. attention map, which quantified the contributions of each monitor, showed that our algorithm utilized data from each monitor with weights ranging from 8 to 22% for determining hypotension.conclusionsa deep learning model utilizing multi-channel non-invasive monitors could predict intraoperative hypotension with high accuracy. future prospective studies are needed to determine whether this model can assist clinicians in preventing hypotension in patients undergoing surgery with non-invasive monitoring.","key_phrases":["purposeintraoperative hypotension","adverse outcomes","hypotension","its incidence","hypotension prediction algorithms","artificial intelligence","invasive arterial blood pressure monitors","this study","routine non-invasive monitors","intraoperative hypotension","algorithms.methodsan open-source database","non-cardiac surgery patients","https:\/\/vitadb.net\/dataset","the deep learning algorithm","the algorithm","external data","a tertiary korean hospital","intraoperative hypotension","a systolic blood pressure","less than 90 mmhg","the input data","five monitors","non-invasive blood pressure","electrocardiography","photoplethysmography","capnography","bispectral index","the primary outcome","the performance","the deep learning model","the area","the receiver operating characteristic curve","auroc).resultsdata","421 patients","algorithm development","external validation","the fully connected model","multi-head attention architecture","the globally attentive locally recurrent model","focal loss function","intraoperative hypotension 5 min","its occurrence","the auroc","the algorithm","95% confidence interval","the original data","(95% ci","the external validation data","attention map","which","the contributions","each monitor","our algorithm utilized data","each monitor","weights","8 to 22%","hypotension.conclusionsa deep learning model","multi-channel non-invasive monitors","intraoperative hypotension","high accuracy","future prospective studies","this model","clinicians","hypotension","patients","surgery","non-invasive monitoring","algorithms.methodsan","tertiary","korean","less than","five","4754","421","5","0.917","95%","0.915\u20130.918","0.833","95%","8 to 22%","hypotension.conclusionsa","clinicians"]},{"title":"Deep learning nomogram for predicting neoadjuvant chemotherapy response in locally advanced gastric cancer patients","authors":["Jingjing Zhang","Qiang Zhang","Bo Zhao","Gaofeng Shi"],"abstract":"PurposeDeveloped and validated a deep learning radiomics nomogram using multi-phase contrast-enhanced computed tomography (CECT) images to predict neoadjuvant chemotherapy (NAC) response in locally advanced gastric cancer (LAGC) patients.MethodsThis multi-center study retrospectively included 322 patients diagnosed with gastric cancer from January 2013 to June 2023 at two hospitals. Handcrafted radiomics technique and the EfficientNet V2 neural network were applied to arterial, portal venous, and delayed phase CT images to extract two-dimensional handcrafted and deep learning features. A nomogram model was built by integrating the handcrafted signature, the deep learning signature, with clinical features. Discriminative ability was assessed using the receiver operating characteristics (ROC) curve and the precision-recall (P-R) curve. Model fitting was evaluated using calibration curves, and clinical utility was assessed through decision curve analysis (DCA).ResultsThe nomogram exhibited excellent performance. The area under the ROC curve (AUC) was 0.848 [95% confidence interval (CI), 0.793\u20130.893)], 0.802 (95% CI 0.688\u20130.889), and 0.751 (95% CI 0.652\u20130.833) for the training, internal validation, and external validation sets, respectively. The AUCs of the P-R curves were 0.838 (95% CI 0.756\u20130.895), 0.541 (95% CI 0.329\u20130.740), and 0.556 (95% CI 0.376\u20130.722) for the corresponding sets. The nomogram outperformed the clinical model and handcrafted signature across all sets (all P\u2009<\u20090.05). The nomogram model demonstrated good calibration and provided greater net benefit within the relevant threshold range compared to other models.ConclusionThis study created a deep learning nomogram using CECT images and clinical data to predict NAC response in LAGC patients undergoing surgical resection, offering personalized treatment insights.Graphical abstract","doi":"10.1007\/s00261-024-04331-7","cleaned_title":"deep learning nomogram for predicting neoadjuvant chemotherapy response in locally advanced gastric cancer patients","cleaned_abstract":"purposedeveloped and validated a deep learning radiomics nomogram using multi-phase contrast-enhanced computed tomography (cect) images to predict neoadjuvant chemotherapy (nac) response in locally advanced gastric cancer (lagc) patients.methodsthis multi-center study retrospectively included 322 patients diagnosed with gastric cancer from january 2013 to june 2023 at two hospitals. handcrafted radiomics technique and the efficientnet v2 neural network were applied to arterial, portal venous, and delayed phase ct images to extract two-dimensional handcrafted and deep learning features. a nomogram model was built by integrating the handcrafted signature, the deep learning signature, with clinical features. discriminative ability was assessed using the receiver operating characteristics (roc) curve and the precision-recall (p-r) curve. model fitting was evaluated using calibration curves, and clinical utility was assessed through decision curve analysis (dca).resultsthe nomogram exhibited excellent performance. the area under the roc curve (auc) was 0.848 [95% confidence interval (ci), 0.793\u20130.893)], 0.802 (95% ci 0.688\u20130.889), and 0.751 (95% ci 0.652\u20130.833) for the training, internal validation, and external validation sets, respectively. the aucs of the p-r curves were 0.838 (95% ci 0.756\u20130.895), 0.541 (95% ci 0.329\u20130.740), and 0.556 (95% ci 0.376\u20130.722) for the corresponding sets. the nomogram outperformed the clinical model and handcrafted signature across all sets (all p < 0.05). the nomogram model demonstrated good calibration and provided greater net benefit within the relevant threshold range compared to other models.conclusionthis study created a deep learning nomogram using cect images and clinical data to predict nac response in lagc patients undergoing surgical resection, offering personalized treatment insights.graphical abstract","key_phrases":["a deep learning radiomics nomogram","multi-phase contrast-enhanced computed tomography","cect","neoadjuvant chemotherapy (nac) response","locally advanced gastric cancer","lagc","patients.methodsthis multi-center study","322 patients","gastric cancer","january","june","two hospitals","handcrafted radiomics technique","the efficientnet v2 neural network","arterial, portal venous, and delayed phase ct images","two-dimensional handcrafted and deep learning features","a nomogram model","the handcrafted signature","the deep learning signature","clinical features","discriminative ability","the receiver operating characteristics","(roc) curve","the precision-recall (p-r) curve","model fitting","calibration curves","clinical utility","decision curve analysis","dca).resultsthe nomogram","excellent performance","the area","the roc curve","auc","[95% confidence interval","ci","0.793\u20130.893","0.802 (95%","ci 0.688\u20130.889","(95%","ci","the training","internal validation","external validation sets","the aucs","the p-r curves","0.756\u20130.895","(95%","ci 0.329\u20130.740","ci 0.376\u20130.722","the corresponding sets","the nomogram","the clinical model","signature","all sets","all p","the nomogram model","good calibration","greater net benefit","the relevant threshold range","other models.conclusionthis study","a deep learning nomogram","cect images","clinical data","nac response","lagc patients","surgical resection","personalized treatment insights.graphical abstract","322","january 2013 to june 2023","two","two","roc","roc","0.848","95%","0.802","95%","0.688\u20130.889","0.751","95%","0.652\u20130.833","0.838","95%","0.541","95%","0.329\u20130.740","0.556","95%"]},{"title":"Deep learning in cancer genomics and histopathology","authors":["Michaela Unger","Jakob Nikolas Kather"],"abstract":"Histopathology and genomic profiling are cornerstones of precision oncology and are routinely obtained for patients with cancer. Traditionally, histopathology slides are manually reviewed by highly trained pathologists. Genomic data, on the other hand, is evaluated by engineered computational pipelines. In both applications, the advent of modern artificial intelligence methods, specifically machine learning (ML) and deep learning (DL), have opened up a fundamentally new way of extracting actionable insights from raw data, which could augment and potentially replace some aspects of traditional evaluation workflows. In this review, we summarize current and emerging applications of DL in histopathology and genomics, including basic diagnostic as well as advanced prognostic tasks. Based on a growing body of evidence, we suggest that DL could be the groundwork for a new kind of workflow in oncology and cancer research. However, we also point out that DL models can have biases and other flaws that users in healthcare and research need to know about, and we propose ways to address them.","doi":"10.1186\/s13073-024-01315-6","cleaned_title":"deep learning in cancer genomics and histopathology","cleaned_abstract":"histopathology and genomic profiling are cornerstones of precision oncology and are routinely obtained for patients with cancer. traditionally, histopathology slides are manually reviewed by highly trained pathologists. genomic data, on the other hand, is evaluated by engineered computational pipelines. in both applications, the advent of modern artificial intelligence methods, specifically machine learning (ml) and deep learning (dl), have opened up a fundamentally new way of extracting actionable insights from raw data, which could augment and potentially replace some aspects of traditional evaluation workflows. in this review, we summarize current and emerging applications of dl in histopathology and genomics, including basic diagnostic as well as advanced prognostic tasks. based on a growing body of evidence, we suggest that dl could be the groundwork for a new kind of workflow in oncology and cancer research. however, we also point out that dl models can have biases and other flaws that users in healthcare and research need to know about, and we propose ways to address them.","key_phrases":["histopathology","genomic profiling","cornerstones","precision oncology","patients","cancer","histopathology slides","highly trained pathologists","genomic data","the other hand","engineered computational pipelines","both applications","the advent","modern artificial intelligence methods","specifically machine learning","deep learning","dl","a fundamentally new way","actionable insights","raw data","which","some aspects","traditional evaluation workflows","this review","we","current and emerging applications","dl","histopathology","genomics","basic diagnostic as well as advanced prognostic tasks","a growing body","evidence","we","dl","the groundwork","a new kind","workflow","oncology and cancer research","we","dl models","biases","other flaws","that","users","healthcare","research","we","ways","them"]},{"title":"Phase unwrapping based on deep learning in light field fringe projection 3D measurement","authors":["Xinjun Zhu","Haichuan Zhao","Mengkai Yuan","Zhizhi Zhang","Hongyi Wang","Limei Song"],"abstract":"Phase unwrapping is one of the key roles in fringe projection three-dimensional (3D) measurement technology. We propose a new method to achieve phase unwrapping in camera array light filed fringe projection 3D measurement based on deep learning. A multi-stream convolutional neural network (CNN) is proposed to learn the mapping relationship between camera array light filed wrapped phases and fringe orders of the expected central view, and is used to predict the fringe order to achieve the phase unwrapping. Experiments are performed on the light field fringe projection data generated by the simulated camera array fringe projection measurement system in Blender and by the experimental 3\u00d73 camera array light field fringe projection system. The performance of the proposed network with light field wrapped phases using multiple directions as network input data is studied, and the advantages of phase unwrapping based on deep learning in light filed fringe projection are demonstrated.","doi":"10.1007\/s11801-023-3002-4","cleaned_title":"phase unwrapping based on deep learning in light field fringe projection 3d measurement","cleaned_abstract":"phase unwrapping is one of the key roles in fringe projection three-dimensional (3d) measurement technology. we propose a new method to achieve phase unwrapping in camera array light filed fringe projection 3d measurement based on deep learning. a multi-stream convolutional neural network (cnn) is proposed to learn the mapping relationship between camera array light filed wrapped phases and fringe orders of the expected central view, and is used to predict the fringe order to achieve the phase unwrapping. experiments are performed on the light field fringe projection data generated by the simulated camera array fringe projection measurement system in blender and by the experimental 3\u00d73 camera array light field fringe projection system. the performance of the proposed network with light field wrapped phases using multiple directions as network input data is studied, and the advantages of phase unwrapping based on deep learning in light filed fringe projection are demonstrated.","key_phrases":["phase","the key roles","fringe projection three-dimensional (3d) measurement technology","we","a new method","phase","camera array light","fringe projection 3d measurement","deep learning","a multi-stream convolutional neural network","cnn","the mapping relationship","camera array light","phases","fringe orders","the expected central view","the fringe order","the phase","experiments","the light field fringe projection data","the simulated camera array fringe projection measurement system","blender","the experimental 3\u00d73 camera array light field fringe projection system","the performance","the proposed network","light field","phases","multiple directions","network input data","the advantages","phase","deep learning","light filed fringe projection","three","3d","3d","cnn","3\u00d73"]},{"title":"Radiological age assessment based on clavicle ossification in CT: enhanced accuracy through deep learning","authors":["Philipp Wesp","Balthasar Maria Schachtner","Katharina Jeblick","Johanna Topalis","Marvin Weber","Florian Fischer","Randolph Penning","Jens Ricke","Michael Ingrisch","Bastian Oliver Sabel"],"abstract":"BackgroundRadiological age assessment using reference studies is inherently limited in accuracy due to a finite number of assignable skeletal maturation stages. To overcome this limitation, we present a deep learning approach for continuous age assessment based on clavicle ossification in computed tomography (CT).MethodsThoracic CT scans were retrospectively collected from the picture archiving and communication system. Individuals aged 15.0 to 30.0\u00a0years examined in routine clinical practice were included. All scans were automatically cropped around the medial clavicular epiphyseal cartilages. A deep learning model was trained to predict a person\u2019s chronological age based on these scans. Performance was evaluated using mean absolute error (MAE). Model performance was compared to an optimistic human reader performance estimate for an established reference study method.ResultsThe deep learning model was trained on 4,400 scans of 1,935 patients (training set: mean age\u2009=\u200924.2\u00a0years\u2009\u00b1\u20094.0, 1132 female) and evaluated on 300 scans of 300 patients with a balanced age and sex distribution (test set: mean age\u2009=\u200922.5\u00a0years\u2009\u00b1\u20094.4, 150 female). Model MAE was 1.65\u00a0years, and the highest absolute error was 6.40\u00a0years for females and 7.32\u00a0years for males. However, performance could be attributed to norm-variants or pathologic disorders. Human reader estimate MAE was 1.84\u00a0years and the highest absolute error was 3.40\u00a0years for females and 3.78\u00a0years for males.ConclusionsWe present a deep learning approach for continuous age predictions using CT volumes highlighting the medial clavicular epiphyseal cartilage with performance comparable to the human reader estimate.","doi":"10.1007\/s00414-024-03167-6","cleaned_title":"radiological age assessment based on clavicle ossification in ct: enhanced accuracy through deep learning","cleaned_abstract":"backgroundradiological age assessment using reference studies is inherently limited in accuracy due to a finite number of assignable skeletal maturation stages. to overcome this limitation, we present a deep learning approach for continuous age assessment based on clavicle ossification in computed tomography (ct).methodsthoracic ct scans were retrospectively collected from the picture archiving and communication system. individuals aged 15.0 to 30.0 years examined in routine clinical practice were included. all scans were automatically cropped around the medial clavicular epiphyseal cartilages. a deep learning model was trained to predict a person\u2019s chronological age based on these scans. performance was evaluated using mean absolute error (mae). model performance was compared to an optimistic human reader performance estimate for an established reference study method.resultsthe deep learning model was trained on 4,400 scans of 1,935 patients (training set: mean age = 24.2 years \u00b1 4.0, 1132 female) and evaluated on 300 scans of 300 patients with a balanced age and sex distribution (test set: mean age = 22.5 years \u00b1 4.4, 150 female). model mae was 1.65 years, and the highest absolute error was 6.40 years for females and 7.32 years for males. however, performance could be attributed to norm-variants or pathologic disorders. human reader estimate mae was 1.84 years and the highest absolute error was 3.40 years for females and 3.78 years for males.conclusionswe present a deep learning approach for continuous age predictions using ct volumes highlighting the medial clavicular epiphyseal cartilage with performance comparable to the human reader estimate.","key_phrases":["backgroundradiological age assessment","reference studies","accuracy","a finite number","assignable skeletal maturation stages","this limitation","we","a deep learning approach","continuous age assessment","clavicle ossification","computed tomography","(ct).methodsthoracic ct scans","the picture archiving and communication system","individuals","15.0 to 30.0 years","routine clinical practice","all scans","the medial clavicular epiphyseal cartilages","a deep learning model","a person\u2019s chronological age","these scans","performance","mean absolute error","mae","model performance","an optimistic human reader performance estimate","an established reference study","method.resultsthe deep learning model","4,400 scans","1,935 patients","training set","mean age","\u00b1","1132 female","300 scans","300 patients","a balanced age","sex distribution","mean age","150 female","model mae","1.65 years","the highest absolute error","6.40 years","females","7.32 years","males","performance","norm-variants","pathologic disorders","human reader","mae","1.84 years","the highest absolute error","3.40 years","females","3.78 years","a deep learning approach","continuous age predictions","ct volumes","the medial clavicular epiphyseal cartilage","performance","the human reader estimate","15.0 to 30.0 years","4,400","1,935","24.2 years","4.0, 1132","300","300","22.5 years","4.4","150","mae","1.65 years","6.40 years","7.32 years","1.84 years","3.40 years","3.78 years"]},{"title":"Development and application of a deep learning-based comprehensive early diagnostic model for chronic obstructive pulmonary disease","authors":["Zecheng Zhu","Shunjin Zhao","Jiahui Li","Yuting Wang","Luopiao Xu","Yubing Jia","Zihan Li","Wenyuan Li","Gang Chen","Xifeng Wu"],"abstract":"BackgroundChronic obstructive pulmonary disease (COPD) is a frequently diagnosed yet treatable condition, provided it is identified early and managed effectively. This study aims to develop an advanced COPD diagnostic model by integrating deep learning and radiomics features.MethodsWe utilized a dataset comprising CT images from 2,983 participants, of which 2,317 participants also provided epidemiological data through questionnaires. Deep learning features were extracted using a Variational Autoencoder, and radiomics features were obtained using the PyRadiomics package. Multi-Layer Perceptrons were used to construct models based on deep learning and radiomics features independently, as well as a fusion model integrating both. Subsequently, epidemiological questionnaire data were incorporated to establish a more comprehensive model. The diagnostic performance of standalone models, the fusion model and the comprehensive model was evaluated and compared using metrics including accuracy, precision, recall, F1-score, Brier score, receiver operating characteristic curves, and area under the curve (AUC).ResultsThe fusion model exhibited outstanding performance with an AUC of 0.952, surpassing the standalone models based solely on deep learning features (AUC\u2009=\u20090.844) or radiomics features (AUC\u2009=\u20090.944). Notably, the comprehensive model, incorporating deep learning features, radiomics features, and questionnaire variables demonstrated the highest diagnostic performance among all models, yielding an AUC of 0.971.ConclusionWe developed and implemented a data fusion strategy to construct a state-of-the-art COPD diagnostic model integrating deep learning features, radiomics features, and questionnaire variables. Our data fusion strategy proved effective, and the model can be easily deployed in clinical settings.Trial registrationNot applicable. This study is NOT a clinical trial, it does not report the results of a health care intervention on human participants.","doi":"10.1186\/s12931-024-02793-3","cleaned_title":"development and application of a deep learning-based comprehensive early diagnostic model for chronic obstructive pulmonary disease","cleaned_abstract":"backgroundchronic obstructive pulmonary disease (copd) is a frequently diagnosed yet treatable condition, provided it is identified early and managed effectively. this study aims to develop an advanced copd diagnostic model by integrating deep learning and radiomics features.methodswe utilized a dataset comprising ct images from 2,983 participants, of which 2,317 participants also provided epidemiological data through questionnaires. deep learning features were extracted using a variational autoencoder, and radiomics features were obtained using the pyradiomics package. multi-layer perceptrons were used to construct models based on deep learning and radiomics features independently, as well as a fusion model integrating both. subsequently, epidemiological questionnaire data were incorporated to establish a more comprehensive model. the diagnostic performance of standalone models, the fusion model and the comprehensive model was evaluated and compared using metrics including accuracy, precision, recall, f1-score, brier score, receiver operating characteristic curves, and area under the curve (auc).resultsthe fusion model exhibited outstanding performance with an auc of 0.952, surpassing the standalone models based solely on deep learning features (auc = 0.844) or radiomics features (auc = 0.944). notably, the comprehensive model, incorporating deep learning features, radiomics features, and questionnaire variables demonstrated the highest diagnostic performance among all models, yielding an auc of 0.971.conclusionwe developed and implemented a data fusion strategy to construct a state-of-the-art copd diagnostic model integrating deep learning features, radiomics features, and questionnaire variables. our data fusion strategy proved effective, and the model can be easily deployed in clinical settings.trial registrationnot applicable. this study is not a clinical trial, it does not report the results of a health care intervention on human participants.","key_phrases":["backgroundchronic obstructive pulmonary disease","copd","a frequently diagnosed yet treatable condition","it","this study","an advanced copd diagnostic model","deep learning and radiomics features.methodswe","a dataset","ct images","2,983 participants","which","2,317 participants","epidemiological data","questionnaires","deep learning features","a variational autoencoder","radiomics features","the pyradiomics package","multi-layer perceptrons","models","deep learning","radiomics","a fusion model","both","epidemiological questionnaire data","a more comprehensive model","the diagnostic performance","standalone models","the fusion model","the comprehensive model","metrics","accuracy","precision","recall, f1-score","brier score","the curve","auc).resultsthe fusion model","outstanding performance","an auc","the standalone models","deep learning features","auc","radiomics features","auc =","the comprehensive model","deep learning features","radiomics features","questionnaire variables","the highest diagnostic performance","all models","an auc","0.971.conclusionwe","a data fusion strategy","the-art","deep learning features","radiomics features","questionnaire variables","our data fusion strategy","the model","clinical settings.trial registrationnot","this study","a clinical trial","it","the results","a health care intervention","human participants","2,983","2,317","0.952","0.844","0.944","0.971.conclusionwe"]},{"title":"Deep learning-based phenotype imputation on population-scale biobank data increases genetic discoveries","authors":["Ulzee An","Ali Pazokitoroudi","Marcus Alvarez","Lianyun Huang","Silviu Bacanu","Andrew J. Schork","Kenneth Kendler","P\u00e4ivi Pajukanta","Jonathan Flint","Noah Zaitlen","Na Cai","Andy Dahl","Sriram Sankararaman"],"abstract":"Biobanks that collect deep phenotypic and genomic data across many individuals have emerged as a key resource in human genetics. However, phenotypes in biobanks are often missing across many individuals, limiting their utility. We propose AutoComplete, a deep learning-based imputation method to impute or \u2018fill-in\u2019 missing phenotypes in population-scale biobank datasets. When applied to collections of phenotypes measured across ~300,000 individuals from the UK Biobank, AutoComplete substantially improved imputation accuracy over existing methods. On three traits with notable amounts of missingness, we show that AutoComplete yields imputed phenotypes that are genetically similar to the originally observed phenotypes while increasing the effective sample size by about twofold on average. Further, genome-wide association analyses on the resulting imputed phenotypes led to a substantial increase in the number of associated loci. Our results demonstrate the utility of deep learning-based phenotype imputation to increase power for genetic discoveries in existing biobank datasets.","doi":"10.1038\/s41588-023-01558-w","cleaned_title":"deep learning-based phenotype imputation on population-scale biobank data increases genetic discoveries","cleaned_abstract":"biobanks that collect deep phenotypic and genomic data across many individuals have emerged as a key resource in human genetics. however, phenotypes in biobanks are often missing across many individuals, limiting their utility. we propose autocomplete, a deep learning-based imputation method to impute or \u2018fill-in\u2019 missing phenotypes in population-scale biobank datasets. when applied to collections of phenotypes measured across ~300,000 individuals from the uk biobank, autocomplete substantially improved imputation accuracy over existing methods. on three traits with notable amounts of missingness, we show that autocomplete yields imputed phenotypes that are genetically similar to the originally observed phenotypes while increasing the effective sample size by about twofold on average. further, genome-wide association analyses on the resulting imputed phenotypes led to a substantial increase in the number of associated loci. our results demonstrate the utility of deep learning-based phenotype imputation to increase power for genetic discoveries in existing biobank datasets.","key_phrases":["that","deep phenotypic and genomic data","many individuals","a key resource","human genetics","phenotypes","biobanks","many individuals","their utility","we","a deep learning-based imputation method","population-scale biobank datasets","collections","phenotypes","individuals","the uk biobank","existing methods","three traits","notable amounts","missingness","we","autocomplete yields","phenotypes","that","the originally observed phenotypes","the effective sample size","genome-wide association","the resulting imputed phenotypes","a substantial increase","the number","associated loci","our results","the utility","deep learning-based phenotype imputation","power","genetic discoveries","existing biobank datasets","uk","three","about twofold"]},{"title":"Automatic early detection of rice leaf diseases using hybrid deep learning and machine learning methods","authors":["Vikram Rajpoot","Akhilesh Tiwari","Anand Singh Jalal"],"abstract":"Plant leaf disease detection is critical for long-term agricultural viability. Numerous Artificial Intelligence (AI) and Machine Learning (ML) technologies have been implemented for detecting rice diseases. However, such methods failed to identify or have slow recognition causing severe output loss. Therefore, an advanced and precise detection method has become necessary to overcome this issue. This study analyzes plant diseases that affect rice, comprising three different forms of diseases. Bacterial leaf blight, Brown spot, and Leaf smut are three of the six diseases that can affect rice plants. In the proposed approach a VGG-16 transfer learning with Faster R-CNN deep architecture is used to extract features. After completing the transfer learning step, the gathered characteristics are categorized using the random forest method. The random forest classifier divided the radish field into three distinct regions. The images of rice plant leaves are taken from UCI Machine Learning Repository. The proposed approach obtains an average predicting accuracy of 97.3% for rice disease imagery class prediction. The extensive experiment outcomes demonstrate the suggested technique\u2019s validity, so it effectively detects rice diseases.","doi":"10.1007\/s11042-023-14969-y","cleaned_title":"automatic early detection of rice leaf diseases using hybrid deep learning and machine learning methods","cleaned_abstract":"plant leaf disease detection is critical for long-term agricultural viability. numerous artificial intelligence (ai) and machine learning (ml) technologies have been implemented for detecting rice diseases. however, such methods failed to identify or have slow recognition causing severe output loss. therefore, an advanced and precise detection method has become necessary to overcome this issue. this study analyzes plant diseases that affect rice, comprising three different forms of diseases. bacterial leaf blight, brown spot, and leaf smut are three of the six diseases that can affect rice plants. in the proposed approach a vgg-16 transfer learning with faster r-cnn deep architecture is used to extract features. after completing the transfer learning step, the gathered characteristics are categorized using the random forest method. the random forest classifier divided the radish field into three distinct regions. the images of rice plant leaves are taken from uci machine learning repository. the proposed approach obtains an average predicting accuracy of 97.3% for rice disease imagery class prediction. the extensive experiment outcomes demonstrate the suggested technique\u2019s validity, so it effectively detects rice diseases.","key_phrases":["plant leaf disease detection","long-term agricultural viability","numerous artificial intelligence","ai","machine learning (ml) technologies","rice diseases","such methods","slow recognition","severe output loss","an advanced and precise detection method","this issue","this study","plant diseases","that","rice","three different forms","diseases","bacterial leaf blight","brown spot","leaf smut","the six diseases","that","rice plants","the proposed approach","a vgg-16 transfer","faster r-cnn deep architecture","features","the transfer learning step","the gathered characteristics","the random forest method","the random forest classifier","the radish field","three distinct regions","the images","rice plant leaves","uci machine learning repository","the proposed approach","an average predicting accuracy","97.3%","rice disease imagery class prediction","the extensive experiment outcomes","the suggested technique\u2019s validity","it","rice diseases","three","three","six","three","uci","97.3%"]},{"title":"Enhancing Weld Inspection Through Comparative Analysis of Traditional Algorithms and Deep Learning Approaches","authors":["Baoxin Zhang","Xiaopeng Wang","Jinhan Cui","Juntao Wu","Zhi Xiong","Wenpin Zhang","Xinghua Yu"],"abstract":"Automated inspection is vital in modern industrial manufacturing, optimizing production processes and ensuring product quality. Welding, a widely used joining technique, is susceptible to defects like porosity and cracks, compromising product reliability. Traditional nondestructive testing (NDT) methods suffer from inefficiency and limited accuracy. Many researchers have tried to apply deep learning for defect detection to address these limitations. This study compares traditional algorithms with deep learning methods, specifically evaluating the SwinUNet model for weld segmentation. The model achieves an impressive F1 score of 96.31, surpassing traditional algorithms. Feature analysis utilizing class activation maps confirms the model's robust recognition and generalization capabilities. Additionally, segmentation results for different welding defects were compared among various models, further substantiating the recognition capabilities of SwinUNet. The findings contribute to the automation of weld identification and segmentation, driving industrial production efficiency and enhancing defect detection.","doi":"10.1007\/s10921-024-01047-y","cleaned_title":"enhancing weld inspection through comparative analysis of traditional algorithms and deep learning approaches","cleaned_abstract":"automated inspection is vital in modern industrial manufacturing, optimizing production processes and ensuring product quality. welding, a widely used joining technique, is susceptible to defects like porosity and cracks, compromising product reliability. traditional nondestructive testing (ndt) methods suffer from inefficiency and limited accuracy. many researchers have tried to apply deep learning for defect detection to address these limitations. this study compares traditional algorithms with deep learning methods, specifically evaluating the swinunet model for weld segmentation. the model achieves an impressive f1 score of 96.31, surpassing traditional algorithms. feature analysis utilizing class activation maps confirms the model's robust recognition and generalization capabilities. additionally, segmentation results for different welding defects were compared among various models, further substantiating the recognition capabilities of swinunet. the findings contribute to the automation of weld identification and segmentation, driving industrial production efficiency and enhancing defect detection.","key_phrases":["automated inspection","modern industrial manufacturing","production processes","product quality","welding","a widely used joining technique","defects","porosity","cracks","product reliability","traditional nondestructive testing (ndt) methods","inefficiency","limited accuracy","many researchers","deep learning","defect detection","these limitations","this study","traditional algorithms","deep learning methods","the swinunet model","weld segmentation","the model","an impressive f1 score","traditional algorithms","feature analysis","class activation maps","the model's robust recognition","generalization capabilities","segmentation results","different welding defects","various models","the recognition capabilities","swinunet","the findings","the automation","weld identification","segmentation","industrial production efficiency","defect detection","96.31"]},{"title":"Deep learning in cancer genomics and histopathology","authors":["Michaela Unger","Jakob Nikolas Kather"],"abstract":"Histopathology and genomic profiling are cornerstones of precision oncology and are routinely obtained for patients with cancer. Traditionally, histopathology slides are manually reviewed by highly trained pathologists. Genomic data, on the other hand, is evaluated by engineered computational pipelines. In both applications, the advent of modern artificial intelligence methods, specifically machine learning (ML) and deep learning (DL), have opened up a fundamentally new way of extracting actionable insights from raw data, which could augment and potentially replace some aspects of traditional evaluation workflows. In this review, we summarize current and emerging applications of DL in histopathology and genomics, including basic diagnostic as well as advanced prognostic tasks. Based on a growing body of evidence, we suggest that DL could be the groundwork for a new kind of workflow in oncology and cancer research. However, we also point out that DL models can have biases and other flaws that users in healthcare and research need to know about, and we propose ways to address them.","doi":"10.1186\/s13073-024-01315-6","cleaned_title":"deep learning in cancer genomics and histopathology","cleaned_abstract":"histopathology and genomic profiling are cornerstones of precision oncology and are routinely obtained for patients with cancer. traditionally, histopathology slides are manually reviewed by highly trained pathologists. genomic data, on the other hand, is evaluated by engineered computational pipelines. in both applications, the advent of modern artificial intelligence methods, specifically machine learning (ml) and deep learning (dl), have opened up a fundamentally new way of extracting actionable insights from raw data, which could augment and potentially replace some aspects of traditional evaluation workflows. in this review, we summarize current and emerging applications of dl in histopathology and genomics, including basic diagnostic as well as advanced prognostic tasks. based on a growing body of evidence, we suggest that dl could be the groundwork for a new kind of workflow in oncology and cancer research. however, we also point out that dl models can have biases and other flaws that users in healthcare and research need to know about, and we propose ways to address them.","key_phrases":["histopathology","genomic profiling","cornerstones","precision oncology","patients","cancer","histopathology slides","highly trained pathologists","genomic data","the other hand","engineered computational pipelines","both applications","the advent","modern artificial intelligence methods","specifically machine learning","deep learning","dl","a fundamentally new way","actionable insights","raw data","which","some aspects","traditional evaluation workflows","this review","we","current and emerging applications","dl","histopathology","genomics","basic diagnostic as well as advanced prognostic tasks","a growing body","evidence","we","dl","the groundwork","a new kind","workflow","oncology and cancer research","we","dl models","biases","other flaws","that","users","healthcare","research","we","ways","them"]},{"title":"Phase unwrapping based on deep learning in light field fringe projection 3D measurement","authors":["Xinjun Zhu","Haichuan Zhao","Mengkai Yuan","Zhizhi Zhang","Hongyi Wang","Limei Song"],"abstract":"Phase unwrapping is one of the key roles in fringe projection three-dimensional (3D) measurement technology. We propose a new method to achieve phase unwrapping in camera array light filed fringe projection 3D measurement based on deep learning. A multi-stream convolutional neural network (CNN) is proposed to learn the mapping relationship between camera array light filed wrapped phases and fringe orders of the expected central view, and is used to predict the fringe order to achieve the phase unwrapping. Experiments are performed on the light field fringe projection data generated by the simulated camera array fringe projection measurement system in Blender and by the experimental 3\u00d73 camera array light field fringe projection system. The performance of the proposed network with light field wrapped phases using multiple directions as network input data is studied, and the advantages of phase unwrapping based on deep learning in light filed fringe projection are demonstrated.","doi":"10.1007\/s11801-023-3002-4","cleaned_title":"phase unwrapping based on deep learning in light field fringe projection 3d measurement","cleaned_abstract":"phase unwrapping is one of the key roles in fringe projection three-dimensional (3d) measurement technology. we propose a new method to achieve phase unwrapping in camera array light filed fringe projection 3d measurement based on deep learning. a multi-stream convolutional neural network (cnn) is proposed to learn the mapping relationship between camera array light filed wrapped phases and fringe orders of the expected central view, and is used to predict the fringe order to achieve the phase unwrapping. experiments are performed on the light field fringe projection data generated by the simulated camera array fringe projection measurement system in blender and by the experimental 3\u00d73 camera array light field fringe projection system. the performance of the proposed network with light field wrapped phases using multiple directions as network input data is studied, and the advantages of phase unwrapping based on deep learning in light filed fringe projection are demonstrated.","key_phrases":["phase","the key roles","fringe projection three-dimensional (3d) measurement technology","we","a new method","phase","camera array light","fringe projection 3d measurement","deep learning","a multi-stream convolutional neural network","cnn","the mapping relationship","camera array light","phases","fringe orders","the expected central view","the fringe order","the phase","experiments","the light field fringe projection data","the simulated camera array fringe projection measurement system","blender","the experimental 3\u00d73 camera array light field fringe projection system","the performance","the proposed network","light field","phases","multiple directions","network input data","the advantages","phase","deep learning","light filed fringe projection","three","3d","3d","cnn","3\u00d73"]},{"title":"Radiological age assessment based on clavicle ossification in CT: enhanced accuracy through deep learning","authors":["Philipp Wesp","Balthasar Maria Schachtner","Katharina Jeblick","Johanna Topalis","Marvin Weber","Florian Fischer","Randolph Penning","Jens Ricke","Michael Ingrisch","Bastian Oliver Sabel"],"abstract":"BackgroundRadiological age assessment using reference studies is inherently limited in accuracy due to a finite number of assignable skeletal maturation stages. To overcome this limitation, we present a deep learning approach for continuous age assessment based on clavicle ossification in computed tomography (CT).MethodsThoracic CT scans were retrospectively collected from the picture archiving and communication system. Individuals aged 15.0 to 30.0\u00a0years examined in routine clinical practice were included. All scans were automatically cropped around the medial clavicular epiphyseal cartilages. A deep learning model was trained to predict a person\u2019s chronological age based on these scans. Performance was evaluated using mean absolute error (MAE). Model performance was compared to an optimistic human reader performance estimate for an established reference study method.ResultsThe deep learning model was trained on 4,400 scans of 1,935 patients (training set: mean age\u2009=\u200924.2\u00a0years\u2009\u00b1\u20094.0, 1132 female) and evaluated on 300 scans of 300 patients with a balanced age and sex distribution (test set: mean age\u2009=\u200922.5\u00a0years\u2009\u00b1\u20094.4, 150 female). Model MAE was 1.65\u00a0years, and the highest absolute error was 6.40\u00a0years for females and 7.32\u00a0years for males. However, performance could be attributed to norm-variants or pathologic disorders. Human reader estimate MAE was 1.84\u00a0years and the highest absolute error was 3.40\u00a0years for females and 3.78\u00a0years for males.ConclusionsWe present a deep learning approach for continuous age predictions using CT volumes highlighting the medial clavicular epiphyseal cartilage with performance comparable to the human reader estimate.","doi":"10.1007\/s00414-024-03167-6","cleaned_title":"radiological age assessment based on clavicle ossification in ct: enhanced accuracy through deep learning","cleaned_abstract":"backgroundradiological age assessment using reference studies is inherently limited in accuracy due to a finite number of assignable skeletal maturation stages. to overcome this limitation, we present a deep learning approach for continuous age assessment based on clavicle ossification in computed tomography (ct).methodsthoracic ct scans were retrospectively collected from the picture archiving and communication system. individuals aged 15.0 to 30.0 years examined in routine clinical practice were included. all scans were automatically cropped around the medial clavicular epiphyseal cartilages. a deep learning model was trained to predict a person\u2019s chronological age based on these scans. performance was evaluated using mean absolute error (mae). model performance was compared to an optimistic human reader performance estimate for an established reference study method.resultsthe deep learning model was trained on 4,400 scans of 1,935 patients (training set: mean age = 24.2 years \u00b1 4.0, 1132 female) and evaluated on 300 scans of 300 patients with a balanced age and sex distribution (test set: mean age = 22.5 years \u00b1 4.4, 150 female). model mae was 1.65 years, and the highest absolute error was 6.40 years for females and 7.32 years for males. however, performance could be attributed to norm-variants or pathologic disorders. human reader estimate mae was 1.84 years and the highest absolute error was 3.40 years for females and 3.78 years for males.conclusionswe present a deep learning approach for continuous age predictions using ct volumes highlighting the medial clavicular epiphyseal cartilage with performance comparable to the human reader estimate.","key_phrases":["backgroundradiological age assessment","reference studies","accuracy","a finite number","assignable skeletal maturation stages","this limitation","we","a deep learning approach","continuous age assessment","clavicle ossification","computed tomography","(ct).methodsthoracic ct scans","the picture archiving and communication system","individuals","15.0 to 30.0 years","routine clinical practice","all scans","the medial clavicular epiphyseal cartilages","a deep learning model","a person\u2019s chronological age","these scans","performance","mean absolute error","mae","model performance","an optimistic human reader performance estimate","an established reference study","method.resultsthe deep learning model","4,400 scans","1,935 patients","training set","mean age","\u00b1","1132 female","300 scans","300 patients","a balanced age","sex distribution","mean age","150 female","model mae","1.65 years","the highest absolute error","6.40 years","females","7.32 years","males","performance","norm-variants","pathologic disorders","human reader","mae","1.84 years","the highest absolute error","3.40 years","females","3.78 years","a deep learning approach","continuous age predictions","ct volumes","the medial clavicular epiphyseal cartilage","performance","the human reader estimate","15.0 to 30.0 years","4,400","1,935","24.2 years","4.0, 1132","300","300","22.5 years","4.4","150","mae","1.65 years","6.40 years","7.32 years","1.84 years","3.40 years","3.78 years"]},{"title":"Development and application of a deep learning-based comprehensive early diagnostic model for chronic obstructive pulmonary disease","authors":["Zecheng Zhu","Shunjin Zhao","Jiahui Li","Yuting Wang","Luopiao Xu","Yubing Jia","Zihan Li","Wenyuan Li","Gang Chen","Xifeng Wu"],"abstract":"BackgroundChronic obstructive pulmonary disease (COPD) is a frequently diagnosed yet treatable condition, provided it is identified early and managed effectively. This study aims to develop an advanced COPD diagnostic model by integrating deep learning and radiomics features.MethodsWe utilized a dataset comprising CT images from 2,983 participants, of which 2,317 participants also provided epidemiological data through questionnaires. Deep learning features were extracted using a Variational Autoencoder, and radiomics features were obtained using the PyRadiomics package. Multi-Layer Perceptrons were used to construct models based on deep learning and radiomics features independently, as well as a fusion model integrating both. Subsequently, epidemiological questionnaire data were incorporated to establish a more comprehensive model. The diagnostic performance of standalone models, the fusion model and the comprehensive model was evaluated and compared using metrics including accuracy, precision, recall, F1-score, Brier score, receiver operating characteristic curves, and area under the curve (AUC).ResultsThe fusion model exhibited outstanding performance with an AUC of 0.952, surpassing the standalone models based solely on deep learning features (AUC\u2009=\u20090.844) or radiomics features (AUC\u2009=\u20090.944). Notably, the comprehensive model, incorporating deep learning features, radiomics features, and questionnaire variables demonstrated the highest diagnostic performance among all models, yielding an AUC of 0.971.ConclusionWe developed and implemented a data fusion strategy to construct a state-of-the-art COPD diagnostic model integrating deep learning features, radiomics features, and questionnaire variables. Our data fusion strategy proved effective, and the model can be easily deployed in clinical settings.Trial registrationNot applicable. This study is NOT a clinical trial, it does not report the results of a health care intervention on human participants.","doi":"10.1186\/s12931-024-02793-3","cleaned_title":"development and application of a deep learning-based comprehensive early diagnostic model for chronic obstructive pulmonary disease","cleaned_abstract":"backgroundchronic obstructive pulmonary disease (copd) is a frequently diagnosed yet treatable condition, provided it is identified early and managed effectively. this study aims to develop an advanced copd diagnostic model by integrating deep learning and radiomics features.methodswe utilized a dataset comprising ct images from 2,983 participants, of which 2,317 participants also provided epidemiological data through questionnaires. deep learning features were extracted using a variational autoencoder, and radiomics features were obtained using the pyradiomics package. multi-layer perceptrons were used to construct models based on deep learning and radiomics features independently, as well as a fusion model integrating both. subsequently, epidemiological questionnaire data were incorporated to establish a more comprehensive model. the diagnostic performance of standalone models, the fusion model and the comprehensive model was evaluated and compared using metrics including accuracy, precision, recall, f1-score, brier score, receiver operating characteristic curves, and area under the curve (auc).resultsthe fusion model exhibited outstanding performance with an auc of 0.952, surpassing the standalone models based solely on deep learning features (auc = 0.844) or radiomics features (auc = 0.944). notably, the comprehensive model, incorporating deep learning features, radiomics features, and questionnaire variables demonstrated the highest diagnostic performance among all models, yielding an auc of 0.971.conclusionwe developed and implemented a data fusion strategy to construct a state-of-the-art copd diagnostic model integrating deep learning features, radiomics features, and questionnaire variables. our data fusion strategy proved effective, and the model can be easily deployed in clinical settings.trial registrationnot applicable. this study is not a clinical trial, it does not report the results of a health care intervention on human participants.","key_phrases":["backgroundchronic obstructive pulmonary disease","copd","a frequently diagnosed yet treatable condition","it","this study","an advanced copd diagnostic model","deep learning and radiomics features.methodswe","a dataset","ct images","2,983 participants","which","2,317 participants","epidemiological data","questionnaires","deep learning features","a variational autoencoder","radiomics features","the pyradiomics package","multi-layer perceptrons","models","deep learning","radiomics","a fusion model","both","epidemiological questionnaire data","a more comprehensive model","the diagnostic performance","standalone models","the fusion model","the comprehensive model","metrics","accuracy","precision","recall, f1-score","brier score","the curve","auc).resultsthe fusion model","outstanding performance","an auc","the standalone models","deep learning features","auc","radiomics features","auc =","the comprehensive model","deep learning features","radiomics features","questionnaire variables","the highest diagnostic performance","all models","an auc","0.971.conclusionwe","a data fusion strategy","the-art","deep learning features","radiomics features","questionnaire variables","our data fusion strategy","the model","clinical settings.trial registrationnot","this study","a clinical trial","it","the results","a health care intervention","human participants","2,983","2,317","0.952","0.844","0.944","0.971.conclusionwe"]},{"title":"Deep learning-based phenotype imputation on population-scale biobank data increases genetic discoveries","authors":["Ulzee An","Ali Pazokitoroudi","Marcus Alvarez","Lianyun Huang","Silviu Bacanu","Andrew J. Schork","Kenneth Kendler","P\u00e4ivi Pajukanta","Jonathan Flint","Noah Zaitlen","Na Cai","Andy Dahl","Sriram Sankararaman"],"abstract":"Biobanks that collect deep phenotypic and genomic data across many individuals have emerged as a key resource in human genetics. However, phenotypes in biobanks are often missing across many individuals, limiting their utility. We propose AutoComplete, a deep learning-based imputation method to impute or \u2018fill-in\u2019 missing phenotypes in population-scale biobank datasets. When applied to collections of phenotypes measured across ~300,000 individuals from the UK Biobank, AutoComplete substantially improved imputation accuracy over existing methods. On three traits with notable amounts of missingness, we show that AutoComplete yields imputed phenotypes that are genetically similar to the originally observed phenotypes while increasing the effective sample size by about twofold on average. Further, genome-wide association analyses on the resulting imputed phenotypes led to a substantial increase in the number of associated loci. Our results demonstrate the utility of deep learning-based phenotype imputation to increase power for genetic discoveries in existing biobank datasets.","doi":"10.1038\/s41588-023-01558-w","cleaned_title":"deep learning-based phenotype imputation on population-scale biobank data increases genetic discoveries","cleaned_abstract":"biobanks that collect deep phenotypic and genomic data across many individuals have emerged as a key resource in human genetics. however, phenotypes in biobanks are often missing across many individuals, limiting their utility. we propose autocomplete, a deep learning-based imputation method to impute or \u2018fill-in\u2019 missing phenotypes in population-scale biobank datasets. when applied to collections of phenotypes measured across ~300,000 individuals from the uk biobank, autocomplete substantially improved imputation accuracy over existing methods. on three traits with notable amounts of missingness, we show that autocomplete yields imputed phenotypes that are genetically similar to the originally observed phenotypes while increasing the effective sample size by about twofold on average. further, genome-wide association analyses on the resulting imputed phenotypes led to a substantial increase in the number of associated loci. our results demonstrate the utility of deep learning-based phenotype imputation to increase power for genetic discoveries in existing biobank datasets.","key_phrases":["that","deep phenotypic and genomic data","many individuals","a key resource","human genetics","phenotypes","biobanks","many individuals","their utility","we","a deep learning-based imputation method","population-scale biobank datasets","collections","phenotypes","individuals","the uk biobank","existing methods","three traits","notable amounts","missingness","we","autocomplete yields","phenotypes","that","the originally observed phenotypes","the effective sample size","genome-wide association","the resulting imputed phenotypes","a substantial increase","the number","associated loci","our results","the utility","deep learning-based phenotype imputation","power","genetic discoveries","existing biobank datasets","uk","three","about twofold"]},{"title":"Active Exploration Deep Reinforcement Learning for Continuous Action Space with Forward Prediction","authors":["Dongfang Zhao","Xu Huanshi","Zhang Xun"],"abstract":"The application of reinforcement learning (RL) to the field of autonomous robotics has high requirements about sample efficiency, since the agent expends for interaction with the environment. One method for sample efficiency is to extract knowledge from existing samples and used to exploration. Typical RL algorithms achieve exploration using task-specific knowledge or adding exploration noise. These methods are limited to current policy improvement level and lack of long-term planning. We propose a novel active exploration deep RL algorithm for the continuous action space problem named active exploration deep reinforcement learning (AEDRL). Our method uses the Gaussian process to model dynamic model, enabling the probability description of prediction sample. Action selection is formulated as the solution of the optimization problem. Thus, the optimization objective is specifically designed for selecting samples that can minimize the uncertainty of the dynamic model. Active exploration is achieved through long-term optimized action selection. This long-term considered action exploration method is more guidance for learning. Enable intelligent agents to explore more interesting action spaces. The proposed AEDRL algorithm is evaluated on several robotic control task including classic pendulum problem and five complex articulated robots. The AEDRL can learn a controller using fewer episodes and demonstrates performance and sample efficiency.","doi":"10.1007\/s44196-023-00389-1","cleaned_title":"active exploration deep reinforcement learning for continuous action space with forward prediction","cleaned_abstract":"the application of reinforcement learning (rl) to the field of autonomous robotics has high requirements about sample efficiency, since the agent expends for interaction with the environment. one method for sample efficiency is to extract knowledge from existing samples and used to exploration. typical rl algorithms achieve exploration using task-specific knowledge or adding exploration noise. these methods are limited to current policy improvement level and lack of long-term planning. we propose a novel active exploration deep rl algorithm for the continuous action space problem named active exploration deep reinforcement learning (aedrl). our method uses the gaussian process to model dynamic model, enabling the probability description of prediction sample. action selection is formulated as the solution of the optimization problem. thus, the optimization objective is specifically designed for selecting samples that can minimize the uncertainty of the dynamic model. active exploration is achieved through long-term optimized action selection. this long-term considered action exploration method is more guidance for learning. enable intelligent agents to explore more interesting action spaces. the proposed aedrl algorithm is evaluated on several robotic control task including classic pendulum problem and five complex articulated robots. the aedrl can learn a controller using fewer episodes and demonstrates performance and sample efficiency.","key_phrases":["the application","reinforcement learning","the field","autonomous robotics","high requirements","sample efficiency","the agent","interaction","the environment","one method","sample efficiency","knowledge","existing samples","typical rl algorithms","exploration","task-specific knowledge","exploration noise","these methods","current policy improvement level","lack","long-term planning","we","a novel active exploration","deep rl algorithm","the continuous action space problem","active exploration deep reinforcement learning","our method","the gaussian process","dynamic model","the probability description","prediction sample","action selection","the solution","the optimization problem","the optimization objective","samples","that","the uncertainty","the dynamic model","active exploration","long-term optimized action selection","this long-term considered action exploration method","more guidance","intelligent agents","more interesting action spaces","the proposed aedrl algorithm","several robotic control task","classic pendulum problem","five complex articulated robots","the aedrl","a controller","fewer episodes","performance","sample efficiency","one","five"]},{"title":"Automatic early detection of rice leaf diseases using hybrid deep learning and machine learning methods","authors":["Vikram Rajpoot","Akhilesh Tiwari","Anand Singh Jalal"],"abstract":"Plant leaf disease detection is critical for long-term agricultural viability. Numerous Artificial Intelligence (AI) and Machine Learning (ML) technologies have been implemented for detecting rice diseases. However, such methods failed to identify or have slow recognition causing severe output loss. Therefore, an advanced and precise detection method has become necessary to overcome this issue. This study analyzes plant diseases that affect rice, comprising three different forms of diseases. Bacterial leaf blight, Brown spot, and Leaf smut are three of the six diseases that can affect rice plants. In the proposed approach a VGG-16 transfer learning with Faster R-CNN deep architecture is used to extract features. After completing the transfer learning step, the gathered characteristics are categorized using the random forest method. The random forest classifier divided the radish field into three distinct regions. The images of rice plant leaves are taken from UCI Machine Learning Repository. The proposed approach obtains an average predicting accuracy of 97.3% for rice disease imagery class prediction. The extensive experiment outcomes demonstrate the suggested technique\u2019s validity, so it effectively detects rice diseases.","doi":"10.1007\/s11042-023-14969-y","cleaned_title":"automatic early detection of rice leaf diseases using hybrid deep learning and machine learning methods","cleaned_abstract":"plant leaf disease detection is critical for long-term agricultural viability. numerous artificial intelligence (ai) and machine learning (ml) technologies have been implemented for detecting rice diseases. however, such methods failed to identify or have slow recognition causing severe output loss. therefore, an advanced and precise detection method has become necessary to overcome this issue. this study analyzes plant diseases that affect rice, comprising three different forms of diseases. bacterial leaf blight, brown spot, and leaf smut are three of the six diseases that can affect rice plants. in the proposed approach a vgg-16 transfer learning with faster r-cnn deep architecture is used to extract features. after completing the transfer learning step, the gathered characteristics are categorized using the random forest method. the random forest classifier divided the radish field into three distinct regions. the images of rice plant leaves are taken from uci machine learning repository. the proposed approach obtains an average predicting accuracy of 97.3% for rice disease imagery class prediction. the extensive experiment outcomes demonstrate the suggested technique\u2019s validity, so it effectively detects rice diseases.","key_phrases":["plant leaf disease detection","long-term agricultural viability","numerous artificial intelligence","ai","machine learning (ml) technologies","rice diseases","such methods","slow recognition","severe output loss","an advanced and precise detection method","this issue","this study","plant diseases","that","rice","three different forms","diseases","bacterial leaf blight","brown spot","leaf smut","the six diseases","that","rice plants","the proposed approach","a vgg-16 transfer","faster r-cnn deep architecture","features","the transfer learning step","the gathered characteristics","the random forest method","the random forest classifier","the radish field","three distinct regions","the images","rice plant leaves","uci machine learning repository","the proposed approach","an average predicting accuracy","97.3%","rice disease imagery class prediction","the extensive experiment outcomes","the suggested technique\u2019s validity","it","rice diseases","three","three","six","three","uci","97.3%"]},{"title":"A literature review on deep learning algorithms for analysis of X-ray images","authors":["Gokhan Seyfi","Engin Esme","Merve Yilmaz","Mustafa Servet Kiran"],"abstract":"Since the invention of the X-ray beam, it has been used for useful applications in various fields, such as medical diagnosis, fluoroscopy, radiation therapy, and computed tomography. In addition, it is also widely used to identify prohibited or illegal materials using X-ray imaging in the security field. However, these procedures are generally dependent on the human factor. An operator detects prohibited objects by projecting pseudo-color images onto a computer screen. Because these processes are prone to error, much work has gone into automating the processes involved. Initial research on this topic consisted mainly of machine learning and methods using hand-crafted features. The newly developed deep learning methods have subsequently been more successful. For this reason, deep learning algorithms are a trend in recent studies and the number of publications has increased in areas such as X-ray imaging. Therefore, we surveyed the studies published in the literature on Deep Learning-based X-ray imaging to attract new readers and provide new perspectives.","doi":"10.1007\/s13042-023-01961-z","cleaned_title":"a literature review on deep learning algorithms for analysis of x-ray images","cleaned_abstract":"since the invention of the x-ray beam, it has been used for useful applications in various fields, such as medical diagnosis, fluoroscopy, radiation therapy, and computed tomography. in addition, it is also widely used to identify prohibited or illegal materials using x-ray imaging in the security field. however, these procedures are generally dependent on the human factor. an operator detects prohibited objects by projecting pseudo-color images onto a computer screen. because these processes are prone to error, much work has gone into automating the processes involved. initial research on this topic consisted mainly of machine learning and methods using hand-crafted features. the newly developed deep learning methods have subsequently been more successful. for this reason, deep learning algorithms are a trend in recent studies and the number of publications has increased in areas such as x-ray imaging. therefore, we surveyed the studies published in the literature on deep learning-based x-ray imaging to attract new readers and provide new perspectives.","key_phrases":["the invention","the x-ray beam","it","useful applications","various fields","medical diagnosis","radiation therapy","tomography","addition","it","prohibited or illegal materials","x-ray imaging","the security field","these procedures","the human factor","an operator detects","objects","pseudo-color images","a computer screen","these processes","error","much work","the processes","initial research","this topic","machine learning","methods","hand-crafted features","the newly developed deep learning methods","this reason","deep learning algorithms","a trend","recent studies","the number","publications","areas","x-ray imaging","we","the studies","the literature","deep learning-based x","-ray imaging","new readers","new perspectives"]},{"title":"Systematic review of deep learning image analyses for the diagnosis and monitoring of skin disease","authors":["Shern Ping Choy","Byung Jin Kim","Alexandra Paolino","Wei Ren Tan","Sarah Man Lin Lim","Jessica Seo","Sze Ping Tan","Luc Francis","Teresa Tsakok","Michael Simpson","Jonathan N. W. N. Barker","Magnus D. Lynch","Mark S. Corbett","Catherine H. Smith","Satveer K. Mahil"],"abstract":"Skin diseases affect one-third of the global population, posing a major healthcare burden. Deep learning may optimise healthcare workflows through processing skin images via neural networks to make predictions. A focus of deep learning research is skin lesion triage to detect cancer, but this may not translate to the wider scope of >2000 other skin diseases. We searched for studies applying deep learning to skin images, excluding benign\/malignant lesions (1\/1\/2000-23\/6\/2022, PROSPERO CRD42022309935). The primary outcome was accuracy of deep learning algorithms in disease diagnosis or severity assessment. We modified QUADAS-2 for quality assessment. Of 13,857 references identified, 64 were included. The most studied diseases were acne, psoriasis, eczema, rosacea, vitiligo, urticaria. Deep learning algorithms had high specificity and variable sensitivity in diagnosing these conditions. Accuracy of algorithms in diagnosing acne (median 94%, IQR 86\u201398; n\u2009=\u200911), rosacea (94%, 90\u201397; n\u2009=\u20094), eczema (93%, 90\u201399; n\u2009=\u20099) and psoriasis (89%, 78\u201392; n\u2009=\u20098) was high. Accuracy for grading severity was highest for psoriasis (range 93\u2013100%, n\u2009=\u20092), eczema (88%, n\u2009=\u20091), and acne (67\u201386%, n\u2009=\u20094). However, 59 (92%) studies had high risk-of-bias judgements and 62 (97%) had high-level applicability concerns. Only 12 (19%) reported participant ethnicity\/skin type. Twenty-four (37.5%) evaluated the algorithm in an independent dataset, clinical setting or prospectively. These data indicate potential of deep learning image analysis in diagnosing and monitoring common skin diseases. Current research has important methodological\/reporting limitations. Real-world, prospectively-acquired image datasets with external validation\/testing will advance deep learning beyond the current experimental phase towards clinically-useful tools to mitigate rising health and cost impacts of skin disease.","doi":"10.1038\/s41746-023-00914-8","cleaned_title":"systematic review of deep learning image analyses for the diagnosis and monitoring of skin disease","cleaned_abstract":"skin diseases affect one-third of the global population, posing a major healthcare burden. deep learning may optimise healthcare workflows through processing skin images via neural networks to make predictions. a focus of deep learning research is skin lesion triage to detect cancer, but this may not translate to the wider scope of >2000 other skin diseases. we searched for studies applying deep learning to skin images, excluding benign\/malignant lesions (1\/1\/2000-23\/6\/2022, prospero crd42022309935). the primary outcome was accuracy of deep learning algorithms in disease diagnosis or severity assessment. we modified quadas-2 for quality assessment. of 13,857 references identified, 64 were included. the most studied diseases were acne, psoriasis, eczema, rosacea, vitiligo, urticaria. deep learning algorithms had high specificity and variable sensitivity in diagnosing these conditions. accuracy of algorithms in diagnosing acne (median 94%, iqr 86\u201398; n = 11), rosacea (94%, 90\u201397; n = 4), eczema (93%, 90\u201399; n = 9) and psoriasis (89%, 78\u201392; n = 8) was high. accuracy for grading severity was highest for psoriasis (range 93\u2013100%, n = 2), eczema (88%, n = 1), and acne (67\u201386%, n = 4). however, 59 (92%) studies had high risk-of-bias judgements and 62 (97%) had high-level applicability concerns. only 12 (19%) reported participant ethnicity\/skin type. twenty-four (37.5%) evaluated the algorithm in an independent dataset, clinical setting or prospectively. these data indicate potential of deep learning image analysis in diagnosing and monitoring common skin diseases. current research has important methodological\/reporting limitations. real-world, prospectively-acquired image datasets with external validation\/testing will advance deep learning beyond the current experimental phase towards clinically-useful tools to mitigate rising health and cost impacts of skin disease.","key_phrases":["skin diseases","one-third","the global population","a major healthcare burden","deep learning","healthcare workflows","skin images","neural networks","predictions","a focus","deep learning research","skin lesion triage","cancer","this","the wider scope","other skin diseases","we","studies","deep learning","skin images","benign\/malignant lesions","prospero","the primary outcome","accuracy","deep learning algorithms","disease diagnosis","severity assessment","we","quadas-2","quality assessment","13,857 references","the most studied diseases","urticaria","deep learning algorithms","high specificity","variable sensitivity","these conditions","accuracy","algorithms","acne","iqr","rosacea","94%","90\u201397","eczema","93%","90\u201399","psoriasis","89%","n = 8)","accuracy","grading severity","psoriasis","n","eczema","88%","acne","67\u201386%","92%","bias","62 (97%","high-level applicability concerns","only 12 (19%","participant ethnicity\/skin type","twenty-four (37.5%","the algorithm","an independent dataset, clinical setting","these data","potential","deep learning image analysis","common skin diseases","current research","important methodological\/reporting limitations","real-world, prospectively-acquired image datasets","external validation\/testing","deep learning","the current experimental phase","clinically-useful tools","rising health and cost impacts","skin disease","one-third","2000","1\/1\/2000-23\/6\/2022","13,857","64","94%","11","94%","90\u201397","4","93%","90\u201399","9","89%","78\u201392","8)","93\u2013100%","2","88%","1","67\u201386%","4","59","92%","62","97%","only 12","19%","twenty-four","37.5%"]},{"title":"Deep learning-based automatic measurement system for patellar height: a multicenter retrospective study","authors":["Zeyu Liu","Jiangjiang Wu","Xu Gao","Zhipeng Qin","Run Tian","Chunsheng Wang"],"abstract":"BackgroundThe patellar height index is important; however, the measurement procedures are time-consuming and prone to significant variability among and within observers. We developed a deep learning-based automatic measurement system for the patellar height and evaluated its performance and generalization ability to accurately measure the patellar height index.MethodsWe developed a dataset containing 3,923 lateral knee X-ray images. Notably, all X-ray images were from three tertiary level A hospitals, and 2,341 cases were included in the analysis after screening. By manually labeling key points, the model was trained using the residual network (ResNet) and high-resolution network (HRNet) for human pose estimation architectures to measure the patellar height index. Various data enhancement techniques were used to enhance the robustness of the model. The root mean square error (RMSE), object keypoint similarity (OKS), and percentage of correct keypoint (PCK) metrics were used to evaluate the training results. In addition, we used the intraclass correlation coefficient (ICC) to assess the consistency between manual and automatic measurements.ResultsThe HRNet model performed excellently in keypoint detection tasks by comparing different deep learning models. Furthermore, the pose_hrnet_w48 model was particularly outstanding in the RMSE, OKS, and PCK metrics, and the Insall\u2013Salvati index (ISI) automatically calculated by this model was also highly consistent with the manual measurements (intraclass correlation coefficient [ICC], 0.809\u20130.885). This evidence demonstrates the accuracy and generalizability of this deep learning system in practical applications.ConclusionWe successfully developed a deep learning-based automatic measurement system for the patellar height. The system demonstrated accuracy comparable to that of experienced radiologists and a strong generalizability across different datasets. It provides an essential tool for assessing and treating knee diseases early and monitoring and rehabilitation after knee surgery. Due to the potential bias in the selection of datasets in this study, different datasets should be examined in the future to optimize the model so that it can be reliably applied in clinical practice.Trial registrationThe study was registered at the Medical Research Registration and Filing Information System (medicalresearch.org.cn) MR-61-23-013065. Date of registration: May 04, 2023 (retrospectively registered).","doi":"10.1186\/s13018-024-04809-6","cleaned_title":"deep learning-based automatic measurement system for patellar height: a multicenter retrospective study","cleaned_abstract":"backgroundthe patellar height index is important; however, the measurement procedures are time-consuming and prone to significant variability among and within observers. we developed a deep learning-based automatic measurement system for the patellar height and evaluated its performance and generalization ability to accurately measure the patellar height index.methodswe developed a dataset containing 3,923 lateral knee x-ray images. notably, all x-ray images were from three tertiary level a hospitals, and 2,341 cases were included in the analysis after screening. by manually labeling key points, the model was trained using the residual network (resnet) and high-resolution network (hrnet) for human pose estimation architectures to measure the patellar height index. various data enhancement techniques were used to enhance the robustness of the model. the root mean square error (rmse), object keypoint similarity (oks), and percentage of correct keypoint (pck) metrics were used to evaluate the training results. in addition, we used the intraclass correlation coefficient (icc) to assess the consistency between manual and automatic measurements.resultsthe hrnet model performed excellently in keypoint detection tasks by comparing different deep learning models. furthermore, the pose_hrnet_w48 model was particularly outstanding in the rmse, oks, and pck metrics, and the insall\u2013salvati index (isi) automatically calculated by this model was also highly consistent with the manual measurements (intraclass correlation coefficient [icc], 0.809\u20130.885). this evidence demonstrates the accuracy and generalizability of this deep learning system in practical applications.conclusionwe successfully developed a deep learning-based automatic measurement system for the patellar height. the system demonstrated accuracy comparable to that of experienced radiologists and a strong generalizability across different datasets. it provides an essential tool for assessing and treating knee diseases early and monitoring and rehabilitation after knee surgery. due to the potential bias in the selection of datasets in this study, different datasets should be examined in the future to optimize the model so that it can be reliably applied in clinical practice.trial registrationthe study was registered at the medical research registration and filing information system (medicalresearch.org.cn) mr-61-23-013065. date of registration: may 04, 2023 (retrospectively registered).","key_phrases":["backgroundthe patellar height index","the measurement procedures","significant variability","observers","we","a deep learning-based automatic measurement system","the patellar height","its performance and generalization ability","the patellar height index.methodswe","a dataset","3,923 lateral knee x-ray images","all x-ray images","three tertiary level","a hospitals","2,341 cases","the analysis","key points","the model","the residual network","resnet","high-resolution network","hrnet","human pose estimation","the patellar height index","various data enhancement techniques","the robustness","the model","the root mean square error","rmse","object keypoint similarity","oks","percentage","pck","the training results","addition","we","the intraclass correlation","coefficient","icc","the consistency","manual and automatic measurements.resultsthe hrnet model","keypoint detection tasks","different deep learning models","the pose_hrnet_w48 model","the rmse","oks","pck metrics","salvati index","isi","this model","the manual measurements","intraclass correlation","this evidence","the accuracy","generalizability","this deep learning system","a deep learning-based automatic measurement system","the patellar height","the system","accuracy","that","experienced radiologists","a strong generalizability","different datasets","it","an essential tool","knee diseases","knee surgery","the potential bias","the selection","datasets","this study","different datasets","the future","the model","it","clinical practice.trial registrationthe study","the medical research registration and filing information system","date","registration","3,923","three","2,341","rmse","oks","isi","mr-61","may 04, 2023"]},{"title":"Addressing data imbalance challenges in oral cavity histopathological whole slide images with advanced deep learning techniques","authors":["Tabasum Majeed","Tariq Ahmad Masoodi","Muzafar Ahmad Macha","Muzafar Rasool Bhat","Khalid Muzaffar","Assif Assad"],"abstract":"Oral Cavity Squamous Cell Carcinoma (OCSCC) represents a common form of head and neck cancer originating from the mucosal lining of the oral cavity, often detected in advanced stages. Traditional detection methods rely on analyzing hematoxylin and eosin (H&E)-stained histopathological whole-slide images, which are time-consuming and require expert pathology skills. Hence, automated analysis is urgently needed to expedite diagnosis and improve patient outcomes. Deep learning, through automated feature extraction, offers a promising avenue for capturing high-level abstract features with greater accuracy than traditional methods. However, the imbalance in class distribution within datasets significantly affects the performance of deep learning models during training, necessitating specialized approaches. To address the issue, various methods have been proposed at both data and algorithmic levels. This study investigates strategies to mitigate class imbalance by employing a publicly available OCSCC imbalance dataset. We evaluated undersampling methods (Near Miss, Edited Nearest Neighbors) and oversampling techniques (SMOTE, Deep SMOTE, ADASYN) integrated with transfer learning across different imbalance ratios (0.1, 0.15, 0.20, 0.30). Our findings demonstrate the effectiveness of SMOTE in improving test performance, highlighting the efficacy of strategic oversampling combined with transfer learning in classifying imbalanced medical datasets. This enhances OCSCC diagnostic accuracy, streamlines clinical decisions, and reduces reliance on costly histopathological tests.","doi":"10.1007\/s13198-024-02440-6","cleaned_title":"addressing data imbalance challenges in oral cavity histopathological whole slide images with advanced deep learning techniques","cleaned_abstract":"oral cavity squamous cell carcinoma (ocscc) represents a common form of head and neck cancer originating from the mucosal lining of the oral cavity, often detected in advanced stages. traditional detection methods rely on analyzing hematoxylin and eosin (h&e)-stained histopathological whole-slide images, which are time-consuming and require expert pathology skills. hence, automated analysis is urgently needed to expedite diagnosis and improve patient outcomes. deep learning, through automated feature extraction, offers a promising avenue for capturing high-level abstract features with greater accuracy than traditional methods. however, the imbalance in class distribution within datasets significantly affects the performance of deep learning models during training, necessitating specialized approaches. to address the issue, various methods have been proposed at both data and algorithmic levels. this study investigates strategies to mitigate class imbalance by employing a publicly available ocscc imbalance dataset. we evaluated undersampling methods (near miss, edited nearest neighbors) and oversampling techniques (smote, deep smote, adasyn) integrated with transfer learning across different imbalance ratios (0.1, 0.15, 0.20, 0.30). our findings demonstrate the effectiveness of smote in improving test performance, highlighting the efficacy of strategic oversampling combined with transfer learning in classifying imbalanced medical datasets. this enhances ocscc diagnostic accuracy, streamlines clinical decisions, and reduces reliance on costly histopathological tests.","key_phrases":["oral cavity squamous cell carcinoma","ocscc","a common form","head and neck cancer","the mucosal lining","the oral cavity","advanced stages","traditional detection methods","hematoxylin","eosin","histopathological whole-slide images","which","expert pathology skills","automated analysis","diagnosis","patient outcomes","deep learning","automated feature extraction","a promising avenue","high-level abstract features","greater accuracy","traditional methods","the imbalance","class distribution","datasets","the performance","deep learning models","training","specialized approaches","the issue","various methods","both data","algorithmic levels","this study","strategies","class imbalance","a publicly available ocscc imbalance dataset","we","undersampling methods","near miss","neighbors","techniques","smote","deep smote","adasyn","different imbalance ratios","our findings","the effectiveness","smote","test performance","the efficacy","strategic oversampling","imbalanced medical datasets","diagnostic accuracy","clinical decisions","reliance","costly histopathological tests","hematoxylin","0.1","0.15","0.20","0.30"]},{"title":"Heat-vision based drone surveillance augmented by deep learning for critical industrial monitoring","authors":["Do Yeong Lim","Ik Jae Jin","In Cheol Bang"],"abstract":"This study examines the application of drone-assisted infrared (IR) imaging with vision\u00a0grayscale imaging and deep learning for enhanced abnormal detection in nuclear power plants. A scaled model, replicating the modern pressurized water reactor, facilitated the data collection for normal and abnormal conditions. A drone, equipped with dual vision and IR cameras, captured detailed operational imagery, crucial for detecting subtle anomalies within the plant's primary systems. Deep learning algorithms were deployed to interpret these images, aiming to identify component abnormals not easily discernible by traditional monitoring. The object detection model was trained to classify normal and abnormal component states within the facility, marked by color-coded bounding boxes for clarity. Models like YOLO and Mask R-CNN were evaluated for their precision in anomaly detection. Results indicated that the YOLO v8m model was particularly effective, showcasing high accuracy in both detecting and adapting to system anomalies, as validated by high mAP scores. The integration of drone technology with IR imaging and deep learning illustrates a significant stride toward automating abnormal detection in complex industrial environments, enhancing operational safety and efficiency. This approach has the potential to revolutionize real-time monitoring in safety\u2013critical settings by providing a comprehensive, automated solution to abnormal detection.","doi":"10.1038\/s41598-023-49589-x","cleaned_title":"heat-vision based drone surveillance augmented by deep learning for critical industrial monitoring","cleaned_abstract":"this study examines the application of drone-assisted infrared (ir) imaging with vision grayscale imaging and deep learning for enhanced abnormal detection in nuclear power plants. a scaled model, replicating the modern pressurized water reactor, facilitated the data collection for normal and abnormal conditions. a drone, equipped with dual vision and ir cameras, captured detailed operational imagery, crucial for detecting subtle anomalies within the plant's primary systems. deep learning algorithms were deployed to interpret these images, aiming to identify component abnormals not easily discernible by traditional monitoring. the object detection model was trained to classify normal and abnormal component states within the facility, marked by color-coded bounding boxes for clarity. models like yolo and mask r-cnn were evaluated for their precision in anomaly detection. results indicated that the yolo v8m model was particularly effective, showcasing high accuracy in both detecting and adapting to system anomalies, as validated by high map scores. the integration of drone technology with ir imaging and deep learning illustrates a significant stride toward automating abnormal detection in complex industrial environments, enhancing operational safety and efficiency. this approach has the potential to revolutionize real-time monitoring in safety\u2013critical settings by providing a comprehensive, automated solution to abnormal detection.","key_phrases":["this study","the application","vision grayscale imaging","deep learning","enhanced abnormal detection","nuclear power plants","a scaled model","the modern pressurized water reactor","the data collection","normal and abnormal conditions","a drone","dual vision","ir cameras","detailed operational imagery","subtle anomalies","the plant's primary systems","deep learning algorithms","these images","component abnormals","traditional monitoring","the object detection model","normal and abnormal component states","the facility","color-coded bounding boxes","clarity","models","yolo","mask r-cnn","their precision","anomaly detection","results","the yolo v8m model","high accuracy","system anomalies","high map scores","the integration","drone technology","ir imaging","deep learning","a significant stride","abnormal detection","complex industrial environments","operational safety","efficiency","this approach","the potential","real-time monitoring","safety","critical settings","a comprehensive, automated solution","abnormal detection","anomaly detection"]},{"title":"COVID-19 Fake News Detection using Deep Learning Model","authors":["Mahabuba Akhter","Syed Md. Minhaz Hossain","Rizma Sijana Nigar","Srabanti Paul","Khaleque Md. Aashiq Kamal","Anik Sen","Iqbal H. Sarker"],"abstract":"People may now receive and share information more quickly and easily than ever due to the widespread use of mobile networked devices. However, this can occasionally lead to the spread of false information. Such information is being disseminated widely, which may cause people to make incorrect decisions about potentially crucial topics. This occurred in 2020, the year of the fatal and extremely contagious Coronavirus Disease (COVID-19) outbreak. The spread of false information about COVID-19 on social media has already been labeled as an \u201cinfodemic\u201d by the World Health Organization (WHO), causing serious difficulties for governments attempting to control the pandemic. Consequently, it is crucial to have a model for detecting fake news related to COVID-19. In this paper, we present an effective Convolutional Neural Network (CNN)-based deep learning model using word embedding. For selecting the best CNN architecture, we take into account the optimal values of model hyper-parameters using grid search. Further, for measuring the effectiveness of our proposed CNN model, various state-of-the-art machine learning algorithms are conducted for COVID-19 fake news detection. Among them, CNN outperforms with 96.19% mean accuracy, 95% mean F1-score, and 0.985 area under ROC curve (AUC).","doi":"10.1007\/s40745-023-00507-y","cleaned_title":"covid-19 fake news detection using deep learning model","cleaned_abstract":"people may now receive and share information more quickly and easily than ever due to the widespread use of mobile networked devices. however, this can occasionally lead to the spread of false information. such information is being disseminated widely, which may cause people to make incorrect decisions about potentially crucial topics. this occurred in 2020, the year of the fatal and extremely contagious coronavirus disease (covid-19) outbreak. the spread of false information about covid-19 on social media has already been labeled as an \u201cinfodemic\u201d by the world health organization (who), causing serious difficulties for governments attempting to control the pandemic. consequently, it is crucial to have a model for detecting fake news related to covid-19. in this paper, we present an effective convolutional neural network (cnn)-based deep learning model using word embedding. for selecting the best cnn architecture, we take into account the optimal values of model hyper-parameters using grid search. further, for measuring the effectiveness of our proposed cnn model, various state-of-the-art machine learning algorithms are conducted for covid-19 fake news detection. among them, cnn outperforms with 96.19% mean accuracy, 95% mean f1-score, and 0.985 area under roc curve (auc).","key_phrases":["people","information","the widespread use","mobile networked devices","this","the spread","false information","such information","which","people","incorrect decisions","potentially crucial topics","this","the year","the fatal and extremely contagious coronavirus disease","covid-19","the spread","false information","covid-19","social media","the world health organization","who","serious difficulties","governments","it","a model","fake news","covid-19","this paper","we","an effective convolutional neural network","cnn)-based deep learning model","word","the best cnn architecture","we","account","the optimal values","model hyper-parameters","grid search","the effectiveness","our proposed cnn model","the-art","covid-19 fake news detection","them","cnn","96.19%","mean accuracy","95%","0.985 area","roc curve","auc","2020","the year","covid-19","covid-19","the world health organization","covid-19","cnn","cnn","covid-19","cnn","96.19%","95%","0.985","roc"]},{"title":"Deep learning-based magnetic resonance image super-resolution: a survey","authors":["Zexin Ji","Beiji Zou","Xiaoyan Kui","Jun Liu","Wei Zhao","Chengzhang Zhu","Peishan Dai","Yulan Dai"],"abstract":"Magnetic resonance imaging (MRI) is a medical imaging technique used to show anatomical structures and physiological processes of the human body. Due to limitations like image acquisition time, hardware capabilities, or uncooperative patients, the resolution of MR images is insufficient.  Super-resolution (SR) is a crucial method to enhance the resolution of images without expensive scanning equipment. Recent years have witnessed significant progress in MR image super-resolution. Therefore, this survey presents a thorough overview of current developments in deep learning-based MR image super-resolution methods. In general, we can roughly divide the MRI super-resolution methods into single-contrast MR image SR methods and multi-contrast MR image SR methods. Additionally, we introduce the multi-task learning approaches about the MR image super-resolution. We also summarize other crucial topics, such as the degradation model, the definition of the super-resolution problem, the dataset, loss functions, and image quality assessment. Lastly, we indicate the challenges in the field of super-resolution and draw a conclusion to our survey.","doi":"10.1007\/s00521-024-09890-w","cleaned_title":"deep learning-based magnetic resonance image super-resolution: a survey","cleaned_abstract":"magnetic resonance imaging (mri) is a medical imaging technique used to show anatomical structures and physiological processes of the human body. due to limitations like image acquisition time, hardware capabilities, or uncooperative patients, the resolution of mr images is insufficient. super-resolution (sr) is a crucial method to enhance the resolution of images without expensive scanning equipment. recent years have witnessed significant progress in mr image super-resolution. therefore, this survey presents a thorough overview of current developments in deep learning-based mr image super-resolution methods. in general, we can roughly divide the mri super-resolution methods into single-contrast mr image sr methods and multi-contrast mr image sr methods. additionally, we introduce the multi-task learning approaches about the mr image super-resolution. we also summarize other crucial topics, such as the degradation model, the definition of the super-resolution problem, the dataset, loss functions, and image quality assessment. lastly, we indicate the challenges in the field of super-resolution and draw a conclusion to our survey.","key_phrases":["magnetic resonance imaging","mri","a medical imaging technique","anatomical structures","physiological processes","the human body","limitations","image acquisition time","hardware capabilities","uncooperative patients","the resolution","mr images","-","resolution","sr","a crucial method","the resolution","images","expensive scanning equipment","recent years","significant progress","mr image","super","-","resolution","this survey","a thorough overview","current developments","deep learning-based mr image super-resolution methods","we","the mri super-resolution methods","single-contrast mr image sr methods","-","contrast mr image sr methods","we","the multi-task learning approaches","the mr image","super","-","resolution","we","other crucial topics","the degradation model","the definition","the super-resolution problem","the dataset","loss functions","image quality assessment","we","the challenges","the field","super","-","resolution","a conclusion","our survey","recent years"]},{"title":"sqFm: a novel adaptive optimization scheme for deep learning model","authors":["Shubhankar Bhakta","Utpal Nandi","Madhab Mondal","Kuheli Ray Mahapatra","Partha Chowdhuri","Pabitra Pal"],"abstract":"For deep model training, an optimization technique is required that minimizes loss and maximizes accuracy. The development of an effective optimization method is one of the most important study areas. The diffGrad optimization method uses gradient changes during optimization phases but does not update 2nd order moments based on 1st order moments, and the AngularGrad optimization method uses the angular value of the gradient, which necessitates additional calculation. Due to these factors, both of those approaches result in zigzag trajectories that take a long time and require additional calculations to attain a global minimum. To overcome those limitations, a novel adaptive deep learning optimization method based on square of first momentum (sqFm) has been proposed. By adjusting 2nd order moments depending on 1st order moments and changing step size according to the present gradient on the non-negative function, the suggested sqFm delivers a smoother trajectory and better image classification accuracy. The empirical research comparing the performance of the proposed sqFm with Adam, diffGrad, and AngularGrad applying non-convex functions demonstrates that the suggested method delivers the best convergence and parameter values. In comparison to SGD, Adam, diffGrad, RAdam, and AngularGrad(tan) using the Rosenbrock function, the proposed sqFm method can attain the global minima gradually with less overshoot. Additionally, it is demonstrated that the proposed sqFm gives consistently good classification accuracy when training CNN networks (ResNet16, ResNet50, VGG34, ResNet18, and DenseNet121) on the CIFAR10, CIFAR100, and MNIST datasets, in contrast to SGDM, diffGrad, Adam, AngularGrad(Cos), and AngularGrad(Tan). The proposed method also gives the best classification accuracy than SGD, Adam, AdaBelief, Yogi, RAdam, and AngularGrad using the ImageNet dataset on the ResNet18 network. Source code link: https:\/\/github.com\/UtpalNandi\/sqFm-A-novel-adaptive-optimization-scheme-for-deep-learning-model.","doi":"10.1007\/s12065-023-00897-1","cleaned_title":"sqfm: a novel adaptive optimization scheme for deep learning model","cleaned_abstract":"for deep model training, an optimization technique is required that minimizes loss and maximizes accuracy. the development of an effective optimization method is one of the most important study areas. the diffgrad optimization method uses gradient changes during optimization phases but does not update 2nd order moments based on 1st order moments, and the angulargrad optimization method uses the angular value of the gradient, which necessitates additional calculation. due to these factors, both of those approaches result in zigzag trajectories that take a long time and require additional calculations to attain a global minimum. to overcome those limitations, a novel adaptive deep learning optimization method based on square of first momentum (sqfm) has been proposed. by adjusting 2nd order moments depending on 1st order moments and changing step size according to the present gradient on the non-negative function, the suggested sqfm delivers a smoother trajectory and better image classification accuracy. the empirical research comparing the performance of the proposed sqfm with adam, diffgrad, and angulargrad applying non-convex functions demonstrates that the suggested method delivers the best convergence and parameter values. in comparison to sgd, adam, diffgrad, radam, and angulargrad(tan) using the rosenbrock function, the proposed sqfm method can attain the global minima gradually with less overshoot. additionally, it is demonstrated that the proposed sqfm gives consistently good classification accuracy when training cnn networks (resnet16, resnet50, vgg34, resnet18, and densenet121) on the cifar10, cifar100, and mnist datasets, in contrast to sgdm, diffgrad, adam, angulargrad(cos), and angulargrad(tan). the proposed method also gives the best classification accuracy than sgd, adam, adabelief, yogi, radam, and angulargrad using the imagenet dataset on the resnet18 network. source code link: https:\/\/github.com\/utpalnandi\/sqfm-a-novel-adaptive-optimization-scheme-for-deep-learning-model.","key_phrases":["deep model training","an optimization technique","the development","an effective optimization method","the most important study areas","the diffgrad optimization method","gradient changes","optimization phases","2nd order moments","1st order moments","the angulargrad optimization method","the angular value","the gradient","which","additional calculation","these factors","both","those approaches","zigzag trajectories","that","a long time","additional calculations","a global minimum","those limitations","a novel adaptive deep learning optimization method","square","first momentum","(sqfm","2nd order moments","1st order moments","step size","the present gradient","the non-negative function","the suggested sqfm","a smoother trajectory","better image classification accuracy","the empirical research","the performance","the proposed sqfm","adam","diffgrad","angulargrad","non-convex functions","the suggested method","the best convergence and parameter values","comparison","the rosenbrock function","the proposed sqfm method","the global minima","less overshoot","it","the proposed sqfm","consistently good classification accuracy","cnn networks","resnet16","resnet50","vgg34","resnet18","densenet121","the cifar10","cifar100","mnist datasets","contrast","diffgrad","adam","angulargrad(cos","angulargrad(tan","the proposed method","the best classification accuracy","sgd","adam","adabelief","yogi","radam","angulargrad","the imagenet","the resnet18 network","source code link","https:\/\/github.com\/utpalnandi\/sqfm-a-novel-adaptive-optimization-scheme-for-deep-learning-model","2nd","1st","first","2nd","1st","diffgrad","adam","diffgrad","angulargrad(tan","cnn","resnet16","resnet50","resnet18","cifar10","diffgrad","adam","angulargrad(cos","resnet18"]},{"title":"Deep source transfer learning for the estimation of internal brain dynamics using scalp EEG","authors":["Haitao Yu","Zhiwen Hu","Quanfa Zhao","Jing Liu"],"abstract":"Electroencephalography (EEG) provides high temporal resolution neural data for brain-computer interfacing via noninvasive electrophysiological recording. Estimating the internal brain activity by means of source imaging techniques can further improve the spatial resolution of EEG and enhance the reliability of neural decoding and brain-computer interaction. In this work, we propose a novel EEG data-driven source imaging scheme for precise and efficient estimation of macroscale spatiotemporal brain dynamics across thalamus and cortical regions with deep learning methods. A deep source imaging framework with a convolutional-recurrent neural network is designed to estimate the internal brain dynamics from high-density EEG recordings. Moreover, a brain model including 210 cortical regions and 16 thalamic nuclei is established based on human brain connectome to provide synthetic training data, which manifests intrinsic characteristics of underlying brain dynamics in spontaneous, stimulation-evoked, and pathological states. Transfer learning algorithm is further applied to the trained network to reduce the dynamical differences between synthetic and realistic EEG. Extensive experiments exhibit that the proposed deep-learning method can accurately estimate the spatial and temporal activity of brain sources and achieves superior performance compared to the state-of-the-art approaches. Moreover, the EEG data-driven source imaging framework is effective in the location of seizure onset zone in epilepsy and reconstruction of dynamical thalamocortical interactions during sensory processing of acupuncture stimulation, implying its applicability in brain-computer interfacing for neuroscience research and clinical applications.","doi":"10.1007\/s11571-024-10149-2","cleaned_title":"deep source transfer learning for the estimation of internal brain dynamics using scalp eeg","cleaned_abstract":"electroencephalography (eeg) provides high temporal resolution neural data for brain-computer interfacing via noninvasive electrophysiological recording. estimating the internal brain activity by means of source imaging techniques can further improve the spatial resolution of eeg and enhance the reliability of neural decoding and brain-computer interaction. in this work, we propose a novel eeg data-driven source imaging scheme for precise and efficient estimation of macroscale spatiotemporal brain dynamics across thalamus and cortical regions with deep learning methods. a deep source imaging framework with a convolutional-recurrent neural network is designed to estimate the internal brain dynamics from high-density eeg recordings. moreover, a brain model including 210 cortical regions and 16 thalamic nuclei is established based on human brain connectome to provide synthetic training data, which manifests intrinsic characteristics of underlying brain dynamics in spontaneous, stimulation-evoked, and pathological states. transfer learning algorithm is further applied to the trained network to reduce the dynamical differences between synthetic and realistic eeg. extensive experiments exhibit that the proposed deep-learning method can accurately estimate the spatial and temporal activity of brain sources and achieves superior performance compared to the state-of-the-art approaches. moreover, the eeg data-driven source imaging framework is effective in the location of seizure onset zone in epilepsy and reconstruction of dynamical thalamocortical interactions during sensory processing of acupuncture stimulation, implying its applicability in brain-computer interfacing for neuroscience research and clinical applications.","key_phrases":["electroencephalography","(eeg","high temporal resolution neural data","brain-computer","noninvasive electrophysiological recording","the internal brain activity","means","source","imaging techniques","the spatial resolution","eeg","the reliability","neural decoding and brain-computer interaction","this work","we","a novel eeg data-driven source","scheme","precise and efficient estimation","macroscale spatiotemporal brain dynamics","thalamus and cortical regions","deep learning methods","a deep source imaging framework","a convolutional-recurrent neural network","the internal brain dynamics","high-density eeg recordings","a brain model","210 cortical regions","16 thalamic nuclei","human brain connectome","synthetic training data","which","intrinsic characteristics","underlying brain dynamics","spontaneous, stimulation-evoked, and pathological states","transfer learning algorithm","the trained network","the dynamical differences","synthetic and realistic eeg","extensive experiments","the proposed deep-learning method","the spatial and temporal activity","brain sources","superior performance","the-art","the eeg data-driven source imaging framework","the location","seizure onset zone","epilepsy","reconstruction","dynamical thalamocortical interactions","sensory processing","acupuncture stimulation","its applicability","brain-computer","neuroscience research and clinical applications","electroencephalography","210","16"]},{"title":"New hybrid deep learning models for multi-target NILM disaggregation","authors":["Jamila Ouzine","Manal Marzouq","Saad Dosse Bennani","Khadija Lahrech","Hakim EL Fadili"],"abstract":"Non-Intrusive Load Monitoring (NILM) technique or energy disaggregation is a technique used to detect the appliance\u2019s states and estimate their individual energy consumption, given the aggregated data through the main smart meter. Indeed, energy efficiency is the main goal of the NILM techniques, which can be achieved by providing energy disaggregation feedback to the consumers. Unlike single models where training must be performed for each appliance, this work proposes multi-target disaggregation which is more appropriate due to the drastic reduction of resources when training is performed for all target appliances simultaneously. For this purpose, new hybrid models are proposed by combining well-known deep learning models: Convolutional Neural Network (CNN), Denoising Autoencoder (DAE), Recurrent Neural Network (RNN), and Long Short-Term Memory network (LSTM). An implementation and detailed comparative study is then suggested between the proposed hybrid deep learning models and conventional single models in terms of various performance metrics on the UK-Domestic Appliance-Level Electricity (UKDALE) benchmarking database. The experimental results show that the proposed hybrid models provide the best disaggregation performances for multi-target disaggregation compared to single models. Specifically, the CNN-LSTM and the DAE-LSTM are the best hybrid models with the highest overall F1-score of 78.90% and 72.94% respectively.","doi":"10.1007\/s12053-023-10161-1","cleaned_title":"new hybrid deep learning models for multi-target nilm disaggregation","cleaned_abstract":"non-intrusive load monitoring (nilm) technique or energy disaggregation is a technique used to detect the appliance\u2019s states and estimate their individual energy consumption, given the aggregated data through the main smart meter. indeed, energy efficiency is the main goal of the nilm techniques, which can be achieved by providing energy disaggregation feedback to the consumers. unlike single models where training must be performed for each appliance, this work proposes multi-target disaggregation which is more appropriate due to the drastic reduction of resources when training is performed for all target appliances simultaneously. for this purpose, new hybrid models are proposed by combining well-known deep learning models: convolutional neural network (cnn), denoising autoencoder (dae), recurrent neural network (rnn), and long short-term memory network (lstm). an implementation and detailed comparative study is then suggested between the proposed hybrid deep learning models and conventional single models in terms of various performance metrics on the uk-domestic appliance-level electricity (ukdale) benchmarking database. the experimental results show that the proposed hybrid models provide the best disaggregation performances for multi-target disaggregation compared to single models. specifically, the cnn-lstm and the dae-lstm are the best hybrid models with the highest overall f1-score of 78.90% and 72.94% respectively.","key_phrases":["non-intrusive load monitoring (nilm) technique or energy disaggregation","a technique","the appliance\u2019s states","their individual energy consumption","the aggregated data","the main smart meter","energy efficiency","the main goal","the nilm techniques","which","energy disaggregation feedback","the consumers","single models","training","each appliance","this work","multi-target disaggregation","which","the drastic reduction","resources","training","all target appliances","this purpose","new hybrid models","well-known deep learning models","convolutional neural network","cnn","autoencoder","dae","recurrent neural network","rnn","long short-term memory network","lstm","an implementation","detailed comparative study","the proposed hybrid deep learning models","conventional single models","terms","various performance metrics","the uk-domestic appliance-level electricity","(ukdale","benchmarking database","the experimental results","the proposed hybrid models","the best disaggregation performances","multi-target disaggregation","single models","the cnn-lstm","the dae-lstm","the best hybrid models","the highest overall f1-score","78.90%","72.94%","cnn","dae","uk","cnn","78.90%","72.94%"]},{"title":"Advancing brain tumor classification accuracy through deep learning: harnessing radimagenet pre-trained convolutional neural networks, ensemble learning, and machine learning classifiers on MRI brain images","authors":["Nihal Remzan","Karim Tahiry","Abdelmajid Farchi"],"abstract":"Brain tumors, a severe health concern across all age groups, present challenges for accurate grading in health monitoring and automated diagnosis. Choosing MRI scans for their superior quality and comprehensive anatomical insight, this study navigates the complexities of brain tumor classification. Deep learning (DL) algorithms revolutionize this field, empowering radiologists with enhanced diagnostic precision. Despite the pivotal role of extensive training data in DL success, medical image datasets often suffer from limitations in diversity and quantity. Transfer learning emerges as a solution, bridging gaps between similar yet distinct areas. The inadequacy of general pre-trained models on ImageNet for medical imaging characteristics prompts our recommendation for RadImageNet. This tailored solution aligns pre-trained models with the specific demands of medical imaging datasets, addressing the current mismatch. Our dataset includes 7023 MR images of Normal brain, Glioma, Meningioma, and Pituitary tumors. We employ two approaches: 'Feature Ensemble,' combining 2-top features, and 'Stacking Ensemble,' utilizing SVM, k-NN, Extra Trees, and MLP as base learners, and Logistic Regression (LR) as a meta-learner. In our initial approach with ResNet-50 and DenseNet121 features guided by MLP, we attained an impressive 97.71% accuracy and a remarkable Area Under the Curve (AUC) of 99.87%. Shifting to the Stacking Ensemble method, where DenseNet121 serves as the optimal feature extractor, we secured a commendable 97.40% accuracy, accompanied by an exceptional AUC of 99.83%. These outcomes highlight the efficacy of our ensemble learning strategies in enhancing both accuracy and the overall discriminative power of the model.","doi":"10.1007\/s11042-024-18780-1","cleaned_title":"advancing brain tumor classification accuracy through deep learning: harnessing radimagenet pre-trained convolutional neural networks, ensemble learning, and machine learning classifiers on mri brain images","cleaned_abstract":"brain tumors, a severe health concern across all age groups, present challenges for accurate grading in health monitoring and automated diagnosis. choosing mri scans for their superior quality and comprehensive anatomical insight, this study navigates the complexities of brain tumor classification. deep learning (dl) algorithms revolutionize this field, empowering radiologists with enhanced diagnostic precision. despite the pivotal role of extensive training data in dl success, medical image datasets often suffer from limitations in diversity and quantity. transfer learning emerges as a solution, bridging gaps between similar yet distinct areas. the inadequacy of general pre-trained models on imagenet for medical imaging characteristics prompts our recommendation for radimagenet. this tailored solution aligns pre-trained models with the specific demands of medical imaging datasets, addressing the current mismatch. our dataset includes 7023 mr images of normal brain, glioma, meningioma, and pituitary tumors. we employ two approaches: 'feature ensemble,' combining 2-top features, and 'stacking ensemble,' utilizing svm, k-nn, extra trees, and mlp as base learners, and logistic regression (lr) as a meta-learner. in our initial approach with resnet-50 and densenet121 features guided by mlp, we attained an impressive 97.71% accuracy and a remarkable area under the curve (auc) of 99.87%. shifting to the stacking ensemble method, where densenet121 serves as the optimal feature extractor, we secured a commendable 97.40% accuracy, accompanied by an exceptional auc of 99.83%. these outcomes highlight the efficacy of our ensemble learning strategies in enhancing both accuracy and the overall discriminative power of the model.","key_phrases":["brain tumors","a severe health concern","all age groups","present challenges","accurate grading","health monitoring","automated diagnosis","mri scans","their superior quality","comprehensive anatomical insight","this study","the complexities","brain tumor classification","deep learning (dl) algorithms","this field","radiologists","enhanced diagnostic precision","the pivotal role","extensive training data","dl success","medical image datasets","limitations","diversity","quantity","transfer learning","a solution","gaps","similar yet distinct areas","the inadequacy","general pre-trained models","imagenet","medical imaging characteristics","our recommendation","radimagenet","this tailored solution","pre-trained models","the specific demands","medical imaging datasets","the current mismatch","our dataset","mr images","normal brain","glioma","meningioma","pituitary tumors","we","two approaches","feature ensemble","2-top features","ensemble","svm","extra trees","mlp","base learners","logistic regression","lr","a meta-learner","our initial approach","resnet-50 and densenet121 features","mlp","we","an impressive 97.71% accuracy","a remarkable area","the curve","auc","99.87%","the stacking ensemble method","densenet121","the optimal feature extractor","we","a commendable 97.40% accuracy","an exceptional auc","99.83%","these outcomes","the efficacy","our ensemble learning strategies","both accuracy","the overall discriminative power","the model","7023","glioma","two","2","resnet-50","97.71%","99.87%","97.40%","99.83%"]},{"title":"Optimization Based Deep Learning for COVID-19 Detection Using Respiratory Sound Signals","authors":["Jawad Ahmad Dar","Kamal Kr Srivastava","Sajaad Ahmed Lone"],"abstract":"The COVID-19 prediction process is more indispensable to handle the spread and death occurred rate because of COVID-19. However, early and precise prediction of COVID-19 is more difficult, because of different sizes and resolutions of input image. Thus, these challenges and problems experienced by traditional COVID-19 detection methods are considered as major motivation to develop SJHBO-based Deep Q Network. The classification issue of respiratory sound has perceived a great focus from the clinical scientists as well as the community of medical researcher in the previous year for the identification of COVID-19 disease. The major contribution of this research is to design an effectual COVID-19 detection model using devised SJHBO-based Deep Q Network. In this paper, the COVID-19 detection is carried out by the deep learning with optimization technique, namely Snake Jaya Honey Badger Optimization (SJHBO) algorithm-driven Deep Q Network. Here, the SJHBO algorithm is the incorporation of Jaya Honey Badger Optimization (JHBO) along with Snake optimization (SO). Here, the COVID-19 is detected by the Deep Q Network wherein the weights of Deep Q Network are tuned by the SJHBO algorithm. Moreover, JHBO is modelled by hybrids, which are the Jaya algorithm and Honey Badger Optimization (HBO) algorithm. Furthermore, the features, such as spectral contrast, Mel frequency cepstral coefficients (MFCC), empirical mode decomposition (EMD) algorithm, spectral flux, fast Fourier transform (FFT), spectral roll-off, spectral centroid, zero-crossing rate, root mean square energy, spectral bandwidth, spectral flatness, power spectral density, mobility complexity, fluctuation index and relative amplitude, are mined for enlightening the detection performance. The developed method realized the better performance based on the accuracy, sensitivity and specificity of 0.9511, 0.9506 and 0.9469. All test results are validated with the k-fold cross validation method in order to make an assessment of the generalizability of these results. Statistical analysis is performed to analyze the performance of the proposed method based on testing accuracy, sensitivity and specificity. Hence, this paper presents the newly devised SJHBO-based Deep Q-Net for COVID-19 detection. This research considers the audio samples as an input, which is acquired from the Coswara dataset. The SJHBO-based Deep Q network approach is developed for COVID-19 detection. The developed approach can be extended by including other hybrid optimization algorithms as well as other features that can be extracted for further improving the detection performance. The proposed COVID-19 detection method is useful in various applications, like medical and so on. Developed SJHBO-enabled Deep Q network for COVID-19 detection: An effective COVID-19 detection technique is introduced based on hybrid optimization\u2013driven deep learning model. The Deep Q Network is used for detecting COVID-19, which classifies the feature vector as COVID-19 or non-COVID-19. Moreover, the Deep Q Network is trained by devised SJHBO approach, which is the incorporation of Jaya Honey Badger Optimization (JHBO) along with Snake optimization (SO).","doi":"10.1007\/s12559-024-10300-5","cleaned_title":"optimization based deep learning for covid-19 detection using respiratory sound signals","cleaned_abstract":"the covid-19 prediction process is more indispensable to handle the spread and death occurred rate because of covid-19. however, early and precise prediction of covid-19 is more difficult, because of different sizes and resolutions of input image. thus, these challenges and problems experienced by traditional covid-19 detection methods are considered as major motivation to develop sjhbo-based deep q network. the classification issue of respiratory sound has perceived a great focus from the clinical scientists as well as the community of medical researcher in the previous year for the identification of covid-19 disease. the major contribution of this research is to design an effectual covid-19 detection model using devised sjhbo-based deep q network. in this paper, the covid-19 detection is carried out by the deep learning with optimization technique, namely snake jaya honey badger optimization (sjhbo) algorithm-driven deep q network. here, the sjhbo algorithm is the incorporation of jaya honey badger optimization (jhbo) along with snake optimization (so). here, the covid-19 is detected by the deep q network wherein the weights of deep q network are tuned by the sjhbo algorithm. moreover, jhbo is modelled by hybrids, which are the jaya algorithm and honey badger optimization (hbo) algorithm. furthermore, the features, such as spectral contrast, mel frequency cepstral coefficients (mfcc), empirical mode decomposition (emd) algorithm, spectral flux, fast fourier transform (fft), spectral roll-off, spectral centroid, zero-crossing rate, root mean square energy, spectral bandwidth, spectral flatness, power spectral density, mobility complexity, fluctuation index and relative amplitude, are mined for enlightening the detection performance. the developed method realized the better performance based on the accuracy, sensitivity and specificity of 0.9511, 0.9506 and 0.9469. all test results are validated with the k-fold cross validation method in order to make an assessment of the generalizability of these results. statistical analysis is performed to analyze the performance of the proposed method based on testing accuracy, sensitivity and specificity. hence, this paper presents the newly devised sjhbo-based deep q-net for covid-19 detection. this research considers the audio samples as an input, which is acquired from the coswara dataset. the sjhbo-based deep q network approach is developed for covid-19 detection. the developed approach can be extended by including other hybrid optimization algorithms as well as other features that can be extracted for further improving the detection performance. the proposed covid-19 detection method is useful in various applications, like medical and so on. developed sjhbo-enabled deep q network for covid-19 detection: an effective covid-19 detection technique is introduced based on hybrid optimization\u2013driven deep learning model. the deep q network is used for detecting covid-19, which classifies the feature vector as covid-19 or non-covid-19. moreover, the deep q network is trained by devised sjhbo approach, which is the incorporation of jaya honey badger optimization (jhbo) along with snake optimization (so).","key_phrases":["the covid-19 prediction process","the spread","death","occurred rate","covid-19","early and precise prediction","covid-19","different sizes","resolutions","input image","these challenges","problems","traditional covid-19 detection methods","major motivation","sjhbo-based deep q network","the classification issue","respiratory sound","a great focus","the clinical scientists","the community","medical researcher","the previous year","the identification","covid-19 disease","the major contribution","this research","an effectual covid-19 detection model","devised sjhbo-based deep q network","this paper","the covid-19 detection","the deep learning","optimization technique","namely snake jaya honey badger optimization","sjhbo","algorithm-driven deep q network","the sjhbo algorithm","the incorporation","jaya honey badger optimization","jhbo","snake optimization","the covid-19","the deep q network","the weights","deep q network","the sjhbo algorithm","jhbo","hybrids","which","the jaya algorithm","honey badger optimization","hbo","algorithm","the features","spectral contrast","empirical mode decomposition","emd","algorithm","spectral flux","fast fourier transform","fft","spectral roll-off","spectral centroid","zero-crossing rate","root mean square energy","spectral flatness","power spectral density","mobility complexity","fluctuation index","relative amplitude","the detection performance","the developed method","the better performance","the accuracy","sensitivity","specificity","all test results","the k-fold cross validation method","order","an assessment","the generalizability","these results","statistical analysis","the performance","the proposed method","testing accuracy","sensitivity","specificity","this paper","the newly devised sjhbo-based deep q-net","covid-19 detection","this research","the audio samples","an input","which","the coswara dataset","the sjhbo-based deep q network approach","covid-19 detection","the developed approach","other features","that","the detection performance","the proposed covid-19 detection method","various applications","developed sjhbo-enabled deep q network","covid-19 detection","an effective covid-19 detection technique","hybrid optimization","driven deep learning model","the deep q network","covid-19","which","the feature vector","covid-19","non","-","covid-19","the deep q network","devised sjhbo approach","which","the incorporation","jaya honey badger optimization","jhbo","snake optimization","covid-19","covid-19","covid-19","covid-19","the previous year","covid-19","covid-19","covid-19","jaya honey badger optimization","covid-19","jhbo","mel","zero","0.9511","0.9506","0.9469","the k-fold","covid-19","coswara","covid-19","covid-19","covid-19","covid-19","covid-19","covid-19","non-covid-19","jaya honey badger optimization"]},{"title":"Image Analysis of the Automatic Welding Defects Detection Based on Deep Learning","authors":["Xiaopeng Wang","Baoxin Zhang","Jinhan Cui","Juntao Wu","Yan Li","Jinhang Li","Yunhua Tan","Xiaoming Chen","Wenliang Wu","Xinghua Yu"],"abstract":"Automatic detection of welding flaws based on deep learning methods has aroused great interest in the non-destructive testing. However, few studies focus on the characteristics of welding flaws in the X-ray image. This study uses four deep learning models to train and test on a dataset containing 15,194 X-ray images. A hybrid prediction based on OR logic is proposed to avoid the miss detection as much as possible and reduce the miss detection rate to 0.61%, which is state of the art. Quantitative analysis of flaws\u2019 characteristics, including the area, aspect ratio, mean, and variance, suggests the aspect ratios of miss detected flaws are smaller than 2, and the coefficient variances of miss detected flaws are smaller than 0.2. Tracking the critical pixels of X-ray images show that salt noises lead to false alarmed predictions. Error analysis indicates that when using the deep learning method for automatic welding flaws detection, the characteristics of flaws and the factors caused by inappropriate X-ray exposure techniques also should be noted.","doi":"10.1007\/s10921-023-00992-4","cleaned_title":"image analysis of the automatic welding defects detection based on deep learning","cleaned_abstract":"automatic detection of welding flaws based on deep learning methods has aroused great interest in the non-destructive testing. however, few studies focus on the characteristics of welding flaws in the x-ray image. this study uses four deep learning models to train and test on a dataset containing 15,194 x-ray images. a hybrid prediction based on or logic is proposed to avoid the miss detection as much as possible and reduce the miss detection rate to 0.61%, which is state of the art. quantitative analysis of flaws\u2019 characteristics, including the area, aspect ratio, mean, and variance, suggests the aspect ratios of miss detected flaws are smaller than 2, and the coefficient variances of miss detected flaws are smaller than 0.2. tracking the critical pixels of x-ray images show that salt noises lead to false alarmed predictions. error analysis indicates that when using the deep learning method for automatic welding flaws detection, the characteristics of flaws and the factors caused by inappropriate x-ray exposure techniques also should be noted.","key_phrases":["automatic detection","welding flaws","deep learning methods","great interest","the non-destructive testing","few studies","the characteristics","welding flaws","the x-ray image","this study","four deep learning models","a dataset","15,194 x-ray images","a hybrid prediction","logic","the miss detection","the miss detection rate","0.61%","which","state","the art","quantitative analysis","flaws\u2019 characteristics","the area","aspect ratio","variance","the aspect ratios","miss detected flaws","the coefficient variances","miss detected flaws","the critical pixels","x-ray images","salt noises","false alarmed predictions","error analysis","the deep learning method","automatic welding flaws detection","the characteristics","flaws","the factors","inappropriate x-ray exposure techniques","four","15,194","0.61%","2","0.2"]},{"title":"Ensemble Deep Learning Approach for Turbidity Prediction of Dooskal Lake Using Remote Sensing Data","authors":["Janjhyam Venkata Naga Ramesh","Pavithra Roy Patibandla","Manjula Shanbhog","Srinivas Ambala","Mohd Ashraf","Ajmeera Kiran"],"abstract":"The summer season in India is marked by a severe shortage of water, which poses significant challenges for daily usage and agricultural practices. With unpredictable weather patterns and irregular rainfall, it is crucial to monitor and maintain water bodies such as domestic ponds and lakes in urban areas to ensure they provide clean and safe water for regular use, free from industrial pollutants. In this research paper, we propose an innovative ensemble deep learning approach (e-DLA) that leverages deep learning models to predict the turbidity of Dooskal Lake, located in Telangana, India, using remote sensing data. The proposed approach utilizes various deep learning models, including bagging, boosting, and stacking, to analyze the complex relationships between remote sensing data and turbidity levels in the lake. The study aims to provide accurate and efficient predictions of turbidity levels, which can aid in the management and conservation of water resources in the region. Hyperparameter tuning is employed, and dynamic climatic features are extracted and integrated with the ensemble learning global protective intelligent algorithm to reveal the complex relationship between in situ and measured values of turbidity during the measuring timeline. The proposed approach provides accurate predictions of turbidity levels, enabling the implementation of effective control measures to maintain water quality standards. Experimental results demonstrate that the proposed approach significantly reduces prediction errors compared to existing deep learning models. Overall, this research highlights the potential of machine learning techniques in monitoring and maintaining water resources, particularly in urban areas, to support sustainable water management and usage, and addresses an urgent and pressing issue in India and around the world.","doi":"10.1007\/s41976-023-00098-5","cleaned_title":"ensemble deep learning approach for turbidity prediction of dooskal lake using remote sensing data","cleaned_abstract":"the summer season in india is marked by a severe shortage of water, which poses significant challenges for daily usage and agricultural practices. with unpredictable weather patterns and irregular rainfall, it is crucial to monitor and maintain water bodies such as domestic ponds and lakes in urban areas to ensure they provide clean and safe water for regular use, free from industrial pollutants. in this research paper, we propose an innovative ensemble deep learning approach (e-dla) that leverages deep learning models to predict the turbidity of dooskal lake, located in telangana, india, using remote sensing data. the proposed approach utilizes various deep learning models, including bagging, boosting, and stacking, to analyze the complex relationships between remote sensing data and turbidity levels in the lake. the study aims to provide accurate and efficient predictions of turbidity levels, which can aid in the management and conservation of water resources in the region. hyperparameter tuning is employed, and dynamic climatic features are extracted and integrated with the ensemble learning global protective intelligent algorithm to reveal the complex relationship between in situ and measured values of turbidity during the measuring timeline. the proposed approach provides accurate predictions of turbidity levels, enabling the implementation of effective control measures to maintain water quality standards. experimental results demonstrate that the proposed approach significantly reduces prediction errors compared to existing deep learning models. overall, this research highlights the potential of machine learning techniques in monitoring and maintaining water resources, particularly in urban areas, to support sustainable water management and usage, and addresses an urgent and pressing issue in india and around the world.","key_phrases":["the summer season","india","a severe shortage","water","which","significant challenges","daily usage","agricultural practices","unpredictable weather patterns","irregular rainfall","it","water bodies","domestic ponds","lakes","urban areas","they","clean and safe water","regular use","industrial pollutants","this research paper","we","an innovative ensemble deep learning approach","e","-","dla","that","deep learning models","the turbidity","dooskal lake","telangana","india","remote sensing data","the proposed approach","various deep learning models","bagging","stacking","the complex relationships","remote sensing data","turbidity levels","the lake","the study","accurate and efficient predictions","turbidity levels","which","the management","conservation","water resources","the region","hyperparameter tuning","dynamic climatic features","global protective intelligent algorithm","the complex relationship","situ","measured values","turbidity","the measuring timeline","the proposed approach","accurate predictions","turbidity levels","the implementation","effective control measures","water quality standards","experimental results","the proposed approach","prediction errors","existing deep learning models","this research","the potential","machine learning techniques","monitoring","water resources","urban areas","sustainable water management","usage","an urgent and pressing issue","india","the world","the summer season","india","daily","telangana","india","india"]},{"title":"Analyzing tourism reviews using Deep Learning and AI to predict sentiments","authors":["Piergiorgio Marigliano"],"abstract":"In this study, we investigate the application of Artificial Intelligence (AI), specifically through Deep Learning and neural networks, in analyzing and predicting sentiments expressed in tourism reviews. Our dataset comprised various hotel reviews, with the objective to predict whether each textual review indicates positive or negative feedback. The primary challenge was to use solely the textual data of the reviews for this prediction. Through meticulous data processing and analysis, we developed neural network-based models that highlight the efficacy of Deep Learning in the accurate interpretation of reviews. Our findings reveal a significant correlation between the content of reviews and their overall ratings, thereby providing new insights into the application of AI in automating and enhancing understanding of customer needs and perceptions in the tourism sector. The principal contribution of this study is the practical demonstration of how AI techniques can be effectively employed to analyze large volumes of textual data, opening new avenues for marketing strategies and service optimization in the hospitality industry. Each review represents a client\u2019s assessment of a hotel. For each textual review, we aim to predict whether it corresponds to a positive review (the customer is satisfied) or a negative review (the customer is dissatisfied). The overall ratings of the reviews can range from 2.5\/10 to 10\/10. To simplify the issue, we\u2019ll categorize them as follows: negative reviews have overall ratings of less than 5; positive reviews have ratings of 5 or higher. The challenge lies in predicting this information using only the raw textual data of the review.","doi":"10.1007\/s11135-024-01840-x","cleaned_title":"analyzing tourism reviews using deep learning and ai to predict sentiments","cleaned_abstract":"in this study, we investigate the application of artificial intelligence (ai), specifically through deep learning and neural networks, in analyzing and predicting sentiments expressed in tourism reviews. our dataset comprised various hotel reviews, with the objective to predict whether each textual review indicates positive or negative feedback. the primary challenge was to use solely the textual data of the reviews for this prediction. through meticulous data processing and analysis, we developed neural network-based models that highlight the efficacy of deep learning in the accurate interpretation of reviews. our findings reveal a significant correlation between the content of reviews and their overall ratings, thereby providing new insights into the application of ai in automating and enhancing understanding of customer needs and perceptions in the tourism sector. the principal contribution of this study is the practical demonstration of how ai techniques can be effectively employed to analyze large volumes of textual data, opening new avenues for marketing strategies and service optimization in the hospitality industry. each review represents a client\u2019s assessment of a hotel. for each textual review, we aim to predict whether it corresponds to a positive review (the customer is satisfied) or a negative review (the customer is dissatisfied). the overall ratings of the reviews can range from 2.5\/10 to 10\/10. to simplify the issue, we\u2019ll categorize them as follows: negative reviews have overall ratings of less than 5; positive reviews have ratings of 5 or higher. the challenge lies in predicting this information using only the raw textual data of the review.","key_phrases":["this study","we","the application","artificial intelligence","(ai","deep learning","neural networks","sentiments","tourism reviews","our dataset","various hotel reviews","the objective","each textual review","positive or negative feedback","the primary challenge","the textual data","the reviews","this prediction","meticulous data processing","analysis","we","neural network-based models","that","the efficacy","deep learning","the accurate interpretation","reviews","our findings","a significant correlation","the content","reviews","their overall ratings","new insights","the application","ai","understanding","customer needs","perceptions","the tourism sector","the principal contribution","this study","the practical demonstration","techniques","large volumes","textual data","new avenues","marketing strategies","service optimization","the hospitality industry","each review","a client\u2019s assessment","a hotel","each textual review","we","it","a positive review","the customer","a negative review","the customer","the overall ratings","the reviews","the issue","we","them","negative reviews","overall ratings","positive reviews","ratings","the challenge","this information","only the raw textual data","the review","2.5\/10","10\/10","less than 5","5"]},{"title":"Automated assembly quality inspection by deep learning with 2D and 3D synthetic CAD data","authors":["Xiaomeng Zhu","P\u00e4r M\u00e5rtensson","Lars Hanson","M\u00e5rten Bj\u00f6rkman","Atsuto Maki"],"abstract":"In the manufacturing industry, automatic quality inspections can lead to improved product quality and productivity. Deep learning-based computer vision technologies, with their superior performance in many applications, can be a possible solution for automatic quality inspections. However, collecting a large amount of annotated training data for deep learning is expensive and time-consuming, especially for processes involving various products and human activities such as assembly. To address this challenge, we propose a method for automated assembly quality inspection using synthetic data generated from computer-aided design (CAD) models. The method involves two steps: automatic data generation and model implementation. In the first step, we generate synthetic data in two formats: two-dimensional (2D) images and three-dimensional (3D) point clouds. In the second step, we apply different state-of-the-art deep learning approaches to the data for quality inspection, including unsupervised domain adaptation, i.e., a method of adapting models across different data distributions, and transfer learning, which transfers knowledge between related tasks. We evaluate the methods in a case study of pedal car front-wheel assembly quality inspection to identify the possible optimal approach for assembly quality inspection. Our results show that the method using Transfer Learning on 2D synthetic images achieves superior performance compared with others. Specifically, it attained 95% accuracy through fine-tuning with only five annotated real images per class. With promising results, our method may be suggested for other similar quality inspection use cases. By utilizing synthetic CAD data, our method reduces the need for manual data collection and annotation. Furthermore, our method performs well on test data with different backgrounds, making it suitable for different manufacturing environments.","doi":"10.1007\/s10845-024-02375-6","cleaned_title":"automated assembly quality inspection by deep learning with 2d and 3d synthetic cad data","cleaned_abstract":"in the manufacturing industry, automatic quality inspections can lead to improved product quality and productivity. deep learning-based computer vision technologies, with their superior performance in many applications, can be a possible solution for automatic quality inspections. however, collecting a large amount of annotated training data for deep learning is expensive and time-consuming, especially for processes involving various products and human activities such as assembly. to address this challenge, we propose a method for automated assembly quality inspection using synthetic data generated from computer-aided design (cad) models. the method involves two steps: automatic data generation and model implementation. in the first step, we generate synthetic data in two formats: two-dimensional (2d) images and three-dimensional (3d) point clouds. in the second step, we apply different state-of-the-art deep learning approaches to the data for quality inspection, including unsupervised domain adaptation, i.e., a method of adapting models across different data distributions, and transfer learning, which transfers knowledge between related tasks. we evaluate the methods in a case study of pedal car front-wheel assembly quality inspection to identify the possible optimal approach for assembly quality inspection. our results show that the method using transfer learning on 2d synthetic images achieves superior performance compared with others. specifically, it attained 95% accuracy through fine-tuning with only five annotated real images per class. with promising results, our method may be suggested for other similar quality inspection use cases. by utilizing synthetic cad data, our method reduces the need for manual data collection and annotation. furthermore, our method performs well on test data with different backgrounds, making it suitable for different manufacturing environments.","key_phrases":["the manufacturing industry","automatic quality inspections","improved product quality","productivity","deep learning-based computer vision technologies","their superior performance","many applications","a possible solution","automatic quality inspections","a large amount","annotated training data","deep learning","processes","various products","human activities","assembly","this challenge","we","a method","automated assembly quality inspection","synthetic data","computer-aided design (cad) models","the method","two steps","automatic data generation and model implementation","the first step","we","synthetic data","two formats","two-dimensional (2d) images","three-dimensional (3d","the second step","we","the-art","the data","quality inspection","unsupervised domain adaptation","i.e., a method","adapting models","different data distributions","transfer learning","which","knowledge","related tasks","we","the methods","a case study","pedal car front-wheel assembly quality inspection","the possible optimal approach","assembly quality inspection","our results","the method","transfer learning","2d synthetic images","superior performance","others","it","95% accuracy","fine-tuning","only five annotated real images","class","promising results","our method","other similar quality inspection use cases","synthetic cad data","our method","the need","manual data collection","annotation","our method","test data","different backgrounds","it","different manufacturing environments","two","first","two","two","2d","three","3d","second","2d","95%","only five"]},{"title":"ConIQA: A deep learning method for perceptual image quality assessment with limited data","authors":["M. Hossein Eybposh","Changjia Cai","Aram Moossavi","Jose Rodriguez-Romaguera","Nicolas C. P\u00e9gard"],"abstract":"Effectively assessing the realism and naturalness of images in virtual (VR) and augmented (AR) reality applications requires Full Reference Image Quality Assessment (FR-IQA) metrics that closely align with human perception. Deep learning-based IQAs that are trained on human-labeled data have recently shown promise in generic computer vision tasks. However, their performance decreases in applications where perfect matches between the reference and the distorted images should not be expected, or whenever distortion patterns are restricted to specific domains. Tackling this issue necessitates training a task-specific neural network, yet generating human-labeled FR-IQAs is costly, and deep learning typically demands substantial labeled data. To address these challenges, we developed ConIQA, a deep learning-based IQA that leverages consistency training and a novel data augmentation method to learn from both labeled and unlabeled data. This makes ConIQA well-suited for contexts with scarce labeled data. To validate ConIQA, we considered the example application of Computer-Generated Holography (CGH) where specific artifacts such as ringing, speckle, and quantization errors routinely occur, yet are not explicitly accounted for by existing IQAs. We developed a new dataset, HQA1k, that comprises 1000 natural images each paired with an image rendered using various popular CGH algorithms, and quality-rated by thirteen human participants. Our results show that ConIQA achieves superior Pearson (0.98), Spearman (0.965), and Kendall\u2019s tau (0.86) correlations over fifteen FR-IQA metrics by up to 5%, showcasing significant improvements in aligning with human perception on the HQA1k dataset.","doi":"10.1038\/s41598-024-70469-5","cleaned_title":"coniqa: a deep learning method for perceptual image quality assessment with limited data","cleaned_abstract":"effectively assessing the realism and naturalness of images in virtual (vr) and augmented (ar) reality applications requires full reference image quality assessment (fr-iqa) metrics that closely align with human perception. deep learning-based iqas that are trained on human-labeled data have recently shown promise in generic computer vision tasks. however, their performance decreases in applications where perfect matches between the reference and the distorted images should not be expected, or whenever distortion patterns are restricted to specific domains. tackling this issue necessitates training a task-specific neural network, yet generating human-labeled fr-iqas is costly, and deep learning typically demands substantial labeled data. to address these challenges, we developed coniqa, a deep learning-based iqa that leverages consistency training and a novel data augmentation method to learn from both labeled and unlabeled data. this makes coniqa well-suited for contexts with scarce labeled data. to validate coniqa, we considered the example application of computer-generated holography (cgh) where specific artifacts such as ringing, speckle, and quantization errors routinely occur, yet are not explicitly accounted for by existing iqas. we developed a new dataset, hqa1k, that comprises 1000 natural images each paired with an image rendered using various popular cgh algorithms, and quality-rated by thirteen human participants. our results show that coniqa achieves superior pearson (0.98), spearman (0.965), and kendall\u2019s tau (0.86) correlations over fifteen fr-iqa metrics by up to 5%, showcasing significant improvements in aligning with human perception on the hqa1k dataset.","key_phrases":["the realism","naturalness","images","virtual (vr","augmented (ar) reality applications","full reference image quality assessment","fr-iqa","metrics","that","human perception","deep learning-based iqas","that","human-labeled data","promise","generic computer vision tasks","their performance","applications","perfect matches","the reference","the distorted images","distortion patterns","specific domains","this issue necessitates","a task-specific neural network","human-labeled fr-iqas","deep learning","substantial labeled data","these challenges","we","coniqa","that","consistency training","a novel data augmentation method","both labeled and unlabeled data","this","contexts","scarce labeled data","validate coniqa","we","the example application","computer-generated holography","cgh","where specific artifacts","ringing","speckle","quantization errors","existing iqas","we","a new dataset","hqa1k","that","1000 natural images","each","an image","various popular cgh algorithms","thirteen human participants","our results","coniqa","superior pearson","spearman","kendall\u2019s tau (0.86) correlations","fifteen fr-iqa metrics","up to 5%","significant improvements","human perception","the hqa1k dataset","1000","thirteen","0.98","0.965","0.86","fifteen","up to 5%"]},{"title":"DCLGM: Fusion Recommendation Model Based on LightGBM and Deep Learning","authors":["Bin Zhao","Bin Li","Jiqun Zhang","Wei Cao","Yilong Gao"],"abstract":"The recommendation system can mine valuable information according to user preferences, so it is widely used in various industries. However, the performance of recommendation systems is generally affected by the problem of data sparsity, and LightGBM can alleviate the impact caused by data sparsity to a certain extent. To this end, this paper proposes a fusion recommendation model based on the LightGBM and deep learning\u2014CLGM model. The model is composed of LighGBM, cross network and deep neural network. First, the features in the dataset are fused and extracted through LightGBM, and the feature with the highest classification accuracy is selected as the input of the neural network layer; Then, using the cross network and the deep neural network, the linear cross combination feature relationship and nonlinear correlation relationship between high-order features are respectively obtained; finally, the results obtained by the pre-order network are linearly weighted and combined to obtain the final recommendation result. In this paper, AUC and Logloss are used as evaluation indicators to verify the model on the public dataset Criteo and dataset Avazu. The simulation experiment results show that, compared with the four typical recommendation models, the recommendation effect of this model is better.","doi":"10.1007\/s11063-024-11504-4","cleaned_title":"dclgm: fusion recommendation model based on lightgbm and deep learning","cleaned_abstract":"the recommendation system can mine valuable information according to user preferences, so it is widely used in various industries. however, the performance of recommendation systems is generally affected by the problem of data sparsity, and lightgbm can alleviate the impact caused by data sparsity to a certain extent. to this end, this paper proposes a fusion recommendation model based on the lightgbm and deep learning\u2014clgm model. the model is composed of lighgbm, cross network and deep neural network. first, the features in the dataset are fused and extracted through lightgbm, and the feature with the highest classification accuracy is selected as the input of the neural network layer; then, using the cross network and the deep neural network, the linear cross combination feature relationship and nonlinear correlation relationship between high-order features are respectively obtained; finally, the results obtained by the pre-order network are linearly weighted and combined to obtain the final recommendation result. in this paper, auc and logloss are used as evaluation indicators to verify the model on the public dataset criteo and dataset avazu. the simulation experiment results show that, compared with the four typical recommendation models, the recommendation effect of this model is better.","key_phrases":["the recommendation system","valuable information","user preferences","it","various industries","the performance","recommendation systems","the problem","data sparsity","lightgbm","the impact","data sparsity","a certain extent","this end","this paper","a fusion recommendation model","the lightgbm","deep learning","clgm model","the model","lighgbm","cross network","deep neural network","the features","the dataset","lightgbm","the feature","the highest classification accuracy","the input","the neural network layer","the cross network","the deep neural network","the linear cross combination feature relationship","nonlinear correlation relationship","high-order features","the results","the pre-order network","the final recommendation result","this paper","auc","logloss","evaluation indicators","the model","the public dataset criteo","dataset avazu","the simulation experiment results","the four typical recommendation models","the recommendation effect","this model","first","four"]},{"title":"Deep learning versus manual morphology-based embryo selection in IVF: a randomized, double-blind noninferiority trial","authors":["Peter J. Illingworth","Christos Venetis","David K. Gardner","Scott M. Nelson","J\u00f8rgen Berntsen","Mark G. Larman","Franca Agresta","Saran Ahitan","Aisling Ahlstr\u00f6m","Fleur Cattrall","Simon Cooke","Kristy Demmers","Anette Gabrielsen","Johnny Hindkj\u00e6r","Rebecca L. Kelley","Charlotte Knight","Lisa Lee","Robert Lahoud","Manveen Mangat","Hannah Park","Anthony Price","Geoffrey Trew","Bettina Troest","Anna Vincent","Susanne Wennerstr\u00f6m","Lyndsey Zujovic","Thorir Hardarson"],"abstract":"To assess the value of deep learning in selecting the optimal embryo for in vitro fertilization, a multicenter, randomized, double-blind, noninferiority parallel-group trial was conducted across 14 in vitro fertilization clinics in Australia and Europe. Women under 42 years of age with at least two early-stage blastocysts on day 5 were randomized to either the control arm, using standard morphological assessment, or the study arm, employing a deep learning algorithm, intelligent Data Analysis Score (iDAScore), for embryo selection. The primary endpoint was a clinical pregnancy rate with a noninferiority margin of 5%. The trial included 1,066 patients (533 in the iDAScore group and 533 in the morphology group). The iDAScore group exhibited a clinical pregnancy rate of 46.5% (248 of 533 patients), compared to 48.2% (257 of 533 patients) in the morphology arm (risk difference \u22121.7%; 95% confidence interval \u22127.7, 4.3; P\u2009=\u20090.62). This study was not able to demonstrate noninferiority of deep learning for clinical pregnancy rate when compared to standard morphology and a predefined prioritization scheme. Australian New Zealand Clinical Trials Registry (ANZCTR) registration: 379161.","doi":"10.1038\/s41591-024-03166-5","cleaned_title":"deep learning versus manual morphology-based embryo selection in ivf: a randomized, double-blind noninferiority trial","cleaned_abstract":"to assess the value of deep learning in selecting the optimal embryo for in vitro fertilization, a multicenter, randomized, double-blind, noninferiority parallel-group trial was conducted across 14 in vitro fertilization clinics in australia and europe. women under 42 years of age with at least two early-stage blastocysts on day 5 were randomized to either the control arm, using standard morphological assessment, or the study arm, employing a deep learning algorithm, intelligent data analysis score (idascore), for embryo selection. the primary endpoint was a clinical pregnancy rate with a noninferiority margin of 5%. the trial included 1,066 patients (533 in the idascore group and 533 in the morphology group). the idascore group exhibited a clinical pregnancy rate of 46.5% (248 of 533 patients), compared to 48.2% (257 of 533 patients) in the morphology arm (risk difference \u22121.7%; 95% confidence interval \u22127.7, 4.3; p = 0.62). this study was not able to demonstrate noninferiority of deep learning for clinical pregnancy rate when compared to standard morphology and a predefined prioritization scheme. australian new zealand clinical trials registry (anzctr) registration: 379161.","key_phrases":["the value","deep learning","the optimal embryo","in vitro fertilization","a multicenter","randomized, double-blind, noninferiority parallel-group trial","in vitro fertilization clinics","australia","europe","women","42 years","age","at least two early-stage blastocysts","day","either the control arm","standard morphological assessment","the study arm","a deep learning algorithm","intelligent data analysis score","(idascore","embryo selection","the primary endpoint","a clinical pregnancy rate","a noninferiority margin","5%","the trial","1,066 patients","the idascore group","the morphology group","the idascore group","a clinical pregnancy rate","46.5%","533 patients","48.2%","533 patients","the morphology arm","risk difference","\u22121.7%","95% confidence interval \u22127.7","p","this study","noninferiority","deep learning","clinical pregnancy rate","standard morphology","a predefined prioritization scheme","australian new zealand clinical trials registry (anzctr) registration","14","australia","europe","42 years of age","at least two","day 5","5%","1,066","533","533","46.5%","248","533","48.2%","257","533","\u22121.7%","95%","\u22127.7","4.3","0.62","australian","new zealand","379161"]},{"title":"Cyber Digital Twin with Deep Learning Model for Enterprise Products Management","authors":["Ziqian Wang"],"abstract":"Time series abnormalities might be signs of upcoming problems; thus, new computational anomaly detection techniques are needed for early warning systems and real-time system condition monitoring. Security and intrusion detection systems (IDS) are critical components of Internet of Things (IoT) devices. Current approaches are inadequate for handling complex data and unique intrusion detection systems (IDSs) in today\u2019s network security platforms; deep learning techniques are needed. The Cybertwin-Enhanced self-attention mechanisms with long short-term memory anomaly detection (DL-Cyberwin-Enhanced SAM-LSTM-AD) model for business solutions that may achieve higher prediction accuracy for IoT devices is the main component of this suggested study. It is based on deep learning. To find the absolute error rate threshold of a new model, this model examines assaults against the Cybertwin-neural network. We looked at the CSE-CIC-IDS-2018 dataset to gauge the classifiers\u2019 performance. Utilising the model's capacity to do time series analysis, this research combines the processed data into its suggested framework. These models show that the suggested model is feasible based on the high true positive rate (TPR) and low false positive rate (FPR) that were achieved. The model is evaluated using the test dataset using important metrics including F1-score, ROC-AUC, TPR, FPR, accuracy, and precision.","doi":"10.1007\/s11277-024-11146-8","cleaned_title":"cyber digital twin with deep learning model for enterprise products management","cleaned_abstract":"time series abnormalities might be signs of upcoming problems; thus, new computational anomaly detection techniques are needed for early warning systems and real-time system condition monitoring. security and intrusion detection systems (ids) are critical components of internet of things (iot) devices. current approaches are inadequate for handling complex data and unique intrusion detection systems (idss) in today\u2019s network security platforms; deep learning techniques are needed. the cybertwin-enhanced self-attention mechanisms with long short-term memory anomaly detection (dl-cyberwin-enhanced sam-lstm-ad) model for business solutions that may achieve higher prediction accuracy for iot devices is the main component of this suggested study. it is based on deep learning. to find the absolute error rate threshold of a new model, this model examines assaults against the cybertwin-neural network. we looked at the cse-cic-ids-2018 dataset to gauge the classifiers\u2019 performance. utilising the model's capacity to do time series analysis, this research combines the processed data into its suggested framework. these models show that the suggested model is feasible based on the high true positive rate (tpr) and low false positive rate (fpr) that were achieved. the model is evaluated using the test dataset using important metrics including f1-score, roc-auc, tpr, fpr, accuracy, and precision.","key_phrases":["series abnormalities","signs","upcoming problems","new computational anomaly detection techniques","early warning systems","real-time system condition monitoring","security and intrusion detection systems","ids","critical components","internet","things","(iot) devices","current approaches","complex data","unique intrusion detection systems","idss","today\u2019s network security platforms","deep learning techniques","the cybertwin-enhanced self-attention mechanisms","long short-term memory anomaly detection","(dl-cyberwin-enhanced sam-lstm-ad) model","business solutions","that","higher prediction accuracy","iot devices","the main component","this suggested study","it","deep learning","the absolute error rate threshold","a new model","this model","assaults","the cybertwin-neural network","we","the cse-cic-ids-2018 dataset","the classifiers\u2019 performance","the model's capacity","time series analysis","this research","the processed data","its suggested framework","these models","the suggested model","the high true positive rate","tpr","low false positive rate","fpr","that","the model","the test dataset","important metrics","f1-score","roc-auc","tpr","fpr","accuracy","precision","today","anomaly detection (dl-cyberwin","sam-lstm-ad","roc"]},{"title":"Deep transfer learning-based automated detection of blast disease in paddy crop","authors":["Amandeep Singh","Jaspreet Kaur","Kuldeep Singh","Maninder Lal Singh"],"abstract":"A major proportion of the loss faced by the agricultural industry originates from the diseases of the crop during cultivation. Paddy crop is one of the dominant crops which provides food to a huge population. In this crop, the losses caused by such diseases vary from 30 to 90% of the yield. Therefore, the automated detection of different diseases in paddy crops seeks the attention of the research community. In this context, the present work proposes a deep transfer learning solution for the automated detection of blast disease of paddy, which is the major cause of its yield reduction. For this purpose, an image dataset of healthy and blast disease-infected leave images of paddy crop has been developed. These images are fed to five convolutional neural network-based deep transfer learning algorithms, viz., LeNet, AlexNet, VGG 16, Inception v1, and Xception models for binary classification. The performance analysis of given algorithms reveals that AlexNet provides better results for binary classification with an average accuracy of 98.7% followed by VGG 16 and LeNet architectures having accuracies of 98.2% and 97.8%. So, this deep transfer learning-based approach may assist in reducing the gap between experts and farmers by providing an automated expert advice platform for the timely detection of diseases in paddy crop.","doi":"10.1007\/s11760-023-02735-4","cleaned_title":"deep transfer learning-based automated detection of blast disease in paddy crop","cleaned_abstract":"a major proportion of the loss faced by the agricultural industry originates from the diseases of the crop during cultivation. paddy crop is one of the dominant crops which provides food to a huge population. in this crop, the losses caused by such diseases vary from 30 to 90% of the yield. therefore, the automated detection of different diseases in paddy crops seeks the attention of the research community. in this context, the present work proposes a deep transfer learning solution for the automated detection of blast disease of paddy, which is the major cause of its yield reduction. for this purpose, an image dataset of healthy and blast disease-infected leave images of paddy crop has been developed. these images are fed to five convolutional neural network-based deep transfer learning algorithms, viz., lenet, alexnet, vgg 16, inception v1, and xception models for binary classification. the performance analysis of given algorithms reveals that alexnet provides better results for binary classification with an average accuracy of 98.7% followed by vgg 16 and lenet architectures having accuracies of 98.2% and 97.8%. so, this deep transfer learning-based approach may assist in reducing the gap between experts and farmers by providing an automated expert advice platform for the timely detection of diseases in paddy crop.","key_phrases":["a major proportion","the loss","the agricultural industry originates","the diseases","the crop","cultivation","paddy crop","the dominant crops","which","food","a huge population","this crop","the losses","such diseases","30 to 90%","the yield","the automated detection","different diseases","paddy crops","the attention","the research community","this context","the present work","a deep transfer learning solution","the automated detection","blast disease","paddy","which","the major cause","its yield reduction","this purpose","paddy crop","these images","five convolutional neural network-based deep transfer learning algorithms","viz","lenet","alexnet","vgg","inception v1","xception models","binary classification","the performance analysis","given algorithms","alexnet","better results","binary classification","an average accuracy","98.7%","accuracies","98.2%","97.8%","this deep transfer learning-based approach","the gap","experts","farmers","an automated expert advice platform","the timely detection","diseases","paddy crop","30 to 90%","fed","five","16","98.7%","16","98.2%","97.8%"]},{"title":"Network intrusion detection and mitigation in SDN using deep learning models","authors":["Mamatha Maddu","Yamarthi Narasimha Rao"],"abstract":"Software-Defined Networking (SDN) is a contemporary network strategy utilized instead of a traditional network structure. It provides significantly more administrative efficiency and ease than traditional networks. However, the centralized control used in SDN entails an elevated risk of single-point failure that is more susceptible to different kinds of network assaults like Distributed Denial of Service (DDoS), DoS, spoofing, and API exploitation which are very complex to identify and mitigate. Thus, a powerful intrusion detection system (IDS) based on deep learning is created in this study for the detection and mitigation of network intrusions. This system contains several stages and begins with the data augmentation method named Deep Convolutional Generative Adversarial Networks (DCGAN) to over the data imbalance problem. Then, the features are extracted from the input data using a CenterNet-based approach. After extracting effective characteristics, ResNet152V2 with Slime Mold Algorithm (SMA) based deep learning is implemented to categorize the assaults in InSDN and Edge IIoT datasets. Once the network intrusion is detected, the proposed defense module is activated to restore regular network connectivity quickly. Finally, several experiments are carried out to validate the algorithm's robustness, and the outcomes reveal that the proposed system can successfully detect and mitigate network intrusions.","doi":"10.1007\/s10207-023-00771-2","cleaned_title":"network intrusion detection and mitigation in sdn using deep learning models","cleaned_abstract":"software-defined networking (sdn) is a contemporary network strategy utilized instead of a traditional network structure. it provides significantly more administrative efficiency and ease than traditional networks. however, the centralized control used in sdn entails an elevated risk of single-point failure that is more susceptible to different kinds of network assaults like distributed denial of service (ddos), dos, spoofing, and api exploitation which are very complex to identify and mitigate. thus, a powerful intrusion detection system (ids) based on deep learning is created in this study for the detection and mitigation of network intrusions. this system contains several stages and begins with the data augmentation method named deep convolutional generative adversarial networks (dcgan) to over the data imbalance problem. then, the features are extracted from the input data using a centernet-based approach. after extracting effective characteristics, resnet152v2 with slime mold algorithm (sma) based deep learning is implemented to categorize the assaults in insdn and edge iiot datasets. once the network intrusion is detected, the proposed defense module is activated to restore regular network connectivity quickly. finally, several experiments are carried out to validate the algorithm's robustness, and the outcomes reveal that the proposed system can successfully detect and mitigate network intrusions.","key_phrases":["software-defined networking","sdn","a contemporary network strategy","a traditional network structure","it","significantly more administrative efficiency","ease","traditional networks","the centralized control","sdn","an elevated risk","single-point failure","that","different kinds","network assaults","distributed denial","service","ddos","api exploitation","which","a powerful intrusion detection system","ids","deep learning","this study","the detection","mitigation","network intrusions","this system","several stages","the data augmentation method","deep convolutional generative adversarial networks","dcgan","the data imbalance problem","the features","the input data","a centernet-based approach","effective characteristics","slime mold","algorithm","sma","based deep learning","the assaults","insdn","iiot datasets","the network intrusion","the proposed defense module","regular network connectivity","several experiments","the algorithm's robustness","the outcomes","the proposed system","network intrusions"]},{"title":"Crop disease detection via ensembled-deep-learning paradigm and ABC Coyote pack optimization algorithm (ABC-CPOA)","authors":["M. Chithambarathanu","M. K. Jeyakumar"],"abstract":"Crop disease is a significant issue that affects the growth and yield of crops, leading to financial loss for farmers. Identification and treatment of crop diseases have become challenging due to the increase in the variety of diseases and the lack of knowledge among farmers. To address this issue, this investigate uses an ensembled-deep-learning paradigm to propose a deep learning-based model for crop disease identification trained with an ABC-CPOA. Initially, collected raw images are pre-processed via Bilateral filter and gamma correction Feature Extraction: Then, from the pre-processed images, the features like texture feature (Local Quinary Pattern (LQP), Local Gradient Pattern (LGP), Enriched Local Binary Pattern (E-LBP), color features (Color Histogram, Color Moments), shape features (Contour-based features, Convex Hull). Optimal feature selection- Among the extracted features, the optimal features is designated by means of a self-improved meta-heuristic optimization model referred as ABC-CPOA. This ABC-CPOA model is an extended version of standard Coyote Optimization Algorithm (COA). Crop disease detection phase is modelled with a new ensembled-deep-learning paradigm. Ensembled-deep-learning paradigm comprises Attention-based Bi-LSTM, Recurrent Neural Networks (RNNs) and Optimized Deep Neural Network (O-DNN). The weight function of O-DNN is fine-tuned using the new ABC-CPOA. Precision, recall, sensitivity, and specificity, in addition to TPR, FPR, FNR, and TNR, F1-score, and accuracy are used to assess the suggested approach. The implementation was performed by the MATLAB tool (version: 2022B).","doi":"10.1007\/s11042-024-19329-y","cleaned_title":"crop disease detection via ensembled-deep-learning paradigm and abc coyote pack optimization algorithm (abc-cpoa)","cleaned_abstract":"crop disease is a significant issue that affects the growth and yield of crops, leading to financial loss for farmers. identification and treatment of crop diseases have become challenging due to the increase in the variety of diseases and the lack of knowledge among farmers. to address this issue, this investigate uses an ensembled-deep-learning paradigm to propose a deep learning-based model for crop disease identification trained with an abc-cpoa. initially, collected raw images are pre-processed via bilateral filter and gamma correction feature extraction: then, from the pre-processed images, the features like texture feature (local quinary pattern (lqp), local gradient pattern (lgp), enriched local binary pattern (e-lbp), color features (color histogram, color moments), shape features (contour-based features, convex hull). optimal feature selection- among the extracted features, the optimal features is designated by means of a self-improved meta-heuristic optimization model referred as abc-cpoa. this abc-cpoa model is an extended version of standard coyote optimization algorithm (coa). crop disease detection phase is modelled with a new ensembled-deep-learning paradigm. ensembled-deep-learning paradigm comprises attention-based bi-lstm, recurrent neural networks (rnns) and optimized deep neural network (o-dnn). the weight function of o-dnn is fine-tuned using the new abc-cpoa. precision, recall, sensitivity, and specificity, in addition to tpr, fpr, fnr, and tnr, f1-score, and accuracy are used to assess the suggested approach. the implementation was performed by the matlab tool (version: 2022b).","key_phrases":["crop disease","a significant issue","that","the growth","yield","crops","financial loss","farmers","identification","treatment","crop diseases","the increase","the variety","diseases","the lack","knowledge","farmers","this issue","this investigate","an ensembled-deep-learning paradigm","a deep learning-based model","crop disease identification","an abc-cpoa","raw images","bilateral filter","gamma correction","feature extraction","the pre-processed images","texture feature","local quinary pattern","lqp","local gradient pattern","lgp","enriched local binary pattern","e","-","lbp","color features","color histogram","color moments","shape features","(contour-based features","convex hull","optimal feature","the extracted features","the optimal features","means","a self-improved meta-heuristic optimization model","abc-cpoa","this abc-cpoa model","an extended version","standard coyote optimization algorithm","coa","crop disease detection phase","a new ensembled-deep-learning paradigm","ensembled-deep-learning paradigm","attention-based bi","-","lstm","recurrent neural networks","rnns","deep neural network","o","dnn","the weight function","o","-","dnn","the new abc-cpoa","precision","recall","sensitivity","specificity","addition","tpr","fpr","fnr","tnr","f1-score","accuracy","the suggested approach","the implementation","the matlab tool","version","abc","abc","abc","abc"]},{"title":"Development of a deep learning model for cancer diagnosis by inspecting cell-free DNA end-motifs","authors":["Hongru Shen","Meng Yang","Jilei Liu","Kexin Chen","Xiangchun Li"],"abstract":"Accurate discrimination between patients with and without cancer from cfDNA is crucial for early cancer diagnosis. Herein, we develop and validate a deep-learning-based model entitled end-motif inspection via transformer (EMIT) for discriminating individuals with and without cancer by learning feature representations from cfDNA end-motifs. EMIT is a self-supervised learning approach that models rankings of cfDNA end-motifs. We include 4606 samples subjected to different types of cfDNA sequencing to develop EIMIT, and subsequently evaluate classification performance of linear projections of EMIT on six datasets and an additional inhouse testing set encopassing whole-genome, whole-genome bisulfite and 5-hydroxymethylcytosine sequencing. The linear projection of representations from EMIT achieved area under the receiver operating curve (AUROC) values ranged from 0.895 (0.835\u20130.955) to 0.996 (0.994\u20130.997) across these six datasets, outperforming its baseline by significant margins. Additionally, we showed that linear projection of EMIT representations can achieve an AUROC of 0.962 (0.914\u20131.0) in identification of lung cancer on an independent testing set subjected to whole-exome sequencing. The findings of this study indicate that a transformer-based deep learning model can learn cancer-discrimative representations from cfDNA end-motifs. The representations of this deep learning model can be exploited for discriminating patients with and without cancer.","doi":"10.1038\/s41698-024-00635-5","cleaned_title":"development of a deep learning model for cancer diagnosis by inspecting cell-free dna end-motifs","cleaned_abstract":"accurate discrimination between patients with and without cancer from cfdna is crucial for early cancer diagnosis. herein, we develop and validate a deep-learning-based model entitled end-motif inspection via transformer (emit) for discriminating individuals with and without cancer by learning feature representations from cfdna end-motifs. emit is a self-supervised learning approach that models rankings of cfdna end-motifs. we include 4606 samples subjected to different types of cfdna sequencing to develop eimit, and subsequently evaluate classification performance of linear projections of emit on six datasets and an additional inhouse testing set encopassing whole-genome, whole-genome bisulfite and 5-hydroxymethylcytosine sequencing. the linear projection of representations from emit achieved area under the receiver operating curve (auroc) values ranged from 0.895 (0.835\u20130.955) to 0.996 (0.994\u20130.997) across these six datasets, outperforming its baseline by significant margins. additionally, we showed that linear projection of emit representations can achieve an auroc of 0.962 (0.914\u20131.0) in identification of lung cancer on an independent testing set subjected to whole-exome sequencing. the findings of this study indicate that a transformer-based deep learning model can learn cancer-discrimative representations from cfdna end-motifs. the representations of this deep learning model can be exploited for discriminating patients with and without cancer.","key_phrases":["accurate discrimination","patients","cancer","cfdna","early cancer diagnosis","we","a deep-learning-based model entitled end-motif inspection","transformer","individuals","cancer","feature representations","cfdna end-motifs","emit","a self-supervised learning approach","that","cfdna end-motifs","we","4606 samples","different types","cfdna","eimit","classification performance","linear projections","six datasets","an additional inhouse testing","whole-genome, whole-genome bisulfite","5-hydroxymethylcytosine","the linear projection","representations","emit","area","the receiver operating curve (auroc) values","(0.835\u20130.955","0.994\u20130.997","these six datasets","its baseline","significant margins","we","linear projection","emit representations","an auroc","identification","lung cancer","an independent testing","the findings","this study","a transformer-based deep learning model","cancer-discrimative representations","cfdna end-motifs","the representations","this deep learning model","patients","cancer","emit","linear","six","5","0.895","0.996","six","linear","0.962"]},{"title":"An automated weed detection approach using deep learning and UAV imagery in smart agriculture system","authors":["Baozhong Liu"],"abstract":"Weed detection plays a critical role in smart and precise agriculture systems by enabling targeted weed management and reducing environmental impact. Unmanned aerial vehicles (UAVs) and their associated imagery have emerged as powerful tools for weed detection. Traditional and deep learning methods have been explored for weed detection, with deep learning methods being favored due to their ability to handle complex patterns. However, accuracy rate and computation cost challenges persist in deep learning-based weed detection methods. To address this, we propose a method on the basis of the YOLOv5 algorithm to deal with high accuracy demand and low computation cost requirements. The approach involves model generation using a custom dataset and training, validation, and testing sets. Experimental results and performance evaluation validate the proposed method that indicates the research contributes to advancing weed detection in smart and precise agriculture systems, leveraging deep learning techniques for enhanced accuracy and efficiency.","doi":"10.1007\/s12596-023-01445-x","cleaned_title":"an automated weed detection approach using deep learning and uav imagery in smart agriculture system","cleaned_abstract":"weed detection plays a critical role in smart and precise agriculture systems by enabling targeted weed management and reducing environmental impact. unmanned aerial vehicles (uavs) and their associated imagery have emerged as powerful tools for weed detection. traditional and deep learning methods have been explored for weed detection, with deep learning methods being favored due to their ability to handle complex patterns. however, accuracy rate and computation cost challenges persist in deep learning-based weed detection methods. to address this, we propose a method on the basis of the yolov5 algorithm to deal with high accuracy demand and low computation cost requirements. the approach involves model generation using a custom dataset and training, validation, and testing sets. experimental results and performance evaluation validate the proposed method that indicates the research contributes to advancing weed detection in smart and precise agriculture systems, leveraging deep learning techniques for enhanced accuracy and efficiency.","key_phrases":["weed detection","a critical role","smart and precise agriculture systems","targeted weed management","environmental impact","unmanned aerial vehicles","their associated imagery","powerful tools","weed detection","traditional and deep learning methods","weed detection","deep learning methods","their ability","complex patterns","accuracy rate","computation cost challenges","deep learning-based weed detection methods","this","we","a method","the basis","the yolov5 algorithm","high accuracy demand","low computation cost requirements","the approach","model generation","a custom dataset","training","validation","testing sets","experimental results","performance evaluation","the proposed method","that","the research","weed detection","smart and precise agriculture systems","deep learning techniques","enhanced accuracy","efficiency","yolov5"]},{"title":"Performance prediction in online academic course: a deep learning approach with time series imaging","authors":["Ahmed Ben Said","Abdel-Salam G. Abdel-Salam","Khalifa A. Hazaa"],"abstract":"With the COVID-19 outbreak, schools and universities have massively adopted online learning to ensure the continuation of the learning process. However, in such setting, instructors lack efficient mechanisms to evaluate the learning gains and get insights about difficulties learners encounter. In this research work, we tackle the problem of predicting learner performance in online learning using a deep learning-based approach. Our proposed solution allows stakeholders involved in the online learning to anticipate the learner outcome ahead of the final assessment hence offering the opportunity for proactive measures to assist the learners. We propose a two-pathway deep learning model to classify learner performance using their interaction during the online sessions in the form of clickstreams. We also propose to transform these time series of clicks into images using the Gramian Angular Field. The learning model makes use of the available extra demographic and assessment information. We evaluate our approach on the Open University Learning Analytics Dataset. Comprehensive comparative study is conducted with evaluation against state-of-art approaches under different experimental settings. We also demonstrate the importance of including extra demographic and assessment data in the prediction process.","doi":"10.1007\/s11042-023-17596-9","cleaned_title":"performance prediction in online academic course: a deep learning approach with time series imaging","cleaned_abstract":"with the covid-19 outbreak, schools and universities have massively adopted online learning to ensure the continuation of the learning process. however, in such setting, instructors lack efficient mechanisms to evaluate the learning gains and get insights about difficulties learners encounter. in this research work, we tackle the problem of predicting learner performance in online learning using a deep learning-based approach. our proposed solution allows stakeholders involved in the online learning to anticipate the learner outcome ahead of the final assessment hence offering the opportunity for proactive measures to assist the learners. we propose a two-pathway deep learning model to classify learner performance using their interaction during the online sessions in the form of clickstreams. we also propose to transform these time series of clicks into images using the gramian angular field. the learning model makes use of the available extra demographic and assessment information. we evaluate our approach on the open university learning analytics dataset. comprehensive comparative study is conducted with evaluation against state-of-art approaches under different experimental settings. we also demonstrate the importance of including extra demographic and assessment data in the prediction process.","key_phrases":["the covid-19 outbreak","schools","universities","online learning","the continuation","the learning process","such setting","instructors","efficient mechanisms","the learning gains","insights","difficulties","learners","this research work","we","the problem","learner performance","online learning","a deep learning-based approach","our proposed solution","stakeholders","the online learning","the learner outcome","the final assessment","the opportunity","proactive measures","the learners","we","a two-pathway deep learning model","learner performance","their interaction","the online sessions","the form","clickstreams","we","these time series","clicks","images","the gramian angular field","the learning model","use","the available extra demographic and assessment information","we","our approach","the open university learning analytics","comprehensive comparative study","evaluation","art","different experimental settings","we","the importance","extra demographic and assessment data","the prediction process","covid-19","two","gramian"]},{"title":"Deep Learning Enhanced Snapshot Generation for Efficient Hyper-reduction in Nonlinear Structural Dynamics","authors":["Hossein Najafi","Morteza Karamooz Mahdiabadi"],"abstract":"PurposeThis study presents a novel approach to enhancing hyper-reduction in nonlinear structural dynamics by utilizing the predictive capabilities of stacked Long Short-Term Memory (LSTM) neural networks. Hyper-reduction methods are crucial for overcoming the limitations of traditional model order reduction techniques, particularly in accurately capturing the nonlinear behavior of internal force vectors in complex structures.Method The proposed technique employs stacked LSTM neural networks to generate training snapshots for the Energy Conserving Mesh Sampling and Weighting (ECSW) hyper-reduction method. By training the model on a well-defined dataset, we achieve an impressive accuracy of 97.5%. The effectiveness of our method is demonstrated through a geometrically nonlinear dynamic analysis of a leaf spring, resulting in only a 3.24% error when compared to full simulation results. This study emphasizes the potential of deep learning techniques in improving hyper-reduction methods and underscores the importance of computational efficiency in simulations of complex structural dynamics.ResultsThe findings reveal significant advancements in the application of deep learning for hyper-reduction methods, showcasing the ability to accurately model nonlinear structural behaviors while maintaining computational efficiency. This research contributes valuable insights into the integration of advanced machine learning techniques within the field of structural dynamics.","doi":"10.1007\/s42417-024-01528-4","cleaned_title":"deep learning enhanced snapshot generation for efficient hyper-reduction in nonlinear structural dynamics","cleaned_abstract":"purposethis study presents a novel approach to enhancing hyper-reduction in nonlinear structural dynamics by utilizing the predictive capabilities of stacked long short-term memory (lstm) neural networks. hyper-reduction methods are crucial for overcoming the limitations of traditional model order reduction techniques, particularly in accurately capturing the nonlinear behavior of internal force vectors in complex structures.method the proposed technique employs stacked lstm neural networks to generate training snapshots for the energy conserving mesh sampling and weighting (ecsw) hyper-reduction method. by training the model on a well-defined dataset, we achieve an impressive accuracy of 97.5%. the effectiveness of our method is demonstrated through a geometrically nonlinear dynamic analysis of a leaf spring, resulting in only a 3.24% error when compared to full simulation results. this study emphasizes the potential of deep learning techniques in improving hyper-reduction methods and underscores the importance of computational efficiency in simulations of complex structural dynamics.resultsthe findings reveal significant advancements in the application of deep learning for hyper-reduction methods, showcasing the ability to accurately model nonlinear structural behaviors while maintaining computational efficiency. this research contributes valuable insights into the integration of advanced machine learning techniques within the field of structural dynamics.","key_phrases":["purposethis study","a novel approach","hyper-reduction","nonlinear structural dynamics","the predictive capabilities","stacked long short-term memory","lstm","neural networks","hyper-reduction methods","the limitations","traditional model order reduction techniques","the nonlinear behavior","internal force vectors","complex structures.method","the proposed technique employs","lstm neural networks","training snapshots","the energy conserving mesh sampling","weighting","hyper-reduction method","the model","a well-defined dataset","we","an impressive accuracy","97.5%","the effectiveness","our method","a geometrically nonlinear dynamic analysis","a leaf spring","only a 3.24% error","full simulation results","this study","the potential","deep learning techniques","hyper-reduction methods","the importance","computational efficiency","simulations","complex structural dynamics.resultsthe findings","significant advancements","the application","deep learning","hyper-reduction methods","the ability","nonlinear structural behaviors","computational efficiency","this research","valuable insights","the integration","advanced machine learning techniques","the field","structural dynamics","97.5%","a leaf spring","3.24%"]},{"title":"Optimization of job shop scheduling problem based on deep reinforcement learning","authors":["Dongping Qiao","Lvqi Duan","HongLei Li","Yanqiu Xiao"],"abstract":"Aiming at the optimization problem of minimizing the maximum completion time in job shop scheduling, a deep reinforcement learning optimization algorithm is proposed. First, a deep reinforcement learning scheduling environment is built based on the disjunctive graph model, and three channels of state characteristics are established. The action space consists of 20 designed combination scheduling rules. The reward function is designed based on the proportional relationship between the total work of the scheduled operation and the current maximum completion time. The deep convolutional neural network is used to construct action network and target network, and the state features are used as inputs to output the Q value of each action. Then, the action is selected by using the action validity exploration and exploitation strategy. Finally, the immediate reward is calculated and the scheduling environment is updated. Experiments are carried out using benchmark instances to verify the algorithm. The results show that it can balance solution quality and computation time effectively, and the trained agent has good generalization ability to the scheduling problem in the non-zero initial state.","doi":"10.1007\/s12065-023-00885-5","cleaned_title":"optimization of job shop scheduling problem based on deep reinforcement learning","cleaned_abstract":"aiming at the optimization problem of minimizing the maximum completion time in job shop scheduling, a deep reinforcement learning optimization algorithm is proposed. first, a deep reinforcement learning scheduling environment is built based on the disjunctive graph model, and three channels of state characteristics are established. the action space consists of 20 designed combination scheduling rules. the reward function is designed based on the proportional relationship between the total work of the scheduled operation and the current maximum completion time. the deep convolutional neural network is used to construct action network and target network, and the state features are used as inputs to output the q value of each action. then, the action is selected by using the action validity exploration and exploitation strategy. finally, the immediate reward is calculated and the scheduling environment is updated. experiments are carried out using benchmark instances to verify the algorithm. the results show that it can balance solution quality and computation time effectively, and the trained agent has good generalization ability to the scheduling problem in the non-zero initial state.","key_phrases":["the optimization problem","the maximum completion time","job shop scheduling","a deep reinforcement learning optimization algorithm","a deep reinforcement learning scheduling environment","the disjunctive graph model","three channels","state characteristics","the action space","20 designed combination scheduling rules","the reward function","the proportional relationship","the total work","the scheduled operation","the current maximum completion time","the deep convolutional neural network","action network","target network","the state features","inputs","the q value","each action","the action","the action validity exploration and exploitation strategy","the immediate reward","the scheduling environment","experiments","benchmark instances","the algorithm","the results","it","solution quality and computation time","the trained agent","good generalization ability","the scheduling problem","the non-zero initial state","first","three","20"]},{"title":"A Study on Different Deep Learning Algorithms Used in Deep Neural Nets: MLP SOM and DBN","authors":["J. Naskath","G. Sivakamasundari","A. Alif Siddiqua Begum"],"abstract":"Deep learning is a wildly popular topic in machine learning and is structured as a series of nonlinear layers that learns various levels of data representations. Deep learning employs numerous layers to represent data abstractions to implement various computer models. Deep learning approaches like generative, discriminative models and model transfer have transformed information processing. This article proposes a comprehensive review of various deep learning algorithms Multi layer perception, Self-organizing map and deep belief networks algorithms. It first briefly introduces historical and recent state-of-the-art reviews with suitable architectures and implementation steps. Moreover, the various applications of those algorithms in various fields such as wireless networks, Adhoc networks, Mobile ad-hoc and vehicular ad-hoc networks, speech recognition engineering, medical applications, natural language processing, material science and remote sensing applications, etc. are classified.","doi":"10.1007\/s11277-022-10079-4","cleaned_title":"a study on different deep learning algorithms used in deep neural nets: mlp som and dbn","cleaned_abstract":"deep learning is a wildly popular topic in machine learning and is structured as a series of nonlinear layers that learns various levels of data representations. deep learning employs numerous layers to represent data abstractions to implement various computer models. deep learning approaches like generative, discriminative models and model transfer have transformed information processing. this article proposes a comprehensive review of various deep learning algorithms multi layer perception, self-organizing map and deep belief networks algorithms. it first briefly introduces historical and recent state-of-the-art reviews with suitable architectures and implementation steps. moreover, the various applications of those algorithms in various fields such as wireless networks, adhoc networks, mobile ad-hoc and vehicular ad-hoc networks, speech recognition engineering, medical applications, natural language processing, material science and remote sensing applications, etc. are classified.","key_phrases":["deep learning","a wildly popular topic","machine learning","a series","nonlinear layers","that","various levels","data representations","deep learning","numerous layers","data abstractions","various computer models","approaches","discriminative models","model transfer","information processing","this article","a comprehensive review","various deep learning algorithms","multi layer perception","self-organizing map","deep belief networks","it","the-art","suitable architectures","implementation steps","the various applications","those algorithms","various fields","wireless networks","adhoc networks","mobile ad-hoc and vehicular ad-hoc networks","speech recognition engineering","medical applications","natural language processing","material science","remote sensing applications","first"]},{"title":"A Review of Deep Learning Techniques for Glaucoma Detection","authors":["Takfarines Guergueb","Moulay A. Akhloufi"],"abstract":"Glaucoma is one of the major reasons for visual impairment all across the globe. The recent advancements in machine learning techniques have greatly facilitated ophthalmologists in the early diagnosis of ocular diseases through the employment of automated systems. Several studies have been published lately to address the timely detection of glaucoma using deep learning approaches. A comprehensive review of the deep learning approaches employed for glaucoma detection using retinal fundus images is presented in this paper. The available retinal image datasets, image pre-processing techniques, state-of-the-art models, and performance evaluation metrics used in the recent studies are reviewed. This systematic review aims to provide critical insights and potential research directions to the ophthalmologists and researchers in this domain.","doi":"10.1007\/s42979-023-01734-z","cleaned_title":"a review of deep learning techniques for glaucoma detection","cleaned_abstract":"glaucoma is one of the major reasons for visual impairment all across the globe. the recent advancements in machine learning techniques have greatly facilitated ophthalmologists in the early diagnosis of ocular diseases through the employment of automated systems. several studies have been published lately to address the timely detection of glaucoma using deep learning approaches. a comprehensive review of the deep learning approaches employed for glaucoma detection using retinal fundus images is presented in this paper. the available retinal image datasets, image pre-processing techniques, state-of-the-art models, and performance evaluation metrics used in the recent studies are reviewed. this systematic review aims to provide critical insights and potential research directions to the ophthalmologists and researchers in this domain.","key_phrases":["glaucoma","the major reasons","visual impairment","the globe","the recent advancements","machine learning techniques","the early diagnosis","ocular diseases","the employment","automated systems","several studies","the timely detection","glaucoma","deep learning approaches","a comprehensive review","the deep learning approaches","glaucoma detection","retinal fundus images","this paper","the available retinal image datasets","image pre-processing techniques","the-art","performance evaluation metrics","the recent studies","this systematic review","critical insights","potential research directions","the ophthalmologists","researchers","this domain","glaucoma"]},{"title":"Investigating Deep Learning for Early Detection and Decision-Making in Alzheimer\u2019s Disease: A Comprehensive Review","authors":["Ghazala Hcini","Imen Jdey","Habib Dhahri"],"abstract":"Alzheimer\u2019s disease (AD) is a neurodegenerative disorder that affects millions of people worldwide, making early detection essential for effective intervention. This review paper provides a comprehensive analysis of the use of deep learning techniques, specifically convolutional neural networks (CNN) and vision transformers (ViT), for the classification of AD using brain imaging data. While previous reviews have covered similar topics, this paper offers a unique perspective by providing a detailed comparison of CNN and ViT for AD classification, highlighting the strengths and limitations of each approach. Additionally, the review presents an updated and thorough analysis of the most recent studies in the field, including the latest advancements in CNN and ViT architectures, training methods, and performance evaluation metrics. Furthermore, the paper discusses the ethical considerations and challenges associated with the use of deep learning models for AD classification, such as the need for interpretability and the potential for bias. By addressing these issues, this review aims to provide valuable insights for future research and clinical applications, ultimately advancing the field of AD classification using deep learning techniques.","doi":"10.1007\/s11063-024-11600-5","cleaned_title":"investigating deep learning for early detection and decision-making in alzheimer\u2019s disease: a comprehensive review","cleaned_abstract":"alzheimer\u2019s disease (ad) is a neurodegenerative disorder that affects millions of people worldwide, making early detection essential for effective intervention. this review paper provides a comprehensive analysis of the use of deep learning techniques, specifically convolutional neural networks (cnn) and vision transformers (vit), for the classification of ad using brain imaging data. while previous reviews have covered similar topics, this paper offers a unique perspective by providing a detailed comparison of cnn and vit for ad classification, highlighting the strengths and limitations of each approach. additionally, the review presents an updated and thorough analysis of the most recent studies in the field, including the latest advancements in cnn and vit architectures, training methods, and performance evaluation metrics. furthermore, the paper discusses the ethical considerations and challenges associated with the use of deep learning models for ad classification, such as the need for interpretability and the potential for bias. by addressing these issues, this review aims to provide valuable insights for future research and clinical applications, ultimately advancing the field of ad classification using deep learning techniques.","key_phrases":["alzheimer\u2019s disease","ad","a neurodegenerative disorder","that","millions","people","early detection","effective intervention","this review paper","a comprehensive analysis","the use","deep learning techniques","specifically convolutional neural networks","cnn","vision transformers","vit","the classification","ad","brain imaging data","previous reviews","similar topics","this paper","a unique perspective","a detailed comparison","cnn","vit","ad classification","the strengths","limitations","each approach","the review","an updated and thorough analysis","the most recent studies","the field","the latest advancements","cnn","vit architectures","training methods","performance evaluation metrics","the paper","the ethical considerations","challenges","the use","deep learning models","ad classification","the need","interpretability","the potential","bias","these issues","this review","valuable insights","future research and clinical applications","the field","ad classification","deep learning techniques","millions","cnn","cnn","vit","cnn","vit"]},{"title":"Phase unwrapping based on deep learning in light field fringe projection 3D measurement","authors":["Xinjun Zhu","Haichuan Zhao","Mengkai Yuan","Zhizhi Zhang","Hongyi Wang","Limei Song"],"abstract":"Phase unwrapping is one of the key roles in fringe projection three-dimensional (3D) measurement technology. We propose a new method to achieve phase unwrapping in camera array light filed fringe projection 3D measurement based on deep learning. A multi-stream convolutional neural network (CNN) is proposed to learn the mapping relationship between camera array light filed wrapped phases and fringe orders of the expected central view, and is used to predict the fringe order to achieve the phase unwrapping. Experiments are performed on the light field fringe projection data generated by the simulated camera array fringe projection measurement system in Blender and by the experimental 3\u00d73 camera array light field fringe projection system. The performance of the proposed network with light field wrapped phases using multiple directions as network input data is studied, and the advantages of phase unwrapping based on deep learning in light filed fringe projection are demonstrated.","doi":"10.1007\/s11801-023-3002-4","cleaned_title":"phase unwrapping based on deep learning in light field fringe projection 3d measurement","cleaned_abstract":"phase unwrapping is one of the key roles in fringe projection three-dimensional (3d) measurement technology. we propose a new method to achieve phase unwrapping in camera array light filed fringe projection 3d measurement based on deep learning. a multi-stream convolutional neural network (cnn) is proposed to learn the mapping relationship between camera array light filed wrapped phases and fringe orders of the expected central view, and is used to predict the fringe order to achieve the phase unwrapping. experiments are performed on the light field fringe projection data generated by the simulated camera array fringe projection measurement system in blender and by the experimental 3\u00d73 camera array light field fringe projection system. the performance of the proposed network with light field wrapped phases using multiple directions as network input data is studied, and the advantages of phase unwrapping based on deep learning in light filed fringe projection are demonstrated.","key_phrases":["phase","the key roles","fringe projection three-dimensional (3d) measurement technology","we","a new method","phase","camera array light","fringe projection 3d measurement","deep learning","a multi-stream convolutional neural network","cnn","the mapping relationship","camera array light","phases","fringe orders","the expected central view","the fringe order","the phase","experiments","the light field fringe projection data","the simulated camera array fringe projection measurement system","blender","the experimental 3\u00d73 camera array light field fringe projection system","the performance","the proposed network","light field","phases","multiple directions","network input data","the advantages","phase","deep learning","light filed fringe projection","three","3d","3d","cnn","3\u00d73"]},{"title":"Development and application of a deep learning-based comprehensive early diagnostic model for chronic obstructive pulmonary disease","authors":["Zecheng Zhu","Shunjin Zhao","Jiahui Li","Yuting Wang","Luopiao Xu","Yubing Jia","Zihan Li","Wenyuan Li","Gang Chen","Xifeng Wu"],"abstract":"BackgroundChronic obstructive pulmonary disease (COPD) is a frequently diagnosed yet treatable condition, provided it is identified early and managed effectively. This study aims to develop an advanced COPD diagnostic model by integrating deep learning and radiomics features.MethodsWe utilized a dataset comprising CT images from 2,983 participants, of which 2,317 participants also provided epidemiological data through questionnaires. Deep learning features were extracted using a Variational Autoencoder, and radiomics features were obtained using the PyRadiomics package. Multi-Layer Perceptrons were used to construct models based on deep learning and radiomics features independently, as well as a fusion model integrating both. Subsequently, epidemiological questionnaire data were incorporated to establish a more comprehensive model. The diagnostic performance of standalone models, the fusion model and the comprehensive model was evaluated and compared using metrics including accuracy, precision, recall, F1-score, Brier score, receiver operating characteristic curves, and area under the curve (AUC).ResultsThe fusion model exhibited outstanding performance with an AUC of 0.952, surpassing the standalone models based solely on deep learning features (AUC\u2009=\u20090.844) or radiomics features (AUC\u2009=\u20090.944). Notably, the comprehensive model, incorporating deep learning features, radiomics features, and questionnaire variables demonstrated the highest diagnostic performance among all models, yielding an AUC of 0.971.ConclusionWe developed and implemented a data fusion strategy to construct a state-of-the-art COPD diagnostic model integrating deep learning features, radiomics features, and questionnaire variables. Our data fusion strategy proved effective, and the model can be easily deployed in clinical settings.Trial registrationNot applicable. This study is NOT a clinical trial, it does not report the results of a health care intervention on human participants.","doi":"10.1186\/s12931-024-02793-3","cleaned_title":"development and application of a deep learning-based comprehensive early diagnostic model for chronic obstructive pulmonary disease","cleaned_abstract":"backgroundchronic obstructive pulmonary disease (copd) is a frequently diagnosed yet treatable condition, provided it is identified early and managed effectively. this study aims to develop an advanced copd diagnostic model by integrating deep learning and radiomics features.methodswe utilized a dataset comprising ct images from 2,983 participants, of which 2,317 participants also provided epidemiological data through questionnaires. deep learning features were extracted using a variational autoencoder, and radiomics features were obtained using the pyradiomics package. multi-layer perceptrons were used to construct models based on deep learning and radiomics features independently, as well as a fusion model integrating both. subsequently, epidemiological questionnaire data were incorporated to establish a more comprehensive model. the diagnostic performance of standalone models, the fusion model and the comprehensive model was evaluated and compared using metrics including accuracy, precision, recall, f1-score, brier score, receiver operating characteristic curves, and area under the curve (auc).resultsthe fusion model exhibited outstanding performance with an auc of 0.952, surpassing the standalone models based solely on deep learning features (auc = 0.844) or radiomics features (auc = 0.944). notably, the comprehensive model, incorporating deep learning features, radiomics features, and questionnaire variables demonstrated the highest diagnostic performance among all models, yielding an auc of 0.971.conclusionwe developed and implemented a data fusion strategy to construct a state-of-the-art copd diagnostic model integrating deep learning features, radiomics features, and questionnaire variables. our data fusion strategy proved effective, and the model can be easily deployed in clinical settings.trial registrationnot applicable. this study is not a clinical trial, it does not report the results of a health care intervention on human participants.","key_phrases":["backgroundchronic obstructive pulmonary disease","copd","a frequently diagnosed yet treatable condition","it","this study","an advanced copd diagnostic model","deep learning and radiomics features.methodswe","a dataset","ct images","2,983 participants","which","2,317 participants","epidemiological data","questionnaires","deep learning features","a variational autoencoder","radiomics features","the pyradiomics package","multi-layer perceptrons","models","deep learning","radiomics","a fusion model","both","epidemiological questionnaire data","a more comprehensive model","the diagnostic performance","standalone models","the fusion model","the comprehensive model","metrics","accuracy","precision","recall, f1-score","brier score","the curve","auc).resultsthe fusion model","outstanding performance","an auc","the standalone models","deep learning features","auc","radiomics features","auc =","the comprehensive model","deep learning features","radiomics features","questionnaire variables","the highest diagnostic performance","all models","an auc","0.971.conclusionwe","a data fusion strategy","the-art","deep learning features","radiomics features","questionnaire variables","our data fusion strategy","the model","clinical settings.trial registrationnot","this study","a clinical trial","it","the results","a health care intervention","human participants","2,983","2,317","0.952","0.844","0.944","0.971.conclusionwe"]},{"title":"A multi-agent adaptive deep learning framework for online intrusion detection","authors":["Mahdi Soltani","Khashayar Khajavi","Mahdi Jafari Siavoshani","Amir Hossein Jahangir"],"abstract":"The network security analyzers use intrusion detection systems (IDSes) to distinguish malicious traffic from benign ones. The deep learning-based (DL-based) IDSes are proposed to auto-extract high-level features and eliminate the time-consuming and costly signature extraction process. However, this new generation of IDSes still needs to overcome a number of challenges to be employed in practical environments. One of the main issues of an applicable IDS is facing traffic concept drift, which manifests itself as new (i.e.\u00a0, zero-day) attacks, in addition to the changing behavior of benign users\/applications. Furthermore, a practical DL-based IDS needs to be conformed to a distributed (i.e.\u00a0, multi-sensor) architecture in order to yield more accurate detections, create a collective attack knowledge based on the observations of different sensors, and also handle big data challenges for supporting high throughput networks. This paper proposes a novel multi-agent network intrusion detection framework to address the above shortcomings, considering a more practical scenario (i.e., online adaptable IDSes). This framework employs continual deep anomaly detectors for adapting each agent to the changing attack\/benign patterns in its local traffic. In addition, a federated learning approach is proposed for sharing and exchanging local knowledge between different agents. Furthermore, the proposed framework implements sequential packet labeling for each flow, which provides an attack probability score for the flow by gradually observing each flow packet and updating its estimation. We evaluate the proposed framework by employing different deep models (including CNN-based and LSTM-based) over the CIC-IDS2017 and CSE-CIC-IDS2018 datasets. Through extensive evaluations and experiments, we show that the proposed distributed framework is well adapted to the traffic concept drift. More precisely, our results indicate that the CNN-based models are well suited for continually adapting to the traffic concept drift (i.e.\u00a0, achieving an average detection rate of above 95% while needing just 128 new flows for the updating phase), and the LSTM-based models are a good candidate for sequential packet labeling in practical online IDSes (i.e.\u00a0, detecting intrusions by just observing their first 15 packets).","doi":"10.1186\/s42400-023-00199-0","cleaned_title":"a multi-agent adaptive deep learning framework for online intrusion detection","cleaned_abstract":"the network security analyzers use intrusion detection systems (idses) to distinguish malicious traffic from benign ones. the deep learning-based (dl-based) idses are proposed to auto-extract high-level features and eliminate the time-consuming and costly signature extraction process. however, this new generation of idses still needs to overcome a number of challenges to be employed in practical environments. one of the main issues of an applicable ids is facing traffic concept drift, which manifests itself as new (i.e. , zero-day) attacks, in addition to the changing behavior of benign users\/applications. furthermore, a practical dl-based ids needs to be conformed to a distributed (i.e. , multi-sensor) architecture in order to yield more accurate detections, create a collective attack knowledge based on the observations of different sensors, and also handle big data challenges for supporting high throughput networks. this paper proposes a novel multi-agent network intrusion detection framework to address the above shortcomings, considering a more practical scenario (i.e., online adaptable idses). this framework employs continual deep anomaly detectors for adapting each agent to the changing attack\/benign patterns in its local traffic. in addition, a federated learning approach is proposed for sharing and exchanging local knowledge between different agents. furthermore, the proposed framework implements sequential packet labeling for each flow, which provides an attack probability score for the flow by gradually observing each flow packet and updating its estimation. we evaluate the proposed framework by employing different deep models (including cnn-based and lstm-based) over the cic-ids2017 and cse-cic-ids2018 datasets. through extensive evaluations and experiments, we show that the proposed distributed framework is well adapted to the traffic concept drift. more precisely, our results indicate that the cnn-based models are well suited for continually adapting to the traffic concept drift (i.e. , achieving an average detection rate of above 95% while needing just 128 new flows for the updating phase), and the lstm-based models are a good candidate for sequential packet labeling in practical online idses (i.e. , detecting intrusions by just observing their first 15 packets).","key_phrases":["the network security analyzers","intrusion detection systems","(idses","malicious traffic","benign ones","the deep learning-based (dl-based) idses","auto-extract high-level features","the time-consuming and costly signature extraction process","this new generation","idses","a number","challenges","practical environments","the main issues","an applicable ids","traffic concept drift","which","itself","new (i.e. , zero-day","attacks","addition","the changing behavior","benign users\/applications","a practical dl-based ids","a distributed (i.e. , multi-sensor) architecture","order","more accurate detections","a collective attack knowledge","the observations","different sensors","big data challenges","high throughput networks","this paper","a novel multi-agent network intrusion detection framework","the above shortcomings","a more practical scenario","(i.e., online adaptable idses","this framework","continual deep anomaly detectors","each agent","the changing attack\/benign patterns","its local traffic","addition","a federated learning approach","local knowledge","different agents","the proposed framework","sequential packet labeling","each flow","which","an attack probability score","the flow","each flow packet","its estimation","we","the proposed framework","different deep models","the cic","-ids2017","cse-cic-ids2018","datasets","extensive evaluations","experiments","we","the proposed distributed framework","the traffic concept drift","our results","the cnn-based models","the traffic concept drift","an average detection rate","above 95%","just 128 new flows","the updating phase","the lstm-based models","a good candidate","sequential packet labeling","practical online idses","intrusions","their first 15 packets","one","zero-day","cnn","cnn","95%","just 128","first","15"]},{"title":"Addressing data imbalance challenges in oral cavity histopathological whole slide images with advanced deep learning techniques","authors":["Tabasum Majeed","Tariq Ahmad Masoodi","Muzafar Ahmad Macha","Muzafar Rasool Bhat","Khalid Muzaffar","Assif Assad"],"abstract":"Oral Cavity Squamous Cell Carcinoma (OCSCC) represents a common form of head and neck cancer originating from the mucosal lining of the oral cavity, often detected in advanced stages. Traditional detection methods rely on analyzing hematoxylin and eosin (H&E)-stained histopathological whole-slide images, which are time-consuming and require expert pathology skills. Hence, automated analysis is urgently needed to expedite diagnosis and improve patient outcomes. Deep learning, through automated feature extraction, offers a promising avenue for capturing high-level abstract features with greater accuracy than traditional methods. However, the imbalance in class distribution within datasets significantly affects the performance of deep learning models during training, necessitating specialized approaches. To address the issue, various methods have been proposed at both data and algorithmic levels. This study investigates strategies to mitigate class imbalance by employing a publicly available OCSCC imbalance dataset. We evaluated undersampling methods (Near Miss, Edited Nearest Neighbors) and oversampling techniques (SMOTE, Deep SMOTE, ADASYN) integrated with transfer learning across different imbalance ratios (0.1, 0.15, 0.20, 0.30). Our findings demonstrate the effectiveness of SMOTE in improving test performance, highlighting the efficacy of strategic oversampling combined with transfer learning in classifying imbalanced medical datasets. This enhances OCSCC diagnostic accuracy, streamlines clinical decisions, and reduces reliance on costly histopathological tests.","doi":"10.1007\/s13198-024-02440-6","cleaned_title":"addressing data imbalance challenges in oral cavity histopathological whole slide images with advanced deep learning techniques","cleaned_abstract":"oral cavity squamous cell carcinoma (ocscc) represents a common form of head and neck cancer originating from the mucosal lining of the oral cavity, often detected in advanced stages. traditional detection methods rely on analyzing hematoxylin and eosin (h&e)-stained histopathological whole-slide images, which are time-consuming and require expert pathology skills. hence, automated analysis is urgently needed to expedite diagnosis and improve patient outcomes. deep learning, through automated feature extraction, offers a promising avenue for capturing high-level abstract features with greater accuracy than traditional methods. however, the imbalance in class distribution within datasets significantly affects the performance of deep learning models during training, necessitating specialized approaches. to address the issue, various methods have been proposed at both data and algorithmic levels. this study investigates strategies to mitigate class imbalance by employing a publicly available ocscc imbalance dataset. we evaluated undersampling methods (near miss, edited nearest neighbors) and oversampling techniques (smote, deep smote, adasyn) integrated with transfer learning across different imbalance ratios (0.1, 0.15, 0.20, 0.30). our findings demonstrate the effectiveness of smote in improving test performance, highlighting the efficacy of strategic oversampling combined with transfer learning in classifying imbalanced medical datasets. this enhances ocscc diagnostic accuracy, streamlines clinical decisions, and reduces reliance on costly histopathological tests.","key_phrases":["oral cavity squamous cell carcinoma","ocscc","a common form","head and neck cancer","the mucosal lining","the oral cavity","advanced stages","traditional detection methods","hematoxylin","eosin","histopathological whole-slide images","which","expert pathology skills","automated analysis","diagnosis","patient outcomes","deep learning","automated feature extraction","a promising avenue","high-level abstract features","greater accuracy","traditional methods","the imbalance","class distribution","datasets","the performance","deep learning models","training","specialized approaches","the issue","various methods","both data","algorithmic levels","this study","strategies","class imbalance","a publicly available ocscc imbalance dataset","we","undersampling methods","near miss","neighbors","techniques","smote","deep smote","adasyn","different imbalance ratios","our findings","the effectiveness","smote","test performance","the efficacy","strategic oversampling","imbalanced medical datasets","diagnostic accuracy","clinical decisions","reliance","costly histopathological tests","hematoxylin","0.1","0.15","0.20","0.30"]},{"title":"Deep learning implementation of image segmentation in agricultural applications: a comprehensive review","authors":["Lian Lei","Qiliang Yang","Ling Yang","Tao Shen","Ruoxi Wang","Chengbiao Fu"],"abstract":"Image segmentation is a crucial task in computer vision, which divides a digital image into multiple segments and objects. In agriculture, image segmentation is extensively used for crop and soil monitoring, predicting the best times to sow, fertilize, and harvest, estimating crop yield, and detecting plant diseases. However, image segmentation faces difficulties in agriculture, such as the challenges of disease staging recognition, labeling inconsistency, and changes in plant morphology with the environment. Consequently, we have conducted a comprehensive review of image segmentation techniques based on deep learning, exploring the development and prospects of image segmentation in agriculture. Deep learning-based image segmentation solutions widely used in agriculture are categorized into eight main groups: encoder-decoder structures, multi-scale and pyramid-based methods, dilated convolutional networks, visual attention models, generative adversarial networks, graph neural networks, instance segmentation networks, and transformer-based models. In addition, the applications of image segmentation methods in agriculture are presented, such as plant disease detection, weed identification, crop growth monitoring, crop yield estimation, and counting. Furthermore, a collection of publicly available plant image segmentation datasets has been reviewed, and the evaluation and comparison of performance for image segmentation algorithms have been conducted on benchmark datasets. Finally, there is a discussion of the challenges and future prospects of image segmentation in agriculture.","doi":"10.1007\/s10462-024-10775-6","cleaned_title":"deep learning implementation of image segmentation in agricultural applications: a comprehensive review","cleaned_abstract":"image segmentation is a crucial task in computer vision, which divides a digital image into multiple segments and objects. in agriculture, image segmentation is extensively used for crop and soil monitoring, predicting the best times to sow, fertilize, and harvest, estimating crop yield, and detecting plant diseases. however, image segmentation faces difficulties in agriculture, such as the challenges of disease staging recognition, labeling inconsistency, and changes in plant morphology with the environment. consequently, we have conducted a comprehensive review of image segmentation techniques based on deep learning, exploring the development and prospects of image segmentation in agriculture. deep learning-based image segmentation solutions widely used in agriculture are categorized into eight main groups: encoder-decoder structures, multi-scale and pyramid-based methods, dilated convolutional networks, visual attention models, generative adversarial networks, graph neural networks, instance segmentation networks, and transformer-based models. in addition, the applications of image segmentation methods in agriculture are presented, such as plant disease detection, weed identification, crop growth monitoring, crop yield estimation, and counting. furthermore, a collection of publicly available plant image segmentation datasets has been reviewed, and the evaluation and comparison of performance for image segmentation algorithms have been conducted on benchmark datasets. finally, there is a discussion of the challenges and future prospects of image segmentation in agriculture.","key_phrases":["image segmentation","a crucial task","computer vision","which","a digital image","multiple segments","objects","agriculture","image segmentation","crop","soil monitoring","the best times","crop yield","plant diseases","image segmentation","difficulties","agriculture","the challenges","disease staging recognition","labeling inconsistency","changes","plant morphology","the environment","we","a comprehensive review","image segmentation techniques","deep learning","the development","prospects","image segmentation","agriculture","deep learning-based image segmentation solutions","agriculture","eight main groups","encoder-decoder structures","-scale and pyramid-based methods","dilated convolutional networks","visual attention models","generative adversarial networks","graph neural networks","instance segmentation networks","transformer-based models","addition","the applications","image segmentation methods","agriculture","plant disease detection","identification","crop growth monitoring","crop yield estimation","a collection","publicly available plant image segmentation datasets","the evaluation","comparison","performance","image segmentation algorithms","benchmark datasets","a discussion","the challenges","future prospects","image segmentation","agriculture","eight"]},{"title":"COVID-19 Fake News Detection using Deep Learning Model","authors":["Mahabuba Akhter","Syed Md. Minhaz Hossain","Rizma Sijana Nigar","Srabanti Paul","Khaleque Md. Aashiq Kamal","Anik Sen","Iqbal H. Sarker"],"abstract":"People may now receive and share information more quickly and easily than ever due to the widespread use of mobile networked devices. However, this can occasionally lead to the spread of false information. Such information is being disseminated widely, which may cause people to make incorrect decisions about potentially crucial topics. This occurred in 2020, the year of the fatal and extremely contagious Coronavirus Disease (COVID-19) outbreak. The spread of false information about COVID-19 on social media has already been labeled as an \u201cinfodemic\u201d by the World Health Organization (WHO), causing serious difficulties for governments attempting to control the pandemic. Consequently, it is crucial to have a model for detecting fake news related to COVID-19. In this paper, we present an effective Convolutional Neural Network (CNN)-based deep learning model using word embedding. For selecting the best CNN architecture, we take into account the optimal values of model hyper-parameters using grid search. Further, for measuring the effectiveness of our proposed CNN model, various state-of-the-art machine learning algorithms are conducted for COVID-19 fake news detection. Among them, CNN outperforms with 96.19% mean accuracy, 95% mean F1-score, and 0.985 area under ROC curve (AUC).","doi":"10.1007\/s40745-023-00507-y","cleaned_title":"covid-19 fake news detection using deep learning model","cleaned_abstract":"people may now receive and share information more quickly and easily than ever due to the widespread use of mobile networked devices. however, this can occasionally lead to the spread of false information. such information is being disseminated widely, which may cause people to make incorrect decisions about potentially crucial topics. this occurred in 2020, the year of the fatal and extremely contagious coronavirus disease (covid-19) outbreak. the spread of false information about covid-19 on social media has already been labeled as an \u201cinfodemic\u201d by the world health organization (who), causing serious difficulties for governments attempting to control the pandemic. consequently, it is crucial to have a model for detecting fake news related to covid-19. in this paper, we present an effective convolutional neural network (cnn)-based deep learning model using word embedding. for selecting the best cnn architecture, we take into account the optimal values of model hyper-parameters using grid search. further, for measuring the effectiveness of our proposed cnn model, various state-of-the-art machine learning algorithms are conducted for covid-19 fake news detection. among them, cnn outperforms with 96.19% mean accuracy, 95% mean f1-score, and 0.985 area under roc curve (auc).","key_phrases":["people","information","the widespread use","mobile networked devices","this","the spread","false information","such information","which","people","incorrect decisions","potentially crucial topics","this","the year","the fatal and extremely contagious coronavirus disease","covid-19","the spread","false information","covid-19","social media","the world health organization","who","serious difficulties","governments","it","a model","fake news","covid-19","this paper","we","an effective convolutional neural network","cnn)-based deep learning model","word","the best cnn architecture","we","account","the optimal values","model hyper-parameters","grid search","the effectiveness","our proposed cnn model","the-art","covid-19 fake news detection","them","cnn","96.19%","mean accuracy","95%","0.985 area","roc curve","auc","2020","the year","covid-19","covid-19","the world health organization","covid-19","cnn","cnn","covid-19","cnn","96.19%","95%","0.985","roc"]},{"title":"Model-based deep learning framework for accelerated optical projection tomography","authors":["Marcos Obando","Andrea Bassi","Nicolas Ducros","Germ\u00e1n Mato","Teresa M. Correia"],"abstract":"In this work, we propose a model-based deep learning reconstruction algorithm for optical projection tomography (ToMoDL), to greatly reduce acquisition and reconstruction times. The proposed method iterates over a data consistency step and an image domain artefact removal step achieved by a convolutional neural network. A preprocessing stage is also included to avoid potential misalignments between the sample center of rotation and the detector. The algorithm is trained using a database of wild-type zebrafish (Danio rerio) at different stages of development to minimise the mean square error for a fixed number of iterations. Using a cross-validation scheme, we compare the results to other reconstruction methods, such as filtered backprojection, compressed sensing and a direct deep learning method where the pseudo-inverse solution is corrected by a U-Net. The proposed method performs equally well or better than the alternatives. For a highly reduced number of projections, only the U-Net method provides images comparable to those obtained with ToMoDL. However, ToMoDL has a much better performance if the amount of data available for training is limited, given that the number of network trainable parameters is smaller.","doi":"10.1038\/s41598-023-47650-3","cleaned_title":"model-based deep learning framework for accelerated optical projection tomography","cleaned_abstract":"in this work, we propose a model-based deep learning reconstruction algorithm for optical projection tomography (tomodl), to greatly reduce acquisition and reconstruction times. the proposed method iterates over a data consistency step and an image domain artefact removal step achieved by a convolutional neural network. a preprocessing stage is also included to avoid potential misalignments between the sample center of rotation and the detector. the algorithm is trained using a database of wild-type zebrafish (danio rerio) at different stages of development to minimise the mean square error for a fixed number of iterations. using a cross-validation scheme, we compare the results to other reconstruction methods, such as filtered backprojection, compressed sensing and a direct deep learning method where the pseudo-inverse solution is corrected by a u-net. the proposed method performs equally well or better than the alternatives. for a highly reduced number of projections, only the u-net method provides images comparable to those obtained with tomodl. however, tomodl has a much better performance if the amount of data available for training is limited, given that the number of network trainable parameters is smaller.","key_phrases":["this work","we","a model-based deep learning reconstruction algorithm","optical projection tomography","acquisition and reconstruction times","the proposed method","a data consistency step","an image domain artefact removal step","a convolutional neural network","a preprocessing stage","potential misalignments","the sample center","rotation","the detector","the algorithm","a database","wild-type zebrafish","danio rerio","different stages","development","the mean square error","a fixed number","iterations","a cross-validation scheme","we","the results","other reconstruction methods","filtered backprojection","compressed sensing","a direct deep learning method","the pseudo-inverse solution","a u","-","net","the proposed method","the alternatives","a highly reduced number","projections","only the u-net method","images","those","a much better performance","the amount","data","training","the number","network trainable parameters","danio rerio"]},{"title":"Mammography using low-frequency electromagnetic fields with deep learning","authors":["Hamid Akbari-Chelaresi","Dawood Alsaedi","Seyed Hossein Mirjahanmardi","Mohamed El Badawe","Ali M. Albishi","Vahid Nayyeri","Omar M. Ramahi"],"abstract":"In this paper, a novel technique for detecting female breast anomalous tissues is presented and validated through numerical simulations. The technique, to a high degree, resembles X-ray mammography; however, instead of using X-rays for obtaining images of the breast, low-frequency electromagnetic fields are leveraged. To capture breast impressions, a metasurface, which can be thought of as analogous to X-rays film, has been employed. To achieve deep and sufficient penetration within the breast tissues, the source of excitation is a simple narrow-band dipole antenna operating at 200\u00a0MHz. The metasurface is designed to operate at the same frequency. The detection mechanism is based on comparing the impressions obtained from the breast under examination to the reference case (healthy breasts) using machine learning techniques. Using this system, not only would it be possible to detect tumors (benign or malignant), but one can also determine the location and size of the tumors. Remarkably, deep learning models were found to achieve very high classification accuracy.","doi":"10.1038\/s41598-023-40494-x","cleaned_title":"mammography using low-frequency electromagnetic fields with deep learning","cleaned_abstract":"in this paper, a novel technique for detecting female breast anomalous tissues is presented and validated through numerical simulations. the technique, to a high degree, resembles x-ray mammography; however, instead of using x-rays for obtaining images of the breast, low-frequency electromagnetic fields are leveraged. to capture breast impressions, a metasurface, which can be thought of as analogous to x-rays film, has been employed. to achieve deep and sufficient penetration within the breast tissues, the source of excitation is a simple narrow-band dipole antenna operating at 200 mhz. the metasurface is designed to operate at the same frequency. the detection mechanism is based on comparing the impressions obtained from the breast under examination to the reference case (healthy breasts) using machine learning techniques. using this system, not only would it be possible to detect tumors (benign or malignant), but one can also determine the location and size of the tumors. remarkably, deep learning models were found to achieve very high classification accuracy.","key_phrases":["this paper","a novel technique","female breast anomalous tissues","numerical simulations","the technique","a high degree","x-ray mammography","x","-","rays","images","the breast, low-frequency electromagnetic fields","breast impressions","a metasurface","which","x-rays film","deep and sufficient penetration","the breast tissues","the source","excitation","a simple narrow-band dipole antenna","200 mhz","the metasurface","the same frequency","the detection mechanism","the impressions","the breast","examination","the reference case","healthy breasts","machine learning techniques","this system","it","tumors","one","the location","size","the tumors","deep learning models","very high classification accuracy","dipole antenna","200"]},{"title":"Taxonomy of deep learning-based intrusion detection system approaches in fog computing: a systematic review","authors":["Sepide Najafli","Abolfazl Toroghi Haghighat","Babak Karasfi"],"abstract":"The Internet of Things (IoT) has been used in various aspects. Fundamental security issues must be addressed to accelerate and develop the Internet of Things. An intrusion detection system (IDS) is an essential element in network security designed to detect and determine the type of attacks. The use of deep learning (DL) shows promising results in the design of IDS based on IoT. DL facilitates analytics and learning in the dynamic IoT domain. Some deep learning-based IDS in IOT sensors cannot be executed, because of resource restrictions. Although cloud computing could overcome limitations, the distance between the cloud and the end IoT sensors causes high communication costs, security problems and delays. Fog computing has been presented to handle these issues and can bring resources to the edge of the network. Many studies have been conducted to investigate IDS based on IoT. Our goal is to investigate and classify deep learning-based IDS on fog processing. In this paper, researchers can access comprehensive resources in this field. Therefore, first, we provide a complete classification of IDS in IoT. Then practical and important proposed IDSs in the fog environment are discussed in three groups (binary, multi-class, and hybrid), and are examined the advantages and disadvantages of each approach. The results show that most of the studied methods consider hybrid strategies (binary and multi-class). In addition, in the reviewed papers the average Accuracy obtained in the binary method is better than the multi-class. Finally, we highlight some challenges and future directions for the next research in IDS techniques.","doi":"10.1007\/s10115-024-02162-y","cleaned_title":"taxonomy of deep learning-based intrusion detection system approaches in fog computing: a systematic review","cleaned_abstract":"the internet of things (iot) has been used in various aspects. fundamental security issues must be addressed to accelerate and develop the internet of things. an intrusion detection system (ids) is an essential element in network security designed to detect and determine the type of attacks. the use of deep learning (dl) shows promising results in the design of ids based on iot. dl facilitates analytics and learning in the dynamic iot domain. some deep learning-based ids in iot sensors cannot be executed, because of resource restrictions. although cloud computing could overcome limitations, the distance between the cloud and the end iot sensors causes high communication costs, security problems and delays. fog computing has been presented to handle these issues and can bring resources to the edge of the network. many studies have been conducted to investigate ids based on iot. our goal is to investigate and classify deep learning-based ids on fog processing. in this paper, researchers can access comprehensive resources in this field. therefore, first, we provide a complete classification of ids in iot. then practical and important proposed idss in the fog environment are discussed in three groups (binary, multi-class, and hybrid), and are examined the advantages and disadvantages of each approach. the results show that most of the studied methods consider hybrid strategies (binary and multi-class). in addition, in the reviewed papers the average accuracy obtained in the binary method is better than the multi-class. finally, we highlight some challenges and future directions for the next research in ids techniques.","key_phrases":["the internet","things","iot","various aspects","fundamental security issues","the internet","things","an intrusion detection system","ids","an essential element","network security","the type","attacks","the use","deep learning","dl","results","the design","ids","iot","dl","analytics","the dynamic iot domain","some deep learning-based ids","iot sensors","resource restrictions","cloud computing","limitations","the distance","the cloud","the end iot sensors","high communication costs","security problems","delays","fog computing","these issues","resources","the edge","the network","many studies","ids","iot","our goal","deep learning-based ids","fog processing","this paper","researchers","comprehensive resources","this field","we","a complete classification","ids","iot","practical and important proposed idss","the fog environment","three groups","the advantages","disadvantages","each approach","the results","the studied methods","hybrid strategies","binary and multi-class","addition","the reviewed papers","the average accuracy","the binary method","the multi","-","class","we","some challenges","future directions","the next research","ids techniques","fog computing","first","three"]},{"title":"Deep learning-based pathway-centric approach to characterize recurrent hepatocellular carcinoma after liver transplantation","authors":["Jeffrey To","Soumita Ghosh","Xun Zhao","Elisa Pasini","Sandra Fischer","Gonzalo Sapisochin","Anand Ghanekar","Elmar Jaeckel","Mamatha Bhat"],"abstract":"BackgroundLiver transplantation (LT) is offered as a cure for Hepatocellular carcinoma (HCC), however 15\u201320% develop recurrence post-transplant which tends to be aggressive. In this study, we examined the transcriptome profiles of patients with recurrent HCC to identify differentially expressed genes (DEGs), the involved pathways, biological functions, and potential gene signatures of recurrent HCC post-transplant using deep machine learning (ML) methodology.Materials and methodsWe analyzed the transcriptomic profiles of primary and recurrent tumor samples from 7 pairs of patients who underwent LT. Following differential gene expression analysis, we performed pathway enrichment, gene ontology (GO) analyses and protein-protein interactions (PPIs) with top 10 hub gene networks. We also predicted the landscape of infiltrating immune cells using Cibersortx. We next develop pathway and GO term-based deep learning models leveraging primary tissue gene expression data from The Cancer Genome Atlas (TCGA) to identify gene signatures in recurrent HCC.ResultsThe PI3K\/Akt signaling pathway and cytokine-mediated signaling pathway were particularly activated in HCC recurrence. The recurrent tumors exhibited upregulation of an immune-escape related gene, CD274, in the top 10 hub gene analysis. Significantly higher infiltration of monocytes and lower M1 macrophages were found in recurrent HCC tumors. Our deep learning approach identified a 20-gene signature in recurrent HCC. Amongst the 20 genes, through multiple analysis, IL6 was found to be significantly associated with HCC recurrence.ConclusionOur deep learning approach identified PI3K\/Akt signaling as potentially regulating cytokine-mediated functions and the expression of immune escape genes, leading to alterations in the pattern of immune cell infiltration. In conclusion, IL6 was identified to play an important role in HCC recurrence.","doi":"10.1186\/s40246-024-00624-6","cleaned_title":"deep learning-based pathway-centric approach to characterize recurrent hepatocellular carcinoma after liver transplantation","cleaned_abstract":"backgroundliver transplantation (lt) is offered as a cure for hepatocellular carcinoma (hcc), however 15\u201320% develop recurrence post-transplant which tends to be aggressive. in this study, we examined the transcriptome profiles of patients with recurrent hcc to identify differentially expressed genes (degs), the involved pathways, biological functions, and potential gene signatures of recurrent hcc post-transplant using deep machine learning (ml) methodology.materials and methodswe analyzed the transcriptomic profiles of primary and recurrent tumor samples from 7 pairs of patients who underwent lt. following differential gene expression analysis, we performed pathway enrichment, gene ontology (go) analyses and protein-protein interactions (ppis) with top 10 hub gene networks. we also predicted the landscape of infiltrating immune cells using cibersortx. we next develop pathway and go term-based deep learning models leveraging primary tissue gene expression data from the cancer genome atlas (tcga) to identify gene signatures in recurrent hcc.resultsthe pi3k\/akt signaling pathway and cytokine-mediated signaling pathway were particularly activated in hcc recurrence. the recurrent tumors exhibited upregulation of an immune-escape related gene, cd274, in the top 10 hub gene analysis. significantly higher infiltration of monocytes and lower m1 macrophages were found in recurrent hcc tumors. our deep learning approach identified a 20-gene signature in recurrent hcc. amongst the 20 genes, through multiple analysis, il6 was found to be significantly associated with hcc recurrence.conclusionour deep learning approach identified pi3k\/akt signaling as potentially regulating cytokine-mediated functions and the expression of immune escape genes, leading to alterations in the pattern of immune cell infiltration. in conclusion, il6 was identified to play an important role in hcc recurrence.","key_phrases":["backgroundliver transplantation","lt","a cure","hepatocellular carcinoma","hcc","15\u201320%","recurrence","transplant","which","this study","we","the transcriptome profiles","patients","recurrent hcc","differentially expressed genes","the involved pathways","biological functions","potential gene signatures","recurrent hcc post","-","deep machine learning","(ml) methodology.materials","methodswe","the transcriptomic profiles","primary and recurrent tumor samples","7 pairs","patients","who","lt","differential gene expression analysis","we","pathway enrichment","gene ontology","analyses","protein-protein interactions","top 10 hub gene networks","we","the landscape","infiltrating immune cells","cibersortx","we","term-based deep learning models","primary tissue gene expression data","the cancer genome atlas","(tcga","gene signatures","recurrent hcc.resultsthe pi3k\/akt","pathway","hcc recurrence","the recurrent tumors","upregulation","an immune-escape related gene","cd274","the top 10 hub gene analysis","significantly higher infiltration","monocytes","lower m1 macrophages","recurrent hcc tumors","our deep learning approach","a 20-gene signature","recurrent hcc","the 20 genes","multiple analysis","il6","hcc","recurrence.conclusionour deep learning approach","pi3k\/akt","cytokine-mediated functions","the expression","immune escape genes","alterations","the pattern","immune cell infiltration","conclusion","il6","an important role","hcc recurrence","15\u201320%","7","10","10","20","20"]},{"title":"Deep learning models for perception of brightness related illusions","authors":["Amrita Mukherjee","Avijit Paul","Kuntal Ghosh"],"abstract":"Illusions are like holes in our effortless visual mechanism through which we can peep into the internal mechanisms of the brain. Scientists attempted to explain underlying physiological, physical, and cognitive mechanisms of illusions by the receptive field hierarchical organizations, information sampling, filtering, etc. Some antagonistic illusions cannot be explained by them and for this, deep learning networks were used recently as a model for illusion perception. To further broaden the scope of the perceptual functionality in the brightness contrast genre, handle the background removal effects on some illusions that reduce the illusory effects, and replicate the antagonistic illusions with the same parameter setup, we have used Convolutional Neural Network, Autoencoder, U-Net, and U-Net++ models for replicating the visual illusions. The networks are specialized in low-level vision tasks like De-noising, De-blurring, and a combination of both. A high number of brightness contrast visual illusions are tested on all the networks and most of the outcomes significantly matched human perceptions. Overall, our method will guide the development of neurobiological frameworks which might enrich the computational neuroscience study by distilling some biological principles. On the other hand, the machine learning community will benefit from knowing the inherent flaws of the networks so that the true image of reality can be taken into consideration, especially in imaging situations where experts too can be deceived.","doi":"10.1007\/s10489-024-05658-w","cleaned_title":"deep learning models for perception of brightness related illusions","cleaned_abstract":"illusions are like holes in our effortless visual mechanism through which we can peep into the internal mechanisms of the brain. scientists attempted to explain underlying physiological, physical, and cognitive mechanisms of illusions by the receptive field hierarchical organizations, information sampling, filtering, etc. some antagonistic illusions cannot be explained by them and for this, deep learning networks were used recently as a model for illusion perception. to further broaden the scope of the perceptual functionality in the brightness contrast genre, handle the background removal effects on some illusions that reduce the illusory effects, and replicate the antagonistic illusions with the same parameter setup, we have used convolutional neural network, autoencoder, u-net, and u-net++ models for replicating the visual illusions. the networks are specialized in low-level vision tasks like de-noising, de-blurring, and a combination of both. a high number of brightness contrast visual illusions are tested on all the networks and most of the outcomes significantly matched human perceptions. overall, our method will guide the development of neurobiological frameworks which might enrich the computational neuroscience study by distilling some biological principles. on the other hand, the machine learning community will benefit from knowing the inherent flaws of the networks so that the true image of reality can be taken into consideration, especially in imaging situations where experts too can be deceived.","key_phrases":["illusions","holes","our effortless visual mechanism","which","we","the internal mechanisms","the brain","scientists","underlying physiological, physical, and cognitive mechanisms","illusions","the receptive field hierarchical organizations","information sampling","filtering","some antagonistic illusions","them","this","deep learning networks","a model","illusion perception","the scope","the perceptual functionality","the brightness contrast genre","the background removal effects","some illusions","that","the illusory effects","the antagonistic illusions","the same parameter setup","we","convolutional neural network","autoencoder","-","net","u-net++ models","the visual illusions","the networks","low-level vision tasks","de","de","-","a combination","both","a high number","brightness contrast visual illusions","all the networks","the outcomes","human perceptions","our method","the development","neurobiological frameworks","which","the computational neuroscience study","some biological principles","the other hand","the machine learning community","the inherent flaws","the networks","the true image","reality","consideration","imaging situations","experts"]},{"title":"Exploiting biochemical data to improve osteosarcoma diagnosis with deep learning","authors":["Shidong Wang","Yangyang Shen","Fanwei Zeng","Meng Wang","Bohan Li","Dian Shen","Xiaodong Tang","Beilun Wang"],"abstract":"Early and accurate diagnosis of osteosarcomas (OS) is of great clinical significance, and machine learning (ML) based methods are increasingly adopted. However, current ML-based methods for osteosarcoma diagnosis consider only X-ray images, usually fail to generalize to new cases, and lack explainability. In this paper, we seek to explore the capability of deep learning models in diagnosing primary OS, with higher accuracy, explainability, and generality. Concretely, we analyze the added value of integrating the biochemical data, i.e., alkaline phosphatase (ALP) and lactate dehydrogenase (LDH), and design a model that incorporates the numerical features of ALP and LDH and the visual features of X-ray imaging through a late fusion approach in the feature space. We evaluate this model on real-world clinic data with 848 patients aged from 4 to 81. The experimental results reveal the effectiveness of incorporating ALP and LDH simultaneously in a late fusion approach, with the accuracy of the considered 2608 cases increased to 97.17%, compared to 94.35% in the baseline. Grad-CAM visualizations consistent with orthopedic specialists further justified the model\u2019s explainability.","doi":"10.1007\/s13755-024-00288-5","cleaned_title":"exploiting biochemical data to improve osteosarcoma diagnosis with deep learning","cleaned_abstract":"early and accurate diagnosis of osteosarcomas (os) is of great clinical significance, and machine learning (ml) based methods are increasingly adopted. however, current ml-based methods for osteosarcoma diagnosis consider only x-ray images, usually fail to generalize to new cases, and lack explainability. in this paper, we seek to explore the capability of deep learning models in diagnosing primary os, with higher accuracy, explainability, and generality. concretely, we analyze the added value of integrating the biochemical data, i.e., alkaline phosphatase (alp) and lactate dehydrogenase (ldh), and design a model that incorporates the numerical features of alp and ldh and the visual features of x-ray imaging through a late fusion approach in the feature space. we evaluate this model on real-world clinic data with 848 patients aged from 4 to 81. the experimental results reveal the effectiveness of incorporating alp and ldh simultaneously in a late fusion approach, with the accuracy of the considered 2608 cases increased to 97.17%, compared to 94.35% in the baseline. grad-cam visualizations consistent with orthopedic specialists further justified the model\u2019s explainability.","key_phrases":["early and accurate diagnosis","great clinical significance","machine learning (ml) based methods","current ml-based methods","osteosarcoma diagnosis","only x-ray images","new cases","lack explainability","this paper","we","the capability","deep learning models","higher accuracy","explainability","generality","we","the added value","the biochemical data","i.e., alkaline phosphatase","alp","lactate dehydrogenase","ldh","a model","that","the numerical features","alp","ldh","the visual features","x","-ray imaging","a late fusion approach","the feature space","we","this model","real-world clinic data","848 patients","the experimental results","the effectiveness","alp","ldh","a late fusion approach","the accuracy","the considered 2608 cases","97.17%","94.35%","the baseline","grad-cam visualizations","orthopedic specialists","the model\u2019s explainability","848","4","81","2608","97.17%","94.35%"]},{"title":"Deep learning-based predictive classification of functional subpopulations of hematopoietic stem cells and multipotent progenitors","authors":["Shen Wang","Jianzhong Han","Jingru Huang","Khayrul Islam","Yuheng Shi","Yuyuan Zhou","Dongwook Kim","Jane Zhou","Zhaorui Lian","Yaling Liu","Jian Huang"],"abstract":"BackgroundHematopoietic stem cells (HSCs) and multipotent progenitors (MPPs) play a pivotal role in maintaining lifelong hematopoiesis. The distinction between stem cells and other progenitors, as well as the assessment of their functions, has long been a central focus in stem cell research. In recent years, deep learning has emerged as a powerful tool for cell image analysis and classification\/prediction.MethodsIn this study, we explored the feasibility of employing deep learning techniques to differentiate murine HSCs and MPPs based solely on their morphology, as observed through light microscopy (DIC) images.ResultsAfter rigorous training and validation using extensive image datasets, we successfully developed a three-class classifier, referred to as the LSM model, capable of reliably distinguishing long-term HSCs, short-term HSCs, and MPPs. The LSM model extracts intrinsic morphological features unique to different cell types, irrespective of the methods used for cell identification and isolation, such as surface markers or intracellular GFP markers. Furthermore, employing the same deep learning framework, we created a two-class classifier that effectively discriminates between aged HSCs and young HSCs. This discovery is particularly significant as both cell types share identical surface markers yet serve distinct functions. This classifier holds the potential to offer a novel, rapid, and efficient means of assessing the functional states of HSCs, thus obviating the need for time-consuming transplantation experiments.ConclusionOur study represents the pioneering use of deep learning to differentiate HSCs and MPPs under steady-state conditions. This novel and robust deep learning-based platform will provide a basis for the future development of a new generation stem cell identification and separation system. It may also provide new insight into the molecular mechanisms underlying stem cell self-renewal.","doi":"10.1186\/s13287-024-03682-8","cleaned_title":"deep learning-based predictive classification of functional subpopulations of hematopoietic stem cells and multipotent progenitors","cleaned_abstract":"backgroundhematopoietic stem cells (hscs) and multipotent progenitors (mpps) play a pivotal role in maintaining lifelong hematopoiesis. the distinction between stem cells and other progenitors, as well as the assessment of their functions, has long been a central focus in stem cell research. in recent years, deep learning has emerged as a powerful tool for cell image analysis and classification\/prediction.methodsin this study, we explored the feasibility of employing deep learning techniques to differentiate murine hscs and mpps based solely on their morphology, as observed through light microscopy (dic) images.resultsafter rigorous training and validation using extensive image datasets, we successfully developed a three-class classifier, referred to as the lsm model, capable of reliably distinguishing long-term hscs, short-term hscs, and mpps. the lsm model extracts intrinsic morphological features unique to different cell types, irrespective of the methods used for cell identification and isolation, such as surface markers or intracellular gfp markers. furthermore, employing the same deep learning framework, we created a two-class classifier that effectively discriminates between aged hscs and young hscs. this discovery is particularly significant as both cell types share identical surface markers yet serve distinct functions. this classifier holds the potential to offer a novel, rapid, and efficient means of assessing the functional states of hscs, thus obviating the need for time-consuming transplantation experiments.conclusionour study represents the pioneering use of deep learning to differentiate hscs and mpps under steady-state conditions. this novel and robust deep learning-based platform will provide a basis for the future development of a new generation stem cell identification and separation system. it may also provide new insight into the molecular mechanisms underlying stem cell self-renewal.","key_phrases":["backgroundhematopoietic stem cells","hscs","multipotent progenitors","mpps","a pivotal role","lifelong hematopoiesis","the distinction","stem cells","other progenitors","the assessment","their functions","a central focus","stem cell research","recent years","deep learning","a powerful tool","cell image analysis","classification\/prediction.methodsin","we","the feasibility","deep learning techniques","murine hscs","mpps","their morphology","light microscopy (dic) images.resultsafter rigorous training","validation","extensive image datasets","we","a three-class classifier","the lsm model","reliably distinguishing long-term hscs","short-term hscs","mpps","the lsm model","intrinsic morphological features","different cell types","the methods","cell identification","isolation","surface markers","intracellular gfp markers","the same deep learning framework","we","a two-class classifier","that","aged hscs","young hscs","this discovery","both cell types","identical surface markers","distinct functions","this classifier","the potential","a novel, rapid, and efficient means","the functional states","hscs","the need","time-consuming transplantation","experiments.conclusionour study","the pioneering use","deep learning","hscs","mpps","steady-state conditions","this novel and robust deep learning-based platform","a basis","the future development","a new generation stem cell identification","separation system","it","new insight","the molecular mechanisms","stem cell self-renewal","recent years","murine hscs","three","two"]},{"title":"Deep source transfer learning for the estimation of internal brain dynamics using scalp EEG","authors":["Haitao Yu","Zhiwen Hu","Quanfa Zhao","Jing Liu"],"abstract":"Electroencephalography (EEG) provides high temporal resolution neural data for brain-computer interfacing via noninvasive electrophysiological recording. Estimating the internal brain activity by means of source imaging techniques can further improve the spatial resolution of EEG and enhance the reliability of neural decoding and brain-computer interaction. In this work, we propose a novel EEG data-driven source imaging scheme for precise and efficient estimation of macroscale spatiotemporal brain dynamics across thalamus and cortical regions with deep learning methods. A deep source imaging framework with a convolutional-recurrent neural network is designed to estimate the internal brain dynamics from high-density EEG recordings. Moreover, a brain model including 210 cortical regions and 16 thalamic nuclei is established based on human brain connectome to provide synthetic training data, which manifests intrinsic characteristics of underlying brain dynamics in spontaneous, stimulation-evoked, and pathological states. Transfer learning algorithm is further applied to the trained network to reduce the dynamical differences between synthetic and realistic EEG. Extensive experiments exhibit that the proposed deep-learning method can accurately estimate the spatial and temporal activity of brain sources and achieves superior performance compared to the state-of-the-art approaches. Moreover, the EEG data-driven source imaging framework is effective in the location of seizure onset zone in epilepsy and reconstruction of dynamical thalamocortical interactions during sensory processing of acupuncture stimulation, implying its applicability in brain-computer interfacing for neuroscience research and clinical applications.","doi":"10.1007\/s11571-024-10149-2","cleaned_title":"deep source transfer learning for the estimation of internal brain dynamics using scalp eeg","cleaned_abstract":"electroencephalography (eeg) provides high temporal resolution neural data for brain-computer interfacing via noninvasive electrophysiological recording. estimating the internal brain activity by means of source imaging techniques can further improve the spatial resolution of eeg and enhance the reliability of neural decoding and brain-computer interaction. in this work, we propose a novel eeg data-driven source imaging scheme for precise and efficient estimation of macroscale spatiotemporal brain dynamics across thalamus and cortical regions with deep learning methods. a deep source imaging framework with a convolutional-recurrent neural network is designed to estimate the internal brain dynamics from high-density eeg recordings. moreover, a brain model including 210 cortical regions and 16 thalamic nuclei is established based on human brain connectome to provide synthetic training data, which manifests intrinsic characteristics of underlying brain dynamics in spontaneous, stimulation-evoked, and pathological states. transfer learning algorithm is further applied to the trained network to reduce the dynamical differences between synthetic and realistic eeg. extensive experiments exhibit that the proposed deep-learning method can accurately estimate the spatial and temporal activity of brain sources and achieves superior performance compared to the state-of-the-art approaches. moreover, the eeg data-driven source imaging framework is effective in the location of seizure onset zone in epilepsy and reconstruction of dynamical thalamocortical interactions during sensory processing of acupuncture stimulation, implying its applicability in brain-computer interfacing for neuroscience research and clinical applications.","key_phrases":["electroencephalography","(eeg","high temporal resolution neural data","brain-computer","noninvasive electrophysiological recording","the internal brain activity","means","source","imaging techniques","the spatial resolution","eeg","the reliability","neural decoding and brain-computer interaction","this work","we","a novel eeg data-driven source","scheme","precise and efficient estimation","macroscale spatiotemporal brain dynamics","thalamus and cortical regions","deep learning methods","a deep source imaging framework","a convolutional-recurrent neural network","the internal brain dynamics","high-density eeg recordings","a brain model","210 cortical regions","16 thalamic nuclei","human brain connectome","synthetic training data","which","intrinsic characteristics","underlying brain dynamics","spontaneous, stimulation-evoked, and pathological states","transfer learning algorithm","the trained network","the dynamical differences","synthetic and realistic eeg","extensive experiments","the proposed deep-learning method","the spatial and temporal activity","brain sources","superior performance","the-art","the eeg data-driven source imaging framework","the location","seizure onset zone","epilepsy","reconstruction","dynamical thalamocortical interactions","sensory processing","acupuncture stimulation","its applicability","brain-computer","neuroscience research and clinical applications","electroencephalography","210","16"]},{"title":"Fully dynamic reorder policies with deep reinforcement learning for multi-echelon inventory management","authors":["Patric Hammler","Nicolas Riesterer","Torsten Braun"],"abstract":"The operation of inventory systems plays an important role in the success of manufacturing companies, making it a\u00a0highly relevant domain for optimization. In particular, the domain lends itself to being approached via Deep Reinforcement Learning (DRL) models due to it requiring sequential reorder decisions based on uncertainty to minimize cost. In this paper, we evaluate state-of-the-art optimization approaches to determine whether Deep Reinforcement Learning can be applied to the multi-echelon inventory optimization (MEIO) framework in a\u00a0practically feasible manner to generate fully dynamic reorder policies. We investigate how it performs in comparison to an optimized static reorder policy, how robust it is when it comes to structural changes in the environment, and whether the use of DRL is safe in terms of risk in real-world applications. Our results show promising performance for DRL with potential for improvement in terms of minimizing risky behavior.","doi":"10.1007\/s00287-023-01556-6","cleaned_title":"fully dynamic reorder policies with deep reinforcement learning for multi-echelon inventory management","cleaned_abstract":"the operation of inventory systems plays an important role in the success of manufacturing companies, making it a highly relevant domain for optimization. in particular, the domain lends itself to being approached via deep reinforcement learning (drl) models due to it requiring sequential reorder decisions based on uncertainty to minimize cost. in this paper, we evaluate state-of-the-art optimization approaches to determine whether deep reinforcement learning can be applied to the multi-echelon inventory optimization (meio) framework in a practically feasible manner to generate fully dynamic reorder policies. we investigate how it performs in comparison to an optimized static reorder policy, how robust it is when it comes to structural changes in the environment, and whether the use of drl is safe in terms of risk in real-world applications. our results show promising performance for drl with potential for improvement in terms of minimizing risky behavior.","key_phrases":["the operation","inventory systems","an important role","the success","manufacturing companies","it","optimization","the domain","itself","deep reinforcement learning (drl) models","it","sequential reorder decisions","uncertainty","cost","this paper","we","the-art","deep reinforcement learning","the multi-echelon inventory optimization","meio) framework","a practically feasible manner","fully dynamic reorder policies","we","it","comparison","an optimized static reorder policy","it","it","structural changes","the environment","the use","drl","terms","risk","real-world applications","our results","promising performance","drl","potential","improvement","terms","risky behavior"]},{"title":"New hybrid deep learning models for multi-target NILM disaggregation","authors":["Jamila Ouzine","Manal Marzouq","Saad Dosse Bennani","Khadija Lahrech","Hakim EL Fadili"],"abstract":"Non-Intrusive Load Monitoring (NILM) technique or energy disaggregation is a technique used to detect the appliance\u2019s states and estimate their individual energy consumption, given the aggregated data through the main smart meter. Indeed, energy efficiency is the main goal of the NILM techniques, which can be achieved by providing energy disaggregation feedback to the consumers. Unlike single models where training must be performed for each appliance, this work proposes multi-target disaggregation which is more appropriate due to the drastic reduction of resources when training is performed for all target appliances simultaneously. For this purpose, new hybrid models are proposed by combining well-known deep learning models: Convolutional Neural Network (CNN), Denoising Autoencoder (DAE), Recurrent Neural Network (RNN), and Long Short-Term Memory network (LSTM). An implementation and detailed comparative study is then suggested between the proposed hybrid deep learning models and conventional single models in terms of various performance metrics on the UK-Domestic Appliance-Level Electricity (UKDALE) benchmarking database. The experimental results show that the proposed hybrid models provide the best disaggregation performances for multi-target disaggregation compared to single models. Specifically, the CNN-LSTM and the DAE-LSTM are the best hybrid models with the highest overall F1-score of 78.90% and 72.94% respectively.","doi":"10.1007\/s12053-023-10161-1","cleaned_title":"new hybrid deep learning models for multi-target nilm disaggregation","cleaned_abstract":"non-intrusive load monitoring (nilm) technique or energy disaggregation is a technique used to detect the appliance\u2019s states and estimate their individual energy consumption, given the aggregated data through the main smart meter. indeed, energy efficiency is the main goal of the nilm techniques, which can be achieved by providing energy disaggregation feedback to the consumers. unlike single models where training must be performed for each appliance, this work proposes multi-target disaggregation which is more appropriate due to the drastic reduction of resources when training is performed for all target appliances simultaneously. for this purpose, new hybrid models are proposed by combining well-known deep learning models: convolutional neural network (cnn), denoising autoencoder (dae), recurrent neural network (rnn), and long short-term memory network (lstm). an implementation and detailed comparative study is then suggested between the proposed hybrid deep learning models and conventional single models in terms of various performance metrics on the uk-domestic appliance-level electricity (ukdale) benchmarking database. the experimental results show that the proposed hybrid models provide the best disaggregation performances for multi-target disaggregation compared to single models. specifically, the cnn-lstm and the dae-lstm are the best hybrid models with the highest overall f1-score of 78.90% and 72.94% respectively.","key_phrases":["non-intrusive load monitoring (nilm) technique or energy disaggregation","a technique","the appliance\u2019s states","their individual energy consumption","the aggregated data","the main smart meter","energy efficiency","the main goal","the nilm techniques","which","energy disaggregation feedback","the consumers","single models","training","each appliance","this work","multi-target disaggregation","which","the drastic reduction","resources","training","all target appliances","this purpose","new hybrid models","well-known deep learning models","convolutional neural network","cnn","autoencoder","dae","recurrent neural network","rnn","long short-term memory network","lstm","an implementation","detailed comparative study","the proposed hybrid deep learning models","conventional single models","terms","various performance metrics","the uk-domestic appliance-level electricity","(ukdale","benchmarking database","the experimental results","the proposed hybrid models","the best disaggregation performances","multi-target disaggregation","single models","the cnn-lstm","the dae-lstm","the best hybrid models","the highest overall f1-score","78.90%","72.94%","cnn","dae","uk","cnn","78.90%","72.94%"]},{"title":"Identifying depression in the United States veterans using deep learning algorithms, NHANES 2005\u20132018","authors":["Zihan Qu","Yashan Wang","Dingjie Guo","Guangliang He","Chuanying Sui","Yuqing Duan","Xin Zhang","Linwei Lan","Hengyu Meng","Yajing Wang","Xin Liu"],"abstract":"BackgroundDepression is a common mental health problem among veterans, with high mortality. Despite the numerous conducted investigations, the prediction and identification of risk factors for depression are still severely limited. This study used a deep learning algorithm to identify depression in veterans and its factors associated with clinical manifestations.MethodsOur data originated from the National Health and Nutrition Examination Survey (2005\u20132018). A dataset of 2,546 veterans was identified using deep learning and five traditional machine learning algorithms with 10-fold cross-validation. Model performance was assessed by examining the area under the subject operating characteristic curve (AUC), accuracy, recall, specificity, precision, and F1 score.ResultsDeep learning had the highest AUC (0.891, 95%CI 0.869\u20130.914) and specificity (0.906) in identifying depression in veterans. Further study on depression among veterans of different ages showed that the AUC values for deep learning were 0.929 (95%CI 0.904\u20130.955) in the middle-aged group and 0.924(95%CI 0.900-0.948) in the older age group. In addition to general health conditions, sleep difficulties, memory impairment, work incapacity, income, BMI, and chronic diseases, factors such as vitamins E and C, and palmitic acid were also identified as important influencing factors.ConclusionsCompared with traditional machine learning methods, deep learning algorithms achieved optimal performance, making it conducive for identifying depression and its risk factors among veterans.","doi":"10.1186\/s12888-023-05109-9","cleaned_title":"identifying depression in the united states veterans using deep learning algorithms, nhanes 2005\u20132018","cleaned_abstract":"backgrounddepression is a common mental health problem among veterans, with high mortality. despite the numerous conducted investigations, the prediction and identification of risk factors for depression are still severely limited. this study used a deep learning algorithm to identify depression in veterans and its factors associated with clinical manifestations.methodsour data originated from the national health and nutrition examination survey (2005\u20132018). a dataset of 2,546 veterans was identified using deep learning and five traditional machine learning algorithms with 10-fold cross-validation. model performance was assessed by examining the area under the subject operating characteristic curve (auc), accuracy, recall, specificity, precision, and f1 score.resultsdeep learning had the highest auc (0.891, 95%ci 0.869\u20130.914) and specificity (0.906) in identifying depression in veterans. further study on depression among veterans of different ages showed that the auc values for deep learning were 0.929 (95%ci 0.904\u20130.955) in the middle-aged group and 0.924(95%ci 0.900-0.948) in the older age group. in addition to general health conditions, sleep difficulties, memory impairment, work incapacity, income, bmi, and chronic diseases, factors such as vitamins e and c, and palmitic acid were also identified as important influencing factors.conclusionscompared with traditional machine learning methods, deep learning algorithms achieved optimal performance, making it conducive for identifying depression and its risk factors among veterans.","key_phrases":["backgrounddepression","a common mental health problem","veterans","high mortality","the numerous conducted investigations","the prediction","identification","risk factors","depression","this study","a deep learning algorithm","depression","veterans","its factors","clinical manifestations.methodsour data","the national health and nutrition examination survey","a dataset","2,546 veterans","deep learning","five traditional machine learning algorithms","10-fold cross","validation","model performance","the area","the subject operating characteristic curve","auc","accuracy","recall","specificity","precision","f1 score.resultsdeep learning","the highest auc","specificity","depression","veterans","further study","depression","veterans","different ages","the auc values","deep learning","the middle-aged group","0.924(95%ci","the older age group","addition","general health conditions","sleep difficulties","memory impairment","work incapacity","income","bmi","chronic diseases","factors","vitamins e","c","palmitic acid","important influencing","traditional machine learning methods","deep learning algorithms","optimal performance","it","depression","its risk factors","veterans","2005\u20132018","2,546","five","10-fold","0.891","95%ci 0.869\u20130.914","0.906","0.929","95%ci 0.904\u20130.955","0.900-0.948"]},{"title":"Designing observables for measurements with deep learning","authors":["Owen Long","Benjamin Nachman"],"abstract":"Many analyses in particle and nuclear physics use simulations to infer fundamental, effective, or phenomenological parameters of the underlying physics models. When the inference is performed with unfolded cross sections, the observables are designed using physics intuition and heuristics. We propose to design targeted observables with machine learning. Unfolded, differential cross sections in a neural network output contain the most information about parameters of interest and can be well-measured by construction. The networks are trained using a custom loss function that rewards outputs that are sensitive to the parameter(s) of interest while simultaneously penalizing outputs that are different between particle-level and detector-level (to minimize detector distortions). We demonstrate this idea in simulation using two physics models for inclusive measurements in deep inelastic scattering. We find that the new approach is more sensitive than classical observables at distinguishing the two models and also has a reduced unfolding uncertainty due to the reduced detector distortions.","doi":"10.1140\/epjc\/s10052-024-13135-4","cleaned_title":"designing observables for measurements with deep learning","cleaned_abstract":"many analyses in particle and nuclear physics use simulations to infer fundamental, effective, or phenomenological parameters of the underlying physics models. when the inference is performed with unfolded cross sections, the observables are designed using physics intuition and heuristics. we propose to design targeted observables with machine learning. unfolded, differential cross sections in a neural network output contain the most information about parameters of interest and can be well-measured by construction. the networks are trained using a custom loss function that rewards outputs that are sensitive to the parameter(s) of interest while simultaneously penalizing outputs that are different between particle-level and detector-level (to minimize detector distortions). we demonstrate this idea in simulation using two physics models for inclusive measurements in deep inelastic scattering. we find that the new approach is more sensitive than classical observables at distinguishing the two models and also has a reduced unfolding uncertainty due to the reduced detector distortions.","key_phrases":["many analyses","particle","nuclear physics","simulations","fundamental, effective, or phenomenological parameters","the underlying physics models","the inference","unfolded cross sections","the observables","physics intuition","heuristics","we","targeted observables","machine learning","unfolded, differential cross sections","a neural network output","the most information","parameters","interest","construction","the networks","a custom loss function","that","outputs","that","the parameter(s","interest","outputs","that","particle-level","detector-level","detector distortions","we","this idea","simulation","two physics models","inclusive measurements","deep inelastic scattering","we","the new approach","classical observables","the two models","a reduced unfolding uncertainty","the reduced detector distortions","two","two"]},{"title":"DEL-Thyroid: deep ensemble learning framework for detection of thyroid cancer progression through genomic mutation","authors":["Asghar Ali Shah","Ali Daud","Amal Bukhari","Bader Alshemaimri","Muhammad Ahsan","Rehmana Younis"],"abstract":"Genes, expressed as sequences of nucleotides, are susceptible to mutations, some of which can lead to cancer. Machine learning and deep learning methods have emerged as vital tools in identifying mutations associated with cancer. Thyroid cancer ranks as the 5th most prevalent cancer in the USA, with thousands diagnosed annually. This paper presents an ensemble learning model leveraging deep learning techniques such as Long Short-Term Memory (LSTM), Gated Recurrent Units (GRUs), and Bi-directional LSTM (Bi-LSTM) to detect thyroid cancer mutations early. The model is trained on a dataset sourced from asia.ensembl.org and IntOGen.org, consisting of 633 samples with 969 mutations across 41 genes, collected from individuals of various demographics. Feature extraction encompasses techniques including Hahn moments, central moments, raw moments, and various matrix-based methods. Evaluation employs three\u00a0testing methods: self-consistency test (SCT), independent set test (IST), and 10-fold cross-validation test (10-FCVT). The proposed ensemble learning model demonstrates promising performance, achieving 96% accuracy in the independent set test (IST). Statistical measures such as training accuracy, testing accuracy, recall, sensitivity, specificity, Mathew's Correlation Coefficient (MCC), loss, training accuracy, F1 Score, and Cohen's kappa are utilized for comprehensive evaluation.","doi":"10.1186\/s12911-024-02604-1","cleaned_title":"del-thyroid: deep ensemble learning framework for detection of thyroid cancer progression through genomic mutation","cleaned_abstract":"genes, expressed as sequences of nucleotides, are susceptible to mutations, some of which can lead to cancer. machine learning and deep learning methods have emerged as vital tools in identifying mutations associated with cancer. thyroid cancer ranks as the 5th most prevalent cancer in the usa, with thousands diagnosed annually. this paper presents an ensemble learning model leveraging deep learning techniques such as long short-term memory (lstm), gated recurrent units (grus), and bi-directional lstm (bi-lstm) to detect thyroid cancer mutations early. the model is trained on a dataset sourced from asia.ensembl.org and intogen.org, consisting of 633 samples with 969 mutations across 41 genes, collected from individuals of various demographics. feature extraction encompasses techniques including hahn moments, central moments, raw moments, and various matrix-based methods. evaluation employs three testing methods: self-consistency test (sct), independent set test (ist), and 10-fold cross-validation test (10-fcvt). the proposed ensemble learning model demonstrates promising performance, achieving 96% accuracy in the independent set test (ist). statistical measures such as training accuracy, testing accuracy, recall, sensitivity, specificity, mathew's correlation coefficient (mcc), loss, training accuracy, f1 score, and cohen's kappa are utilized for comprehensive evaluation.","key_phrases":["genes","sequences","nucleotides","mutations","some","which","cancer","machine learning","deep learning methods","vital tools","mutations","cancer","thyroid cancer","the 5th most prevalent cancer","the usa","thousands","this paper","an ensemble learning model","deep learning techniques","long short-term memory","lstm","gated recurrent units","grus","bi-directional lstm","bi","-","lstm","thyroid cancer mutations","the model","a dataset","intogen.org","633 samples","969 mutations","41 genes","individuals","various demographics","feature extraction","techniques","hahn moments","central moments","raw moments","various matrix-based methods","evaluation","three testing methods","self-consistency test","sct","independent set test","ist","10-fold cross-validation test","the proposed ensemble learning model","performance","96% accuracy","the independent set test","(ist","statistical measures","training accuracy","testing accuracy","recall","sensitivity","specificity","mathew's correlation coefficient","mcc","loss","training accuracy","f1 score","cohen's kappa","comprehensive evaluation","5th","thousands","annually","633","969","41","three","sct","10-fold","10-fcvt","96%","cohen"]},{"title":"Advancing horizons in remote sensing: a comprehensive survey of deep learning models and applications in image classification and beyond","authors":["Sidike Paheding","Ashraf Saleem","Mohammad Faridul Haque Siddiqui","Nathir Rawashdeh","Almabrok Essa","Abel A. Reyes"],"abstract":"In recent years, deep learning has significantly reshaped numerous fields and applications, fundamentally altering how we tackle a variety of challenges. Areas such as natural language processing (NLP), computer vision, healthcare, network security, wide-area surveillance, and precision agriculture have leveraged the merits of the deep learning era. Particularly, deep learning has significantly improved the analysis of remote sensing images, with a\u00a0continuous increase in the number of researchers and\u00a0contributions to the field. The high impact of deep learning development is complemented by rapid advancements and the availability of data from a variety of sensors, including high-resolution RGB, thermal, LiDAR, and multi-\/hyperspectral cameras, as well as emerging sensing platforms such as satellites and aerial vehicles that can be captured by multi-temporal, multi-sensor, and sensing devices with a wider view. This study aims to present an extensive survey that encapsulates widely used deep learning strategies for tackling image classification challenges in remote sensing. It encompasses an exploration of remote sensing imaging platforms, sensor varieties, practical applications, and prospective developments in the field.","doi":"10.1007\/s00521-024-10165-7","cleaned_title":"advancing horizons in remote sensing: a comprehensive survey of deep learning models and applications in image classification and beyond","cleaned_abstract":"in recent years, deep learning has significantly reshaped numerous fields and applications, fundamentally altering how we tackle a variety of challenges. areas such as natural language processing (nlp), computer vision, healthcare, network security, wide-area surveillance, and precision agriculture have leveraged the merits of the deep learning era. particularly, deep learning has significantly improved the analysis of remote sensing images, with a continuous increase in the number of researchers and contributions to the field. the high impact of deep learning development is complemented by rapid advancements and the availability of data from a variety of sensors, including high-resolution rgb, thermal, lidar, and multi-\/hyperspectral cameras, as well as emerging sensing platforms such as satellites and aerial vehicles that can be captured by multi-temporal, multi-sensor, and sensing devices with a wider view. this study aims to present an extensive survey that encapsulates widely used deep learning strategies for tackling image classification challenges in remote sensing. it encompasses an exploration of remote sensing imaging platforms, sensor varieties, practical applications, and prospective developments in the field.","key_phrases":["recent years","deep learning","numerous fields","applications","we","a variety","challenges","areas","natural language processing","nlp","computer vision","healthcare","network security","wide-area surveillance","precision agriculture","the merits","the deep learning era","deep learning","the analysis","remote sensing images","a continuous increase","the number","researchers","contributions","the field","the high impact","deep learning development","rapid advancements","the availability","data","a variety","sensors","high-resolution rgb, thermal, lidar, and multi-\/hyperspectral cameras, as well as emerging sensing platforms","satellites","aerial vehicles","that","-","sensor","devices","a wider view","this study","an extensive survey","that","widely used deep learning strategies","image classification challenges","remote sensing","it","an exploration","remote sensing imaging platforms","sensor varieties","practical applications","prospective developments","the field","recent years","healthcare"]},{"title":"Boosting in-transit entertainment: deep reinforcement learning for intelligent multimedia caching in bus networks","authors":["Dan Lan","Incheol Shin"],"abstract":"Multimedia content delivery in advanced networks faces exponential growth in data volumes, rendering existing solutions obsolete. This research investigates deep reinforcement learning (DRL) for autonomous optimization without extensive datasets. The work analyzes two prominent DRL algorithms, i.e., Dueling Deep Q-Network (DDQN) and Deep Q-Network (DQN) for multimedia delivery in simulated bus networks. DDQN utilizes a novel \u201cdueling\u201d architecture to estimate state value and action advantages, accelerating learning separately. DQN employs deep neural networks to approximate optimal policies. The environment simulates urban buses with passenger file requests and cache sizes modeled on actual data. Comparative analysis evaluates cumulative rewards and losses over 1500 training episodes to analyze learning efficiency, stability, and performance. Results demonstrate DDQN\u2019s superior convergence and 32% higher cumulative rewards than DQN. However, DQN showed potential for gains over successive runs despite inconsistencies. It establishes DRL\u2019s promise for automated decision-making while revealing enhancements to improve DQN. Further research should evaluate generalizability across problem domains, investigate hybrid models, and test physical systems. DDQN emerged as the most efficient algorithm, highlighting DRL\u2019s potential to enable intelligent agents that optimize multimedia delivery.","doi":"10.1007\/s00500-023-09354-8","cleaned_title":"boosting in-transit entertainment: deep reinforcement learning for intelligent multimedia caching in bus networks","cleaned_abstract":"multimedia content delivery in advanced networks faces exponential growth in data volumes, rendering existing solutions obsolete. this research investigates deep reinforcement learning (drl) for autonomous optimization without extensive datasets. the work analyzes two prominent drl algorithms, i.e., dueling deep q-network (ddqn) and deep q-network (dqn) for multimedia delivery in simulated bus networks. ddqn utilizes a novel \u201cdueling\u201d architecture to estimate state value and action advantages, accelerating learning separately. dqn employs deep neural networks to approximate optimal policies. the environment simulates urban buses with passenger file requests and cache sizes modeled on actual data. comparative analysis evaluates cumulative rewards and losses over 1500 training episodes to analyze learning efficiency, stability, and performance. results demonstrate ddqn\u2019s superior convergence and 32% higher cumulative rewards than dqn. however, dqn showed potential for gains over successive runs despite inconsistencies. it establishes drl\u2019s promise for automated decision-making while revealing enhancements to improve dqn. further research should evaluate generalizability across problem domains, investigate hybrid models, and test physical systems. ddqn emerged as the most efficient algorithm, highlighting drl\u2019s potential to enable intelligent agents that optimize multimedia delivery.","key_phrases":["multimedia content delivery","advanced networks","exponential growth","data volumes","this research investigates","deep reinforcement learning","drl","autonomous optimization","extensive datasets","the work","two prominent drl algorithms","deep q-network (ddqn","deep q-network","dqn","multimedia delivery","simulated bus networks","ddqn","a novel \u201cdueling\u201d architecture","state value","action advantages","dqn","deep neural networks","optimal policies","the environment","urban buses","passenger file requests","cache sizes","actual data","comparative analysis","cumulative rewards","losses","1500 training episodes","learning efficiency","stability","performance","results","ddqn\u2019s superior convergence","32% higher cumulative rewards","dqn","dqn","potential","gains","successive runs","inconsistencies","it","drl\u2019s promise","automated decision-making","enhancements","dqn","further research","generalizability","problem domains","hybrid models","physical systems","ddqn","the most efficient algorithm","drl\u2019s potential","intelligent agents","that","multimedia delivery","two","1500","32%"]},{"title":"Parameter-Free Reduction of the Estimation Bias in Deep Reinforcement Learning for Deterministic Policy Gradients","authors":["Baturay Saglam","Furkan Burak Mutlu","Dogan Can Cicek","Suleyman Serdar Kozat"],"abstract":"Approximation of the value functions in value-based deep reinforcement learning induces overestimation bias, resulting in suboptimal policies. We show that when the reinforcement signals received by the agents have a high variance, deep actor-critic approaches that overcome the overestimation bias lead to a substantial underestimation bias. We first address the detrimental issues in the existing approaches that aim to overcome such underestimation error. Then, through extensive statistical analysis, we introduce a novel, parameter-free Deep Q-learning variant to reduce this underestimation bias in deterministic policy gradients. By sampling the weights of a linear combination of two approximate critics from a highly shrunk estimation bias interval, our Q-value update rule is not affected by the variance of the rewards received by the agents throughout learning. We test the performance of the introduced improvement on a set of MuJoCo and Box2D continuous control tasks and demonstrate that it outperforms the existing approaches and improves the baseline actor-critic algorithm in most of the environments tested.","doi":"10.1007\/s11063-024-11461-y","cleaned_title":"parameter-free reduction of the estimation bias in deep reinforcement learning for deterministic policy gradients","cleaned_abstract":"approximation of the value functions in value-based deep reinforcement learning induces overestimation bias, resulting in suboptimal policies. we show that when the reinforcement signals received by the agents have a high variance, deep actor-critic approaches that overcome the overestimation bias lead to a substantial underestimation bias. we first address the detrimental issues in the existing approaches that aim to overcome such underestimation error. then, through extensive statistical analysis, we introduce a novel, parameter-free deep q-learning variant to reduce this underestimation bias in deterministic policy gradients. by sampling the weights of a linear combination of two approximate critics from a highly shrunk estimation bias interval, our q-value update rule is not affected by the variance of the rewards received by the agents throughout learning. we test the performance of the introduced improvement on a set of mujoco and box2d continuous control tasks and demonstrate that it outperforms the existing approaches and improves the baseline actor-critic algorithm in most of the environments tested.","key_phrases":["approximation","the value functions","value-based deep reinforcement learning","overestimation bias","suboptimal policies","we","the reinforcement signals","the agents","a high variance","deep actor-critic approaches","that","the overestimation bias","a substantial underestimation bias","we","the detrimental issues","the existing approaches","that","such underestimation error","extensive statistical analysis","we","a novel, parameter-free deep q-learning variant","this underestimation bias","deterministic policy gradients","the weights","a linear combination","two approximate critics","a highly shrunk estimation bias interval","our q-value update rule","the variance","the rewards","the agents","we","the performance","the introduced improvement","a set","mujoco","box2d continuous control tasks","it","the existing approaches","the baseline actor-critic algorithm","the environments","first","two"]},{"title":"Advanced deep learning and large language models for suicide ideation detection on social media","authors":["Mohammed Qorich","Rajae El Ouazzani"],"abstract":"Recently, suicide ideations represent a worldwide health concern and pose many anticipation challenges. Actually, the prevalence of expressing self-destructive thoughts especially on forums and social media requires effective monitoring for suicide prevention, and early intervention.\u00a0Meanwhile, deep learning techniques and Large Language Models (LLMs) have emerged as promising tools in diverse Natural Language Processing (NLP) tasks, including sentiment analysis and text classification. In this paper, we propose a deep learning model incorporating triple models of word embeddings, as well as various fine-tuned LLMs, to identify suicidal thoughts in Reddit posts. In effect, we implemented a Bidirectional Long Short-Term Memory (BiLSTM), and a Convolutional Neural Network (CNN) model to categorize posts associated with non-suicidal and suicidal thoughts. Besides, through the combination of Word2Vec, FastText and GloVe embeddings, our models learn intricate patterns and prevalent nuances in suicide-related language. Furthermore, we employed a merged version of CNN and BiLSTM models, entitled C-BiLSTM, and several LLMs, including pre-trained Bidirectional Encoder Representations from Transformers (BERT) models and a Generative Pre-training Transformer (GPT) model. The analysis of all our proposed models shows that our C-BiLSTM model with triple word embedding and our GPT model got the best performance compared to deep learning and LLMs baseline models, reaching accuracies of 94.5% and 97.69%, respectively. In fact, our best model\u2019s capacity to extract meaningful interdependencies among words significantly promotes its classification performance. This analysis contributes to a deeper understanding of the psychological factors and linguistic markers indicative of suicidal thoughts, thereby informing future research and intervention strategies.","doi":"10.1007\/s13748-024-00326-z","cleaned_title":"advanced deep learning and large language models for suicide ideation detection on social media","cleaned_abstract":"recently, suicide ideations represent a worldwide health concern and pose many anticipation challenges. actually, the prevalence of expressing self-destructive thoughts especially on forums and social media requires effective monitoring for suicide prevention, and early intervention. meanwhile, deep learning techniques and large language models (llms) have emerged as promising tools in diverse natural language processing (nlp) tasks, including sentiment analysis and text classification. in this paper, we propose a deep learning model incorporating triple models of word embeddings, as well as various fine-tuned llms, to identify suicidal thoughts in reddit posts. in effect, we implemented a bidirectional long short-term memory (bilstm), and a convolutional neural network (cnn) model to categorize posts associated with non-suicidal and suicidal thoughts. besides, through the combination of word2vec, fasttext and glove embeddings, our models learn intricate patterns and prevalent nuances in suicide-related language. furthermore, we employed a merged version of cnn and bilstm models, entitled c-bilstm, and several llms, including pre-trained bidirectional encoder representations from transformers (bert) models and a generative pre-training transformer (gpt) model. the analysis of all our proposed models shows that our c-bilstm model with triple word embedding and our gpt model got the best performance compared to deep learning and llms baseline models, reaching accuracies of 94.5% and 97.69%, respectively. in fact, our best model\u2019s capacity to extract meaningful interdependencies among words significantly promotes its classification performance. this analysis contributes to a deeper understanding of the psychological factors and linguistic markers indicative of suicidal thoughts, thereby informing future research and intervention strategies.","key_phrases":["suicide ideations","a worldwide health concern","many anticipation challenges","the prevalence","self-destructive thoughts","forums","social media","effective monitoring","suicide prevention","early intervention","deep learning techniques","large language models","llms","promising tools","nlp","sentiment analysis","text classification","this paper","we","a deep learning model","triple models","word embeddings","various fine-tuned llms","suicidal thoughts","reddit posts","effect","we","a bidirectional long short-term memory","bilstm","a convolutional neural network (cnn) model","posts","-suicidal and suicidal thoughts","the combination","word2vec, fasttext and glove embeddings","our models","intricate patterns","prevalent nuances","suicide-related language","we","a merged version","cnn","bilstm models","c-bilstm","several llms","pre-trained bidirectional encoder representations","transformers","(bert) models","a generative pre-training transformer","(gpt) model","the analysis","all our proposed models","our c-bilstm model","triple word","our gpt model","the best performance","deep learning and llms baseline models","accuracies","94.5%","97.69%","fact","our best model\u2019s capacity","meaningful interdependencies","words","its classification performance","this analysis","a deeper understanding","the psychological factors","linguistic markers","suicidal thoughts","future research and intervention strategies","cnn","cnn","gpt","gpt","94.5%","97.69%"]},{"title":"Comparative study of machine learning and deep learning techniques for fault diagnosis in suspension system","authors":["P. Arun Balaji","V. Sugumaran"],"abstract":"Comfort and stability are the prime reasons to own an automobile (car). Suspension system of an automobile plays a major role in providing comfort, stability and control. Over a period of time, internal components in the suspension system exhibit faults due to fatigue and wear. Hence, it is essential to perform fault diagnosis such that the performance of the suspension components is restored. However, high instrumentation cost, skilled labor requirement and expertise in the particular field of study are certain drawbacks of traditional fault diagnosis techniques. Such challenges have made industrialists and the research communities look for advanced fault diagnosis techniques. Advancements in machine learning and deep learning techniques can be used to fulfill the need of a high degree intelligent fault diagnosis system. In the current study, the performance of machine learning (ML) classifiers are compared with the performance of deep learning (DL) models and the best performing model among them is adopted to detect faults in the automobile suspension system. A total of eight test conditions, namely strut external damage, strut mount failure, ball joint worn out, control arm bush worn out, control arm ball joint worn out, strut worn out, low wheel pressure and good condition, were considered in the study. The vibration measurements were acquired for three load conditions. Among all the techniques considered for classification, the pre-trained VGG16 model outperformed other DL and ML models with an overall classification accuracy of 98.10%.","doi":"10.1007\/s40430-023-04145-6","cleaned_title":"comparative study of machine learning and deep learning techniques for fault diagnosis in suspension system","cleaned_abstract":"comfort and stability are the prime reasons to own an automobile (car). suspension system of an automobile plays a major role in providing comfort, stability and control. over a period of time, internal components in the suspension system exhibit faults due to fatigue and wear. hence, it is essential to perform fault diagnosis such that the performance of the suspension components is restored. however, high instrumentation cost, skilled labor requirement and expertise in the particular field of study are certain drawbacks of traditional fault diagnosis techniques. such challenges have made industrialists and the research communities look for advanced fault diagnosis techniques. advancements in machine learning and deep learning techniques can be used to fulfill the need of a high degree intelligent fault diagnosis system. in the current study, the performance of machine learning (ml) classifiers are compared with the performance of deep learning (dl) models and the best performing model among them is adopted to detect faults in the automobile suspension system. a total of eight test conditions, namely strut external damage, strut mount failure, ball joint worn out, control arm bush worn out, control arm ball joint worn out, strut worn out, low wheel pressure and good condition, were considered in the study. the vibration measurements were acquired for three load conditions. among all the techniques considered for classification, the pre-trained vgg16 model outperformed other dl and ml models with an overall classification accuracy of 98.10%.","key_phrases":["comfort","stability","the prime reasons","an automobile (car","suspension system","an automobile","a major role","comfort","stability","control","a period","time","the suspension system","exhibit faults","fatigue","it","fault diagnosis","the performance","the suspension components","high instrumentation cost","skilled labor requirement","expertise","the particular field","study","certain drawbacks","traditional fault diagnosis techniques","such challenges","industrialists","the research communities","advanced fault diagnosis techniques","advancements","machine learning","deep learning techniques","the need","a high degree intelligent fault diagnosis system","the current study","the performance","machine learning (ml) classifiers","the performance","deep learning (dl) models","the best performing model","them","faults","the automobile suspension system","a total","eight test conditions","namely strut external damage","strut mount failure","ball joint","control arm bush","control arm ball joint","strut","low wheel pressure","good condition","the study","the vibration measurements","three load conditions","all the techniques","classification","the pre-trained vgg16 model","other dl and ml models","an overall classification accuracy","98.10%","eight","strut mount","bush","three","98.10%"]},{"title":"A Deep Learning Framework for Monitoring Audience Engagement in Online Video Events","authors":["Alexandros Vrochidis","Nikolaos Dimitriou","Stelios Krinidis","Savvas Panagiotidis","Stathis Parcharidis","Dimitrios Tzovaras"],"abstract":"This paper introduces a deep learning methodology for analyzing audience engagement in online video events. The proposed deep learning framework consists of six layers and starts with keyframe extraction from the video stream and the participants\u2019 face detection. Subsequently, the head pose and emotion per participant are estimated using the HopeNet and JAA-Net deep architectures. Complementary to video analysis, the audio signal is also processed using a neural network that follows the DenseNet-121 architecture. Its purpose is to detect events related to audience engagement, including speech, pauses, and applause. With the combined analysis of video and audio streams, the interest and attention of each participant are inferred more accurately. An experimental evaluation is performed on a newly generated dataset consisting of recordings from online video events, where the proposed framework achieves promising results. Concretely, the F1 scores were 79.21% for interest estimation according to pose, 65.38% for emotion estimation, and 80% for sound event detection. The proposed framework has applications in online educational events, where it can help tutors assess audience engagement and comprehension while hinting at points in their lectures that may require further clarification. It is effective for video streaming platforms that want to provide video recommendations to online users according to audience engagement.","doi":"10.1007\/s44196-024-00512-w","cleaned_title":"a deep learning framework for monitoring audience engagement in online video events","cleaned_abstract":"this paper introduces a deep learning methodology for analyzing audience engagement in online video events. the proposed deep learning framework consists of six layers and starts with keyframe extraction from the video stream and the participants\u2019 face detection. subsequently, the head pose and emotion per participant are estimated using the hopenet and jaa-net deep architectures. complementary to video analysis, the audio signal is also processed using a neural network that follows the densenet-121 architecture. its purpose is to detect events related to audience engagement, including speech, pauses, and applause. with the combined analysis of video and audio streams, the interest and attention of each participant are inferred more accurately. an experimental evaluation is performed on a newly generated dataset consisting of recordings from online video events, where the proposed framework achieves promising results. concretely, the f1 scores were 79.21% for interest estimation according to pose, 65.38% for emotion estimation, and 80% for sound event detection. the proposed framework has applications in online educational events, where it can help tutors assess audience engagement and comprehension while hinting at points in their lectures that may require further clarification. it is effective for video streaming platforms that want to provide video recommendations to online users according to audience engagement.","key_phrases":["this paper","a deep learning methodology","audience engagement","online video events","the proposed deep learning framework","six layers","keyframe extraction","the video stream","the participants","face detection","the head","emotion","participant","the hopenet and jaa-net deep architectures","video analysis","the audio signal","a neural network","that","the densenet-121 architecture","its purpose","events","audience engagement","speech","pauses","applause","the combined analysis","video and audio streams","the interest","attention","each participant","an experimental evaluation","a newly generated dataset","recordings","online video events","the proposed framework","promising results","the f1 scores","79.21%","interest estimation","emotion estimation","sound event detection","the proposed framework","applications","online educational events","it","tutors","audience engagement","comprehension","points","their lectures","that","further clarification","it","video streaming platforms","that","video recommendations","online users","audience engagement","six","79.21%","65.38%","80%"]},{"title":"Automated detection and recognition system for chewable food items using advanced deep learning models","authors":["Yogesh Kumar","Apeksha Koul","Kamini","Marcin Wo\u017aniak","Jana Shafi","Muhammad Fazal Ijaz"],"abstract":"Identifying and recognizing the food on the basis of its eating sounds is a challenging task, as it plays an important role in avoiding allergic foods, providing dietary preferences to people who are restricted to a particular diet, showcasing its cultural significance, etc. In this research paper, the aim is to design a novel methodology that helps to identify food items by analyzing their eating sounds using various deep learning models. To achieve this objective, a system has been proposed that extracts meaningful features from food-eating sounds with the help of signal processing techniques and deep learning models for classifying them into their respective food classes. Initially, 1200 audio files for 20 food items labeled have been collected and visualized to find relationships between the sound files of different food items. Later, to extract meaningful features, various techniques such as spectrograms, spectral rolloff, spectral bandwidth, and mel-frequency cepstral coefficients are used for the cleaning of audio files as well as to capture the unique characteristics of different food items. In the next phase, various deep learning models like GRU, LSTM, InceptionResNetV2, and the customized CNN model have been trained to learn spectral and temporal patterns in audio signals. Besides this, the models have also been hybridized i.e. Bidirectional LSTM\u2009+\u2009GRU and RNN\u2009+\u2009Bidirectional LSTM, and RNN\u2009+\u2009Bidirectional GRU to analyze their performance for the same labeled data in order to associate particular patterns of sound with their corresponding class of food item. During evaluation, the highest accuracy, precision,F1 score, and recall have been obtained by GRU with 99.28%, Bidirectional LSTM\u2009+\u2009GRU with 97.7% as well as 97.3%, and RNN\u2009+\u2009Bidirectional LSTM with 97.45%, respectively. The results of this study demonstrate that deep learning models have the potential to precisely identify foods on the basis of their sound by computing the best outcomes.","doi":"10.1038\/s41598-024-57077-z","cleaned_title":"automated detection and recognition system for chewable food items using advanced deep learning models","cleaned_abstract":"identifying and recognizing the food on the basis of its eating sounds is a challenging task, as it plays an important role in avoiding allergic foods, providing dietary preferences to people who are restricted to a particular diet, showcasing its cultural significance, etc. in this research paper, the aim is to design a novel methodology that helps to identify food items by analyzing their eating sounds using various deep learning models. to achieve this objective, a system has been proposed that extracts meaningful features from food-eating sounds with the help of signal processing techniques and deep learning models for classifying them into their respective food classes. initially, 1200 audio files for 20 food items labeled have been collected and visualized to find relationships between the sound files of different food items. later, to extract meaningful features, various techniques such as spectrograms, spectral rolloff, spectral bandwidth, and mel-frequency cepstral coefficients are used for the cleaning of audio files as well as to capture the unique characteristics of different food items. in the next phase, various deep learning models like gru, lstm, inceptionresnetv2, and the customized cnn model have been trained to learn spectral and temporal patterns in audio signals. besides this, the models have also been hybridized i.e. bidirectional lstm + gru and rnn + bidirectional lstm, and rnn + bidirectional gru to analyze their performance for the same labeled data in order to associate particular patterns of sound with their corresponding class of food item. during evaluation, the highest accuracy, precision,f1 score, and recall have been obtained by gru with 99.28%, bidirectional lstm + gru with 97.7% as well as 97.3%, and rnn + bidirectional lstm with 97.45%, respectively. the results of this study demonstrate that deep learning models have the potential to precisely identify foods on the basis of their sound by computing the best outcomes.","key_phrases":["the food","the basis","its eating sounds","a challenging task","it","an important role","allergic foods","dietary preferences","people","who","a particular diet","its cultural significance","this research paper","the aim","a novel methodology","that","food items","their eating sounds","various deep learning models","this objective","a system","meaningful features","the help","signal processing techniques","deep learning models","them","their respective food classes","1200 audio files","20 food items","relationships","the sound files","different food items","meaningful features","various techniques","spectrograms","spectral rolloff","spectral","mel-frequency cepstral coefficients","the cleaning","audio files","the unique characteristics","different food items","the next phase","various deep learning models","gru","lstm","inceptionresnetv2","the customized cnn model","spectral and temporal patterns","audio signals","this","the models","i.e. bidirectional lstm","bidirectional lstm","bidirectional gru","their performance","the same labeled data","order","particular patterns","sound","their corresponding class","food item","evaluation","the highest accuracy","precision","f1 score","recall","gru","99.28%","bidirectional lstm","97.7%","97.3%","bidirectional lstm","97.45%","the results","this study","deep learning models","the potential","foods","the basis","their sound","the best outcomes","1200","20","mel","inceptionresnetv2","cnn","99.28%","97.7%","97.3%","+ bidirectional lstm","97.45%"]},{"title":"Rapid diagnosis of celiac disease based on plasma Raman spectroscopy combined with deep learning","authors":["Tian Shi","Jiahe Li","Na Li","Cheng Chen","Chen Chen","Chenjie Chang","Shenglong Xue","Weidong Liu","Ainur Maimaiti Reyim","Feng Gao","Xiaoyi Lv"],"abstract":"Celiac Disease (CD) is a primary malabsorption syndrome resulting from the interplay of genetic, immune, and dietary factors. CD negatively impacts daily activities and may lead to conditions such as osteoporosis, malignancies in the small intestine, ulcerative jejunitis, and enteritis, ultimately causing severe malnutrition. Therefore, an effective and rapid differentiation between healthy individuals and those with celiac disease is crucial for early diagnosis and treatment. This study utilizes Raman spectroscopy combined with deep learning models to achieve a non-invasive, rapid, and accurate diagnostic method for celiac disease and healthy controls. A total of 59 plasma samples, comprising 29 celiac disease cases and 30 healthy controls, were collected for experimental purposes. Convolutional Neural Network (CNN), Multi-Scale Convolutional Neural Network (MCNN), Residual Network (ResNet), and Deep Residual Shrinkage Network (DRSN) classification models were employed. The accuracy rates for these models were found to be 86.67%, 90.76%, 86.67% and 95.00%, respectively. Comparative validation results revealed that the DRSN model exhibited the best performance, with an AUC value and accuracy of 97.60% and 95%, respectively. This confirms the superiority of Raman spectroscopy combined with deep learning in the diagnosis of celiac disease.","doi":"10.1038\/s41598-024-64621-4","cleaned_title":"rapid diagnosis of celiac disease based on plasma raman spectroscopy combined with deep learning","cleaned_abstract":"celiac disease (cd) is a primary malabsorption syndrome resulting from the interplay of genetic, immune, and dietary factors. cd negatively impacts daily activities and may lead to conditions such as osteoporosis, malignancies in the small intestine, ulcerative jejunitis, and enteritis, ultimately causing severe malnutrition. therefore, an effective and rapid differentiation between healthy individuals and those with celiac disease is crucial for early diagnosis and treatment. this study utilizes raman spectroscopy combined with deep learning models to achieve a non-invasive, rapid, and accurate diagnostic method for celiac disease and healthy controls. a total of 59 plasma samples, comprising 29 celiac disease cases and 30 healthy controls, were collected for experimental purposes. convolutional neural network (cnn), multi-scale convolutional neural network (mcnn), residual network (resnet), and deep residual shrinkage network (drsn) classification models were employed. the accuracy rates for these models were found to be 86.67%, 90.76%, 86.67% and 95.00%, respectively. comparative validation results revealed that the drsn model exhibited the best performance, with an auc value and accuracy of 97.60% and 95%, respectively. this confirms the superiority of raman spectroscopy combined with deep learning in the diagnosis of celiac disease.","key_phrases":["celiac disease","cd","a primary malabsorption syndrome","the interplay","genetic, immune, and dietary factors","cd","daily activities","conditions","osteoporosis","malignancies","the small intestine","ulcerative jejunitis","enteritis","severe malnutrition","an effective and rapid differentiation","healthy individuals","those","celiac disease","early diagnosis","treatment","this study","raman spectroscopy","deep learning models","a non-invasive, rapid, and accurate diagnostic method","celiac disease","healthy controls","a total","59 plasma samples","29 celiac disease cases","30 healthy controls","experimental purposes","convolutional neural network","cnn","multi-scale convolutional neural network","mcnn","residual network","resnet","deep residual shrinkage network","drsn) classification models","the accuracy rates","these models","86.67%","90.76%","86.67%","95.00%","comparative validation results","the drsn model","the best performance","an auc value","accuracy","97.60%","95%","this","the superiority","raman spectroscopy","deep learning","the diagnosis","celiac disease","daily","59","29","30","cnn","86.67%","90.76%","86.67%","95.00%","97.60% and","95%"]},{"title":"A deep learning model for brain segmentation across pediatric and adult populations","authors":["Jaime Simarro","Maria Ines Meyer","Simon Van Eyndhoven","Thanh V\u00e2n Phan","Thibo Billiet","Diana M. Sima","Els Ortibus"],"abstract":"Automated quantification of brain tissues on MR images has greatly contributed to the diagnosis and follow-up of neurological pathologies across various life stages. However, existing solutions are specifically designed for certain age ranges, limiting their applicability in monitoring brain development from infancy to late adulthood. This retrospective study aims to develop and validate a brain segmentation model across pediatric and adult populations. First, we trained a deep learning model to segment tissues and brain structures using T1-weighted MR images from 390 patients (age range: 2\u201381 years) across four different datasets. Subsequently, the model was validated on a cohort of 280 patients from six distinct test datasets (age range: 4\u201390 years). In the initial experiment, the proposed deep learning-based pipeline, icobrain-dl, demonstrated segmentation accuracy comparable to both pediatric and adult-specific models across diverse age groups. Subsequently, we evaluated intra- and inter-scanner variability in measurements of various tissues and structures in both pediatric and adult populations computed by icobrain-dl. Results demonstrated significantly higher reproducibility compared to similar brain quantification tools, including childmetrix, FastSurfer, and the medical device icobrain v5.9 (p-value< 0.01). Finally, we explored the potential clinical applications of icobrain-dl measurements in diagnosing pediatric patients with Cerebral Visual Impairment and adult patients with Alzheimer\u2019s Disease.","doi":"10.1038\/s41598-024-61798-6","cleaned_title":"a deep learning model for brain segmentation across pediatric and adult populations","cleaned_abstract":"automated quantification of brain tissues on mr images has greatly contributed to the diagnosis and follow-up of neurological pathologies across various life stages. however, existing solutions are specifically designed for certain age ranges, limiting their applicability in monitoring brain development from infancy to late adulthood. this retrospective study aims to develop and validate a brain segmentation model across pediatric and adult populations. first, we trained a deep learning model to segment tissues and brain structures using t1-weighted mr images from 390 patients (age range: 2\u201381 years) across four different datasets. subsequently, the model was validated on a cohort of 280 patients from six distinct test datasets (age range: 4\u201390 years). in the initial experiment, the proposed deep learning-based pipeline, icobrain-dl, demonstrated segmentation accuracy comparable to both pediatric and adult-specific models across diverse age groups. subsequently, we evaluated intra- and inter-scanner variability in measurements of various tissues and structures in both pediatric and adult populations computed by icobrain-dl. results demonstrated significantly higher reproducibility compared to similar brain quantification tools, including childmetrix, fastsurfer, and the medical device icobrain v5.9 (p-value< 0.01). finally, we explored the potential clinical applications of icobrain-dl measurements in diagnosing pediatric patients with cerebral visual impairment and adult patients with alzheimer\u2019s disease.","key_phrases":["automated quantification","brain tissues","mr images","the diagnosis","follow-up","neurological pathologies","various life stages","existing solutions","certain age ranges","their applicability","brain development","infancy","late adulthood","this retrospective study","a brain segmentation model","pediatric and adult populations","we","a deep learning model","segment tissues","brain structures","mr images","390 patients","age range","2\u201381 years","four different datasets","the model","a cohort","280 patients","six distinct test datasets","age range","4\u201390 years","the initial experiment","the proposed deep learning-based pipeline","icobrain-dl","segmentation accuracy","both pediatric and adult-specific models","diverse age groups","we","intra- and inter-scanner variability","measurements","various tissues","structures","both pediatric and adult populations","icobrain-dl. results","significantly higher reproducibility","similar brain quantification tools","childmetrix","fastsurfer","the medical device icobrain v5.9","p-value","we","the potential clinical applications","icobrain-dl measurements","pediatric patients","cerebral visual impairment","adult patients","disease","first","390","2\u201381 years","four","280","six","4\u201390 years"]},{"title":"Phase unwrapping based on deep learning in light field fringe projection 3D measurement","authors":["Xinjun Zhu","Haichuan Zhao","Mengkai Yuan","Zhizhi Zhang","Hongyi Wang","Limei Song"],"abstract":"Phase unwrapping is one of the key roles in fringe projection three-dimensional (3D) measurement technology. We propose a new method to achieve phase unwrapping in camera array light filed fringe projection 3D measurement based on deep learning. A multi-stream convolutional neural network (CNN) is proposed to learn the mapping relationship between camera array light filed wrapped phases and fringe orders of the expected central view, and is used to predict the fringe order to achieve the phase unwrapping. Experiments are performed on the light field fringe projection data generated by the simulated camera array fringe projection measurement system in Blender and by the experimental 3\u00d73 camera array light field fringe projection system. The performance of the proposed network with light field wrapped phases using multiple directions as network input data is studied, and the advantages of phase unwrapping based on deep learning in light filed fringe projection are demonstrated.","doi":"10.1007\/s11801-023-3002-4","cleaned_title":"phase unwrapping based on deep learning in light field fringe projection 3d measurement","cleaned_abstract":"phase unwrapping is one of the key roles in fringe projection three-dimensional (3d) measurement technology. we propose a new method to achieve phase unwrapping in camera array light filed fringe projection 3d measurement based on deep learning. a multi-stream convolutional neural network (cnn) is proposed to learn the mapping relationship between camera array light filed wrapped phases and fringe orders of the expected central view, and is used to predict the fringe order to achieve the phase unwrapping. experiments are performed on the light field fringe projection data generated by the simulated camera array fringe projection measurement system in blender and by the experimental 3\u00d73 camera array light field fringe projection system. the performance of the proposed network with light field wrapped phases using multiple directions as network input data is studied, and the advantages of phase unwrapping based on deep learning in light filed fringe projection are demonstrated.","key_phrases":["phase","the key roles","fringe projection three-dimensional (3d) measurement technology","we","a new method","phase","camera array light","fringe projection 3d measurement","deep learning","a multi-stream convolutional neural network","cnn","the mapping relationship","camera array light","phases","fringe orders","the expected central view","the fringe order","the phase","experiments","the light field fringe projection data","the simulated camera array fringe projection measurement system","blender","the experimental 3\u00d73 camera array light field fringe projection system","the performance","the proposed network","light field","phases","multiple directions","network input data","the advantages","phase","deep learning","light filed fringe projection","three","3d","3d","cnn","3\u00d73"]},{"title":"An experimental evaluation of deep reinforcement learning algorithms for HVAC control","authors":["Antonio Manjavacas","Alejandro Campoy-Nieves","Javier Jim\u00e9nez-Raboso","Miguel Molina-Solana","Juan G\u00f3mez-Romero"],"abstract":"Heating, ventilation, and air conditioning (HVAC) systems are a major driver of energy consumption in commercial and residential buildings. Recent studies have shown that Deep Reinforcement Learning (DRL) algorithms can outperform traditional reactive controllers. However, DRL-based solutions are generally designed for ad hoc setups and lack standardization for comparison. To fill this gap, this paper provides a critical and reproducible evaluation, in terms of comfort and energy consumption, of several state-of-the-art DRL algorithms for HVAC control. The study examines the controllers\u2019 robustness, adaptability, and trade-off between optimization goals by using the Sinergym framework. The results obtained confirm the potential of DRL algorithms, such as SAC and TD3, in complex scenarios and reveal several challenges related to generalization and incremental learning.","doi":"10.1007\/s10462-024-10819-x","cleaned_title":"an experimental evaluation of deep reinforcement learning algorithms for hvac control","cleaned_abstract":"heating, ventilation, and air conditioning (hvac) systems are a major driver of energy consumption in commercial and residential buildings. recent studies have shown that deep reinforcement learning (drl) algorithms can outperform traditional reactive controllers. however, drl-based solutions are generally designed for ad hoc setups and lack standardization for comparison. to fill this gap, this paper provides a critical and reproducible evaluation, in terms of comfort and energy consumption, of several state-of-the-art drl algorithms for hvac control. the study examines the controllers\u2019 robustness, adaptability, and trade-off between optimization goals by using the sinergym framework. the results obtained confirm the potential of drl algorithms, such as sac and td3, in complex scenarios and reveal several challenges related to generalization and incremental learning.","key_phrases":["heating, ventilation, and air conditioning (hvac) systems","a major driver","energy consumption","commercial and residential buildings","recent studies","drl","traditional reactive controllers","drl-based solutions","ad hoc setups","lack standardization","comparison","this gap","this paper","a critical and reproducible evaluation","terms","comfort and energy consumption","the-art","hvac control","the study","the controllers\u2019 robustness","adaptability","trade-off","optimization goals","the sinergym framework","the results","the potential","drl algorithms","sac","td3","complex scenarios","several challenges","generalization","incremental learning"]},{"title":"Development and application of a deep learning-based comprehensive early diagnostic model for chronic obstructive pulmonary disease","authors":["Zecheng Zhu","Shunjin Zhao","Jiahui Li","Yuting Wang","Luopiao Xu","Yubing Jia","Zihan Li","Wenyuan Li","Gang Chen","Xifeng Wu"],"abstract":"BackgroundChronic obstructive pulmonary disease (COPD) is a frequently diagnosed yet treatable condition, provided it is identified early and managed effectively. This study aims to develop an advanced COPD diagnostic model by integrating deep learning and radiomics features.MethodsWe utilized a dataset comprising CT images from 2,983 participants, of which 2,317 participants also provided epidemiological data through questionnaires. Deep learning features were extracted using a Variational Autoencoder, and radiomics features were obtained using the PyRadiomics package. Multi-Layer Perceptrons were used to construct models based on deep learning and radiomics features independently, as well as a fusion model integrating both. Subsequently, epidemiological questionnaire data were incorporated to establish a more comprehensive model. The diagnostic performance of standalone models, the fusion model and the comprehensive model was evaluated and compared using metrics including accuracy, precision, recall, F1-score, Brier score, receiver operating characteristic curves, and area under the curve (AUC).ResultsThe fusion model exhibited outstanding performance with an AUC of 0.952, surpassing the standalone models based solely on deep learning features (AUC\u2009=\u20090.844) or radiomics features (AUC\u2009=\u20090.944). Notably, the comprehensive model, incorporating deep learning features, radiomics features, and questionnaire variables demonstrated the highest diagnostic performance among all models, yielding an AUC of 0.971.ConclusionWe developed and implemented a data fusion strategy to construct a state-of-the-art COPD diagnostic model integrating deep learning features, radiomics features, and questionnaire variables. Our data fusion strategy proved effective, and the model can be easily deployed in clinical settings.Trial registrationNot applicable. This study is NOT a clinical trial, it does not report the results of a health care intervention on human participants.","doi":"10.1186\/s12931-024-02793-3","cleaned_title":"development and application of a deep learning-based comprehensive early diagnostic model for chronic obstructive pulmonary disease","cleaned_abstract":"backgroundchronic obstructive pulmonary disease (copd) is a frequently diagnosed yet treatable condition, provided it is identified early and managed effectively. this study aims to develop an advanced copd diagnostic model by integrating deep learning and radiomics features.methodswe utilized a dataset comprising ct images from 2,983 participants, of which 2,317 participants also provided epidemiological data through questionnaires. deep learning features were extracted using a variational autoencoder, and radiomics features were obtained using the pyradiomics package. multi-layer perceptrons were used to construct models based on deep learning and radiomics features independently, as well as a fusion model integrating both. subsequently, epidemiological questionnaire data were incorporated to establish a more comprehensive model. the diagnostic performance of standalone models, the fusion model and the comprehensive model was evaluated and compared using metrics including accuracy, precision, recall, f1-score, brier score, receiver operating characteristic curves, and area under the curve (auc).resultsthe fusion model exhibited outstanding performance with an auc of 0.952, surpassing the standalone models based solely on deep learning features (auc = 0.844) or radiomics features (auc = 0.944). notably, the comprehensive model, incorporating deep learning features, radiomics features, and questionnaire variables demonstrated the highest diagnostic performance among all models, yielding an auc of 0.971.conclusionwe developed and implemented a data fusion strategy to construct a state-of-the-art copd diagnostic model integrating deep learning features, radiomics features, and questionnaire variables. our data fusion strategy proved effective, and the model can be easily deployed in clinical settings.trial registrationnot applicable. this study is not a clinical trial, it does not report the results of a health care intervention on human participants.","key_phrases":["backgroundchronic obstructive pulmonary disease","copd","a frequently diagnosed yet treatable condition","it","this study","an advanced copd diagnostic model","deep learning and radiomics features.methodswe","a dataset","ct images","2,983 participants","which","2,317 participants","epidemiological data","questionnaires","deep learning features","a variational autoencoder","radiomics features","the pyradiomics package","multi-layer perceptrons","models","deep learning","radiomics","a fusion model","both","epidemiological questionnaire data","a more comprehensive model","the diagnostic performance","standalone models","the fusion model","the comprehensive model","metrics","accuracy","precision","recall, f1-score","brier score","the curve","auc).resultsthe fusion model","outstanding performance","an auc","the standalone models","deep learning features","auc","radiomics features","auc =","the comprehensive model","deep learning features","radiomics features","questionnaire variables","the highest diagnostic performance","all models","an auc","0.971.conclusionwe","a data fusion strategy","the-art","deep learning features","radiomics features","questionnaire variables","our data fusion strategy","the model","clinical settings.trial registrationnot","this study","a clinical trial","it","the results","a health care intervention","human participants","2,983","2,317","0.952","0.844","0.944","0.971.conclusionwe"]},{"title":"A multi-agent adaptive deep learning framework for online intrusion detection","authors":["Mahdi Soltani","Khashayar Khajavi","Mahdi Jafari Siavoshani","Amir Hossein Jahangir"],"abstract":"The network security analyzers use intrusion detection systems (IDSes) to distinguish malicious traffic from benign ones. The deep learning-based (DL-based) IDSes are proposed to auto-extract high-level features and eliminate the time-consuming and costly signature extraction process. However, this new generation of IDSes still needs to overcome a number of challenges to be employed in practical environments. One of the main issues of an applicable IDS is facing traffic concept drift, which manifests itself as new (i.e.\u00a0, zero-day) attacks, in addition to the changing behavior of benign users\/applications. Furthermore, a practical DL-based IDS needs to be conformed to a distributed (i.e.\u00a0, multi-sensor) architecture in order to yield more accurate detections, create a collective attack knowledge based on the observations of different sensors, and also handle big data challenges for supporting high throughput networks. This paper proposes a novel multi-agent network intrusion detection framework to address the above shortcomings, considering a more practical scenario (i.e., online adaptable IDSes). This framework employs continual deep anomaly detectors for adapting each agent to the changing attack\/benign patterns in its local traffic. In addition, a federated learning approach is proposed for sharing and exchanging local knowledge between different agents. Furthermore, the proposed framework implements sequential packet labeling for each flow, which provides an attack probability score for the flow by gradually observing each flow packet and updating its estimation. We evaluate the proposed framework by employing different deep models (including CNN-based and LSTM-based) over the CIC-IDS2017 and CSE-CIC-IDS2018 datasets. Through extensive evaluations and experiments, we show that the proposed distributed framework is well adapted to the traffic concept drift. More precisely, our results indicate that the CNN-based models are well suited for continually adapting to the traffic concept drift (i.e.\u00a0, achieving an average detection rate of above 95% while needing just 128 new flows for the updating phase), and the LSTM-based models are a good candidate for sequential packet labeling in practical online IDSes (i.e.\u00a0, detecting intrusions by just observing their first 15 packets).","doi":"10.1186\/s42400-023-00199-0","cleaned_title":"a multi-agent adaptive deep learning framework for online intrusion detection","cleaned_abstract":"the network security analyzers use intrusion detection systems (idses) to distinguish malicious traffic from benign ones. the deep learning-based (dl-based) idses are proposed to auto-extract high-level features and eliminate the time-consuming and costly signature extraction process. however, this new generation of idses still needs to overcome a number of challenges to be employed in practical environments. one of the main issues of an applicable ids is facing traffic concept drift, which manifests itself as new (i.e. , zero-day) attacks, in addition to the changing behavior of benign users\/applications. furthermore, a practical dl-based ids needs to be conformed to a distributed (i.e. , multi-sensor) architecture in order to yield more accurate detections, create a collective attack knowledge based on the observations of different sensors, and also handle big data challenges for supporting high throughput networks. this paper proposes a novel multi-agent network intrusion detection framework to address the above shortcomings, considering a more practical scenario (i.e., online adaptable idses). this framework employs continual deep anomaly detectors for adapting each agent to the changing attack\/benign patterns in its local traffic. in addition, a federated learning approach is proposed for sharing and exchanging local knowledge between different agents. furthermore, the proposed framework implements sequential packet labeling for each flow, which provides an attack probability score for the flow by gradually observing each flow packet and updating its estimation. we evaluate the proposed framework by employing different deep models (including cnn-based and lstm-based) over the cic-ids2017 and cse-cic-ids2018 datasets. through extensive evaluations and experiments, we show that the proposed distributed framework is well adapted to the traffic concept drift. more precisely, our results indicate that the cnn-based models are well suited for continually adapting to the traffic concept drift (i.e. , achieving an average detection rate of above 95% while needing just 128 new flows for the updating phase), and the lstm-based models are a good candidate for sequential packet labeling in practical online idses (i.e. , detecting intrusions by just observing their first 15 packets).","key_phrases":["the network security analyzers","intrusion detection systems","(idses","malicious traffic","benign ones","the deep learning-based (dl-based) idses","auto-extract high-level features","the time-consuming and costly signature extraction process","this new generation","idses","a number","challenges","practical environments","the main issues","an applicable ids","traffic concept drift","which","itself","new (i.e. , zero-day","attacks","addition","the changing behavior","benign users\/applications","a practical dl-based ids","a distributed (i.e. , multi-sensor) architecture","order","more accurate detections","a collective attack knowledge","the observations","different sensors","big data challenges","high throughput networks","this paper","a novel multi-agent network intrusion detection framework","the above shortcomings","a more practical scenario","(i.e., online adaptable idses","this framework","continual deep anomaly detectors","each agent","the changing attack\/benign patterns","its local traffic","addition","a federated learning approach","local knowledge","different agents","the proposed framework","sequential packet labeling","each flow","which","an attack probability score","the flow","each flow packet","its estimation","we","the proposed framework","different deep models","the cic","-ids2017","cse-cic-ids2018","datasets","extensive evaluations","experiments","we","the proposed distributed framework","the traffic concept drift","our results","the cnn-based models","the traffic concept drift","an average detection rate","above 95%","just 128 new flows","the updating phase","the lstm-based models","a good candidate","sequential packet labeling","practical online idses","intrusions","their first 15 packets","one","zero-day","cnn","cnn","95%","just 128","first","15"]},{"title":"Addressing data imbalance challenges in oral cavity histopathological whole slide images with advanced deep learning techniques","authors":["Tabasum Majeed","Tariq Ahmad Masoodi","Muzafar Ahmad Macha","Muzafar Rasool Bhat","Khalid Muzaffar","Assif Assad"],"abstract":"Oral Cavity Squamous Cell Carcinoma (OCSCC) represents a common form of head and neck cancer originating from the mucosal lining of the oral cavity, often detected in advanced stages. Traditional detection methods rely on analyzing hematoxylin and eosin (H&E)-stained histopathological whole-slide images, which are time-consuming and require expert pathology skills. Hence, automated analysis is urgently needed to expedite diagnosis and improve patient outcomes. Deep learning, through automated feature extraction, offers a promising avenue for capturing high-level abstract features with greater accuracy than traditional methods. However, the imbalance in class distribution within datasets significantly affects the performance of deep learning models during training, necessitating specialized approaches. To address the issue, various methods have been proposed at both data and algorithmic levels. This study investigates strategies to mitigate class imbalance by employing a publicly available OCSCC imbalance dataset. We evaluated undersampling methods (Near Miss, Edited Nearest Neighbors) and oversampling techniques (SMOTE, Deep SMOTE, ADASYN) integrated with transfer learning across different imbalance ratios (0.1, 0.15, 0.20, 0.30). Our findings demonstrate the effectiveness of SMOTE in improving test performance, highlighting the efficacy of strategic oversampling combined with transfer learning in classifying imbalanced medical datasets. This enhances OCSCC diagnostic accuracy, streamlines clinical decisions, and reduces reliance on costly histopathological tests.","doi":"10.1007\/s13198-024-02440-6","cleaned_title":"addressing data imbalance challenges in oral cavity histopathological whole slide images with advanced deep learning techniques","cleaned_abstract":"oral cavity squamous cell carcinoma (ocscc) represents a common form of head and neck cancer originating from the mucosal lining of the oral cavity, often detected in advanced stages. traditional detection methods rely on analyzing hematoxylin and eosin (h&e)-stained histopathological whole-slide images, which are time-consuming and require expert pathology skills. hence, automated analysis is urgently needed to expedite diagnosis and improve patient outcomes. deep learning, through automated feature extraction, offers a promising avenue for capturing high-level abstract features with greater accuracy than traditional methods. however, the imbalance in class distribution within datasets significantly affects the performance of deep learning models during training, necessitating specialized approaches. to address the issue, various methods have been proposed at both data and algorithmic levels. this study investigates strategies to mitigate class imbalance by employing a publicly available ocscc imbalance dataset. we evaluated undersampling methods (near miss, edited nearest neighbors) and oversampling techniques (smote, deep smote, adasyn) integrated with transfer learning across different imbalance ratios (0.1, 0.15, 0.20, 0.30). our findings demonstrate the effectiveness of smote in improving test performance, highlighting the efficacy of strategic oversampling combined with transfer learning in classifying imbalanced medical datasets. this enhances ocscc diagnostic accuracy, streamlines clinical decisions, and reduces reliance on costly histopathological tests.","key_phrases":["oral cavity squamous cell carcinoma","ocscc","a common form","head and neck cancer","the mucosal lining","the oral cavity","advanced stages","traditional detection methods","hematoxylin","eosin","histopathological whole-slide images","which","expert pathology skills","automated analysis","diagnosis","patient outcomes","deep learning","automated feature extraction","a promising avenue","high-level abstract features","greater accuracy","traditional methods","the imbalance","class distribution","datasets","the performance","deep learning models","training","specialized approaches","the issue","various methods","both data","algorithmic levels","this study","strategies","class imbalance","a publicly available ocscc imbalance dataset","we","undersampling methods","near miss","neighbors","techniques","smote","deep smote","adasyn","different imbalance ratios","our findings","the effectiveness","smote","test performance","the efficacy","strategic oversampling","imbalanced medical datasets","diagnostic accuracy","clinical decisions","reliance","costly histopathological tests","hematoxylin","0.1","0.15","0.20","0.30"]},{"title":"Deep learning implementation of image segmentation in agricultural applications: a comprehensive review","authors":["Lian Lei","Qiliang Yang","Ling Yang","Tao Shen","Ruoxi Wang","Chengbiao Fu"],"abstract":"Image segmentation is a crucial task in computer vision, which divides a digital image into multiple segments and objects. In agriculture, image segmentation is extensively used for crop and soil monitoring, predicting the best times to sow, fertilize, and harvest, estimating crop yield, and detecting plant diseases. However, image segmentation faces difficulties in agriculture, such as the challenges of disease staging recognition, labeling inconsistency, and changes in plant morphology with the environment. Consequently, we have conducted a comprehensive review of image segmentation techniques based on deep learning, exploring the development and prospects of image segmentation in agriculture. Deep learning-based image segmentation solutions widely used in agriculture are categorized into eight main groups: encoder-decoder structures, multi-scale and pyramid-based methods, dilated convolutional networks, visual attention models, generative adversarial networks, graph neural networks, instance segmentation networks, and transformer-based models. In addition, the applications of image segmentation methods in agriculture are presented, such as plant disease detection, weed identification, crop growth monitoring, crop yield estimation, and counting. Furthermore, a collection of publicly available plant image segmentation datasets has been reviewed, and the evaluation and comparison of performance for image segmentation algorithms have been conducted on benchmark datasets. Finally, there is a discussion of the challenges and future prospects of image segmentation in agriculture.","doi":"10.1007\/s10462-024-10775-6","cleaned_title":"deep learning implementation of image segmentation in agricultural applications: a comprehensive review","cleaned_abstract":"image segmentation is a crucial task in computer vision, which divides a digital image into multiple segments and objects. in agriculture, image segmentation is extensively used for crop and soil monitoring, predicting the best times to sow, fertilize, and harvest, estimating crop yield, and detecting plant diseases. however, image segmentation faces difficulties in agriculture, such as the challenges of disease staging recognition, labeling inconsistency, and changes in plant morphology with the environment. consequently, we have conducted a comprehensive review of image segmentation techniques based on deep learning, exploring the development and prospects of image segmentation in agriculture. deep learning-based image segmentation solutions widely used in agriculture are categorized into eight main groups: encoder-decoder structures, multi-scale and pyramid-based methods, dilated convolutional networks, visual attention models, generative adversarial networks, graph neural networks, instance segmentation networks, and transformer-based models. in addition, the applications of image segmentation methods in agriculture are presented, such as plant disease detection, weed identification, crop growth monitoring, crop yield estimation, and counting. furthermore, a collection of publicly available plant image segmentation datasets has been reviewed, and the evaluation and comparison of performance for image segmentation algorithms have been conducted on benchmark datasets. finally, there is a discussion of the challenges and future prospects of image segmentation in agriculture.","key_phrases":["image segmentation","a crucial task","computer vision","which","a digital image","multiple segments","objects","agriculture","image segmentation","crop","soil monitoring","the best times","crop yield","plant diseases","image segmentation","difficulties","agriculture","the challenges","disease staging recognition","labeling inconsistency","changes","plant morphology","the environment","we","a comprehensive review","image segmentation techniques","deep learning","the development","prospects","image segmentation","agriculture","deep learning-based image segmentation solutions","agriculture","eight main groups","encoder-decoder structures","-scale and pyramid-based methods","dilated convolutional networks","visual attention models","generative adversarial networks","graph neural networks","instance segmentation networks","transformer-based models","addition","the applications","image segmentation methods","agriculture","plant disease detection","identification","crop growth monitoring","crop yield estimation","a collection","publicly available plant image segmentation datasets","the evaluation","comparison","performance","image segmentation algorithms","benchmark datasets","a discussion","the challenges","future prospects","image segmentation","agriculture","eight"]},{"title":"COVID-19 Fake News Detection using Deep Learning Model","authors":["Mahabuba Akhter","Syed Md. Minhaz Hossain","Rizma Sijana Nigar","Srabanti Paul","Khaleque Md. Aashiq Kamal","Anik Sen","Iqbal H. Sarker"],"abstract":"People may now receive and share information more quickly and easily than ever due to the widespread use of mobile networked devices. However, this can occasionally lead to the spread of false information. Such information is being disseminated widely, which may cause people to make incorrect decisions about potentially crucial topics. This occurred in 2020, the year of the fatal and extremely contagious Coronavirus Disease (COVID-19) outbreak. The spread of false information about COVID-19 on social media has already been labeled as an \u201cinfodemic\u201d by the World Health Organization (WHO), causing serious difficulties for governments attempting to control the pandemic. Consequently, it is crucial to have a model for detecting fake news related to COVID-19. In this paper, we present an effective Convolutional Neural Network (CNN)-based deep learning model using word embedding. For selecting the best CNN architecture, we take into account the optimal values of model hyper-parameters using grid search. Further, for measuring the effectiveness of our proposed CNN model, various state-of-the-art machine learning algorithms are conducted for COVID-19 fake news detection. Among them, CNN outperforms with 96.19% mean accuracy, 95% mean F1-score, and 0.985 area under ROC curve (AUC).","doi":"10.1007\/s40745-023-00507-y","cleaned_title":"covid-19 fake news detection using deep learning model","cleaned_abstract":"people may now receive and share information more quickly and easily than ever due to the widespread use of mobile networked devices. however, this can occasionally lead to the spread of false information. such information is being disseminated widely, which may cause people to make incorrect decisions about potentially crucial topics. this occurred in 2020, the year of the fatal and extremely contagious coronavirus disease (covid-19) outbreak. the spread of false information about covid-19 on social media has already been labeled as an \u201cinfodemic\u201d by the world health organization (who), causing serious difficulties for governments attempting to control the pandemic. consequently, it is crucial to have a model for detecting fake news related to covid-19. in this paper, we present an effective convolutional neural network (cnn)-based deep learning model using word embedding. for selecting the best cnn architecture, we take into account the optimal values of model hyper-parameters using grid search. further, for measuring the effectiveness of our proposed cnn model, various state-of-the-art machine learning algorithms are conducted for covid-19 fake news detection. among them, cnn outperforms with 96.19% mean accuracy, 95% mean f1-score, and 0.985 area under roc curve (auc).","key_phrases":["people","information","the widespread use","mobile networked devices","this","the spread","false information","such information","which","people","incorrect decisions","potentially crucial topics","this","the year","the fatal and extremely contagious coronavirus disease","covid-19","the spread","false information","covid-19","social media","the world health organization","who","serious difficulties","governments","it","a model","fake news","covid-19","this paper","we","an effective convolutional neural network","cnn)-based deep learning model","word","the best cnn architecture","we","account","the optimal values","model hyper-parameters","grid search","the effectiveness","our proposed cnn model","the-art","covid-19 fake news detection","them","cnn","96.19%","mean accuracy","95%","0.985 area","roc curve","auc","2020","the year","covid-19","covid-19","the world health organization","covid-19","cnn","cnn","covid-19","cnn","96.19%","95%","0.985","roc"]},{"title":"An Outlook for Deep Learning in Ecosystem Science","authors":["George L. W. Perry","Rupert Seidl","Andr\u00e9 M. Bellv\u00e9","Werner Rammer"],"abstract":"Rapid advances in hardware and software, accompanied by public- and private-sector investment, have led to a new generation of data-driven computational tools. Recently, there has been a particular focus on deep learning\u2014a class of machine learning algorithms that uses deep neural networks to identify patterns in large and heterogeneous datasets. These developments have been accompanied by both hype and scepticism by ecologists and others. This review describes the context in which deep learning methods have emerged, the deep learning methods most relevant to ecosystem ecologists, and some of the problem domains they have been applied to. Deep learning methods have high predictive performance in a range of ecological contexts, leveraging the large data resources now available. Furthermore, deep learning tools offer ecosystem ecologists new ways to learn about ecosystem dynamics. In particular, recent advances in interpretable machine learning and in developing hybrid approaches combining deep learning and mechanistic models provide a bridge between pure prediction and causal explanation. We conclude by looking at the opportunities that deep learning tools offer ecosystem ecologists and assess the challenges in interpretability that deep learning applications pose.","doi":"10.1007\/s10021-022-00789-y","cleaned_title":"an outlook for deep learning in ecosystem science","cleaned_abstract":"rapid advances in hardware and software, accompanied by public- and private-sector investment, have led to a new generation of data-driven computational tools. recently, there has been a particular focus on deep learning\u2014a class of machine learning algorithms that uses deep neural networks to identify patterns in large and heterogeneous datasets. these developments have been accompanied by both hype and scepticism by ecologists and others. this review describes the context in which deep learning methods have emerged, the deep learning methods most relevant to ecosystem ecologists, and some of the problem domains they have been applied to. deep learning methods have high predictive performance in a range of ecological contexts, leveraging the large data resources now available. furthermore, deep learning tools offer ecosystem ecologists new ways to learn about ecosystem dynamics. in particular, recent advances in interpretable machine learning and in developing hybrid approaches combining deep learning and mechanistic models provide a bridge between pure prediction and causal explanation. we conclude by looking at the opportunities that deep learning tools offer ecosystem ecologists and assess the challenges in interpretability that deep learning applications pose.","key_phrases":["rapid advances","hardware","software","public-","private-sector investment","a new generation","data-driven computational tools","a particular focus","deep learning","a class","machine learning algorithms","that","deep neural networks","patterns","large and heterogeneous datasets","these developments","both hype","scepticism","ecologists","others","this review","the context","which","deep learning methods","the deep learning methods","ecosystem ecologists","some","the problem domains","they","deep learning methods","high predictive performance","a range","ecological contexts","the large data resources","deep learning tools","ecosystem","new ways","ecosystem dynamics","recent advances","interpretable machine learning","hybrid approaches","deep learning","mechanistic models","a bridge","pure prediction","causal explanation","we","the opportunities","deep learning tools","ecosystem ecologists","the challenges","interpretability","deep learning applications"]},{"title":"A multi-agent adaptive deep learning framework for online intrusion detection","authors":["Mahdi Soltani","Khashayar Khajavi","Mahdi Jafari Siavoshani","Amir Hossein Jahangir"],"abstract":"The network security analyzers use intrusion detection systems (IDSes) to distinguish malicious traffic from benign ones. The deep learning-based (DL-based) IDSes are proposed to auto-extract high-level features and eliminate the time-consuming and costly signature extraction process. However, this new generation of IDSes still needs to overcome a number of challenges to be employed in practical environments. One of the main issues of an applicable IDS is facing traffic concept drift, which manifests itself as new (i.e.\u00a0, zero-day) attacks, in addition to the changing behavior of benign users\/applications. Furthermore, a practical DL-based IDS needs to be conformed to a distributed (i.e.\u00a0, multi-sensor) architecture in order to yield more accurate detections, create a collective attack knowledge based on the observations of different sensors, and also handle big data challenges for supporting high throughput networks. This paper proposes a novel multi-agent network intrusion detection framework to address the above shortcomings, considering a more practical scenario (i.e., online adaptable IDSes). This framework employs continual deep anomaly detectors for adapting each agent to the changing attack\/benign patterns in its local traffic. In addition, a federated learning approach is proposed for sharing and exchanging local knowledge between different agents. Furthermore, the proposed framework implements sequential packet labeling for each flow, which provides an attack probability score for the flow by gradually observing each flow packet and updating its estimation. We evaluate the proposed framework by employing different deep models (including CNN-based and LSTM-based) over the CIC-IDS2017 and CSE-CIC-IDS2018 datasets. Through extensive evaluations and experiments, we show that the proposed distributed framework is well adapted to the traffic concept drift. More precisely, our results indicate that the CNN-based models are well suited for continually adapting to the traffic concept drift (i.e.\u00a0, achieving an average detection rate of above 95% while needing just 128 new flows for the updating phase), and the LSTM-based models are a good candidate for sequential packet labeling in practical online IDSes (i.e.\u00a0, detecting intrusions by just observing their first 15 packets).","doi":"10.1186\/s42400-023-00199-0","cleaned_title":"a multi-agent adaptive deep learning framework for online intrusion detection","cleaned_abstract":"the network security analyzers use intrusion detection systems (idses) to distinguish malicious traffic from benign ones. the deep learning-based (dl-based) idses are proposed to auto-extract high-level features and eliminate the time-consuming and costly signature extraction process. however, this new generation of idses still needs to overcome a number of challenges to be employed in practical environments. one of the main issues of an applicable ids is facing traffic concept drift, which manifests itself as new (i.e. , zero-day) attacks, in addition to the changing behavior of benign users\/applications. furthermore, a practical dl-based ids needs to be conformed to a distributed (i.e. , multi-sensor) architecture in order to yield more accurate detections, create a collective attack knowledge based on the observations of different sensors, and also handle big data challenges for supporting high throughput networks. this paper proposes a novel multi-agent network intrusion detection framework to address the above shortcomings, considering a more practical scenario (i.e., online adaptable idses). this framework employs continual deep anomaly detectors for adapting each agent to the changing attack\/benign patterns in its local traffic. in addition, a federated learning approach is proposed for sharing and exchanging local knowledge between different agents. furthermore, the proposed framework implements sequential packet labeling for each flow, which provides an attack probability score for the flow by gradually observing each flow packet and updating its estimation. we evaluate the proposed framework by employing different deep models (including cnn-based and lstm-based) over the cic-ids2017 and cse-cic-ids2018 datasets. through extensive evaluations and experiments, we show that the proposed distributed framework is well adapted to the traffic concept drift. more precisely, our results indicate that the cnn-based models are well suited for continually adapting to the traffic concept drift (i.e. , achieving an average detection rate of above 95% while needing just 128 new flows for the updating phase), and the lstm-based models are a good candidate for sequential packet labeling in practical online idses (i.e. , detecting intrusions by just observing their first 15 packets).","key_phrases":["the network security analyzers","intrusion detection systems","(idses","malicious traffic","benign ones","the deep learning-based (dl-based) idses","auto-extract high-level features","the time-consuming and costly signature extraction process","this new generation","idses","a number","challenges","practical environments","the main issues","an applicable ids","traffic concept drift","which","itself","new (i.e. , zero-day","attacks","addition","the changing behavior","benign users\/applications","a practical dl-based ids","a distributed (i.e. , multi-sensor) architecture","order","more accurate detections","a collective attack knowledge","the observations","different sensors","big data challenges","high throughput networks","this paper","a novel multi-agent network intrusion detection framework","the above shortcomings","a more practical scenario","(i.e., online adaptable idses","this framework","continual deep anomaly detectors","each agent","the changing attack\/benign patterns","its local traffic","addition","a federated learning approach","local knowledge","different agents","the proposed framework","sequential packet labeling","each flow","which","an attack probability score","the flow","each flow packet","its estimation","we","the proposed framework","different deep models","the cic","-ids2017","cse-cic-ids2018","datasets","extensive evaluations","experiments","we","the proposed distributed framework","the traffic concept drift","our results","the cnn-based models","the traffic concept drift","an average detection rate","above 95%","just 128 new flows","the updating phase","the lstm-based models","a good candidate","sequential packet labeling","practical online idses","intrusions","their first 15 packets","one","zero-day","cnn","cnn","95%","just 128","first","15"]},{"title":"Deep learning implementation of image segmentation in agricultural applications: a comprehensive review","authors":["Lian Lei","Qiliang Yang","Ling Yang","Tao Shen","Ruoxi Wang","Chengbiao Fu"],"abstract":"Image segmentation is a crucial task in computer vision, which divides a digital image into multiple segments and objects. In agriculture, image segmentation is extensively used for crop and soil monitoring, predicting the best times to sow, fertilize, and harvest, estimating crop yield, and detecting plant diseases. However, image segmentation faces difficulties in agriculture, such as the challenges of disease staging recognition, labeling inconsistency, and changes in plant morphology with the environment. Consequently, we have conducted a comprehensive review of image segmentation techniques based on deep learning, exploring the development and prospects of image segmentation in agriculture. Deep learning-based image segmentation solutions widely used in agriculture are categorized into eight main groups: encoder-decoder structures, multi-scale and pyramid-based methods, dilated convolutional networks, visual attention models, generative adversarial networks, graph neural networks, instance segmentation networks, and transformer-based models. In addition, the applications of image segmentation methods in agriculture are presented, such as plant disease detection, weed identification, crop growth monitoring, crop yield estimation, and counting. Furthermore, a collection of publicly available plant image segmentation datasets has been reviewed, and the evaluation and comparison of performance for image segmentation algorithms have been conducted on benchmark datasets. Finally, there is a discussion of the challenges and future prospects of image segmentation in agriculture.","doi":"10.1007\/s10462-024-10775-6","cleaned_title":"deep learning implementation of image segmentation in agricultural applications: a comprehensive review","cleaned_abstract":"image segmentation is a crucial task in computer vision, which divides a digital image into multiple segments and objects. in agriculture, image segmentation is extensively used for crop and soil monitoring, predicting the best times to sow, fertilize, and harvest, estimating crop yield, and detecting plant diseases. however, image segmentation faces difficulties in agriculture, such as the challenges of disease staging recognition, labeling inconsistency, and changes in plant morphology with the environment. consequently, we have conducted a comprehensive review of image segmentation techniques based on deep learning, exploring the development and prospects of image segmentation in agriculture. deep learning-based image segmentation solutions widely used in agriculture are categorized into eight main groups: encoder-decoder structures, multi-scale and pyramid-based methods, dilated convolutional networks, visual attention models, generative adversarial networks, graph neural networks, instance segmentation networks, and transformer-based models. in addition, the applications of image segmentation methods in agriculture are presented, such as plant disease detection, weed identification, crop growth monitoring, crop yield estimation, and counting. furthermore, a collection of publicly available plant image segmentation datasets has been reviewed, and the evaluation and comparison of performance for image segmentation algorithms have been conducted on benchmark datasets. finally, there is a discussion of the challenges and future prospects of image segmentation in agriculture.","key_phrases":["image segmentation","a crucial task","computer vision","which","a digital image","multiple segments","objects","agriculture","image segmentation","crop","soil monitoring","the best times","crop yield","plant diseases","image segmentation","difficulties","agriculture","the challenges","disease staging recognition","labeling inconsistency","changes","plant morphology","the environment","we","a comprehensive review","image segmentation techniques","deep learning","the development","prospects","image segmentation","agriculture","deep learning-based image segmentation solutions","agriculture","eight main groups","encoder-decoder structures","-scale and pyramid-based methods","dilated convolutional networks","visual attention models","generative adversarial networks","graph neural networks","instance segmentation networks","transformer-based models","addition","the applications","image segmentation methods","agriculture","plant disease detection","identification","crop growth monitoring","crop yield estimation","a collection","publicly available plant image segmentation datasets","the evaluation","comparison","performance","image segmentation algorithms","benchmark datasets","a discussion","the challenges","future prospects","image segmentation","agriculture","eight"]},{"title":"Model-based deep learning framework for accelerated optical projection tomography","authors":["Marcos Obando","Andrea Bassi","Nicolas Ducros","Germ\u00e1n Mato","Teresa M. Correia"],"abstract":"In this work, we propose a model-based deep learning reconstruction algorithm for optical projection tomography (ToMoDL), to greatly reduce acquisition and reconstruction times. The proposed method iterates over a data consistency step and an image domain artefact removal step achieved by a convolutional neural network. A preprocessing stage is also included to avoid potential misalignments between the sample center of rotation and the detector. The algorithm is trained using a database of wild-type zebrafish (Danio rerio) at different stages of development to minimise the mean square error for a fixed number of iterations. Using a cross-validation scheme, we compare the results to other reconstruction methods, such as filtered backprojection, compressed sensing and a direct deep learning method where the pseudo-inverse solution is corrected by a U-Net. The proposed method performs equally well or better than the alternatives. For a highly reduced number of projections, only the U-Net method provides images comparable to those obtained with ToMoDL. However, ToMoDL has a much better performance if the amount of data available for training is limited, given that the number of network trainable parameters is smaller.","doi":"10.1038\/s41598-023-47650-3","cleaned_title":"model-based deep learning framework for accelerated optical projection tomography","cleaned_abstract":"in this work, we propose a model-based deep learning reconstruction algorithm for optical projection tomography (tomodl), to greatly reduce acquisition and reconstruction times. the proposed method iterates over a data consistency step and an image domain artefact removal step achieved by a convolutional neural network. a preprocessing stage is also included to avoid potential misalignments between the sample center of rotation and the detector. the algorithm is trained using a database of wild-type zebrafish (danio rerio) at different stages of development to minimise the mean square error for a fixed number of iterations. using a cross-validation scheme, we compare the results to other reconstruction methods, such as filtered backprojection, compressed sensing and a direct deep learning method where the pseudo-inverse solution is corrected by a u-net. the proposed method performs equally well or better than the alternatives. for a highly reduced number of projections, only the u-net method provides images comparable to those obtained with tomodl. however, tomodl has a much better performance if the amount of data available for training is limited, given that the number of network trainable parameters is smaller.","key_phrases":["this work","we","a model-based deep learning reconstruction algorithm","optical projection tomography","acquisition and reconstruction times","the proposed method","a data consistency step","an image domain artefact removal step","a convolutional neural network","a preprocessing stage","potential misalignments","the sample center","rotation","the detector","the algorithm","a database","wild-type zebrafish","danio rerio","different stages","development","the mean square error","a fixed number","iterations","a cross-validation scheme","we","the results","other reconstruction methods","filtered backprojection","compressed sensing","a direct deep learning method","the pseudo-inverse solution","a u","-","net","the proposed method","the alternatives","a highly reduced number","projections","only the u-net method","images","those","a much better performance","the amount","data","training","the number","network trainable parameters","danio rerio"]},{"title":"Mammography using low-frequency electromagnetic fields with deep learning","authors":["Hamid Akbari-Chelaresi","Dawood Alsaedi","Seyed Hossein Mirjahanmardi","Mohamed El Badawe","Ali M. Albishi","Vahid Nayyeri","Omar M. Ramahi"],"abstract":"In this paper, a novel technique for detecting female breast anomalous tissues is presented and validated through numerical simulations. The technique, to a high degree, resembles X-ray mammography; however, instead of using X-rays for obtaining images of the breast, low-frequency electromagnetic fields are leveraged. To capture breast impressions, a metasurface, which can be thought of as analogous to X-rays film, has been employed. To achieve deep and sufficient penetration within the breast tissues, the source of excitation is a simple narrow-band dipole antenna operating at 200\u00a0MHz. The metasurface is designed to operate at the same frequency. The detection mechanism is based on comparing the impressions obtained from the breast under examination to the reference case (healthy breasts) using machine learning techniques. Using this system, not only would it be possible to detect tumors (benign or malignant), but one can also determine the location and size of the tumors. Remarkably, deep learning models were found to achieve very high classification accuracy.","doi":"10.1038\/s41598-023-40494-x","cleaned_title":"mammography using low-frequency electromagnetic fields with deep learning","cleaned_abstract":"in this paper, a novel technique for detecting female breast anomalous tissues is presented and validated through numerical simulations. the technique, to a high degree, resembles x-ray mammography; however, instead of using x-rays for obtaining images of the breast, low-frequency electromagnetic fields are leveraged. to capture breast impressions, a metasurface, which can be thought of as analogous to x-rays film, has been employed. to achieve deep and sufficient penetration within the breast tissues, the source of excitation is a simple narrow-band dipole antenna operating at 200 mhz. the metasurface is designed to operate at the same frequency. the detection mechanism is based on comparing the impressions obtained from the breast under examination to the reference case (healthy breasts) using machine learning techniques. using this system, not only would it be possible to detect tumors (benign or malignant), but one can also determine the location and size of the tumors. remarkably, deep learning models were found to achieve very high classification accuracy.","key_phrases":["this paper","a novel technique","female breast anomalous tissues","numerical simulations","the technique","a high degree","x-ray mammography","x","-","rays","images","the breast, low-frequency electromagnetic fields","breast impressions","a metasurface","which","x-rays film","deep and sufficient penetration","the breast tissues","the source","excitation","a simple narrow-band dipole antenna","200 mhz","the metasurface","the same frequency","the detection mechanism","the impressions","the breast","examination","the reference case","healthy breasts","machine learning techniques","this system","it","tumors","one","the location","size","the tumors","deep learning models","very high classification accuracy","dipole antenna","200"]},{"title":"Taxonomy of deep learning-based intrusion detection system approaches in fog computing: a systematic review","authors":["Sepide Najafli","Abolfazl Toroghi Haghighat","Babak Karasfi"],"abstract":"The Internet of Things (IoT) has been used in various aspects. Fundamental security issues must be addressed to accelerate and develop the Internet of Things. An intrusion detection system (IDS) is an essential element in network security designed to detect and determine the type of attacks. The use of deep learning (DL) shows promising results in the design of IDS based on IoT. DL facilitates analytics and learning in the dynamic IoT domain. Some deep learning-based IDS in IOT sensors cannot be executed, because of resource restrictions. Although cloud computing could overcome limitations, the distance between the cloud and the end IoT sensors causes high communication costs, security problems and delays. Fog computing has been presented to handle these issues and can bring resources to the edge of the network. Many studies have been conducted to investigate IDS based on IoT. Our goal is to investigate and classify deep learning-based IDS on fog processing. In this paper, researchers can access comprehensive resources in this field. Therefore, first, we provide a complete classification of IDS in IoT. Then practical and important proposed IDSs in the fog environment are discussed in three groups (binary, multi-class, and hybrid), and are examined the advantages and disadvantages of each approach. The results show that most of the studied methods consider hybrid strategies (binary and multi-class). In addition, in the reviewed papers the average Accuracy obtained in the binary method is better than the multi-class. Finally, we highlight some challenges and future directions for the next research in IDS techniques.","doi":"10.1007\/s10115-024-02162-y","cleaned_title":"taxonomy of deep learning-based intrusion detection system approaches in fog computing: a systematic review","cleaned_abstract":"the internet of things (iot) has been used in various aspects. fundamental security issues must be addressed to accelerate and develop the internet of things. an intrusion detection system (ids) is an essential element in network security designed to detect and determine the type of attacks. the use of deep learning (dl) shows promising results in the design of ids based on iot. dl facilitates analytics and learning in the dynamic iot domain. some deep learning-based ids in iot sensors cannot be executed, because of resource restrictions. although cloud computing could overcome limitations, the distance between the cloud and the end iot sensors causes high communication costs, security problems and delays. fog computing has been presented to handle these issues and can bring resources to the edge of the network. many studies have been conducted to investigate ids based on iot. our goal is to investigate and classify deep learning-based ids on fog processing. in this paper, researchers can access comprehensive resources in this field. therefore, first, we provide a complete classification of ids in iot. then practical and important proposed idss in the fog environment are discussed in three groups (binary, multi-class, and hybrid), and are examined the advantages and disadvantages of each approach. the results show that most of the studied methods consider hybrid strategies (binary and multi-class). in addition, in the reviewed papers the average accuracy obtained in the binary method is better than the multi-class. finally, we highlight some challenges and future directions for the next research in ids techniques.","key_phrases":["the internet","things","iot","various aspects","fundamental security issues","the internet","things","an intrusion detection system","ids","an essential element","network security","the type","attacks","the use","deep learning","dl","results","the design","ids","iot","dl","analytics","the dynamic iot domain","some deep learning-based ids","iot sensors","resource restrictions","cloud computing","limitations","the distance","the cloud","the end iot sensors","high communication costs","security problems","delays","fog computing","these issues","resources","the edge","the network","many studies","ids","iot","our goal","deep learning-based ids","fog processing","this paper","researchers","comprehensive resources","this field","we","a complete classification","ids","iot","practical and important proposed idss","the fog environment","three groups","the advantages","disadvantages","each approach","the results","the studied methods","hybrid strategies","binary and multi-class","addition","the reviewed papers","the average accuracy","the binary method","the multi","-","class","we","some challenges","future directions","the next research","ids techniques","fog computing","first","three"]},{"title":"Deep learning-based pathway-centric approach to characterize recurrent hepatocellular carcinoma after liver transplantation","authors":["Jeffrey To","Soumita Ghosh","Xun Zhao","Elisa Pasini","Sandra Fischer","Gonzalo Sapisochin","Anand Ghanekar","Elmar Jaeckel","Mamatha Bhat"],"abstract":"BackgroundLiver transplantation (LT) is offered as a cure for Hepatocellular carcinoma (HCC), however 15\u201320% develop recurrence post-transplant which tends to be aggressive. In this study, we examined the transcriptome profiles of patients with recurrent HCC to identify differentially expressed genes (DEGs), the involved pathways, biological functions, and potential gene signatures of recurrent HCC post-transplant using deep machine learning (ML) methodology.Materials and methodsWe analyzed the transcriptomic profiles of primary and recurrent tumor samples from 7 pairs of patients who underwent LT. Following differential gene expression analysis, we performed pathway enrichment, gene ontology (GO) analyses and protein-protein interactions (PPIs) with top 10 hub gene networks. We also predicted the landscape of infiltrating immune cells using Cibersortx. We next develop pathway and GO term-based deep learning models leveraging primary tissue gene expression data from The Cancer Genome Atlas (TCGA) to identify gene signatures in recurrent HCC.ResultsThe PI3K\/Akt signaling pathway and cytokine-mediated signaling pathway were particularly activated in HCC recurrence. The recurrent tumors exhibited upregulation of an immune-escape related gene, CD274, in the top 10 hub gene analysis. Significantly higher infiltration of monocytes and lower M1 macrophages were found in recurrent HCC tumors. Our deep learning approach identified a 20-gene signature in recurrent HCC. Amongst the 20 genes, through multiple analysis, IL6 was found to be significantly associated with HCC recurrence.ConclusionOur deep learning approach identified PI3K\/Akt signaling as potentially regulating cytokine-mediated functions and the expression of immune escape genes, leading to alterations in the pattern of immune cell infiltration. In conclusion, IL6 was identified to play an important role in HCC recurrence.","doi":"10.1186\/s40246-024-00624-6","cleaned_title":"deep learning-based pathway-centric approach to characterize recurrent hepatocellular carcinoma after liver transplantation","cleaned_abstract":"backgroundliver transplantation (lt) is offered as a cure for hepatocellular carcinoma (hcc), however 15\u201320% develop recurrence post-transplant which tends to be aggressive. in this study, we examined the transcriptome profiles of patients with recurrent hcc to identify differentially expressed genes (degs), the involved pathways, biological functions, and potential gene signatures of recurrent hcc post-transplant using deep machine learning (ml) methodology.materials and methodswe analyzed the transcriptomic profiles of primary and recurrent tumor samples from 7 pairs of patients who underwent lt. following differential gene expression analysis, we performed pathway enrichment, gene ontology (go) analyses and protein-protein interactions (ppis) with top 10 hub gene networks. we also predicted the landscape of infiltrating immune cells using cibersortx. we next develop pathway and go term-based deep learning models leveraging primary tissue gene expression data from the cancer genome atlas (tcga) to identify gene signatures in recurrent hcc.resultsthe pi3k\/akt signaling pathway and cytokine-mediated signaling pathway were particularly activated in hcc recurrence. the recurrent tumors exhibited upregulation of an immune-escape related gene, cd274, in the top 10 hub gene analysis. significantly higher infiltration of monocytes and lower m1 macrophages were found in recurrent hcc tumors. our deep learning approach identified a 20-gene signature in recurrent hcc. amongst the 20 genes, through multiple analysis, il6 was found to be significantly associated with hcc recurrence.conclusionour deep learning approach identified pi3k\/akt signaling as potentially regulating cytokine-mediated functions and the expression of immune escape genes, leading to alterations in the pattern of immune cell infiltration. in conclusion, il6 was identified to play an important role in hcc recurrence.","key_phrases":["backgroundliver transplantation","lt","a cure","hepatocellular carcinoma","hcc","15\u201320%","recurrence","transplant","which","this study","we","the transcriptome profiles","patients","recurrent hcc","differentially expressed genes","the involved pathways","biological functions","potential gene signatures","recurrent hcc post","-","deep machine learning","(ml) methodology.materials","methodswe","the transcriptomic profiles","primary and recurrent tumor samples","7 pairs","patients","who","lt","differential gene expression analysis","we","pathway enrichment","gene ontology","analyses","protein-protein interactions","top 10 hub gene networks","we","the landscape","infiltrating immune cells","cibersortx","we","term-based deep learning models","primary tissue gene expression data","the cancer genome atlas","(tcga","gene signatures","recurrent hcc.resultsthe pi3k\/akt","pathway","hcc recurrence","the recurrent tumors","upregulation","an immune-escape related gene","cd274","the top 10 hub gene analysis","significantly higher infiltration","monocytes","lower m1 macrophages","recurrent hcc tumors","our deep learning approach","a 20-gene signature","recurrent hcc","the 20 genes","multiple analysis","il6","hcc","recurrence.conclusionour deep learning approach","pi3k\/akt","cytokine-mediated functions","the expression","immune escape genes","alterations","the pattern","immune cell infiltration","conclusion","il6","an important role","hcc recurrence","15\u201320%","7","10","10","20","20"]},{"title":"Deep learning models for perception of brightness related illusions","authors":["Amrita Mukherjee","Avijit Paul","Kuntal Ghosh"],"abstract":"Illusions are like holes in our effortless visual mechanism through which we can peep into the internal mechanisms of the brain. Scientists attempted to explain underlying physiological, physical, and cognitive mechanisms of illusions by the receptive field hierarchical organizations, information sampling, filtering, etc. Some antagonistic illusions cannot be explained by them and for this, deep learning networks were used recently as a model for illusion perception. To further broaden the scope of the perceptual functionality in the brightness contrast genre, handle the background removal effects on some illusions that reduce the illusory effects, and replicate the antagonistic illusions with the same parameter setup, we have used Convolutional Neural Network, Autoencoder, U-Net, and U-Net++ models for replicating the visual illusions. The networks are specialized in low-level vision tasks like De-noising, De-blurring, and a combination of both. A high number of brightness contrast visual illusions are tested on all the networks and most of the outcomes significantly matched human perceptions. Overall, our method will guide the development of neurobiological frameworks which might enrich the computational neuroscience study by distilling some biological principles. On the other hand, the machine learning community will benefit from knowing the inherent flaws of the networks so that the true image of reality can be taken into consideration, especially in imaging situations where experts too can be deceived.","doi":"10.1007\/s10489-024-05658-w","cleaned_title":"deep learning models for perception of brightness related illusions","cleaned_abstract":"illusions are like holes in our effortless visual mechanism through which we can peep into the internal mechanisms of the brain. scientists attempted to explain underlying physiological, physical, and cognitive mechanisms of illusions by the receptive field hierarchical organizations, information sampling, filtering, etc. some antagonistic illusions cannot be explained by them and for this, deep learning networks were used recently as a model for illusion perception. to further broaden the scope of the perceptual functionality in the brightness contrast genre, handle the background removal effects on some illusions that reduce the illusory effects, and replicate the antagonistic illusions with the same parameter setup, we have used convolutional neural network, autoencoder, u-net, and u-net++ models for replicating the visual illusions. the networks are specialized in low-level vision tasks like de-noising, de-blurring, and a combination of both. a high number of brightness contrast visual illusions are tested on all the networks and most of the outcomes significantly matched human perceptions. overall, our method will guide the development of neurobiological frameworks which might enrich the computational neuroscience study by distilling some biological principles. on the other hand, the machine learning community will benefit from knowing the inherent flaws of the networks so that the true image of reality can be taken into consideration, especially in imaging situations where experts too can be deceived.","key_phrases":["illusions","holes","our effortless visual mechanism","which","we","the internal mechanisms","the brain","scientists","underlying physiological, physical, and cognitive mechanisms","illusions","the receptive field hierarchical organizations","information sampling","filtering","some antagonistic illusions","them","this","deep learning networks","a model","illusion perception","the scope","the perceptual functionality","the brightness contrast genre","the background removal effects","some illusions","that","the illusory effects","the antagonistic illusions","the same parameter setup","we","convolutional neural network","autoencoder","-","net","u-net++ models","the visual illusions","the networks","low-level vision tasks","de","de","-","a combination","both","a high number","brightness contrast visual illusions","all the networks","the outcomes","human perceptions","our method","the development","neurobiological frameworks","which","the computational neuroscience study","some biological principles","the other hand","the machine learning community","the inherent flaws","the networks","the true image","reality","consideration","imaging situations","experts"]},{"title":"Exploiting biochemical data to improve osteosarcoma diagnosis with deep learning","authors":["Shidong Wang","Yangyang Shen","Fanwei Zeng","Meng Wang","Bohan Li","Dian Shen","Xiaodong Tang","Beilun Wang"],"abstract":"Early and accurate diagnosis of osteosarcomas (OS) is of great clinical significance, and machine learning (ML) based methods are increasingly adopted. However, current ML-based methods for osteosarcoma diagnosis consider only X-ray images, usually fail to generalize to new cases, and lack explainability. In this paper, we seek to explore the capability of deep learning models in diagnosing primary OS, with higher accuracy, explainability, and generality. Concretely, we analyze the added value of integrating the biochemical data, i.e., alkaline phosphatase (ALP) and lactate dehydrogenase (LDH), and design a model that incorporates the numerical features of ALP and LDH and the visual features of X-ray imaging through a late fusion approach in the feature space. We evaluate this model on real-world clinic data with 848 patients aged from 4 to 81. The experimental results reveal the effectiveness of incorporating ALP and LDH simultaneously in a late fusion approach, with the accuracy of the considered 2608 cases increased to 97.17%, compared to 94.35% in the baseline. Grad-CAM visualizations consistent with orthopedic specialists further justified the model\u2019s explainability.","doi":"10.1007\/s13755-024-00288-5","cleaned_title":"exploiting biochemical data to improve osteosarcoma diagnosis with deep learning","cleaned_abstract":"early and accurate diagnosis of osteosarcomas (os) is of great clinical significance, and machine learning (ml) based methods are increasingly adopted. however, current ml-based methods for osteosarcoma diagnosis consider only x-ray images, usually fail to generalize to new cases, and lack explainability. in this paper, we seek to explore the capability of deep learning models in diagnosing primary os, with higher accuracy, explainability, and generality. concretely, we analyze the added value of integrating the biochemical data, i.e., alkaline phosphatase (alp) and lactate dehydrogenase (ldh), and design a model that incorporates the numerical features of alp and ldh and the visual features of x-ray imaging through a late fusion approach in the feature space. we evaluate this model on real-world clinic data with 848 patients aged from 4 to 81. the experimental results reveal the effectiveness of incorporating alp and ldh simultaneously in a late fusion approach, with the accuracy of the considered 2608 cases increased to 97.17%, compared to 94.35% in the baseline. grad-cam visualizations consistent with orthopedic specialists further justified the model\u2019s explainability.","key_phrases":["early and accurate diagnosis","great clinical significance","machine learning (ml) based methods","current ml-based methods","osteosarcoma diagnosis","only x-ray images","new cases","lack explainability","this paper","we","the capability","deep learning models","higher accuracy","explainability","generality","we","the added value","the biochemical data","i.e., alkaline phosphatase","alp","lactate dehydrogenase","ldh","a model","that","the numerical features","alp","ldh","the visual features","x","-ray imaging","a late fusion approach","the feature space","we","this model","real-world clinic data","848 patients","the experimental results","the effectiveness","alp","ldh","a late fusion approach","the accuracy","the considered 2608 cases","97.17%","94.35%","the baseline","grad-cam visualizations","orthopedic specialists","the model\u2019s explainability","848","4","81","2608","97.17%","94.35%"]},{"title":"Deep learning-based predictive classification of functional subpopulations of hematopoietic stem cells and multipotent progenitors","authors":["Shen Wang","Jianzhong Han","Jingru Huang","Khayrul Islam","Yuheng Shi","Yuyuan Zhou","Dongwook Kim","Jane Zhou","Zhaorui Lian","Yaling Liu","Jian Huang"],"abstract":"BackgroundHematopoietic stem cells (HSCs) and multipotent progenitors (MPPs) play a pivotal role in maintaining lifelong hematopoiesis. The distinction between stem cells and other progenitors, as well as the assessment of their functions, has long been a central focus in stem cell research. In recent years, deep learning has emerged as a powerful tool for cell image analysis and classification\/prediction.MethodsIn this study, we explored the feasibility of employing deep learning techniques to differentiate murine HSCs and MPPs based solely on their morphology, as observed through light microscopy (DIC) images.ResultsAfter rigorous training and validation using extensive image datasets, we successfully developed a three-class classifier, referred to as the LSM model, capable of reliably distinguishing long-term HSCs, short-term HSCs, and MPPs. The LSM model extracts intrinsic morphological features unique to different cell types, irrespective of the methods used for cell identification and isolation, such as surface markers or intracellular GFP markers. Furthermore, employing the same deep learning framework, we created a two-class classifier that effectively discriminates between aged HSCs and young HSCs. This discovery is particularly significant as both cell types share identical surface markers yet serve distinct functions. This classifier holds the potential to offer a novel, rapid, and efficient means of assessing the functional states of HSCs, thus obviating the need for time-consuming transplantation experiments.ConclusionOur study represents the pioneering use of deep learning to differentiate HSCs and MPPs under steady-state conditions. This novel and robust deep learning-based platform will provide a basis for the future development of a new generation stem cell identification and separation system. It may also provide new insight into the molecular mechanisms underlying stem cell self-renewal.","doi":"10.1186\/s13287-024-03682-8","cleaned_title":"deep learning-based predictive classification of functional subpopulations of hematopoietic stem cells and multipotent progenitors","cleaned_abstract":"backgroundhematopoietic stem cells (hscs) and multipotent progenitors (mpps) play a pivotal role in maintaining lifelong hematopoiesis. the distinction between stem cells and other progenitors, as well as the assessment of their functions, has long been a central focus in stem cell research. in recent years, deep learning has emerged as a powerful tool for cell image analysis and classification\/prediction.methodsin this study, we explored the feasibility of employing deep learning techniques to differentiate murine hscs and mpps based solely on their morphology, as observed through light microscopy (dic) images.resultsafter rigorous training and validation using extensive image datasets, we successfully developed a three-class classifier, referred to as the lsm model, capable of reliably distinguishing long-term hscs, short-term hscs, and mpps. the lsm model extracts intrinsic morphological features unique to different cell types, irrespective of the methods used for cell identification and isolation, such as surface markers or intracellular gfp markers. furthermore, employing the same deep learning framework, we created a two-class classifier that effectively discriminates between aged hscs and young hscs. this discovery is particularly significant as both cell types share identical surface markers yet serve distinct functions. this classifier holds the potential to offer a novel, rapid, and efficient means of assessing the functional states of hscs, thus obviating the need for time-consuming transplantation experiments.conclusionour study represents the pioneering use of deep learning to differentiate hscs and mpps under steady-state conditions. this novel and robust deep learning-based platform will provide a basis for the future development of a new generation stem cell identification and separation system. it may also provide new insight into the molecular mechanisms underlying stem cell self-renewal.","key_phrases":["backgroundhematopoietic stem cells","hscs","multipotent progenitors","mpps","a pivotal role","lifelong hematopoiesis","the distinction","stem cells","other progenitors","the assessment","their functions","a central focus","stem cell research","recent years","deep learning","a powerful tool","cell image analysis","classification\/prediction.methodsin","we","the feasibility","deep learning techniques","murine hscs","mpps","their morphology","light microscopy (dic) images.resultsafter rigorous training","validation","extensive image datasets","we","a three-class classifier","the lsm model","reliably distinguishing long-term hscs","short-term hscs","mpps","the lsm model","intrinsic morphological features","different cell types","the methods","cell identification","isolation","surface markers","intracellular gfp markers","the same deep learning framework","we","a two-class classifier","that","aged hscs","young hscs","this discovery","both cell types","identical surface markers","distinct functions","this classifier","the potential","a novel, rapid, and efficient means","the functional states","hscs","the need","time-consuming transplantation","experiments.conclusionour study","the pioneering use","deep learning","hscs","mpps","steady-state conditions","this novel and robust deep learning-based platform","a basis","the future development","a new generation stem cell identification","separation system","it","new insight","the molecular mechanisms","stem cell self-renewal","recent years","murine hscs","three","two"]},{"title":"Fully dynamic reorder policies with deep reinforcement learning for multi-echelon inventory management","authors":["Patric Hammler","Nicolas Riesterer","Torsten Braun"],"abstract":"The operation of inventory systems plays an important role in the success of manufacturing companies, making it a\u00a0highly relevant domain for optimization. In particular, the domain lends itself to being approached via Deep Reinforcement Learning (DRL) models due to it requiring sequential reorder decisions based on uncertainty to minimize cost. In this paper, we evaluate state-of-the-art optimization approaches to determine whether Deep Reinforcement Learning can be applied to the multi-echelon inventory optimization (MEIO) framework in a\u00a0practically feasible manner to generate fully dynamic reorder policies. We investigate how it performs in comparison to an optimized static reorder policy, how robust it is when it comes to structural changes in the environment, and whether the use of DRL is safe in terms of risk in real-world applications. Our results show promising performance for DRL with potential for improvement in terms of minimizing risky behavior.","doi":"10.1007\/s00287-023-01556-6","cleaned_title":"fully dynamic reorder policies with deep reinforcement learning for multi-echelon inventory management","cleaned_abstract":"the operation of inventory systems plays an important role in the success of manufacturing companies, making it a highly relevant domain for optimization. in particular, the domain lends itself to being approached via deep reinforcement learning (drl) models due to it requiring sequential reorder decisions based on uncertainty to minimize cost. in this paper, we evaluate state-of-the-art optimization approaches to determine whether deep reinforcement learning can be applied to the multi-echelon inventory optimization (meio) framework in a practically feasible manner to generate fully dynamic reorder policies. we investigate how it performs in comparison to an optimized static reorder policy, how robust it is when it comes to structural changes in the environment, and whether the use of drl is safe in terms of risk in real-world applications. our results show promising performance for drl with potential for improvement in terms of minimizing risky behavior.","key_phrases":["the operation","inventory systems","an important role","the success","manufacturing companies","it","optimization","the domain","itself","deep reinforcement learning (drl) models","it","sequential reorder decisions","uncertainty","cost","this paper","we","the-art","deep reinforcement learning","the multi-echelon inventory optimization","meio) framework","a practically feasible manner","fully dynamic reorder policies","we","it","comparison","an optimized static reorder policy","it","it","structural changes","the environment","the use","drl","terms","risk","real-world applications","our results","promising performance","drl","potential","improvement","terms","risky behavior"]},{"title":"Deep learning models for webcam eye tracking in online experiments","authors":["Shreshth Saxena","Lauren K. Fink","Elke B. Lange"],"abstract":"Eye tracking is prevalent in scientific and commercial applications. Recent computer vision and deep learning methods enable eye tracking with off-the-shelf webcams and reduce dependence on expensive, restrictive hardware. However, such deep learning methods have not yet been applied and evaluated for remote, online psychological experiments. In this study, we tackle critical challenges faced in remote eye tracking setups and systematically evaluate appearance-based deep learning methods of gaze tracking and blink detection. From their own homes and laptops, 65 participants performed a battery of eye\u00a0tracking tasks including (i) fixation, (ii) zone classification, (iii) free viewing, (iv) smooth pursuit, and (v) blink detection. Webcam recordings of the participants performing these tasks were processed offline through appearance-based models of gaze and blink detection. The task battery required different eye movements that characterized gaze and blink prediction accuracy over a comprehensive list of measures. We find the best gaze accuracy to be 2.4\u00b0 and precision of 0.47\u00b0, which outperforms previous online eye\u00a0tracking studies and reduces the gap between laboratory-based and online eye\u00a0tracking performance. We release the experiment template, recorded data, and analysis code with the motivation to escalate affordable, accessible, and scalable eye tracking that has the potential to accelerate research in the fields of psychological science, cognitive neuroscience, user experience design, and human\u2013computer interfaces.","doi":"10.3758\/s13428-023-02190-6","cleaned_title":"deep learning models for webcam eye tracking in online experiments","cleaned_abstract":"eye tracking is prevalent in scientific and commercial applications. recent computer vision and deep learning methods enable eye tracking with off-the-shelf webcams and reduce dependence on expensive, restrictive hardware. however, such deep learning methods have not yet been applied and evaluated for remote, online psychological experiments. in this study, we tackle critical challenges faced in remote eye tracking setups and systematically evaluate appearance-based deep learning methods of gaze tracking and blink detection. from their own homes and laptops, 65 participants performed a battery of eye tracking tasks including (i) fixation, (ii) zone classification, (iii) free viewing, (iv) smooth pursuit, and (v) blink detection. webcam recordings of the participants performing these tasks were processed offline through appearance-based models of gaze and blink detection. the task battery required different eye movements that characterized gaze and blink prediction accuracy over a comprehensive list of measures. we find the best gaze accuracy to be 2.4\u00b0 and precision of 0.47\u00b0, which outperforms previous online eye tracking studies and reduces the gap between laboratory-based and online eye tracking performance. we release the experiment template, recorded data, and analysis code with the motivation to escalate affordable, accessible, and scalable eye tracking that has the potential to accelerate research in the fields of psychological science, cognitive neuroscience, user experience design, and human\u2013computer interfaces.","key_phrases":["eye tracking","scientific and commercial applications","recent computer vision","deep learning methods","eye tracking","the-shelf","dependence","expensive, restrictive hardware","such deep learning methods","remote, online psychological experiments","this study","we","critical challenges","remote eye tracking setups","appearance-based deep learning methods","gaze tracking","blink","detection","their own homes","laptops","65 participants","a battery","eye tracking tasks","(i) fixation","(ii) zone classification","iii","free viewing","smooth pursuit","(v) blink detection","webcam recordings","the participants","these tasks","appearance-based models","gaze","detection","the task battery","different eye movements","that","gaze","prediction accuracy","a comprehensive list","measures","we","the best gaze accuracy","precision","which","previous online eye tracking studies","the gap","laboratory-based and online eye tracking performance","we","the experiment template","recorded data","analysis code","the motivation","affordable, accessible, and scalable eye tracking","that","the potential","research","the fields","psychological science","cognitive neuroscience","user experience design","human\u2013computer interfaces","65","2.4","0.47"]},{"title":"Assessment of indoor risk through deep learning -based object recognition in disaster situations","authors":["Irshad Khan","Ziyi Guo","Kihwan Lim","Jaeseon Kim","Young-Woo Kwon"],"abstract":"Disasters can devastate individuals and their properties, highlighting the importance of risk assessment to promote safety. Recently, deep learning techniques have shown the potential in identifying hazardous situations during disasters. Recognizing potentially dangerous objects in indoor environments can be essential for assisting individuals in responding appropriately to emergencies. In this article, we present an indoor-risk analysis framework for disasters based on deep learning. Our framework utilizes modern deep learning techniques to calculate an indoor risk rating based on dangerous objects\u2019 sizes, enabling comprehensive risk assessment of indoor environments during disasters. To that end, we use (Mask R-CNN) to identify hazardous indoor objects in disaster situations with 94% accuracy. By incorporating object size information, our framework offers a more nuanced and detailed risk assessment than previous approaches. Our proposed system provides a valuable tool for promoting ongoing safety improvement and enhancing indoor safety during natural disasters.","doi":"10.1007\/s11042-023-16711-0","cleaned_title":"assessment of indoor risk through deep learning -based object recognition in disaster situations","cleaned_abstract":"disasters can devastate individuals and their properties, highlighting the importance of risk assessment to promote safety. recently, deep learning techniques have shown the potential in identifying hazardous situations during disasters. recognizing potentially dangerous objects in indoor environments can be essential for assisting individuals in responding appropriately to emergencies. in this article, we present an indoor-risk analysis framework for disasters based on deep learning. our framework utilizes modern deep learning techniques to calculate an indoor risk rating based on dangerous objects\u2019 sizes, enabling comprehensive risk assessment of indoor environments during disasters. to that end, we use (mask r-cnn) to identify hazardous indoor objects in disaster situations with 94% accuracy. by incorporating object size information, our framework offers a more nuanced and detailed risk assessment than previous approaches. our proposed system provides a valuable tool for promoting ongoing safety improvement and enhancing indoor safety during natural disasters.","key_phrases":["disasters","individuals","their properties","the importance","risk assessment","safety","deep learning techniques","the potential","hazardous situations","disasters","potentially dangerous objects","indoor environments","individuals","emergencies","this article","we","an indoor-risk analysis framework","disasters","deep learning","our framework","modern deep learning techniques","an indoor risk rating","dangerous objects\u2019 sizes","comprehensive risk assessment","indoor environments","disasters","that end","we","-cnn","hazardous indoor objects","disaster situations","94% accuracy","object size information","our framework","a more nuanced and detailed risk assessment","previous approaches","our proposed system","a valuable tool","ongoing safety improvement","indoor safety","natural disasters","94%"]},{"title":"A deep learning dataset for sample preparation artefacts detection in multispectral high-content microscopy","authors":["Vaibhav Sharma","Artur Yakimovich"],"abstract":"High-content image-based screening is widely used in Drug Discovery and Systems Biology. However, sample preparation artefacts may significantly deteriorate the quality of image-based screening assays. While detection and circumvention of such artefacts could be addressed using modern-day machine learning and deep learning algorithms, this is widely impeded by the lack of suitable datasets. To address this, here we present a purpose-created open dataset of high-content microscopy sample preparation artefact. It consists of high-content microscopy of laboratory dust titrated on fixed cell culture specimens imaged with fluorescence filters covering the complete spectral range. To ensure this dataset is suitable for supervised machine learning tasks like image classification or segmentation we propose rule-based annotation strategies on categorical and pixel levels. We demonstrate the applicability of our dataset for deep learning by training a convolutional-neural-network-based classifier.","doi":"10.1038\/s41597-024-03064-y","cleaned_title":"a deep learning dataset for sample preparation artefacts detection in multispectral high-content microscopy","cleaned_abstract":"high-content image-based screening is widely used in drug discovery and systems biology. however, sample preparation artefacts may significantly deteriorate the quality of image-based screening assays. while detection and circumvention of such artefacts could be addressed using modern-day machine learning and deep learning algorithms, this is widely impeded by the lack of suitable datasets. to address this, here we present a purpose-created open dataset of high-content microscopy sample preparation artefact. it consists of high-content microscopy of laboratory dust titrated on fixed cell culture specimens imaged with fluorescence filters covering the complete spectral range. to ensure this dataset is suitable for supervised machine learning tasks like image classification or segmentation we propose rule-based annotation strategies on categorical and pixel levels. we demonstrate the applicability of our dataset for deep learning by training a convolutional-neural-network-based classifier.","key_phrases":["high-content image-based screening","drug discovery and systems biology","sample preparation artefacts","the quality","image-based screening assays","detection","circumvention","such artefacts","modern-day machine learning","deep learning algorithms","this","the lack","suitable datasets","this","we","a purpose-created open dataset","high-content microscopy sample preparation artefact","it","high-content microscopy","laboratory dust","fixed cell culture specimens","fluorescence filters","the complete spectral range","this dataset","supervised machine learning tasks","image classification","segmentation","we","rule-based annotation strategies","categorical and pixel levels","we","the applicability","our dataset","deep learning","a convolutional-neural-network-based classifier"]},{"title":"CT-based deep learning model for predicting hospital discharge outcome in spontaneous intracerebral hemorrhage","authors":["Xianjing Zhao","Bijing Zhou","Yong Luo","Lei Chen","Lequn Zhu","Shixin Chang","Xiangming Fang","Zhenwei Yao"],"abstract":"ObjectivesTo predict the functional outcome of patients with intracerebral hemorrhage (ICH) using deep learning models based on computed tomography (CT) images.MethodsA retrospective, bi-center study of ICH patients was conducted. Firstly, a custom 3D convolutional model was built for predicting the functional outcome of ICH patients based on CT scans from randomly selected ICH patients in H training dataset collected from H hospital. Secondly, clinical data and radiological features were collected at admission and the Extreme Gradient Boosting (XGBoost) algorithm was used to establish a second model, named the XGBoost model. Finally, the Convolution model and XGBoost model were fused to build the third \u201cFusion model.\u201d Favorable outcome was defined as modified Rankin Scale score of 0\u20133 at discharge. The prognostic predictive accuracy of the three models was evaluated using an H test dataset and an external Y dataset, and compared with the performance of ICH score and ICH grading scale (ICH-GS).ResultsA total of 604 patients with ICH were included in this study, of which 450 patients were in the H training dataset, 50 patients in the H test dataset, and 104 patients in the Y dataset. In the Y dataset, the areas under the curve (AUCs) of the Convolution model, XGBoost model, and Fusion model were 0.829, 0.871, and 0.905, respectively. The Fusion model prognostic performance exceeded that of ICH score and ICH-GS (p\u2009=\u20090.043 and p\u2009=\u20090.045, respectively).ConclusionsDeep learning models have good accuracy for predicting functional outcome of patients with spontaneous intracerebral hemorrhage.Clinical relevance statementThe proposed deep learning Fusion model may assist clinicians in predicting functional outcome and developing treatment strategies, thereby improving the survival and quality of life of patients with spontaneous intracerebral hemorrhage.Key Points\u2022 Integrating clinical presentations, CT images, and radiological features to establish deep learning model for functional outcome prediction of patients with intracerebral hemorrhage.\u2022 Deep learning applied to CT images provides great help in prognosing functional outcome of intracerebral hemorrhage patients.\u2022 The developed deep learning model performs better than clinical prognostic scores in predicting functional outcome of patients with intracerebral hemorrhage.","doi":"10.1007\/s00330-023-10505-6","cleaned_title":"ct-based deep learning model for predicting hospital discharge outcome in spontaneous intracerebral hemorrhage","cleaned_abstract":"objectivesto predict the functional outcome of patients with intracerebral hemorrhage (ich) using deep learning models based on computed tomography (ct) images.methodsa retrospective, bi-center study of ich patients was conducted. firstly, a custom 3d convolutional model was built for predicting the functional outcome of ich patients based on ct scans from randomly selected ich patients in h training dataset collected from h hospital. secondly, clinical data and radiological features were collected at admission and the extreme gradient boosting (xgboost) algorithm was used to establish a second model, named the xgboost model. finally, the convolution model and xgboost model were fused to build the third \u201cfusion model.\u201d favorable outcome was defined as modified rankin scale score of 0\u20133 at discharge. the prognostic predictive accuracy of the three models was evaluated using an h test dataset and an external y dataset, and compared with the performance of ich score and ich grading scale (ich-gs).resultsa total of 604 patients with ich were included in this study, of which 450 patients were in the h training dataset, 50 patients in the h test dataset, and 104 patients in the y dataset. in the y dataset, the areas under the curve (aucs) of the convolution model, xgboost model, and fusion model were 0.829, 0.871, and 0.905, respectively. the fusion model prognostic performance exceeded that of ich score and ich-gs (p = 0.043 and p = 0.045, respectively).conclusionsdeep learning models have good accuracy for predicting functional outcome of patients with spontaneous intracerebral hemorrhage.clinical relevance statementthe proposed deep learning fusion model may assist clinicians in predicting functional outcome and developing treatment strategies, thereby improving the survival and quality of life of patients with spontaneous intracerebral hemorrhage.key points\u2022 integrating clinical presentations, ct images, and radiological features to establish deep learning model for functional outcome prediction of patients with intracerebral hemorrhage.\u2022 deep learning applied to ct images provides great help in prognosing functional outcome of intracerebral hemorrhage patients.\u2022 the developed deep learning model performs better than clinical prognostic scores in predicting functional outcome of patients with intracerebral hemorrhage.","key_phrases":["objectivesto","the functional outcome","patients","intracerebral hemorrhage","ich","deep learning models","computed tomography","(ct","images.methodsa retrospective, bi-center study","ich patients","a custom 3d convolutional model","the functional outcome","ich patients","ct scans","randomly selected ich patients","h training dataset","h hospital","clinical data","radiological features","admission","the extreme gradient","algorithm","a second model","the xgboost model","the convolution model","xgboost model","the third \u201cfusion model","favorable outcome","modified rankin scale score","0\u20133","discharge","the prognostic predictive accuracy","the three models","an h test dataset","an external y dataset","the performance","ich score","ich grading scale","ich-gs).resultsa total","604 patients","ich","this study","which","450 patients","the h training dataset","the h test dataset","104 patients","the y dataset","the y dataset","the areas","the curve","aucs","the convolution model","xgboost model","fusion model","the fusion model prognostic performance","ich score","ich-gs","respectively).conclusionsdeep learning models","good accuracy","functional outcome","patients","spontaneous intracerebral hemorrhage.clinical relevance statementthe proposed deep learning fusion model","clinicians","functional outcome","treatment strategies","the survival","quality","life","patients","clinical presentations","ct images","radiological features","deep learning model","functional outcome prediction","patients","intracerebral hemorrhage.\u2022 deep learning","ct images","great help","functional outcome","intracerebral hemorrhage patients.\u2022","the developed deep learning model","clinical prognostic scores","functional outcome","patients","intracerebral hemorrhage","images.methodsa","firstly","3d","secondly","second","third","three","604","450","50","104","0.829","0.871","0.905","0.043","0.045","clinicians"]},{"title":"Enhanced variational mode decomposition with deep learning SVM kernels for river streamflow forecasting","authors":["Subramaniam Nachimuthu Deepa","Narayanan Natarajan","Mohanadhas Berlin"],"abstract":"The present scenario of global climatic change challenges the sustainability and existence of water bodies around the globe. Due to which, it is always important and necessary to forecast the streamflow of rivers with respect to natural precipitation process. In this research study, novel enhanced variational mode decomposition (EVMD) with deep support vector machine (DSVM) kernels is proposed to perform forecasting of river streamflow. The developed computational intelligent machine learning model is a hybrid combination of the new enhanced VMD and the novel deep SVM kernels that is trained suitably to forecast the streamflow with respect to their deep learning layers. Initially, singular spectrum analysis (SSA) is employed for noise removal and the enhanced VMD with its features of decomposing and extracting more prominent features from the data are hybridized with the deep SVM kernel models to predict the streamflow of the considered Cahaba River data sets. Deterministic grey wolf optimizer (DGWO) is modelled in this research paper to fine tune the parameters of the deep SVM model. Previous prediction techniques modelled had difficulties in respect of local and global minima occurrences, stagnation, delayed and premature convergence and so on. Hence, in this study, the hybrid deep learning model forecasted the streamflow for the considered data sets and its superiority was validated with the comparative analysis with the previous forecasting techniques adopted. The developed forecasting model shall be used by the hydrologists for predicting the daily streamflow with the highest prediction accuracy rate of 97.54% with respect to training process and in case of testing mechanism the prediction accuracy rate is 96.47% This 97.54% of training prediction accuracy rate confirms the effectiveness of modelled new deep SVM algorithm for the streamflow forecasting of hydrologists.","doi":"10.1007\/s12665-023-11222-5","cleaned_title":"enhanced variational mode decomposition with deep learning svm kernels for river streamflow forecasting","cleaned_abstract":"the present scenario of global climatic change challenges the sustainability and existence of water bodies around the globe. due to which, it is always important and necessary to forecast the streamflow of rivers with respect to natural precipitation process. in this research study, novel enhanced variational mode decomposition (evmd) with deep support vector machine (dsvm) kernels is proposed to perform forecasting of river streamflow. the developed computational intelligent machine learning model is a hybrid combination of the new enhanced vmd and the novel deep svm kernels that is trained suitably to forecast the streamflow with respect to their deep learning layers. initially, singular spectrum analysis (ssa) is employed for noise removal and the enhanced vmd with its features of decomposing and extracting more prominent features from the data are hybridized with the deep svm kernel models to predict the streamflow of the considered cahaba river data sets. deterministic grey wolf optimizer (dgwo) is modelled in this research paper to fine tune the parameters of the deep svm model. previous prediction techniques modelled had difficulties in respect of local and global minima occurrences, stagnation, delayed and premature convergence and so on. hence, in this study, the hybrid deep learning model forecasted the streamflow for the considered data sets and its superiority was validated with the comparative analysis with the previous forecasting techniques adopted. the developed forecasting model shall be used by the hydrologists for predicting the daily streamflow with the highest prediction accuracy rate of 97.54% with respect to training process and in case of testing mechanism the prediction accuracy rate is 96.47% this 97.54% of training prediction accuracy rate confirms the effectiveness of modelled new deep svm algorithm for the streamflow forecasting of hydrologists.","key_phrases":["the present scenario","global climatic change","the sustainability","existence","water bodies","the globe","which","it","the streamflow","rivers","respect","natural precipitation process","this research study","novel enhanced variational mode decomposition","(evmd","deep support vector machine (dsvm) kernels","forecasting","river streamflow","the developed computational intelligent machine learning model","a hybrid combination","the new enhanced vmd","the novel deep svm kernels","that","the streamflow","respect","their deep learning layers","singular spectrum analysis","ssa","noise removal","the enhanced vmd","its features","more prominent features","the data","the deep svm kernel models","the streamflow","the considered cahaba river data sets","deterministic grey wolf optimizer","dgwo","this research paper","fine tune","the parameters","the deep svm model","previous prediction techniques","difficulties","respect","local and global minima occurrences","stagnation","convergence","this study","the hybrid deep learning model","the streamflow","the considered data sets","its superiority","the comparative analysis","the previous forecasting techniques","the developed forecasting model","the hydrologists","the daily streamflow","the highest prediction accuracy rate","97.54%","respect","training process","case","testing mechanism","the prediction accuracy rate","96.47%","this 97.54%","training prediction accuracy rate","the effectiveness","modelled new deep svm algorithm","the streamflow forecasting","hydrologists","kernels","grey wolf","daily","97.54%","96.47%","97.54%"]},{"title":"Sound signal analysis in Japanese speech recognition based on deep learning algorithm","authors":["Yang Xiaoxing"],"abstract":"As an important carrier of information, since sound can be collected quickly and is not limited by angle and light, it is often used to assist in understanding the environment and creating information. Voice signal recognition technology is a typical speech recognition application. This article focuses on the voice signal recognition technology around various deep learning models. By using deep learning neural networks with different structures and different types, information and representations related to the recognition of sound signal samples can be obtained, so as to further improve the detection accuracy of the sound signal recognition detection system. Based on this, this paper proposes an enhanced deep learning model of multi-scale neural convolutional network and uses it to recognize sound signals. The CCCP layer is used to reduce the dimensionality of the underlying feature map, so that the units captured in the network will eventually have internal features in each layer, thereby retaining the feature information to the maximum extent, which will form a convolutional multi-scale model in network deep learning Neurons. Finally, the article discusses the related issues of Japanese speech recognition on this basis. This article first uses the font (gra-phonem), that is, all these Japanese kana and common Chinese characters, using a total of 2795 units for modeling. There is a big gap between the experiment and the (BiLSTM-HMM) system. In addition, when Japanese speech is known, it is incorporated into the end-to-end recognition system to improve the performance of the Japanese speech recognition system. Based on the above-mentioned deep learning and sound signal analysis experiments and principles, the final effect obtained is better than the main effect of the Japanese speech recognition system of the latent Markov model and the long\u2013short memory network, thus promoting its development.","doi":"10.1007\/s13198-023-02025-9","cleaned_title":"sound signal analysis in japanese speech recognition based on deep learning algorithm","cleaned_abstract":"as an important carrier of information, since sound can be collected quickly and is not limited by angle and light, it is often used to assist in understanding the environment and creating information. voice signal recognition technology is a typical speech recognition application. this article focuses on the voice signal recognition technology around various deep learning models. by using deep learning neural networks with different structures and different types, information and representations related to the recognition of sound signal samples can be obtained, so as to further improve the detection accuracy of the sound signal recognition detection system. based on this, this paper proposes an enhanced deep learning model of multi-scale neural convolutional network and uses it to recognize sound signals. the cccp layer is used to reduce the dimensionality of the underlying feature map, so that the units captured in the network will eventually have internal features in each layer, thereby retaining the feature information to the maximum extent, which will form a convolutional multi-scale model in network deep learning neurons. finally, the article discusses the related issues of japanese speech recognition on this basis. this article first uses the font (gra-phonem), that is, all these japanese kana and common chinese characters, using a total of 2795 units for modeling. there is a big gap between the experiment and the (bilstm-hmm) system. in addition, when japanese speech is known, it is incorporated into the end-to-end recognition system to improve the performance of the japanese speech recognition system. based on the above-mentioned deep learning and sound signal analysis experiments and principles, the final effect obtained is better than the main effect of the japanese speech recognition system of the latent markov model and the long\u2013short memory network, thus promoting its development.","key_phrases":["an important carrier","information","sound","angle","light","it","the environment","information","voice signal recognition technology","a typical speech recognition application","this article","the voice signal recognition technology","various deep learning models","deep learning neural networks","different structures","different types","information","representations","the recognition","sound signal samples","the detection accuracy","the sound signal recognition detection system","this","this paper","an enhanced deep learning model","multi-scale neural convolutional network","it","sound signals","the cccp layer","the dimensionality","the underlying feature map","the units","the network","internal features","each layer","the feature information","the maximum extent","which","a convolutional multi-scale model","network deep learning neurons","the article","the related issues","japanese speech recognition","this basis","this article","the font","gra-phonem","all these japanese kana","common chinese characters","a total","2795 units","modeling","a big gap","the experiment","the (bilstm-hmm) system","addition","japanese speech","it","end","recognition","the performance","the japanese speech recognition system","the above-mentioned deep learning and sound signal analysis experiments","principles","the final effect","the main effect","the japanese speech recognition system","the latent markov model","the long\u2013short memory network","its development","japanese","japanese","chinese","2795","japanese","japanese","japanese"]},{"title":"SEFWaM\u2013deep learning based smart ensembled framework for waste management","authors":["Sujal Goel","Anannya Mishra","Garima Dua","Vandana Bhatia"],"abstract":"Waste generation has seen a significant surge in the last decade, presenting an urgent need for efficient and sustainable waste management strategies. The mounting piles of landfill waste underscore the criticality of segregating recyclable from non-recyclable waste, a measure that could alleviate numerous global environmental challenges. The increasing necessity to differentiate waste into specific categories such as paper, plastic, and metal is becoming evident; these categories demand distinct disposal and recycling methods. Addressing this automatic detection and segregation issue calls for an automated garbage classification system, achievable through computer vision. This paper introduces a deep learning and computer vision-based Smart Ensembled Framework for Waste Management (SEFWaM) to categorize garbage into various classes for improved waste management. This innovative model employs a fusion of transfer learning\u2014a deep learning technique and boosting to devise an ensembled approach for efficient waste classification. Our proposed approach, trained on the Trashnet 2.0 dataset, has demonstrated superior performance over competing algorithms in terms of Weighted Macro Precision (WMP), Weighted Macro Recall (WMR), Weighted Macro F1-Score (WMF), and test accuracy. The proposed SEFWaM model achieves an accuracy of 94.2% for the considered dataset, proving its superior efficiency over other deep learning-based models. This research thereby contributes to a pressing environmental issue by offering an automated, efficient solution for waste management.","doi":"10.1007\/s10668-023-03568-4","cleaned_title":"sefwam\u2013deep learning based smart ensembled framework for waste management","cleaned_abstract":"waste generation has seen a significant surge in the last decade, presenting an urgent need for efficient and sustainable waste management strategies. the mounting piles of landfill waste underscore the criticality of segregating recyclable from non-recyclable waste, a measure that could alleviate numerous global environmental challenges. the increasing necessity to differentiate waste into specific categories such as paper, plastic, and metal is becoming evident; these categories demand distinct disposal and recycling methods. addressing this automatic detection and segregation issue calls for an automated garbage classification system, achievable through computer vision. this paper introduces a deep learning and computer vision-based smart ensembled framework for waste management (sefwam) to categorize garbage into various classes for improved waste management. this innovative model employs a fusion of transfer learning\u2014a deep learning technique and boosting to devise an ensembled approach for efficient waste classification. our proposed approach, trained on the trashnet 2.0 dataset, has demonstrated superior performance over competing algorithms in terms of weighted macro precision (wmp), weighted macro recall (wmr), weighted macro f1-score (wmf), and test accuracy. the proposed sefwam model achieves an accuracy of 94.2% for the considered dataset, proving its superior efficiency over other deep learning-based models. this research thereby contributes to a pressing environmental issue by offering an automated, efficient solution for waste management.","key_phrases":["waste generation","a significant surge","the last decade","an urgent need","efficient and sustainable waste management strategies","the mounting piles","landfill waste","the criticality","non-recyclable waste","a measure","that","numerous global environmental challenges","the increasing necessity","waste","specific categories","paper","plastic","metal","these categories","distinct disposal","recycling methods","this automatic detection and segregation issue","an automated garbage classification system","computer vision","this paper","a deep learning and computer vision-based smart ensembled framework","waste management","sefwam","garbage","various classes","improved waste management","this innovative model","a fusion","transfer learning","a deep learning technique","an ensembled approach","efficient waste classification","our proposed approach","the trashnet 2.0 dataset","superior performance","competing algorithms","terms","weighted macro precision","wmp","macro recall","wmr","macro f1-score","wmf","test accuracy","the proposed sefwam model","an accuracy","94.2%","the considered dataset","its superior efficiency","other deep learning-based models","this research","a pressing environmental issue","an automated, efficient solution","waste management","the last decade","2.0","94.2%"]},{"title":"Graph-ensemble fusion for enhanced IoT intrusion detection: leveraging GCN and deep learning","authors":["Kajol Mittal","Payal Khurana Batra"],"abstract":"The proliferation of Internet of Things (IoT) applications has heightened the vulnerability of information security, making it susceptible to attacks that may lead to the compromise of sensitive data. Intrusion Detection System (IDS) is deployed in IoT networks for the detection of attacks and to ensure the security of information. In previous works, the IDS datasets suffer from an imbalanced distribution of data about attacks and, the flow of packets in IDS which hinders the ability of deep learning models for potent and coherent classification. With the emergence of graph convolution neural network (GCN), a new sub-field of deep learning models, the structure of graphs can be leveraged to represent the data effectively. IDS datasets typically consist of flow records of data which can naturally be represented as graph structures capturing both edge features and network topology information for classification of attacks. Hence, in this paper, a novel GCN-Ensemble fusion model is proposed for enhanced IoT IDS. There are three stages in this proposed model: (1) Data processing and attribute graph generation, (2) Feature engineering and (3) Classification. The flow attributes of data packets in IDS datasets are represented as the edges and the corresponding varying attacks as nodes of the attribute graph. Here the GCN model is leveraged for feature engineering of the IDS dataset. Further, a novel Ensemble of Convolution Neural Networks is proposed for the classification task. The evaluation of the proposed model encompasses the utilization of four distinct datasets, namely BoT-IoT, ToN-IoT, CIC-IDS2018, and NF UQ NIDS. In the BoT-IoT dataset, the proposed model demonstrates superior performance compared to state-of-art models like Deep learning and Graph neural network (GNN), achieving accuracy improvements of 3.16 and 0.91%, respectively. The observed superior performance of the model in comparison to the baseline models serves to emphasize its potential to augment IoT network security.","doi":"10.1007\/s10586-024-04404-8","cleaned_title":"graph-ensemble fusion for enhanced iot intrusion detection: leveraging gcn and deep learning","cleaned_abstract":"the proliferation of internet of things (iot) applications has heightened the vulnerability of information security, making it susceptible to attacks that may lead to the compromise of sensitive data. intrusion detection system (ids) is deployed in iot networks for the detection of attacks and to ensure the security of information. in previous works, the ids datasets suffer from an imbalanced distribution of data about attacks and, the flow of packets in ids which hinders the ability of deep learning models for potent and coherent classification. with the emergence of graph convolution neural network (gcn), a new sub-field of deep learning models, the structure of graphs can be leveraged to represent the data effectively. ids datasets typically consist of flow records of data which can naturally be represented as graph structures capturing both edge features and network topology information for classification of attacks. hence, in this paper, a novel gcn-ensemble fusion model is proposed for enhanced iot ids. there are three stages in this proposed model: (1) data processing and attribute graph generation, (2) feature engineering and (3) classification. the flow attributes of data packets in ids datasets are represented as the edges and the corresponding varying attacks as nodes of the attribute graph. here the gcn model is leveraged for feature engineering of the ids dataset. further, a novel ensemble of convolution neural networks is proposed for the classification task. the evaluation of the proposed model encompasses the utilization of four distinct datasets, namely bot-iot, ton-iot, cic-ids2018, and nf uq nids. in the bot-iot dataset, the proposed model demonstrates superior performance compared to state-of-art models like deep learning and graph neural network (gnn), achieving accuracy improvements of 3.16 and 0.91%, respectively. the observed superior performance of the model in comparison to the baseline models serves to emphasize its potential to augment iot network security.","key_phrases":["the proliferation","internet","things","(iot) applications","the vulnerability","information security","it","attacks","that","the compromise","sensitive data","intrusion detection system","ids","iot networks","the detection","attacks","the security","information","previous works","the ids datasets","an imbalanced distribution","data","attacks","the flow","packets","ids","which","the ability","deep learning models","potent and coherent classification","the emergence","graph convolution neural network","gcn","a new sub","-","field","deep learning models","the structure","graphs","the data","ids datasets","flow records","data","which","graph structures","both edge features","network topology information","classification","attacks","this paper","a novel gcn-ensemble fusion model","enhanced iot ids","three stages","this proposed model","graph generation","(3) classification","the flow attributes","data packets","ids datasets","the edges","the corresponding varying attacks","nodes","the attribute graph","the gcn model","feature engineering","the ids dataset","a novel ensemble","convolution neural networks","the classification task","the evaluation","the proposed model","the utilization","four distinct datasets","namely bot-iot, ton-iot","cic-ids2018","uq nids","the bot-iot dataset","the proposed model","superior performance","art","deep learning","graph neural network","gnn","accuracy improvements","0.91%","the observed superior performance","the model","comparison","the baseline models","its potential","iot network security","gcn","three","1","2","3","gcn","four","gnn","3.16","0.91%"]},{"title":"Information set supported deep learning architectures for improving noisy image classification","authors":["Saurabh Bhardwaj","Yizhi Wang","Guoqiang Yu","Yue Wang"],"abstract":"Deep learning models have been widely used in many supervised learning applications. However, these models suffer from overfitting due to various types of uncertainty with deteriorating performance when facing data biases, class imbalance, or noise propagation. The Information-Set Deep learning (ISDL) architectures with four variants are developed by integrating information set theory and deep learning principles to address the critical problem of the absence of robust deep learning models. There is a description of the ISDL architectures, learning algorithms, and analytic workflows. The performance of the ISDL models and standard architectures is evaluated using a noise-corrupted benchmark dataset. The experimental results show that the ISDL models can efficiently handle noise-dominated uncertainty and outperform peer architectures.","doi":"10.1038\/s41598-023-31462-6","cleaned_title":"information set supported deep learning architectures for improving noisy image classification","cleaned_abstract":"deep learning models have been widely used in many supervised learning applications. however, these models suffer from overfitting due to various types of uncertainty with deteriorating performance when facing data biases, class imbalance, or noise propagation. the information-set deep learning (isdl) architectures with four variants are developed by integrating information set theory and deep learning principles to address the critical problem of the absence of robust deep learning models. there is a description of the isdl architectures, learning algorithms, and analytic workflows. the performance of the isdl models and standard architectures is evaluated using a noise-corrupted benchmark dataset. the experimental results show that the isdl models can efficiently handle noise-dominated uncertainty and outperform peer architectures.","key_phrases":["deep learning models","many supervised learning applications","these models","various types","uncertainty","deteriorating performance","data biases","class imbalance","noise propagation","the information-set deep learning","isdl","four variants","information set theory","deep learning principles","the critical problem","the absence","robust deep learning models","a description","the isdl architectures","algorithms","analytic workflows","the performance","the isdl models","standard architectures","a noise-corrupted benchmark dataset","the experimental results","the isdl models","noise-dominated uncertainty","outperform peer architectures","four"]},{"title":"Deep transfer learning-based automated detection of blast disease in paddy crop","authors":["Amandeep Singh","Jaspreet Kaur","Kuldeep Singh","Maninder Lal Singh"],"abstract":"A major proportion of the loss faced by the agricultural industry originates from the diseases of the crop during cultivation. Paddy crop is one of the dominant crops which provides food to a huge population. In this crop, the losses caused by such diseases vary from 30 to 90% of the yield. Therefore, the automated detection of different diseases in paddy crops seeks the attention of the research community. In this context, the present work proposes a deep transfer learning solution for the automated detection of blast disease of paddy, which is the major cause of its yield reduction. For this purpose, an image dataset of healthy and blast disease-infected leave images of paddy crop has been developed. These images are fed to five convolutional neural network-based deep transfer learning algorithms, viz., LeNet, AlexNet, VGG 16, Inception v1, and Xception models for binary classification. The performance analysis of given algorithms reveals that AlexNet provides better results for binary classification with an average accuracy of 98.7% followed by VGG 16 and LeNet architectures having accuracies of 98.2% and 97.8%. So, this deep transfer learning-based approach may assist in reducing the gap between experts and farmers by providing an automated expert advice platform for the timely detection of diseases in paddy crop.","doi":"10.1007\/s11760-023-02735-4","cleaned_title":"deep transfer learning-based automated detection of blast disease in paddy crop","cleaned_abstract":"a major proportion of the loss faced by the agricultural industry originates from the diseases of the crop during cultivation. paddy crop is one of the dominant crops which provides food to a huge population. in this crop, the losses caused by such diseases vary from 30 to 90% of the yield. therefore, the automated detection of different diseases in paddy crops seeks the attention of the research community. in this context, the present work proposes a deep transfer learning solution for the automated detection of blast disease of paddy, which is the major cause of its yield reduction. for this purpose, an image dataset of healthy and blast disease-infected leave images of paddy crop has been developed. these images are fed to five convolutional neural network-based deep transfer learning algorithms, viz., lenet, alexnet, vgg 16, inception v1, and xception models for binary classification. the performance analysis of given algorithms reveals that alexnet provides better results for binary classification with an average accuracy of 98.7% followed by vgg 16 and lenet architectures having accuracies of 98.2% and 97.8%. so, this deep transfer learning-based approach may assist in reducing the gap between experts and farmers by providing an automated expert advice platform for the timely detection of diseases in paddy crop.","key_phrases":["a major proportion","the loss","the agricultural industry originates","the diseases","the crop","cultivation","paddy crop","the dominant crops","which","food","a huge population","this crop","the losses","such diseases","30 to 90%","the yield","the automated detection","different diseases","paddy crops","the attention","the research community","this context","the present work","a deep transfer learning solution","the automated detection","blast disease","paddy","which","the major cause","its yield reduction","this purpose","paddy crop","these images","five convolutional neural network-based deep transfer learning algorithms","viz","lenet","alexnet","vgg","inception v1","xception models","binary classification","the performance analysis","given algorithms","alexnet","better results","binary classification","an average accuracy","98.7%","accuracies","98.2%","97.8%","this deep transfer learning-based approach","the gap","experts","farmers","an automated expert advice platform","the timely detection","diseases","paddy crop","30 to 90%","fed","five","16","98.7%","16","98.2%","97.8%"]},{"title":"Octorotor flight control system design with stochastic optimal tuning, deep learning and differential morphing","authors":["Oguz Kose"],"abstract":"In this paper, simultaneous longitudinal and lateral flight control is investigated for an octorotor by using stochastic optimal tuning and deep learning under differential morphing. Octorotor models for differential morphing were drawn in SOLIDWORKS drawing program. Arm lengths are randomly estimated in the algorithm. Moments of inertia changing according to morphing ratios are estimated with deep neural network. In addition, the proportional\u2013integral\u2013derivative controller coefficients required for both longitudinal and lateral flight according to the morphing ratios are estimated by simultaneous perturbation stochastic approximation. Considering the design performance criteria, 49.95% improvement was achieved in the total cost. The estimation of unknown parameters by optimization method and deep learning was tested in simulations, and the octorotor successfully followed the given reference angle.","doi":"10.1007\/s40430-024-04972-1","cleaned_title":"octorotor flight control system design with stochastic optimal tuning, deep learning and differential morphing","cleaned_abstract":"in this paper, simultaneous longitudinal and lateral flight control is investigated for an octorotor by using stochastic optimal tuning and deep learning under differential morphing. octorotor models for differential morphing were drawn in solidworks drawing program. arm lengths are randomly estimated in the algorithm. moments of inertia changing according to morphing ratios are estimated with deep neural network. in addition, the proportional\u2013integral\u2013derivative controller coefficients required for both longitudinal and lateral flight according to the morphing ratios are estimated by simultaneous perturbation stochastic approximation. considering the design performance criteria, 49.95% improvement was achieved in the total cost. the estimation of unknown parameters by optimization method and deep learning was tested in simulations, and the octorotor successfully followed the given reference angle.","key_phrases":["this paper","simultaneous longitudinal and lateral flight control","an octorotor","stochastic optimal tuning","deep learning","differential morphing","octorotor models","differential morphing","solidworks drawing program","arm lengths","the algorithm","moments","inertia","morphing ratios","deep neural network","addition","the proportional\u2013integral\u2013derivative controller coefficients","both longitudinal and lateral flight","the morphing ratios","simultaneous perturbation stochastic approximation","the design performance criteria","49.95% improvement","the total cost","the estimation","unknown parameters","optimization method","deep learning","simulations","the octorotor","the given reference angle","49.95%"]},{"title":"Network intrusion detection and mitigation in SDN using deep learning models","authors":["Mamatha Maddu","Yamarthi Narasimha Rao"],"abstract":"Software-Defined Networking (SDN) is a contemporary network strategy utilized instead of a traditional network structure. It provides significantly more administrative efficiency and ease than traditional networks. However, the centralized control used in SDN entails an elevated risk of single-point failure that is more susceptible to different kinds of network assaults like Distributed Denial of Service (DDoS), DoS, spoofing, and API exploitation which are very complex to identify and mitigate. Thus, a powerful intrusion detection system (IDS) based on deep learning is created in this study for the detection and mitigation of network intrusions. This system contains several stages and begins with the data augmentation method named Deep Convolutional Generative Adversarial Networks (DCGAN) to over the data imbalance problem. Then, the features are extracted from the input data using a CenterNet-based approach. After extracting effective characteristics, ResNet152V2 with Slime Mold Algorithm (SMA) based deep learning is implemented to categorize the assaults in InSDN and Edge IIoT datasets. Once the network intrusion is detected, the proposed defense module is activated to restore regular network connectivity quickly. Finally, several experiments are carried out to validate the algorithm's robustness, and the outcomes reveal that the proposed system can successfully detect and mitigate network intrusions.","doi":"10.1007\/s10207-023-00771-2","cleaned_title":"network intrusion detection and mitigation in sdn using deep learning models","cleaned_abstract":"software-defined networking (sdn) is a contemporary network strategy utilized instead of a traditional network structure. it provides significantly more administrative efficiency and ease than traditional networks. however, the centralized control used in sdn entails an elevated risk of single-point failure that is more susceptible to different kinds of network assaults like distributed denial of service (ddos), dos, spoofing, and api exploitation which are very complex to identify and mitigate. thus, a powerful intrusion detection system (ids) based on deep learning is created in this study for the detection and mitigation of network intrusions. this system contains several stages and begins with the data augmentation method named deep convolutional generative adversarial networks (dcgan) to over the data imbalance problem. then, the features are extracted from the input data using a centernet-based approach. after extracting effective characteristics, resnet152v2 with slime mold algorithm (sma) based deep learning is implemented to categorize the assaults in insdn and edge iiot datasets. once the network intrusion is detected, the proposed defense module is activated to restore regular network connectivity quickly. finally, several experiments are carried out to validate the algorithm's robustness, and the outcomes reveal that the proposed system can successfully detect and mitigate network intrusions.","key_phrases":["software-defined networking","sdn","a contemporary network strategy","a traditional network structure","it","significantly more administrative efficiency","ease","traditional networks","the centralized control","sdn","an elevated risk","single-point failure","that","different kinds","network assaults","distributed denial","service","ddos","api exploitation","which","a powerful intrusion detection system","ids","deep learning","this study","the detection","mitigation","network intrusions","this system","several stages","the data augmentation method","deep convolutional generative adversarial networks","dcgan","the data imbalance problem","the features","the input data","a centernet-based approach","effective characteristics","slime mold","algorithm","sma","based deep learning","the assaults","insdn","iiot datasets","the network intrusion","the proposed defense module","regular network connectivity","several experiments","the algorithm's robustness","the outcomes","the proposed system","network intrusions"]},{"title":"Development of a remote music teaching system based on facial recognition and deep learning","authors":["Ning Zhang","Huizhong Wang"],"abstract":"With the continuous progress of computer and network technology, teaching methods and educational models are also constantly evolving and improving. The development of facial recognition technology has brought new opportunities and challenges to the development of educational theory and systems. This article establishes a remote music teaching system based on facial recognition and deep learning technology. The system adopts the Java EE framework structure and deep learning technology. By conducting deep learning and training on a large amount of facial data, we can identify students' facial expressions and emotional states, thereby better understanding their learning status and needs. At the same time, the system also supports multiple teaching modes and interactive methods, providing teachers and students with a more convenient and efficient teaching management and learning experience. Subsequently, this article evaluated and explored the effectiveness of the remote music teaching system through a questionnaire survey. The results show that most students believe that the system can help them better master basic music knowledge and professional skills, improve learning effectiveness and achieve learning goals. The use of the system can also stimulate students' interest in music learning, providing new ways and means for teaching.","doi":"10.1007\/s00500-023-09120-w","cleaned_title":"development of a remote music teaching system based on facial recognition and deep learning","cleaned_abstract":"with the continuous progress of computer and network technology, teaching methods and educational models are also constantly evolving and improving. the development of facial recognition technology has brought new opportunities and challenges to the development of educational theory and systems. this article establishes a remote music teaching system based on facial recognition and deep learning technology. the system adopts the java ee framework structure and deep learning technology. by conducting deep learning and training on a large amount of facial data, we can identify students' facial expressions and emotional states, thereby better understanding their learning status and needs. at the same time, the system also supports multiple teaching modes and interactive methods, providing teachers and students with a more convenient and efficient teaching management and learning experience. subsequently, this article evaluated and explored the effectiveness of the remote music teaching system through a questionnaire survey. the results show that most students believe that the system can help them better master basic music knowledge and professional skills, improve learning effectiveness and achieve learning goals. the use of the system can also stimulate students' interest in music learning, providing new ways and means for teaching.","key_phrases":["the continuous progress","computer and network technology","teaching methods","educational models","the development","facial recognition technology","new opportunities","challenges","the development","educational theory","systems","this article","a remote music teaching system","facial recognition","deep learning technology","the system","the java ee framework structure","deep learning technology","deep learning","training","a large amount","facial data","we","students' facial expressions","emotional states","their learning status","needs","the same time","the system","multiple teaching modes","interactive methods","teachers","students","a more convenient and efficient teaching management and learning experience","this article","the effectiveness","the remote music teaching system","a questionnaire survey","the results","most students","the system","them","basic music knowledge","professional skills","effectiveness","learning goals","the use","the system","students' interest","new ways","teaching"]},{"title":"Advances in Deep Learning Techniques for Short-term Energy Load Forecasting Applications: A Review","authors":["Radhika Chandrasekaran","Senthil Kumar Paramasivan"],"abstract":"Today, the majority of the leading power companies place a significant emphasis on forecasting the electricity load in the balance of power and administration. Meanwhile, since electricity is an integral component of every person\u2019s contemporary life, energy load forecasting is necessary to afford the energy demand required. The expansion of the electrical infrastructure is a key factor in increasing sustainable economic growth, and the planning and control of the utility power system rely on accurate load forecasting. Due to uncertainty in energy utilization, forecasting is turning into a complex task, and it makes an impact on applications that include energy scheduling and management, price forecasting, etc. The statistical methods involving time series for regression analysis and machine learning techniques have been used in energy load forecasting extensively over the last few decades to precisely predict future energy demands. However, they have some drawbacks with limited model flexibility, generalization, and overfitting. Deep learning addresses the issues of handling unstructured and unlabeled data, automatic feature learning, non-linear model flexibility, the ability to handle high-dimensional data, and simultaneous computation using GPUs efficiently. This paper investigates factors influencing energy load forecasting, then discusses the most commonly used deep learning approaches in energy load forecasting, as well as evaluation metrics to evaluate the performance of the model, followed by bio-inspired algorithms to optimize the model, and other advanced technologies for energy load forecasting. This study discusses the research findings, challenges, and opportunities in energy load forecasting.","doi":"10.1007\/s11831-024-10155-x","cleaned_title":"advances in deep learning techniques for short-term energy load forecasting applications: a review","cleaned_abstract":"today, the majority of the leading power companies place a significant emphasis on forecasting the electricity load in the balance of power and administration. meanwhile, since electricity is an integral component of every person\u2019s contemporary life, energy load forecasting is necessary to afford the energy demand required. the expansion of the electrical infrastructure is a key factor in increasing sustainable economic growth, and the planning and control of the utility power system rely on accurate load forecasting. due to uncertainty in energy utilization, forecasting is turning into a complex task, and it makes an impact on applications that include energy scheduling and management, price forecasting, etc. the statistical methods involving time series for regression analysis and machine learning techniques have been used in energy load forecasting extensively over the last few decades to precisely predict future energy demands. however, they have some drawbacks with limited model flexibility, generalization, and overfitting. deep learning addresses the issues of handling unstructured and unlabeled data, automatic feature learning, non-linear model flexibility, the ability to handle high-dimensional data, and simultaneous computation using gpus efficiently. this paper investigates factors influencing energy load forecasting, then discusses the most commonly used deep learning approaches in energy load forecasting, as well as evaluation metrics to evaluate the performance of the model, followed by bio-inspired algorithms to optimize the model, and other advanced technologies for energy load forecasting. this study discusses the research findings, challenges, and opportunities in energy load forecasting.","key_phrases":["the majority","the leading power companies","a significant emphasis","the electricity load","the balance","power","administration","electricity","an integral component","every person\u2019s contemporary life","energy load forecasting","the energy demand","the expansion","the electrical infrastructure","a key factor","sustainable economic growth","the planning","control","the utility power system","accurate load forecasting","uncertainty","energy utilization","forecasting","a complex task","it","an impact","applications","that","energy scheduling","management","price forecasting","the statistical methods","time series","regression analysis","machine learning techniques","energy load","the last few decades","future energy demands","they","some drawbacks","limited model flexibility","generalization","deep learning addresses","the issues","unstructured and unlabeled data","automatic feature learning","non-linear model flexibility","the ability","high-dimensional data","simultaneous computation","gpus","this paper investigates","energy load forecasting","the most commonly used deep learning approaches","energy load forecasting","evaluation metrics","the performance","the model","bio-inspired algorithms","the model","other advanced technologies","energy load forecasting","this study","the research findings","challenges","opportunities","energy load forecasting","today","the last few decades"]},{"title":"Advancing differential diagnosis: a comprehensive review of deep learning approaches for differentiating tuberculosis, pneumonia, and COVID-19","authors":["Kajal Kansal","Tej Bahadur Chandra","Akansha Singh"],"abstract":"In the realm of medical diagnostics, particularly in differential diagnosis, where differentiating between illnesses or ailments with comparable symptoms is essential, deep learning has gained importance. Recent developments in deep learning have demonstrated considerable promise for revolutionizing medical diagnostics by using the ability of artificial intelligence (AI) to accurately interpret radiological images. We examine the most cutting-edge deep learning techniques currently being utilized for the differential diagnosis of tuberculosis, pneumonia, and COVID-19 in this in-depth review. The study presents an in-depth critical review of several SOTA (state-of-the-art) studies used for differential diagnosis of different respiratory abnormalities like TB, Pneumonia, and COVID-19. In addition, an overview of various approaches, datasets employed in each method, various diagnosis tests, used assessment measures, and obtained performance is summarized and comprehensively compared to assist future research. We suggest a pathway for future research and development of deep learning solutions for differential diagnosis by critically analyzing the current literature and outlining the limitations and potential in this sector.","doi":"10.1007\/s11042-024-19350-1","cleaned_title":"advancing differential diagnosis: a comprehensive review of deep learning approaches for differentiating tuberculosis, pneumonia, and covid-19","cleaned_abstract":"in the realm of medical diagnostics, particularly in differential diagnosis, where differentiating between illnesses or ailments with comparable symptoms is essential, deep learning has gained importance. recent developments in deep learning have demonstrated considerable promise for revolutionizing medical diagnostics by using the ability of artificial intelligence (ai) to accurately interpret radiological images. we examine the most cutting-edge deep learning techniques currently being utilized for the differential diagnosis of tuberculosis, pneumonia, and covid-19 in this in-depth review. the study presents an in-depth critical review of several sota (state-of-the-art) studies used for differential diagnosis of different respiratory abnormalities like tb, pneumonia, and covid-19. in addition, an overview of various approaches, datasets employed in each method, various diagnosis tests, used assessment measures, and obtained performance is summarized and comprehensively compared to assist future research. we suggest a pathway for future research and development of deep learning solutions for differential diagnosis by critically analyzing the current literature and outlining the limitations and potential in this sector.","key_phrases":["the realm","medical diagnostics","differential diagnosis","illnesses","ailments","comparable symptoms","deep learning","importance","recent developments","deep learning","considerable promise","medical diagnostics","the ability","artificial intelligence","radiological images","we","the most cutting-edge deep learning techniques","the differential diagnosis","tuberculosis","pneumonia","covid-19","-depth","the study","an in-depth critical review","the-art","differential diagnosis","different respiratory abnormalities","tb","pneumonia","covid-19","addition","an overview","various approaches","datasets","each method","various diagnosis tests","assessment measures","performance","future research","we","a pathway","future research","development","deep learning solutions","differential diagnosis","the current literature","the limitations","potential","this sector","covid-19","covid-19"]},{"title":"Molecular design with automated quantum computing-based deep learning and optimization","authors":["Akshay Ajagekar","Fengqi You"],"abstract":"Computer-aided design of novel molecules and compounds is a challenging task that can be addressed with quantum computing (QC) owing to its notable advances in optimization and machine learning. Here, we use QC-assisted learning and optimization techniques implemented with near-term QC devices for molecular property prediction and generation tasks. The proposed probabilistic energy-based deep learning model trained in a generative manner facilitated by QC yields robust latent representations of molecules, while the proposed data-driven QC-based optimization framework performs guided navigation of the target chemical space by exploiting the structure\u2013property relationships captured by the energy-based model. We demonstrate the viability of the proposed molecular design approach by generating several molecular candidates that satisfy specific property target requirements. The proposed QC-based methods exhibit an improved predictive performance while efficiently generating novel molecules that accurately fulfill target conditions and exemplify the potential of QC for automated molecular design, thus accentuating its utility.","doi":"10.1038\/s41524-023-01099-0","cleaned_title":"molecular design with automated quantum computing-based deep learning and optimization","cleaned_abstract":"computer-aided design of novel molecules and compounds is a challenging task that can be addressed with quantum computing (qc) owing to its notable advances in optimization and machine learning. here, we use qc-assisted learning and optimization techniques implemented with near-term qc devices for molecular property prediction and generation tasks. the proposed probabilistic energy-based deep learning model trained in a generative manner facilitated by qc yields robust latent representations of molecules, while the proposed data-driven qc-based optimization framework performs guided navigation of the target chemical space by exploiting the structure\u2013property relationships captured by the energy-based model. we demonstrate the viability of the proposed molecular design approach by generating several molecular candidates that satisfy specific property target requirements. the proposed qc-based methods exhibit an improved predictive performance while efficiently generating novel molecules that accurately fulfill target conditions and exemplify the potential of qc for automated molecular design, thus accentuating its utility.","key_phrases":["computer-aided design","novel molecules","compounds","a challenging task","that","quantum computing","qc","its notable advances","optimization","machine learning","we","qc-assisted learning and optimization techniques","near-term qc devices","molecular property prediction","generation tasks","the proposed probabilistic energy-based deep learning model","a generative manner","qc yields","robust latent representations","molecules","the proposed data-driven qc-based optimization framework","guided navigation","the target chemical space","the structure","property relationships","the energy-based model","we","the viability","the proposed molecular design approach","several molecular candidates","that","specific property target requirements","the proposed qc-based methods","an improved predictive performance","novel molecules","that","target conditions","the potential","qc","automated molecular design","its utility","quantum"]},{"title":"ConvMixer deep learning model for detection of pneumonia disease using chest X-ray images","authors":["Ankit Chaudhary","Sushil Kumar Saroj"],"abstract":"Pneumonia is a common and fatal disease in children nowadays. It infects the lungs resulting in difficulties in breathing. Severe cases of it may lead to death. Therefore, early and accurate detection of pneumonia disease is essential. There are existing various methods for the detection of pneumonia disease today. Deep learning methods are considered more effective for this. We have applied a novel deep learning model i.e. the ConvMixer model for the detection of pneumonia disease. The model replaces traditional convolutional layers with a mixer of channels and spatial dimensions. By mixing channels and spatial dimensions, it reduces the number of parameters and computations required for processing each layer leading to improved efficiency. We have applied the model to the large numbers of chest X-ray images that are publicly available on Kaggle, provided by Guangzhou Women and Children\u2019s Medical Centre in Guangzhou. The model has achieved the highest accuracy of 95.11%. It has also been evaluated for precision, recall, and f1-score parameters.","doi":"10.1007\/s10742-024-00334-5","cleaned_title":"convmixer deep learning model for detection of pneumonia disease using chest x-ray images","cleaned_abstract":"pneumonia is a common and fatal disease in children nowadays. it infects the lungs resulting in difficulties in breathing. severe cases of it may lead to death. therefore, early and accurate detection of pneumonia disease is essential. there are existing various methods for the detection of pneumonia disease today. deep learning methods are considered more effective for this. we have applied a novel deep learning model i.e. the convmixer model for the detection of pneumonia disease. the model replaces traditional convolutional layers with a mixer of channels and spatial dimensions. by mixing channels and spatial dimensions, it reduces the number of parameters and computations required for processing each layer leading to improved efficiency. we have applied the model to the large numbers of chest x-ray images that are publicly available on kaggle, provided by guangzhou women and children\u2019s medical centre in guangzhou. the model has achieved the highest accuracy of 95.11%. it has also been evaluated for precision, recall, and f1-score parameters.","key_phrases":["pneumonia","a common and fatal disease","children","it","the lungs","difficulties","breathing","severe cases","it","death","early and accurate detection","pneumonia disease","various methods","the detection","pneumonia disease","deep learning methods","this","we","a novel deep learning model","the detection","pneumonia disease","the model","traditional convolutional layers","a mixer","channels","spatial dimensions","channels","spatial dimensions","it","the number","parameters","computations","each layer","improved efficiency","we","the model","the large numbers","chest x-ray images","that","kaggle","guangzhou women","children\u2019s medical centre","guangzhou","the model","the highest accuracy","95.11%","it","precision, recall, and f1-score parameters","today","guangzhou","guangzhou","95.11%"]},{"title":"Research on Real-time Detection of Stacked Objects Based on Deep Learning","authors":["Kaiguo Geng","Jinwei Qiao","Na Liu","Zhi Yang","Rongmin Zhang","Huiling Li"],"abstract":"Deep Learning has garnered significant attention in the field of object detection and is widely used in both industry and everyday life. The objective of this study is to investigate the applicability and targeted improvements of Deep Learning-based object detection in complex stacked environments. We analyzed the limitations in practical applications under such conditions, pinpointed the specific problems, and proposed corresponding improvement strategies. First, the study provided an overview of recent advancements in mainstream one-stage object detection algorithms, which included Anchor-based, Anchor-free, and Transformer-based architectures. The high real-time performance of these algorithms holds particular significance in practical engineering applications. It then looked at relevant technologies in three emerging research areas: Parts Recognition, Intelligent Driving, and Agricultural Picking. The study summarized existing limitations in real-time object detection within complex stacked environments and provided a comprehensive analysis of prevalent improvement strategies such as multi-level feature fusion, knowledge distillation, and hyperparameter optimization. Finally, after analyzing the performance of recent advanced one-stage algorithms on official datasets, this paper conducted empirical tests on a self-constructed industrial stacked dataset with algorithms of different structure and analyzed the experimental results in detail. A comprehensive analysis shows that Deep Learning-based object detection algorithms offer extensive applicability in complex stacked environments. In addressing diverse target sizes, overlapping occlusions, real-time constraints, and the need for lightweight solutions in complex stacked environments, each improvement strategy has its own advantages and limitations. Selecting and integrating appropriate enhancement strategies is critical and typically requires holistic evaluation, tailored to specific application contexts and challenges.","doi":"10.1007\/s10846-023-02009-8","cleaned_title":"research on real-time detection of stacked objects based on deep learning","cleaned_abstract":"deep learning has garnered significant attention in the field of object detection and is widely used in both industry and everyday life. the objective of this study is to investigate the applicability and targeted improvements of deep learning-based object detection in complex stacked environments. we analyzed the limitations in practical applications under such conditions, pinpointed the specific problems, and proposed corresponding improvement strategies. first, the study provided an overview of recent advancements in mainstream one-stage object detection algorithms, which included anchor-based, anchor-free, and transformer-based architectures. the high real-time performance of these algorithms holds particular significance in practical engineering applications. it then looked at relevant technologies in three emerging research areas: parts recognition, intelligent driving, and agricultural picking. the study summarized existing limitations in real-time object detection within complex stacked environments and provided a comprehensive analysis of prevalent improvement strategies such as multi-level feature fusion, knowledge distillation, and hyperparameter optimization. finally, after analyzing the performance of recent advanced one-stage algorithms on official datasets, this paper conducted empirical tests on a self-constructed industrial stacked dataset with algorithms of different structure and analyzed the experimental results in detail. a comprehensive analysis shows that deep learning-based object detection algorithms offer extensive applicability in complex stacked environments. in addressing diverse target sizes, overlapping occlusions, real-time constraints, and the need for lightweight solutions in complex stacked environments, each improvement strategy has its own advantages and limitations. selecting and integrating appropriate enhancement strategies is critical and typically requires holistic evaluation, tailored to specific application contexts and challenges.","key_phrases":["deep learning","significant attention","the field","object detection","both industry","everyday life","the objective","this study","the applicability","targeted improvements","deep learning-based object detection","complex stacked environments","we","the limitations","practical applications","such conditions","the specific problems","corresponding improvement strategies","the study","an overview","recent advancements","mainstream one-stage object detection algorithms","which","anchor-based, anchor-free, and transformer-based architectures","the high real-time performance","these algorithms","particular significance","practical engineering applications","it","relevant technologies","three emerging research areas","parts recognition","intelligent driving","agricultural picking","the study","existing limitations","real-time object detection","complex stacked environments","a comprehensive analysis","prevalent improvement strategies","multi-level feature fusion","knowledge distillation","hyperparameter optimization","the performance","recent advanced one-stage algorithms","official datasets","this paper","empirical tests","a self-constructed industrial stacked dataset","algorithms","different structure","the experimental results","detail","a comprehensive analysis","deep learning-based object detection algorithms","extensive applicability","complex stacked environments","diverse target sizes","occlusions","real-time constraints","lightweight solutions","complex stacked environments","each improvement strategy","its own advantages","limitations","appropriate enhancement strategies","holistic evaluation","specific application contexts","challenges","first","one","three","one"]},{"title":"Characterization of fault-karst reservoirs based on deep learning and attribute fusion","authors":["Zhipeng Gui","Junhua Zhang","Yintao Zhang","Chong Sun"],"abstract":"The identification of fault-karst reservoir is crucial for the exploration and development of fault-controlled oil and gas reservoirs. Traditional methods primarily rely on well logging and seismic attribute analysis for karst cave identification. However, these methods often lack the resolution needed to meet practical demands. Deep learning methods offer promising solutions by effectively overcoming the complex response characteristics of seismic wave fields, owing to their high learning capabilities. Therefore, this research proposes a method for fault-karst reservoir identification. Initially, a comparative analysis between the improved U-Net++ network and traditional deep convolutional networks is conducted to select appropriate training parameters for separate training of karst caves and faults. Subsequently, the trained models are applied to actual seismic data to predict karst caves and faults within the research area, followed by attribute fusion to acquire data on fault-karst reservoirs. The results indicate that: (1) The proposed method effectively identifies karst caves and faults, outperforming traditional seismic attribute and coherence methods in terms of identification accuracy, and slightly surpassing U-Net and FCN; (2) The fusion of predicted karst caves and faults yields clear delineation of the relationship between top karst caves and bottom fractures within the research area. In summary, the proposed method for fault-karst reservoirs identification and characterization provides valuable insights for the exploration and development of fault-controlled oil and gas reservoirs in the region.","doi":"10.1007\/s11600-024-01420-5","cleaned_title":"characterization of fault-karst reservoirs based on deep learning and attribute fusion","cleaned_abstract":"the identification of fault-karst reservoir is crucial for the exploration and development of fault-controlled oil and gas reservoirs. traditional methods primarily rely on well logging and seismic attribute analysis for karst cave identification. however, these methods often lack the resolution needed to meet practical demands. deep learning methods offer promising solutions by effectively overcoming the complex response characteristics of seismic wave fields, owing to their high learning capabilities. therefore, this research proposes a method for fault-karst reservoir identification. initially, a comparative analysis between the improved u-net++ network and traditional deep convolutional networks is conducted to select appropriate training parameters for separate training of karst caves and faults. subsequently, the trained models are applied to actual seismic data to predict karst caves and faults within the research area, followed by attribute fusion to acquire data on fault-karst reservoirs. the results indicate that: (1) the proposed method effectively identifies karst caves and faults, outperforming traditional seismic attribute and coherence methods in terms of identification accuracy, and slightly surpassing u-net and fcn; (2) the fusion of predicted karst caves and faults yields clear delineation of the relationship between top karst caves and bottom fractures within the research area. in summary, the proposed method for fault-karst reservoirs identification and characterization provides valuable insights for the exploration and development of fault-controlled oil and gas reservoirs in the region.","key_phrases":["the identification","fault-karst reservoir","the exploration","development","fault-controlled oil and gas reservoirs","traditional methods","well logging and seismic attribute analysis","karst cave identification","these methods","the resolution","practical demands","deep learning methods","promising solutions","the complex response characteristics","seismic wave fields","their high learning capabilities","this research","a method","fault-karst reservoir identification","a comparative analysis","the improved u-net++ network","traditional deep convolutional networks","appropriate training parameters","separate training","karst caves","faults","the trained models","actual seismic data","karst caves","faults","the research area","attribute fusion","data","fault-karst reservoirs","the results","the proposed method","karst caves","faults","traditional seismic attribute and coherence methods","terms","identification accuracy","u","-","net","fcn","(2) the fusion","predicted karst caves","faults","clear delineation","the relationship","top karst caves","bottom fractures","the research area","summary","the proposed method","fault-karst reservoirs identification","characterization","valuable insights","the exploration","development","fault-controlled oil and gas reservoirs","the region","1","2"]},{"title":"Crop disease detection via ensembled-deep-learning paradigm and ABC Coyote pack optimization algorithm (ABC-CPOA)","authors":["M. Chithambarathanu","M. K. Jeyakumar"],"abstract":"Crop disease is a significant issue that affects the growth and yield of crops, leading to financial loss for farmers. Identification and treatment of crop diseases have become challenging due to the increase in the variety of diseases and the lack of knowledge among farmers. To address this issue, this investigate uses an ensembled-deep-learning paradigm to propose a deep learning-based model for crop disease identification trained with an ABC-CPOA. Initially, collected raw images are pre-processed via Bilateral filter and gamma correction Feature Extraction: Then, from the pre-processed images, the features like texture feature (Local Quinary Pattern (LQP), Local Gradient Pattern (LGP), Enriched Local Binary Pattern (E-LBP), color features (Color Histogram, Color Moments), shape features (Contour-based features, Convex Hull). Optimal feature selection- Among the extracted features, the optimal features is designated by means of a self-improved meta-heuristic optimization model referred as ABC-CPOA. This ABC-CPOA model is an extended version of standard Coyote Optimization Algorithm (COA). Crop disease detection phase is modelled with a new ensembled-deep-learning paradigm. Ensembled-deep-learning paradigm comprises Attention-based Bi-LSTM, Recurrent Neural Networks (RNNs) and Optimized Deep Neural Network (O-DNN). The weight function of O-DNN is fine-tuned using the new ABC-CPOA. Precision, recall, sensitivity, and specificity, in addition to TPR, FPR, FNR, and TNR, F1-score, and accuracy are used to assess the suggested approach. The implementation was performed by the MATLAB tool (version: 2022B).","doi":"10.1007\/s11042-024-19329-y","cleaned_title":"crop disease detection via ensembled-deep-learning paradigm and abc coyote pack optimization algorithm (abc-cpoa)","cleaned_abstract":"crop disease is a significant issue that affects the growth and yield of crops, leading to financial loss for farmers. identification and treatment of crop diseases have become challenging due to the increase in the variety of diseases and the lack of knowledge among farmers. to address this issue, this investigate uses an ensembled-deep-learning paradigm to propose a deep learning-based model for crop disease identification trained with an abc-cpoa. initially, collected raw images are pre-processed via bilateral filter and gamma correction feature extraction: then, from the pre-processed images, the features like texture feature (local quinary pattern (lqp), local gradient pattern (lgp), enriched local binary pattern (e-lbp), color features (color histogram, color moments), shape features (contour-based features, convex hull). optimal feature selection- among the extracted features, the optimal features is designated by means of a self-improved meta-heuristic optimization model referred as abc-cpoa. this abc-cpoa model is an extended version of standard coyote optimization algorithm (coa). crop disease detection phase is modelled with a new ensembled-deep-learning paradigm. ensembled-deep-learning paradigm comprises attention-based bi-lstm, recurrent neural networks (rnns) and optimized deep neural network (o-dnn). the weight function of o-dnn is fine-tuned using the new abc-cpoa. precision, recall, sensitivity, and specificity, in addition to tpr, fpr, fnr, and tnr, f1-score, and accuracy are used to assess the suggested approach. the implementation was performed by the matlab tool (version: 2022b).","key_phrases":["crop disease","a significant issue","that","the growth","yield","crops","financial loss","farmers","identification","treatment","crop diseases","the increase","the variety","diseases","the lack","knowledge","farmers","this issue","this investigate","an ensembled-deep-learning paradigm","a deep learning-based model","crop disease identification","an abc-cpoa","raw images","bilateral filter","gamma correction","feature extraction","the pre-processed images","texture feature","local quinary pattern","lqp","local gradient pattern","lgp","enriched local binary pattern","e","-","lbp","color features","color histogram","color moments","shape features","(contour-based features","convex hull","optimal feature","the extracted features","the optimal features","means","a self-improved meta-heuristic optimization model","abc-cpoa","this abc-cpoa model","an extended version","standard coyote optimization algorithm","coa","crop disease detection phase","a new ensembled-deep-learning paradigm","ensembled-deep-learning paradigm","attention-based bi","-","lstm","recurrent neural networks","rnns","deep neural network","o","dnn","the weight function","o","-","dnn","the new abc-cpoa","precision","recall","sensitivity","specificity","addition","tpr","fpr","fnr","tnr","f1-score","accuracy","the suggested approach","the implementation","the matlab tool","version","abc","abc","abc","abc"]},{"title":"An interpretable deep learning framework for genome-informed precision oncology","authors":["Shuangxia Ren","Gregory F. Cooper","Lujia Chen","Xinghua Lu"],"abstract":"Cancers result from aberrations in cellular signalling systems, typically resulting from driver somatic genome alterations (SGAs) in individual tumours. Precision oncology requires understanding the cellular state and selecting medications that induce vulnerability in cancer cells under such conditions. To this end, we developed a computational framework consisting of two components: (1) a representation-learning component, which learns a representation of the cellular signalling systems when perturbed by SGAs and uses a biologically motivated and interpretable deep learning model, and (2) a drug-response prediction component, which predicts drug responses by leveraging the information of the cellular state of the cancer cells derived by the first component. Our cell-state-oriented framework notably improves the accuracy of predictions of drug responses compared to models using SGAs directly in cell lines. Moreover, our model performs well with real patient data. Importantly, our framework enables the prediction of responses to chemotherapy agents based on SGAs, thus expanding genome-informed precision oncology beyond molecularly targeted drugs.","doi":"10.1038\/s42256-024-00866-y","cleaned_title":"an interpretable deep learning framework for genome-informed precision oncology","cleaned_abstract":"cancers result from aberrations in cellular signalling systems, typically resulting from driver somatic genome alterations (sgas) in individual tumours. precision oncology requires understanding the cellular state and selecting medications that induce vulnerability in cancer cells under such conditions. to this end, we developed a computational framework consisting of two components: (1) a representation-learning component, which learns a representation of the cellular signalling systems when perturbed by sgas and uses a biologically motivated and interpretable deep learning model, and (2) a drug-response prediction component, which predicts drug responses by leveraging the information of the cellular state of the cancer cells derived by the first component. our cell-state-oriented framework notably improves the accuracy of predictions of drug responses compared to models using sgas directly in cell lines. moreover, our model performs well with real patient data. importantly, our framework enables the prediction of responses to chemotherapy agents based on sgas, thus expanding genome-informed precision oncology beyond molecularly targeted drugs.","key_phrases":["cancers","aberrations","cellular signalling systems","driver somatic genome alterations","sgas","individual tumours","precision oncology","the cellular state and selecting medications","that","vulnerability","cancer cells","such conditions","this end","we","a computational framework","two components","a representation-learning component","which","a representation","the cellular signalling systems","sgas","a biologically motivated and interpretable deep learning model","(2) a drug-response prediction component","which","drug responses","the information","the cellular state","the cancer cells","the first component","our cell-state-oriented framework","the accuracy","predictions","drug responses","models","sgas","cell lines","our model","real patient data","our framework","the prediction","responses","chemotherapy agents","sgas","genome-informed precision oncology","molecularly targeted drugs","two","1","2","first"]},{"title":"Development of a deep learning model for cancer diagnosis by inspecting cell-free DNA end-motifs","authors":["Hongru Shen","Meng Yang","Jilei Liu","Kexin Chen","Xiangchun Li"],"abstract":"Accurate discrimination between patients with and without cancer from cfDNA is crucial for early cancer diagnosis. Herein, we develop and validate a deep-learning-based model entitled end-motif inspection via transformer (EMIT) for discriminating individuals with and without cancer by learning feature representations from cfDNA end-motifs. EMIT is a self-supervised learning approach that models rankings of cfDNA end-motifs. We include 4606 samples subjected to different types of cfDNA sequencing to develop EIMIT, and subsequently evaluate classification performance of linear projections of EMIT on six datasets and an additional inhouse testing set encopassing whole-genome, whole-genome bisulfite and 5-hydroxymethylcytosine sequencing. The linear projection of representations from EMIT achieved area under the receiver operating curve (AUROC) values ranged from 0.895 (0.835\u20130.955) to 0.996 (0.994\u20130.997) across these six datasets, outperforming its baseline by significant margins. Additionally, we showed that linear projection of EMIT representations can achieve an AUROC of 0.962 (0.914\u20131.0) in identification of lung cancer on an independent testing set subjected to whole-exome sequencing. The findings of this study indicate that a transformer-based deep learning model can learn cancer-discrimative representations from cfDNA end-motifs. The representations of this deep learning model can be exploited for discriminating patients with and without cancer.","doi":"10.1038\/s41698-024-00635-5","cleaned_title":"development of a deep learning model for cancer diagnosis by inspecting cell-free dna end-motifs","cleaned_abstract":"accurate discrimination between patients with and without cancer from cfdna is crucial for early cancer diagnosis. herein, we develop and validate a deep-learning-based model entitled end-motif inspection via transformer (emit) for discriminating individuals with and without cancer by learning feature representations from cfdna end-motifs. emit is a self-supervised learning approach that models rankings of cfdna end-motifs. we include 4606 samples subjected to different types of cfdna sequencing to develop eimit, and subsequently evaluate classification performance of linear projections of emit on six datasets and an additional inhouse testing set encopassing whole-genome, whole-genome bisulfite and 5-hydroxymethylcytosine sequencing. the linear projection of representations from emit achieved area under the receiver operating curve (auroc) values ranged from 0.895 (0.835\u20130.955) to 0.996 (0.994\u20130.997) across these six datasets, outperforming its baseline by significant margins. additionally, we showed that linear projection of emit representations can achieve an auroc of 0.962 (0.914\u20131.0) in identification of lung cancer on an independent testing set subjected to whole-exome sequencing. the findings of this study indicate that a transformer-based deep learning model can learn cancer-discrimative representations from cfdna end-motifs. the representations of this deep learning model can be exploited for discriminating patients with and without cancer.","key_phrases":["accurate discrimination","patients","cancer","cfdna","early cancer diagnosis","we","a deep-learning-based model entitled end-motif inspection","transformer","individuals","cancer","feature representations","cfdna end-motifs","emit","a self-supervised learning approach","that","cfdna end-motifs","we","4606 samples","different types","cfdna","eimit","classification performance","linear projections","six datasets","an additional inhouse testing","whole-genome, whole-genome bisulfite","5-hydroxymethylcytosine","the linear projection","representations","emit","area","the receiver operating curve (auroc) values","(0.835\u20130.955","0.994\u20130.997","these six datasets","its baseline","significant margins","we","linear projection","emit representations","an auroc","identification","lung cancer","an independent testing","the findings","this study","a transformer-based deep learning model","cancer-discrimative representations","cfdna end-motifs","the representations","this deep learning model","patients","cancer","emit","linear","six","5","0.895","0.996","six","linear","0.962"]},{"title":"VResNet: A Deep Learning Architecture for Image Inpainting of Irregular Damaged Images","authors":["Sariva Sharma","Rajneesh Rani"],"abstract":"In computer vision, image inpainting is a famous problem to automatically reconstruct the damaged part of the image according to the undamaged portion of an image. Inpainting irregular damaged areas in the image is still challenging. Deep learning-based techniques have given us a fantastic performance over the last few years. In this paper, we propose VResNet, a deep-learning approach for image inpainting, inspired by U-Net architecture and the residual framework. Since deeper neural networks are extra hard to train, the superficial convolution block in U-Net architecture is replaced by the residual learning block in the proposed approach to simplify the training of deeper neural networks. To develop an effective and adaptable model, an extensive series of experiments was conducted using the Paris-Street-View dataset. Our proposed method achieved notable results, including a PSNR of 20.65, an SSIM of 0.65, an L1 Loss of 6.90, and a total loss (L\\(_{\\hbox {Total}}\\)) of 0.30 on the Paris-Street-View dataset. These outcomes clearly demonstrate the superior performance of our model when compared to other techniques. The paper presents both qualitative and quantitative comparisons to provide a comprehensive assessment of our approach.","doi":"10.1007\/s42979-023-02523-4","cleaned_title":"vresnet: a deep learning architecture for image inpainting of irregular damaged images","cleaned_abstract":"in computer vision, image inpainting is a famous problem to automatically reconstruct the damaged part of the image according to the undamaged portion of an image. inpainting irregular damaged areas in the image is still challenging. deep learning-based techniques have given us a fantastic performance over the last few years. in this paper, we propose vresnet, a deep-learning approach for image inpainting, inspired by u-net architecture and the residual framework. since deeper neural networks are extra hard to train, the superficial convolution block in u-net architecture is replaced by the residual learning block in the proposed approach to simplify the training of deeper neural networks. to develop an effective and adaptable model, an extensive series of experiments was conducted using the paris-street-view dataset. our proposed method achieved notable results, including a psnr of 20.65, an ssim of 0.65, an l1 loss of 6.90, and a total loss (l\\(_{\\hbox {total}}\\)) of 0.30 on the paris-street-view dataset. these outcomes clearly demonstrate the superior performance of our model when compared to other techniques. the paper presents both qualitative and quantitative comparisons to provide a comprehensive assessment of our approach.","key_phrases":["computer vision","a famous problem","the damaged part","the image","the undamaged portion","an image","irregular damaged areas","the image","deep learning-based techniques","us","a fantastic performance","the last few years","this paper","we","vresnet","a deep-learning approach","image","u-net architecture","the residual framework","deeper neural networks","the superficial convolution block","u-net architecture","the residual learning block","the proposed approach","the training","deeper neural networks","an effective and adaptable model","an extensive series","experiments","the paris-street-view dataset","our proposed method","notable results","a psnr","an ssim","an l1 loss","a total loss","l\\(_{\\hbox {total}}\\","the paris-street-view dataset","these outcomes","the superior performance","our model","other techniques","the paper","both qualitative and quantitative comparisons","a comprehensive assessment","our approach","the last few years","paris","20.65","0.65","6.90","0.30","paris"]},{"title":"Deep Learning-based Pilot Adaptation and Channel Estimation in OFDM Systems","authors":["Mohammadamin Shahmohammadi","Mohammadali Sebghati","Hassan Zareian"],"abstract":"Clear communication over wireless channels demands overcoming their disruptive effects. Doubly-selective fading channels, with rapidly changing parameters, pose a particular challenge for accurate channel estimation. Traditional models often falter, lacking solutions or becoming overly complex. This is where deep learning takes center stage. In OFDM systems, pilots assist in estimating the channel response, but they also come at the cost of reduced data throughput. Adaptive adjustment of pilot patterns based on the channel state offers a promising solution. This paper introduces a deep learning-based framework that leverages adaptive pilots for fast-varying channels. We employ two deep neural networks. First, the pilot adaptation network dynamically selects the pilot pattern, reacting to the channel coherence bandwidth. Second, the channel estimation network extracts features from the channel frequency response using a 1D convolutional neural network. It then harnesses the power of long short-term memory layers\u00a0to learn the channel behavior and estimate the response across all pilots and data subcarriers. Training and testing datasets are generated using WINNER II. The entire communication link, equipped with our proposed method, undergoes rigorous simulations, evaluated by both bit error rate and pilot overhead. The simulation results illustrate that the proposed scheme outperforms the previous methods, because it yields the same or better errors with less pilot overhead. This translates to a substantial data rate boost, paving the way for faster wireless communication.","doi":"10.1007\/s11277-024-10937-3","cleaned_title":"deep learning-based pilot adaptation and channel estimation in ofdm systems","cleaned_abstract":"clear communication over wireless channels demands overcoming their disruptive effects. doubly-selective fading channels, with rapidly changing parameters, pose a particular challenge for accurate channel estimation. traditional models often falter, lacking solutions or becoming overly complex. this is where deep learning takes center stage. in ofdm systems, pilots assist in estimating the channel response, but they also come at the cost of reduced data throughput. adaptive adjustment of pilot patterns based on the channel state offers a promising solution. this paper introduces a deep learning-based framework that leverages adaptive pilots for fast-varying channels. we employ two deep neural networks. first, the pilot adaptation network dynamically selects the pilot pattern, reacting to the channel coherence bandwidth. second, the channel estimation network extracts features from the channel frequency response using a 1d convolutional neural network. it then harnesses the power of long short-term memory layers to learn the channel behavior and estimate the response across all pilots and data subcarriers. training and testing datasets are generated using winner ii. the entire communication link, equipped with our proposed method, undergoes rigorous simulations, evaluated by both bit error rate and pilot overhead. the simulation results illustrate that the proposed scheme outperforms the previous methods, because it yields the same or better errors with less pilot overhead. this translates to a substantial data rate boost, paving the way for faster wireless communication.","key_phrases":["clear communication","their disruptive effects","doubly-selective fading channels","rapidly changing parameters","a particular challenge","accurate channel estimation","traditional models","solutions","this","deep learning","center stage","ofdm systems","pilots","the channel response","they","the cost","reduced data throughput","adaptive adjustment","pilot patterns","the channel state","a promising solution","this paper","a deep learning-based framework","that","adaptive pilots","fast-varying channels","we","two deep neural networks","the pilot adaptation network","the pilot pattern","the channel coherence","the channel estimation network","features","the channel frequency response","a 1d convolutional neural network","it","the power","long short-term memory layers","the channel behavior","the response","all pilots and data subcarriers","training and testing datasets","winner ii","the entire communication link","our proposed method","undergoes rigorous simulations","both bit error rate","the simulation results","the proposed scheme","the previous methods","it","the same or better errors","this","a substantial data rate boost","the way","faster wireless communication","two","first","second","1d"]},{"title":"Intelligent upper-limb exoskeleton integrated with soft bioelectronics and deep learning for intention-driven augmentation","authors":["Jinwoo Lee","Kangkyu Kwon","Ira Soltis","Jared Matthews","Yoon Jae Lee","Hojoong Kim","Lissette Romero","Nathan Zavanelli","Youngjin Kwon","Shinjae Kwon","Jimin Lee","Yewon Na","Sung Hoon Lee","Ki Jun Yu","Minoru Shinohara","Frank L. Hammond","Woon-Hong Yeo"],"abstract":"The age and stroke-associated decline in musculoskeletal strength degrades the ability to perform daily human tasks using the upper extremities. Here, we introduce an intelligent upper-limb exoskeleton system that utilizes deep learning to predict human intention for strength augmentation. The embedded soft wearable sensors provide sensory feedback by collecting real-time muscle activities, which are simultaneously computed to determine the user\u2019s intended movement. Cloud-based deep learning predicts four upper-limb joint motions with an average accuracy of 96.2% at a 500\u2013550\u2009ms response rate, suggesting that the exoskeleton operates just by human intention. In addition, an array of soft pneumatics assists the intended movements by providing 897 newtons of force while generating a displacement of 87\u2009mm at maximum. The intent-driven exoskeleton can reduce human muscle activities by 3.7 times on average compared to the unassisted exoskeleton.","doi":"10.1038\/s41528-024-00297-0","cleaned_title":"intelligent upper-limb exoskeleton integrated with soft bioelectronics and deep learning for intention-driven augmentation","cleaned_abstract":"the age and stroke-associated decline in musculoskeletal strength degrades the ability to perform daily human tasks using the upper extremities. here, we introduce an intelligent upper-limb exoskeleton system that utilizes deep learning to predict human intention for strength augmentation. the embedded soft wearable sensors provide sensory feedback by collecting real-time muscle activities, which are simultaneously computed to determine the user\u2019s intended movement. cloud-based deep learning predicts four upper-limb joint motions with an average accuracy of 96.2% at a 500\u2013550 ms response rate, suggesting that the exoskeleton operates just by human intention. in addition, an array of soft pneumatics assists the intended movements by providing 897 newtons of force while generating a displacement of 87 mm at maximum. the intent-driven exoskeleton can reduce human muscle activities by 3.7 times on average compared to the unassisted exoskeleton.","key_phrases":["the age and stroke-associated decline","musculoskeletal strength","the ability","daily human tasks","the upper extremities","we","an intelligent upper-limb exoskeleton system","that","deep learning","human intention","strength augmentation","the embedded soft wearable sensors","sensory feedback","real-time muscle activities","which","the user\u2019s intended movement","cloud-based deep learning","four upper-limb joint motions","an average accuracy","96.2%","a 500\u2013550 ms response rate","the exoskeleton","human intention","addition","an array","soft pneumatics","the intended movements","897 newtons","force","a displacement","87 mm","maximum","the intent-driven exoskeleton","human muscle activities","3.7 times","the unassisted exoskeleton","daily","four","96.2%","500\u2013550","897","87 mm","3.7"]},{"title":"Deep sampling of gRNA in the human genome and deep-learning-informed prediction of gRNA activities","authors":["Heng Zhang","Jianfeng Yan","Zhike Lu","Yangfan Zhou","Qingfeng Zhang","Tingting Cui","Yini Li","Hui Chen","Lijia Ma"],"abstract":"Life science studies involving clustered regularly interspaced short palindromic repeat (CRISPR) editing generally apply the best-performing guide RNA (gRNA) for a gene of interest. Computational models are combined with massive experimental quantification on synthetic gRNA-target libraries to accurately predict gRNA activity and mutational patterns. However, the measurements are inconsistent between studies due to differences in the designs of the gRNA-target pair constructs, and there has not yet been an integrated investigation that concurrently focuses on multiple facets of gRNA capacity. In this study, we analyzed the DNA double-strand break (DSB)-induced repair outcomes and measured SpCas9\/gRNA activities at both matched and mismatched locations using 926,476 gRNAs covering 19,111 protein-coding genes and 20,268 non-coding genes. We developed machine learning models to forecast the on-target cleavage efficiency (AIdit_ON), off-target cleavage specificity (AIdit_OFF), and mutational profiles (AIdit_DSB) of SpCas9\/gRNA from a uniformly collected and processed dataset by deep sampling and massively quantifying gRNA capabilities in K562 cells. Each of these models exhibited superlative performance in predicting SpCas9\/gRNA activities on independent datasets when benchmarked with previous models. A previous unknown parameter was also empirically determined regarding the \u201csweet spot\u201d in the size of datasets used to establish an effective model to predict gRNA capabilities at a manageable experimental scale. In addition, we observed cell type-specific mutational profiles and were able to link nucleotidylexotransferase as the key factor driving these outcomes. These massive datasets and deep learning algorithms have been implemented into the user-friendly web service http:\/\/crispr-aidit.com to evaluate and rank gRNAs for life science studies.","doi":"10.1038\/s41421-023-00549-9","cleaned_title":"deep sampling of grna in the human genome and deep-learning-informed prediction of grna activities","cleaned_abstract":"life science studies involving clustered regularly interspaced short palindromic repeat (crispr) editing generally apply the best-performing guide rna (grna) for a gene of interest. computational models are combined with massive experimental quantification on synthetic grna-target libraries to accurately predict grna activity and mutational patterns. however, the measurements are inconsistent between studies due to differences in the designs of the grna-target pair constructs, and there has not yet been an integrated investigation that concurrently focuses on multiple facets of grna capacity. in this study, we analyzed the dna double-strand break (dsb)-induced repair outcomes and measured spcas9\/grna activities at both matched and mismatched locations using 926,476 grnas covering 19,111 protein-coding genes and 20,268 non-coding genes. we developed machine learning models to forecast the on-target cleavage efficiency (aidit_on), off-target cleavage specificity (aidit_off), and mutational profiles (aidit_dsb) of spcas9\/grna from a uniformly collected and processed dataset by deep sampling and massively quantifying grna capabilities in k562 cells. each of these models exhibited superlative performance in predicting spcas9\/grna activities on independent datasets when benchmarked with previous models. a previous unknown parameter was also empirically determined regarding the \u201csweet spot\u201d in the size of datasets used to establish an effective model to predict grna capabilities at a manageable experimental scale. in addition, we observed cell type-specific mutational profiles and were able to link nucleotidylexotransferase as the key factor driving these outcomes. these massive datasets and deep learning algorithms have been implemented into the user-friendly web service http:\/\/crispr-aidit.com to evaluate and rank grnas for life science studies.","key_phrases":["life science studies","regularly interspaced short palindromic repeat","the best-performing guide","rna","grna","a gene","interest","computational models","massive experimental quantification","synthetic grna-target libraries","grna activity","mutational patterns","the measurements","studies","differences","the designs","the grna-target pair constructs","an integrated investigation","that","multiple facets","grna capacity","this study","we","the dna double-strand break","dsb)-induced repair outcomes","measured spcas9\/grna activities","both matched and mismatched locations","926,476 grnas","19,111 protein-coding genes","20,268 non-coding genes","we","machine learning models","target","aidit_on","target","aidit_off","mutational profiles","aidit_dsb","spcas9\/grna","a uniformly collected and processed dataset","deep sampling","grna capabilities","k562 cells","each","these models","superlative performance","spcas9\/grna activities","independent datasets","previous models","a previous unknown parameter","the \u201csweet spot","the size","datasets","an effective model","grna capabilities","a manageable experimental scale","addition","we","cell type-specific mutational profiles","nucleotidylexotransferase","the key factor","these outcomes","these massive datasets","deep learning algorithms","the user-friendly web service","http:\/\/crispr-aidit.com","rank grnas","life science studies","926,476","19,111","20,268"]},{"title":"DEL-Thyroid: deep ensemble learning framework for detection of thyroid cancer progression through genomic mutation","authors":["Asghar Ali Shah","Ali Daud","Amal Bukhari","Bader Alshemaimri","Muhammad Ahsan","Rehmana Younis"],"abstract":"Genes, expressed as sequences of nucleotides, are susceptible to mutations, some of which can lead to cancer. Machine learning and deep learning methods have emerged as vital tools in identifying mutations associated with cancer. Thyroid cancer ranks as the 5th most prevalent cancer in the USA, with thousands diagnosed annually. This paper presents an ensemble learning model leveraging deep learning techniques such as Long Short-Term Memory (LSTM), Gated Recurrent Units (GRUs), and Bi-directional LSTM (Bi-LSTM) to detect thyroid cancer mutations early. The model is trained on a dataset sourced from asia.ensembl.org and IntOGen.org, consisting of 633 samples with 969 mutations across 41 genes, collected from individuals of various demographics. Feature extraction encompasses techniques including Hahn moments, central moments, raw moments, and various matrix-based methods. Evaluation employs three\u00a0testing methods: self-consistency test (SCT), independent set test (IST), and 10-fold cross-validation test (10-FCVT). The proposed ensemble learning model demonstrates promising performance, achieving 96% accuracy in the independent set test (IST). Statistical measures such as training accuracy, testing accuracy, recall, sensitivity, specificity, Mathew's Correlation Coefficient (MCC), loss, training accuracy, F1 Score, and Cohen's kappa are utilized for comprehensive evaluation.","doi":"10.1186\/s12911-024-02604-1","cleaned_title":"del-thyroid: deep ensemble learning framework for detection of thyroid cancer progression through genomic mutation","cleaned_abstract":"genes, expressed as sequences of nucleotides, are susceptible to mutations, some of which can lead to cancer. machine learning and deep learning methods have emerged as vital tools in identifying mutations associated with cancer. thyroid cancer ranks as the 5th most prevalent cancer in the usa, with thousands diagnosed annually. this paper presents an ensemble learning model leveraging deep learning techniques such as long short-term memory (lstm), gated recurrent units (grus), and bi-directional lstm (bi-lstm) to detect thyroid cancer mutations early. the model is trained on a dataset sourced from asia.ensembl.org and intogen.org, consisting of 633 samples with 969 mutations across 41 genes, collected from individuals of various demographics. feature extraction encompasses techniques including hahn moments, central moments, raw moments, and various matrix-based methods. evaluation employs three testing methods: self-consistency test (sct), independent set test (ist), and 10-fold cross-validation test (10-fcvt). the proposed ensemble learning model demonstrates promising performance, achieving 96% accuracy in the independent set test (ist). statistical measures such as training accuracy, testing accuracy, recall, sensitivity, specificity, mathew's correlation coefficient (mcc), loss, training accuracy, f1 score, and cohen's kappa are utilized for comprehensive evaluation.","key_phrases":["genes","sequences","nucleotides","mutations","some","which","cancer","machine learning","deep learning methods","vital tools","mutations","cancer","thyroid cancer","the 5th most prevalent cancer","the usa","thousands","this paper","an ensemble learning model","deep learning techniques","long short-term memory","lstm","gated recurrent units","grus","bi-directional lstm","bi","-","lstm","thyroid cancer mutations","the model","a dataset","intogen.org","633 samples","969 mutations","41 genes","individuals","various demographics","feature extraction","techniques","hahn moments","central moments","raw moments","various matrix-based methods","evaluation","three testing methods","self-consistency test","sct","independent set test","ist","10-fold cross-validation test","the proposed ensemble learning model","performance","96% accuracy","the independent set test","(ist","statistical measures","training accuracy","testing accuracy","recall","sensitivity","specificity","mathew's correlation coefficient","mcc","loss","training accuracy","f1 score","cohen's kappa","comprehensive evaluation","5th","thousands","annually","633","969","41","three","sct","10-fold","10-fcvt","96%","cohen"]},{"title":"Deep learning model for intravascular ultrasound image segmentation with temporal consistency","authors":["Hyeonmin Kim","June-Goo Lee","Gyu-Jun Jeong","Geunyoung Lee","Hyunseok Min","Hyungjoo Cho","Daegyu Min","Seung-Whan Lee","Jun Hwan Cho","Sungsoo Cho","Soo-Jin Kang"],"abstract":"This study was conducted to develop and validate a deep learning model for delineating intravascular ultrasound (IVUS) images of coronary arteries.Using a total of 1240 40-MHz IVUS pullbacks with 191,407 frames, the model for lumen and external elastic membrane (EEM) segmentation was developed. Both frame- and vessel-level performances and clinical impact of the model on 3-year cardiovascular events were evaluated in the independent data sets. In the test set, the Dice similarity coefficients (DSC) were 0.966\u2009\u00b1\u20090.025 and 0.982\u2009\u00b1\u20090.017 for the lumen and EEM, respectively. Even at sites of extensive attenuation, the frame-level performance was excellent (DSCs\u2009>\u20090.96 for the lumen and EEM). The model (vs. the expert) showed a better temporal consistency for contouring the EEM. The agreement between the model- vs. the expert-derived cross-sectional and volumetric measurements was excellent in the independent retrospective cohort (all, intra-class coefficients\u2009>\u20090.94). The model-derived percent atheroma volume\u2009>\u200952.5% (area under curve 0.70, sensitivity 71% and specificity 67%) and plaque burden at the minimal lumen area site (area under curve 0.72, sensitivity 72% and specificity 66%) best predicted 3-year cardiac death and nonculprit-related target vessel revascularization, respectively. In the stented segment, the DSCs\u2009>\u20090.96 for contouring lumen and EEM were achieved. Applied to the 60-MHz IVUS images, the DSCs were >\u20090.97. In the external cohort with 45-MHz IVUS, the DSCs were >\u20090.96. The deep learning model accurately delineated vascular geometry, which may be cost-saving and support clinical decision-making.","doi":"10.1007\/s10554-024-03221-9","cleaned_title":"deep learning model for intravascular ultrasound image segmentation with temporal consistency","cleaned_abstract":"this study was conducted to develop and validate a deep learning model for delineating intravascular ultrasound (ivus) images of coronary arteries.using a total of 1240 40-mhz ivus pullbacks with 191,407 frames, the model for lumen and external elastic membrane (eem) segmentation was developed. both frame- and vessel-level performances and clinical impact of the model on 3-year cardiovascular events were evaluated in the independent data sets. in the test set, the dice similarity coefficients (dsc) were 0.966 \u00b1 0.025 and 0.982 \u00b1 0.017 for the lumen and eem, respectively. even at sites of extensive attenuation, the frame-level performance was excellent (dscs > 0.96 for the lumen and eem). the model (vs. the expert) showed a better temporal consistency for contouring the eem. the agreement between the model- vs. the expert-derived cross-sectional and volumetric measurements was excellent in the independent retrospective cohort (all, intra-class coefficients > 0.94). the model-derived percent atheroma volume > 52.5% (area under curve 0.70, sensitivity 71% and specificity 67%) and plaque burden at the minimal lumen area site (area under curve 0.72, sensitivity 72% and specificity 66%) best predicted 3-year cardiac death and nonculprit-related target vessel revascularization, respectively. in the stented segment, the dscs > 0.96 for contouring lumen and eem were achieved. applied to the 60-mhz ivus images, the dscs were > 0.97. in the external cohort with 45-mhz ivus, the dscs were > 0.96. the deep learning model accurately delineated vascular geometry, which may be cost-saving and support clinical decision-making.","key_phrases":["this study","a deep learning model","ivus","a total","1240 40-mhz ivus pullbacks","191,407 frames","the model","lumen","external elastic membrane","(eem) segmentation","both frame- and vessel-level performances","clinical impact","the model","3-year cardiovascular events","the independent data sets","the test set","the dice similarity coefficients","0.966 \u00b1","0.982 \u00b1","the lumen","eem","sites","extensive attenuation","the frame-level performance","the lumen","eem","the model","the expert","a better temporal consistency","the eem","the agreement","the expert-derived cross","the independent retrospective cohort",", intra-class coefficients","the model-derived percent atheroma volume","52.5%","area","curve","sensitivity","specificity","plaque burden","the minimal lumen area site","area","curve","sensitivity","specificity","3-year cardiac death","nonculprit-related target vessel revascularization","the stented segment","lumen","eem","the 60-mhz ivus images","the dscs","the external cohort","45-mhz ivus","the dscs","the deep learning model","vascular geometry","which","clinical decision-making","1240","40","191,407","frame-","3-year","0.966","0.025","0.982","0.96","0.94","52.5%","0.70","71%","67%","0.72","72%","66%","3-year","0.96","60","0.97","45","0.96"]},{"title":"CT-based deep learning model for predicting hospital discharge outcome in spontaneous intracerebral hemorrhage","authors":["Xianjing Zhao","Bijing Zhou","Yong Luo","Lei Chen","Lequn Zhu","Shixin Chang","Xiangming Fang","Zhenwei Yao"],"abstract":"ObjectivesTo predict the functional outcome of patients with intracerebral hemorrhage (ICH) using deep learning models based on computed tomography (CT) images.MethodsA retrospective, bi-center study of ICH patients was conducted. Firstly, a custom 3D convolutional model was built for predicting the functional outcome of ICH patients based on CT scans from randomly selected ICH patients in H training dataset collected from H hospital. Secondly, clinical data and radiological features were collected at admission and the Extreme Gradient Boosting (XGBoost) algorithm was used to establish a second model, named the XGBoost model. Finally, the Convolution model and XGBoost model were fused to build the third \u201cFusion model.\u201d Favorable outcome was defined as modified Rankin Scale score of 0\u20133 at discharge. The prognostic predictive accuracy of the three models was evaluated using an H test dataset and an external Y dataset, and compared with the performance of ICH score and ICH grading scale (ICH-GS).ResultsA total of 604 patients with ICH were included in this study, of which 450 patients were in the H training dataset, 50 patients in the H test dataset, and 104 patients in the Y dataset. In the Y dataset, the areas under the curve (AUCs) of the Convolution model, XGBoost model, and Fusion model were 0.829, 0.871, and 0.905, respectively. The Fusion model prognostic performance exceeded that of ICH score and ICH-GS (p\u2009=\u20090.043 and p\u2009=\u20090.045, respectively).ConclusionsDeep learning models have good accuracy for predicting functional outcome of patients with spontaneous intracerebral hemorrhage.Clinical relevance statementThe proposed deep learning Fusion model may assist clinicians in predicting functional outcome and developing treatment strategies, thereby improving the survival and quality of life of patients with spontaneous intracerebral hemorrhage.Key Points\u2022 Integrating clinical presentations, CT images, and radiological features to establish deep learning model for functional outcome prediction of patients with intracerebral hemorrhage.\u2022 Deep learning applied to CT images provides great help in prognosing functional outcome of intracerebral hemorrhage patients.\u2022 The developed deep learning model performs better than clinical prognostic scores in predicting functional outcome of patients with intracerebral hemorrhage.","doi":"10.1007\/s00330-023-10505-6","cleaned_title":"ct-based deep learning model for predicting hospital discharge outcome in spontaneous intracerebral hemorrhage","cleaned_abstract":"objectivesto predict the functional outcome of patients with intracerebral hemorrhage (ich) using deep learning models based on computed tomography (ct) images.methodsa retrospective, bi-center study of ich patients was conducted. firstly, a custom 3d convolutional model was built for predicting the functional outcome of ich patients based on ct scans from randomly selected ich patients in h training dataset collected from h hospital. secondly, clinical data and radiological features were collected at admission and the extreme gradient boosting (xgboost) algorithm was used to establish a second model, named the xgboost model. finally, the convolution model and xgboost model were fused to build the third \u201cfusion model.\u201d favorable outcome was defined as modified rankin scale score of 0\u20133 at discharge. the prognostic predictive accuracy of the three models was evaluated using an h test dataset and an external y dataset, and compared with the performance of ich score and ich grading scale (ich-gs).resultsa total of 604 patients with ich were included in this study, of which 450 patients were in the h training dataset, 50 patients in the h test dataset, and 104 patients in the y dataset. in the y dataset, the areas under the curve (aucs) of the convolution model, xgboost model, and fusion model were 0.829, 0.871, and 0.905, respectively. the fusion model prognostic performance exceeded that of ich score and ich-gs (p = 0.043 and p = 0.045, respectively).conclusionsdeep learning models have good accuracy for predicting functional outcome of patients with spontaneous intracerebral hemorrhage.clinical relevance statementthe proposed deep learning fusion model may assist clinicians in predicting functional outcome and developing treatment strategies, thereby improving the survival and quality of life of patients with spontaneous intracerebral hemorrhage.key points\u2022 integrating clinical presentations, ct images, and radiological features to establish deep learning model for functional outcome prediction of patients with intracerebral hemorrhage.\u2022 deep learning applied to ct images provides great help in prognosing functional outcome of intracerebral hemorrhage patients.\u2022 the developed deep learning model performs better than clinical prognostic scores in predicting functional outcome of patients with intracerebral hemorrhage.","key_phrases":["objectivesto","the functional outcome","patients","intracerebral hemorrhage","ich","deep learning models","computed tomography","(ct","images.methodsa retrospective, bi-center study","ich patients","a custom 3d convolutional model","the functional outcome","ich patients","ct scans","randomly selected ich patients","h training dataset","h hospital","clinical data","radiological features","admission","the extreme gradient","algorithm","a second model","the xgboost model","the convolution model","xgboost model","the third \u201cfusion model","favorable outcome","modified rankin scale score","0\u20133","discharge","the prognostic predictive accuracy","the three models","an h test dataset","an external y dataset","the performance","ich score","ich grading scale","ich-gs).resultsa total","604 patients","ich","this study","which","450 patients","the h training dataset","the h test dataset","104 patients","the y dataset","the y dataset","the areas","the curve","aucs","the convolution model","xgboost model","fusion model","the fusion model prognostic performance","ich score","ich-gs","respectively).conclusionsdeep learning models","good accuracy","functional outcome","patients","spontaneous intracerebral hemorrhage.clinical relevance statementthe proposed deep learning fusion model","clinicians","functional outcome","treatment strategies","the survival","quality","life","patients","clinical presentations","ct images","radiological features","deep learning model","functional outcome prediction","patients","intracerebral hemorrhage.\u2022 deep learning","ct images","great help","functional outcome","intracerebral hemorrhage patients.\u2022","the developed deep learning model","clinical prognostic scores","functional outcome","patients","intracerebral hemorrhage","images.methodsa","firstly","3d","secondly","second","third","three","604","450","50","104","0.829","0.871","0.905","0.043","0.045","clinicians"]},{"title":"Deep Learning for Perfusion Cerebral Blood Flow (CBF) and Volume (CBV) Predictions and Diagnostics","authors":["Salmonn Talebi","Siyu Gai","Aaron Sossin","Vivian Zhu","Elizabeth Tong","Mohammad R. K. Mofrad"],"abstract":"Dynamic susceptibility contrast magnetic resonance perfusion (DSC-MRP) is a non-invasive imaging technique for hemodynamic measurements. Various perfusion parameters, such as cerebral blood volume (CBV) and cerebral blood flow (CBF), can be derived from DSC-MRP, hence this non-invasive imaging protocol is widely used clinically for the diagnosis and assessment of intracranial pathologies. Currently, most institutions use commercially available software to compute the perfusion parametric maps. However, these conventional methods often have limitations, such as being time-consuming and sensitive to user input, which can lead to inconsistent results; this highlights the need for a more robust and efficient approach like deep learning. Using the relative cerebral blood volume (rCBV) and relative cerebral blood flow (rCBF) perfusion maps generated by FDA-approved software, we trained a multistage deep learning model. The model, featuring a combination of a 1D convolutional neural network (CNN) and a 2D U-Net encoder-decoder network, processes each 4D MRP dataset by integrating temporal and spatial features of the brain for voxel-wise perfusion parameters prediction. An auxiliary model, with similar architecture, but trained with truncated datasets that had fewer time-points, was designed to explore the contribution of temporal features. Both qualitatively and quantitatively evaluated, deep learning-generated rCBV and rCBF maps showcased effective integration of temporal and spatial data, producing comprehensive predictions for the entire brain volume. Our deep learning model provides a robust and efficient approach for calculating perfusion parameters, demonstrating comparable performance to FDA-approved commercial software, and potentially mitigating the challenges inherent to traditional techniques.","doi":"10.1007\/s10439-024-03471-7","cleaned_title":"deep learning for perfusion cerebral blood flow (cbf) and volume (cbv) predictions and diagnostics","cleaned_abstract":"dynamic susceptibility contrast magnetic resonance perfusion (dsc-mrp) is a non-invasive imaging technique for hemodynamic measurements. various perfusion parameters, such as cerebral blood volume (cbv) and cerebral blood flow (cbf), can be derived from dsc-mrp, hence this non-invasive imaging protocol is widely used clinically for the diagnosis and assessment of intracranial pathologies. currently, most institutions use commercially available software to compute the perfusion parametric maps. however, these conventional methods often have limitations, such as being time-consuming and sensitive to user input, which can lead to inconsistent results; this highlights the need for a more robust and efficient approach like deep learning. using the relative cerebral blood volume (rcbv) and relative cerebral blood flow (rcbf) perfusion maps generated by fda-approved software, we trained a multistage deep learning model. the model, featuring a combination of a 1d convolutional neural network (cnn) and a 2d u-net encoder-decoder network, processes each 4d mrp dataset by integrating temporal and spatial features of the brain for voxel-wise perfusion parameters prediction. an auxiliary model, with similar architecture, but trained with truncated datasets that had fewer time-points, was designed to explore the contribution of temporal features. both qualitatively and quantitatively evaluated, deep learning-generated rcbv and rcbf maps showcased effective integration of temporal and spatial data, producing comprehensive predictions for the entire brain volume. our deep learning model provides a robust and efficient approach for calculating perfusion parameters, demonstrating comparable performance to fda-approved commercial software, and potentially mitigating the challenges inherent to traditional techniques.","key_phrases":["dynamic susceptibility contrast magnetic resonance perfusion","dsc-mrp","a non-invasive imaging technique","hemodynamic measurements","various perfusion parameters","cerebral blood volume","cbv","cerebral blood flow","cbf","dsc-mrp","this non-invasive imaging protocol","the diagnosis","assessment","intracranial pathologies","most institutions","commercially available software","the perfusion parametric maps","these conventional methods","limitations","user input","which","inconsistent results","this","the need","a more robust and efficient approach","deep learning","the relative cerebral blood volume","rcbv","relative cerebral blood flow","(rcbf) perfusion maps","fda-approved software","we","a multistage deep learning model","the model","a combination","a 1d convolutional neural network","cnn","a 2d u-net encoder-decoder network","each 4d mrp dataset","temporal and spatial features","the brain","voxel-wise perfusion parameters prediction","an auxiliary model","similar architecture","truncated datasets","that","fewer time-points","the contribution","temporal features","both qualitatively and quantitatively evaluated, deep learning-generated rcbv","rcbf maps","effective integration","temporal and spatial data","comprehensive predictions","the entire brain volume","our deep learning model","a robust and efficient approach","perfusion parameters","comparable performance","fda-approved commercial software","the challenges","traditional techniques","fda","1d","cnn","2d","4d","fda"]},{"title":"Advancing horizons in remote sensing: a comprehensive survey of deep learning models and applications in image classification and beyond","authors":["Sidike Paheding","Ashraf Saleem","Mohammad Faridul Haque Siddiqui","Nathir Rawashdeh","Almabrok Essa","Abel A. Reyes"],"abstract":"In recent years, deep learning has significantly reshaped numerous fields and applications, fundamentally altering how we tackle a variety of challenges. Areas such as natural language processing (NLP), computer vision, healthcare, network security, wide-area surveillance, and precision agriculture have leveraged the merits of the deep learning era. Particularly, deep learning has significantly improved the analysis of remote sensing images, with a\u00a0continuous increase in the number of researchers and\u00a0contributions to the field. The high impact of deep learning development is complemented by rapid advancements and the availability of data from a variety of sensors, including high-resolution RGB, thermal, LiDAR, and multi-\/hyperspectral cameras, as well as emerging sensing platforms such as satellites and aerial vehicles that can be captured by multi-temporal, multi-sensor, and sensing devices with a wider view. This study aims to present an extensive survey that encapsulates widely used deep learning strategies for tackling image classification challenges in remote sensing. It encompasses an exploration of remote sensing imaging platforms, sensor varieties, practical applications, and prospective developments in the field.","doi":"10.1007\/s00521-024-10165-7","cleaned_title":"advancing horizons in remote sensing: a comprehensive survey of deep learning models and applications in image classification and beyond","cleaned_abstract":"in recent years, deep learning has significantly reshaped numerous fields and applications, fundamentally altering how we tackle a variety of challenges. areas such as natural language processing (nlp), computer vision, healthcare, network security, wide-area surveillance, and precision agriculture have leveraged the merits of the deep learning era. particularly, deep learning has significantly improved the analysis of remote sensing images, with a continuous increase in the number of researchers and contributions to the field. the high impact of deep learning development is complemented by rapid advancements and the availability of data from a variety of sensors, including high-resolution rgb, thermal, lidar, and multi-\/hyperspectral cameras, as well as emerging sensing platforms such as satellites and aerial vehicles that can be captured by multi-temporal, multi-sensor, and sensing devices with a wider view. this study aims to present an extensive survey that encapsulates widely used deep learning strategies for tackling image classification challenges in remote sensing. it encompasses an exploration of remote sensing imaging platforms, sensor varieties, practical applications, and prospective developments in the field.","key_phrases":["recent years","deep learning","numerous fields","applications","we","a variety","challenges","areas","natural language processing","nlp","computer vision","healthcare","network security","wide-area surveillance","precision agriculture","the merits","the deep learning era","deep learning","the analysis","remote sensing images","a continuous increase","the number","researchers","contributions","the field","the high impact","deep learning development","rapid advancements","the availability","data","a variety","sensors","high-resolution rgb, thermal, lidar, and multi-\/hyperspectral cameras, as well as emerging sensing platforms","satellites","aerial vehicles","that","-","sensor","devices","a wider view","this study","an extensive survey","that","widely used deep learning strategies","image classification challenges","remote sensing","it","an exploration","remote sensing imaging platforms","sensor varieties","practical applications","prospective developments","the field","recent years","healthcare"]},{"title":"Comparative study of machine learning and deep learning techniques for fault diagnosis in suspension system","authors":["P. Arun Balaji","V. Sugumaran"],"abstract":"Comfort and stability are the prime reasons to own an automobile (car). Suspension system of an automobile plays a major role in providing comfort, stability and control. Over a period of time, internal components in the suspension system exhibit faults due to fatigue and wear. Hence, it is essential to perform fault diagnosis such that the performance of the suspension components is restored. However, high instrumentation cost, skilled labor requirement and expertise in the particular field of study are certain drawbacks of traditional fault diagnosis techniques. Such challenges have made industrialists and the research communities look for advanced fault diagnosis techniques. Advancements in machine learning and deep learning techniques can be used to fulfill the need of a high degree intelligent fault diagnosis system. In the current study, the performance of machine learning (ML) classifiers are compared with the performance of deep learning (DL) models and the best performing model among them is adopted to detect faults in the automobile suspension system. A total of eight test conditions, namely strut external damage, strut mount failure, ball joint worn out, control arm bush worn out, control arm ball joint worn out, strut worn out, low wheel pressure and good condition, were considered in the study. The vibration measurements were acquired for three load conditions. Among all the techniques considered for classification, the pre-trained VGG16 model outperformed other DL and ML models with an overall classification accuracy of 98.10%.","doi":"10.1007\/s40430-023-04145-6","cleaned_title":"comparative study of machine learning and deep learning techniques for fault diagnosis in suspension system","cleaned_abstract":"comfort and stability are the prime reasons to own an automobile (car). suspension system of an automobile plays a major role in providing comfort, stability and control. over a period of time, internal components in the suspension system exhibit faults due to fatigue and wear. hence, it is essential to perform fault diagnosis such that the performance of the suspension components is restored. however, high instrumentation cost, skilled labor requirement and expertise in the particular field of study are certain drawbacks of traditional fault diagnosis techniques. such challenges have made industrialists and the research communities look for advanced fault diagnosis techniques. advancements in machine learning and deep learning techniques can be used to fulfill the need of a high degree intelligent fault diagnosis system. in the current study, the performance of machine learning (ml) classifiers are compared with the performance of deep learning (dl) models and the best performing model among them is adopted to detect faults in the automobile suspension system. a total of eight test conditions, namely strut external damage, strut mount failure, ball joint worn out, control arm bush worn out, control arm ball joint worn out, strut worn out, low wheel pressure and good condition, were considered in the study. the vibration measurements were acquired for three load conditions. among all the techniques considered for classification, the pre-trained vgg16 model outperformed other dl and ml models with an overall classification accuracy of 98.10%.","key_phrases":["comfort","stability","the prime reasons","an automobile (car","suspension system","an automobile","a major role","comfort","stability","control","a period","time","the suspension system","exhibit faults","fatigue","it","fault diagnosis","the performance","the suspension components","high instrumentation cost","skilled labor requirement","expertise","the particular field","study","certain drawbacks","traditional fault diagnosis techniques","such challenges","industrialists","the research communities","advanced fault diagnosis techniques","advancements","machine learning","deep learning techniques","the need","a high degree intelligent fault diagnosis system","the current study","the performance","machine learning (ml) classifiers","the performance","deep learning (dl) models","the best performing model","them","faults","the automobile suspension system","a total","eight test conditions","namely strut external damage","strut mount failure","ball joint","control arm bush","control arm ball joint","strut","low wheel pressure","good condition","the study","the vibration measurements","three load conditions","all the techniques","classification","the pre-trained vgg16 model","other dl and ml models","an overall classification accuracy","98.10%","eight","strut mount","bush","three","98.10%"]},{"title":"Deep representation learning and reinforcement learning for workpiece setup optimization in CNC milling","authors":["Vladimir Samsonov","Enslin Chrismarie","Hans-Georg K\u00f6pken","Schirin B\u00e4r","Daniel L\u00fctticke","Tobias Meisen"],"abstract":"Computer Numerical Control (CNC) milling is a commonly used manufacturing process with a high level of automation. Nevertheless, setting up a new CNC milling process involves multiple development steps relying heavily on human expertise. In this work, we focus on positioning and orientation of the workpiece (WP) in the working space of a CNC milling machine and propose a deep learning approach to speed up this process significantly. The selection of the WP\u2019s setup depends on the chosen milling technological process, the geometry of the WP, and the capabilities of the considered CNC machining. It directly impacts the milling quality, machine wear, and overall energy consumption. Our approach relies on representation learning of the milling technological process with the subsequent use of reinforcement learning (RL) for the WP positioning and orientation. Solutions proposed by the RL agent are used as a warm start for simple hill-climbing heuristics, which boosts overall performance while keeping the overall number of search iterations low. The novelty of the developed approach is the ability to conduct the WP setup optimization covering both WP positioning and orientation while ensuring the axis collision avoidance, minimization of the axis traveled distances and improving the dynamic characteristics of the milling process with no input from human experts. Experiments show the potential of the proposed learning-based approach to generate almost comparably good WP setups order of magnitude faster than common metaheuristics, such as genetic algorithms (GA) and Particle Swarm Optimisation (PSA).","doi":"10.1007\/s11740-023-01209-3","cleaned_title":"deep representation learning and reinforcement learning for workpiece setup optimization in cnc milling","cleaned_abstract":"computer numerical control (cnc) milling is a commonly used manufacturing process with a high level of automation. nevertheless, setting up a new cnc milling process involves multiple development steps relying heavily on human expertise. in this work, we focus on positioning and orientation of the workpiece (wp) in the working space of a cnc milling machine and propose a deep learning approach to speed up this process significantly. the selection of the wp\u2019s setup depends on the chosen milling technological process, the geometry of the wp, and the capabilities of the considered cnc machining. it directly impacts the milling quality, machine wear, and overall energy consumption. our approach relies on representation learning of the milling technological process with the subsequent use of reinforcement learning (rl) for the wp positioning and orientation. solutions proposed by the rl agent are used as a warm start for simple hill-climbing heuristics, which boosts overall performance while keeping the overall number of search iterations low. the novelty of the developed approach is the ability to conduct the wp setup optimization covering both wp positioning and orientation while ensuring the axis collision avoidance, minimization of the axis traveled distances and improving the dynamic characteristics of the milling process with no input from human experts. experiments show the potential of the proposed learning-based approach to generate almost comparably good wp setups order of magnitude faster than common metaheuristics, such as genetic algorithms (ga) and particle swarm optimisation (psa).","key_phrases":["cnc","a commonly used manufacturing process","a high level","automation","a new cnc milling process","multiple development steps","human expertise","this work","we","positioning","orientation","the workpiece","wp","the working space","a cnc milling machine","a deep learning approach","this process","the selection","the wp\u2019s setup","the chosen milling technological process","the geometry","the wp","the capabilities","the considered cnc machining","it","the milling quality","machine wear","overall energy consumption","our approach","representation learning","the milling technological process","the subsequent use","reinforcement learning","the wp positioning","orientation","solutions","the rl agent","a warm start","simple hill-climbing heuristics","which","overall performance","the overall number","search iterations","the novelty","the developed approach","the ability","the wp setup optimization","both wp positioning","orientation","the axis collision avoidance","minimization","the axis","distances","the dynamic characteristics","the milling process","no input","human experts","experiments","the potential","the proposed learning-based approach","almost comparably good wp setups order","magnitude","common metaheuristics","genetic algorithms","ga","particle swarm optimisation","psa","simple hill","ga"]},{"title":"Boosting in-transit entertainment: deep reinforcement learning for intelligent multimedia caching in bus networks","authors":["Dan Lan","Incheol Shin"],"abstract":"Multimedia content delivery in advanced networks faces exponential growth in data volumes, rendering existing solutions obsolete. This research investigates deep reinforcement learning (DRL) for autonomous optimization without extensive datasets. The work analyzes two prominent DRL algorithms, i.e., Dueling Deep Q-Network (DDQN) and Deep Q-Network (DQN) for multimedia delivery in simulated bus networks. DDQN utilizes a novel \u201cdueling\u201d architecture to estimate state value and action advantages, accelerating learning separately. DQN employs deep neural networks to approximate optimal policies. The environment simulates urban buses with passenger file requests and cache sizes modeled on actual data. Comparative analysis evaluates cumulative rewards and losses over 1500 training episodes to analyze learning efficiency, stability, and performance. Results demonstrate DDQN\u2019s superior convergence and 32% higher cumulative rewards than DQN. However, DQN showed potential for gains over successive runs despite inconsistencies. It establishes DRL\u2019s promise for automated decision-making while revealing enhancements to improve DQN. Further research should evaluate generalizability across problem domains, investigate hybrid models, and test physical systems. DDQN emerged as the most efficient algorithm, highlighting DRL\u2019s potential to enable intelligent agents that optimize multimedia delivery.","doi":"10.1007\/s00500-023-09354-8","cleaned_title":"boosting in-transit entertainment: deep reinforcement learning for intelligent multimedia caching in bus networks","cleaned_abstract":"multimedia content delivery in advanced networks faces exponential growth in data volumes, rendering existing solutions obsolete. this research investigates deep reinforcement learning (drl) for autonomous optimization without extensive datasets. the work analyzes two prominent drl algorithms, i.e., dueling deep q-network (ddqn) and deep q-network (dqn) for multimedia delivery in simulated bus networks. ddqn utilizes a novel \u201cdueling\u201d architecture to estimate state value and action advantages, accelerating learning separately. dqn employs deep neural networks to approximate optimal policies. the environment simulates urban buses with passenger file requests and cache sizes modeled on actual data. comparative analysis evaluates cumulative rewards and losses over 1500 training episodes to analyze learning efficiency, stability, and performance. results demonstrate ddqn\u2019s superior convergence and 32% higher cumulative rewards than dqn. however, dqn showed potential for gains over successive runs despite inconsistencies. it establishes drl\u2019s promise for automated decision-making while revealing enhancements to improve dqn. further research should evaluate generalizability across problem domains, investigate hybrid models, and test physical systems. ddqn emerged as the most efficient algorithm, highlighting drl\u2019s potential to enable intelligent agents that optimize multimedia delivery.","key_phrases":["multimedia content delivery","advanced networks","exponential growth","data volumes","this research investigates","deep reinforcement learning","drl","autonomous optimization","extensive datasets","the work","two prominent drl algorithms","deep q-network (ddqn","deep q-network","dqn","multimedia delivery","simulated bus networks","ddqn","a novel \u201cdueling\u201d architecture","state value","action advantages","dqn","deep neural networks","optimal policies","the environment","urban buses","passenger file requests","cache sizes","actual data","comparative analysis","cumulative rewards","losses","1500 training episodes","learning efficiency","stability","performance","results","ddqn\u2019s superior convergence","32% higher cumulative rewards","dqn","dqn","potential","gains","successive runs","inconsistencies","it","drl\u2019s promise","automated decision-making","enhancements","dqn","further research","generalizability","problem domains","hybrid models","physical systems","ddqn","the most efficient algorithm","drl\u2019s potential","intelligent agents","that","multimedia delivery","two","1500","32%"]},{"title":"A novel deep facenet framework for real-time face detection based on deep learning model","authors":["B Lakshmanan","A Vaishnavi","R Ananthapriya","A K Aananthalakshmi"],"abstract":"Real-time face detection has many challenges, such as non-frontal faces, tiny faces, occlusions, and multifarious backgrounds. Real-time face detection can be done by using Convolutional Neural Network (CNN) models, which result in elevated performance but have a huge computation time. It needs to be implemented on high-end computational devices to produce more accurate face detection results for high resolution images. We proposed a light architecture based on CNN for deep learning-based feature extraction and detection of human faces. The challenges faced during real-time face detection, such as occlusions, different scales, different backgrounds, varying positions, lighting, and poses, are resolved, and faces are detected accurately using the proposed framework. The amount of computation required for real-time face detection is reduced. This light architecture consists of two modules: the backbone module is used to contract the input size of the image and extract the features; the detection module transforms the feature map between prediction layers and detects faces at various scales. In our architecture, we use mini-inception blocks that minimize the computation cost and are implemented using available low-end system configurations without the need for external hardware. The proposed model uses anchor boxes to predict bounding boxes using dimensional clusters. The model is trained and tested using images from the WIDER Face dataset, which has images of various challenging conditions. Finally, images with multiple faces detected are displayed as output. The proposed work shows an increased accuracy rate with reduced computation cost over state-of-the-art performance on the benchmark dataset.","doi":"10.1007\/s12046-023-02329-3","cleaned_title":"a novel deep facenet framework for real-time face detection based on deep learning model","cleaned_abstract":"real-time face detection has many challenges, such as non-frontal faces, tiny faces, occlusions, and multifarious backgrounds. real-time face detection can be done by using convolutional neural network (cnn) models, which result in elevated performance but have a huge computation time. it needs to be implemented on high-end computational devices to produce more accurate face detection results for high resolution images. we proposed a light architecture based on cnn for deep learning-based feature extraction and detection of human faces. the challenges faced during real-time face detection, such as occlusions, different scales, different backgrounds, varying positions, lighting, and poses, are resolved, and faces are detected accurately using the proposed framework. the amount of computation required for real-time face detection is reduced. this light architecture consists of two modules: the backbone module is used to contract the input size of the image and extract the features; the detection module transforms the feature map between prediction layers and detects faces at various scales. in our architecture, we use mini-inception blocks that minimize the computation cost and are implemented using available low-end system configurations without the need for external hardware. the proposed model uses anchor boxes to predict bounding boxes using dimensional clusters. the model is trained and tested using images from the wider face dataset, which has images of various challenging conditions. finally, images with multiple faces detected are displayed as output. the proposed work shows an increased accuracy rate with reduced computation cost over state-of-the-art performance on the benchmark dataset.","key_phrases":["real-time face detection","many challenges","non-frontal faces","tiny faces","occlusions","multifarious backgrounds","real-time face detection","convolutional neural network (cnn) models","which","elevated performance","a huge computation time","it","high-end computational devices","more accurate face detection results","high resolution images","we","a light architecture","cnn","deep learning-based feature extraction","detection","human faces","the challenges","real-time face detection","occlusions","different scales","different backgrounds","varying positions","lighting","poses","faces","the proposed framework","the amount","computation","real-time face detection","this light architecture","two modules","the backbone module","the input size","the image","the features","the detection module","the feature map","prediction layers","detects","various scales","our architecture","we","mini","-inception blocks","that","the computation cost","available low-end system configurations","the need","external hardware","the proposed model","anchor boxes","bounding boxes","dimensional clusters","the model","images","the wider face dataset","which","images","various challenging conditions","images","multiple faces","output","the proposed work","an increased accuracy rate","reduced computation cost","the-art","the benchmark dataset","cnn","cnn","two modules"]},{"title":"Development of a remote music teaching system based on facial recognition and deep learning","authors":["Ning Zhang","Huizhong Wang"],"abstract":"With the continuous progress of computer and network technology, teaching methods and educational models are also constantly evolving and improving. The development of facial recognition technology has brought new opportunities and challenges to the development of educational theory and systems. This article establishes a remote music teaching system based on facial recognition and deep learning technology. The system adopts the Java EE framework structure and deep learning technology. By conducting deep learning and training on a large amount of facial data, we can identify students' facial expressions and emotional states, thereby better understanding their learning status and needs. At the same time, the system also supports multiple teaching modes and interactive methods, providing teachers and students with a more convenient and efficient teaching management and learning experience. Subsequently, this article evaluated and explored the effectiveness of the remote music teaching system through a questionnaire survey. The results show that most students believe that the system can help them better master basic music knowledge and professional skills, improve learning effectiveness and achieve learning goals. The use of the system can also stimulate students' interest in music learning, providing new ways and means for teaching.","doi":"10.1007\/s00500-023-09120-w","cleaned_title":"development of a remote music teaching system based on facial recognition and deep learning","cleaned_abstract":"with the continuous progress of computer and network technology, teaching methods and educational models are also constantly evolving and improving. the development of facial recognition technology has brought new opportunities and challenges to the development of educational theory and systems. this article establishes a remote music teaching system based on facial recognition and deep learning technology. the system adopts the java ee framework structure and deep learning technology. by conducting deep learning and training on a large amount of facial data, we can identify students' facial expressions and emotional states, thereby better understanding their learning status and needs. at the same time, the system also supports multiple teaching modes and interactive methods, providing teachers and students with a more convenient and efficient teaching management and learning experience. subsequently, this article evaluated and explored the effectiveness of the remote music teaching system through a questionnaire survey. the results show that most students believe that the system can help them better master basic music knowledge and professional skills, improve learning effectiveness and achieve learning goals. the use of the system can also stimulate students' interest in music learning, providing new ways and means for teaching.","key_phrases":["the continuous progress","computer and network technology","teaching methods","educational models","the development","facial recognition technology","new opportunities","challenges","the development","educational theory","systems","this article","a remote music teaching system","facial recognition","deep learning technology","the system","the java ee framework structure","deep learning technology","deep learning","training","a large amount","facial data","we","students' facial expressions","emotional states","their learning status","needs","the same time","the system","multiple teaching modes","interactive methods","teachers","students","a more convenient and efficient teaching management and learning experience","this article","the effectiveness","the remote music teaching system","a questionnaire survey","the results","most students","the system","them","basic music knowledge","professional skills","effectiveness","learning goals","the use","the system","students' interest","new ways","teaching"]},{"title":"Parameter-Free Reduction of the Estimation Bias in Deep Reinforcement Learning for Deterministic Policy Gradients","authors":["Baturay Saglam","Furkan Burak Mutlu","Dogan Can Cicek","Suleyman Serdar Kozat"],"abstract":"Approximation of the value functions in value-based deep reinforcement learning induces overestimation bias, resulting in suboptimal policies. We show that when the reinforcement signals received by the agents have a high variance, deep actor-critic approaches that overcome the overestimation bias lead to a substantial underestimation bias. We first address the detrimental issues in the existing approaches that aim to overcome such underestimation error. Then, through extensive statistical analysis, we introduce a novel, parameter-free Deep Q-learning variant to reduce this underestimation bias in deterministic policy gradients. By sampling the weights of a linear combination of two approximate critics from a highly shrunk estimation bias interval, our Q-value update rule is not affected by the variance of the rewards received by the agents throughout learning. We test the performance of the introduced improvement on a set of MuJoCo and Box2D continuous control tasks and demonstrate that it outperforms the existing approaches and improves the baseline actor-critic algorithm in most of the environments tested.","doi":"10.1007\/s11063-024-11461-y","cleaned_title":"parameter-free reduction of the estimation bias in deep reinforcement learning for deterministic policy gradients","cleaned_abstract":"approximation of the value functions in value-based deep reinforcement learning induces overestimation bias, resulting in suboptimal policies. we show that when the reinforcement signals received by the agents have a high variance, deep actor-critic approaches that overcome the overestimation bias lead to a substantial underestimation bias. we first address the detrimental issues in the existing approaches that aim to overcome such underestimation error. then, through extensive statistical analysis, we introduce a novel, parameter-free deep q-learning variant to reduce this underestimation bias in deterministic policy gradients. by sampling the weights of a linear combination of two approximate critics from a highly shrunk estimation bias interval, our q-value update rule is not affected by the variance of the rewards received by the agents throughout learning. we test the performance of the introduced improvement on a set of mujoco and box2d continuous control tasks and demonstrate that it outperforms the existing approaches and improves the baseline actor-critic algorithm in most of the environments tested.","key_phrases":["approximation","the value functions","value-based deep reinforcement learning","overestimation bias","suboptimal policies","we","the reinforcement signals","the agents","a high variance","deep actor-critic approaches","that","the overestimation bias","a substantial underestimation bias","we","the detrimental issues","the existing approaches","that","such underestimation error","extensive statistical analysis","we","a novel, parameter-free deep q-learning variant","this underestimation bias","deterministic policy gradients","the weights","a linear combination","two approximate critics","a highly shrunk estimation bias interval","our q-value update rule","the variance","the rewards","the agents","we","the performance","the introduced improvement","a set","mujoco","box2d continuous control tasks","it","the existing approaches","the baseline actor-critic algorithm","the environments","first","two"]},{"title":"C-net: a deep learning-based Jujube grading approach","authors":["Atif Mahmood","Amod Kumar Tiwari","Sanjay Kumar Singh"],"abstract":"Jujube grading is a crucial process in the jujube-associated industry to ascertain the quality, ripeness, value, and security of the product. Traditionally, jujube grading has been done manually, which may be expensive, time-consuming, and prone to human mistakes. With the expansion of innovation, Machine Learning (ML)\/Deep Learning (DL) turned out as a potent technique for automating the fruits grading process. Within this work, we deployed and analyzed the Concatenated-Convolutional Neural Network (C-Net) based on the residual network concept and seven cutting-edge CNNs for sorting the Indian jujube into six classes. To train and evaluate the models, we collected and assembled the dataset of jujube images. The performance analysis of the model relies upon two varying hyperparameters, batch size, and epochs as well as some performance metrics like F1-score, precision, and recall. The finding indicates that the proposed C-Net model was able to classify jujube images with high precision of 98.61% which surpasses other models but lags slightly behind the EfficientNet-B0 model. Our C-Net model has several advantages over most of the cutting-edge CNN models for jujube grading including increased accuracy, efficiency, cost-effectiveness, better decision-making, scalability, and real-time grading. The use of a C-Net model for jujube grading has the capability to revolutionize the jujube grading task and improve the fruit\u2019s overall quality.","doi":"10.1007\/s11694-024-02765-7","cleaned_title":"c-net: a deep learning-based jujube grading approach","cleaned_abstract":"jujube grading is a crucial process in the jujube-associated industry to ascertain the quality, ripeness, value, and security of the product. traditionally, jujube grading has been done manually, which may be expensive, time-consuming, and prone to human mistakes. with the expansion of innovation, machine learning (ml)\/deep learning (dl) turned out as a potent technique for automating the fruits grading process. within this work, we deployed and analyzed the concatenated-convolutional neural network (c-net) based on the residual network concept and seven cutting-edge cnns for sorting the indian jujube into six classes. to train and evaluate the models, we collected and assembled the dataset of jujube images. the performance analysis of the model relies upon two varying hyperparameters, batch size, and epochs as well as some performance metrics like f1-score, precision, and recall. the finding indicates that the proposed c-net model was able to classify jujube images with high precision of 98.61% which surpasses other models but lags slightly behind the efficientnet-b0 model. our c-net model has several advantages over most of the cutting-edge cnn models for jujube grading including increased accuracy, efficiency, cost-effectiveness, better decision-making, scalability, and real-time grading. the use of a c-net model for jujube grading has the capability to revolutionize the jujube grading task and improve the fruit\u2019s overall quality.","key_phrases":["a crucial process","the jujube-associated industry","the quality","ripeness","value","security","the product","jujube grading","which","human mistakes","the expansion","innovation","machine learning","dl","a potent technique","the fruits grading process","this work","we","the concatenated-convolutional neural network","c-net","the residual network concept","seven cutting-edge cnns","the indian jujube","six classes","the models","we","the dataset","jujube images","the performance analysis","the model","two varying hyperparameters","batch size","epochs","some performance metrics","f1-score","precision","the finding","the proposed c-net model","jujube images","high precision","98.61%","which","other models","lags","the efficientnet-b0 model","our c-net model","several advantages","the cutting-edge cnn models","increased accuracy","efficiency","cost-effectiveness","better decision-making","scalability","real-time grading","the use","a c-net model","the capability","the jujube","task","the fruit\u2019s overall quality","jujube","jujube","seven","indian","six","two","98.61%","cnn"]},{"title":"Hierarchical Goal-Guided Learning for the Evasive Maneuver of Fixed-Wing UAVs based on Deep Reinforcement Learning","authors":["Yinlong Yuan","Jian Yang","Zhu Liang Yu","Yun Cheng","Pengpeng Jiao","Liang Hua"],"abstract":"Fixed-wing unmanned aerial vehicles (UAVs) will play a vital role in forthcoming military conflicts. Effectively avoiding threats and improving the survivability of fixed-wing UAV in dynamic hostile environments are the keys to the success of combat missions. Hence, endowing fixed-wing UAVs with the ability to autonomously generate evasive maneuver is the primary problem that should be solved. With considering the threat of air-to-air missile attacks, this paper designs a novel hierarchical goal-guided learning (HGGL) method, which combines with traditional off-policy deep reinforcement learning (DRL) algorithms and endows the agent with the ability to evade a series of air-to-air missiles. The pivotal idea of the proposed algorithm is to use the hierarchical features of the goal, it improves the availability of training data to eliminate the limitation of the convergence rate of traditional DRL algorithms owing to sparse rewards. We demonstrate the performance of our algorithm in several simulation experiments. All experiments are applied on the XSimStudio platform. The results demonstrate that the proposed algorithm improves the convergence speed and outperforms the state-of-the-art traditional algorithms.","doi":"10.1007\/s10846-023-01953-9","cleaned_title":"hierarchical goal-guided learning for the evasive maneuver of fixed-wing uavs based on deep reinforcement learning","cleaned_abstract":"fixed-wing unmanned aerial vehicles (uavs) will play a vital role in forthcoming military conflicts. effectively avoiding threats and improving the survivability of fixed-wing uav in dynamic hostile environments are the keys to the success of combat missions. hence, endowing fixed-wing uavs with the ability to autonomously generate evasive maneuver is the primary problem that should be solved. with considering the threat of air-to-air missile attacks, this paper designs a novel hierarchical goal-guided learning (hggl) method, which combines with traditional off-policy deep reinforcement learning (drl) algorithms and endows the agent with the ability to evade a series of air-to-air missiles. the pivotal idea of the proposed algorithm is to use the hierarchical features of the goal, it improves the availability of training data to eliminate the limitation of the convergence rate of traditional drl algorithms owing to sparse rewards. we demonstrate the performance of our algorithm in several simulation experiments. all experiments are applied on the xsimstudio platform. the results demonstrate that the proposed algorithm improves the convergence speed and outperforms the state-of-the-art traditional algorithms.","key_phrases":["fixed-wing unmanned aerial vehicles","a vital role","forthcoming military conflicts","threats","the survivability","fixed-wing uav","dynamic hostile environments","the keys","the success","combat missions","fixed-wing uavs","the ability","evasive maneuver","the primary problem","that","the threat","air","this paper","a novel hierarchical goal-guided learning (hggl) method","which","drl","the agent","the ability","a series","air","the pivotal idea","the proposed algorithm","the hierarchical features","the goal","it","the availability","training data","the limitation","the convergence rate","traditional drl algorithms","sparse rewards","we","the performance","our algorithm","several simulation experiments","all experiments","the xsimstudio platform","the results","the proposed algorithm","the convergence speed","the-art"]},{"title":"Advanced deep learning and large language models for suicide ideation detection on social media","authors":["Mohammed Qorich","Rajae El Ouazzani"],"abstract":"Recently, suicide ideations represent a worldwide health concern and pose many anticipation challenges. Actually, the prevalence of expressing self-destructive thoughts especially on forums and social media requires effective monitoring for suicide prevention, and early intervention.\u00a0Meanwhile, deep learning techniques and Large Language Models (LLMs) have emerged as promising tools in diverse Natural Language Processing (NLP) tasks, including sentiment analysis and text classification. In this paper, we propose a deep learning model incorporating triple models of word embeddings, as well as various fine-tuned LLMs, to identify suicidal thoughts in Reddit posts. In effect, we implemented a Bidirectional Long Short-Term Memory (BiLSTM), and a Convolutional Neural Network (CNN) model to categorize posts associated with non-suicidal and suicidal thoughts. Besides, through the combination of Word2Vec, FastText and GloVe embeddings, our models learn intricate patterns and prevalent nuances in suicide-related language. Furthermore, we employed a merged version of CNN and BiLSTM models, entitled C-BiLSTM, and several LLMs, including pre-trained Bidirectional Encoder Representations from Transformers (BERT) models and a Generative Pre-training Transformer (GPT) model. The analysis of all our proposed models shows that our C-BiLSTM model with triple word embedding and our GPT model got the best performance compared to deep learning and LLMs baseline models, reaching accuracies of 94.5% and 97.69%, respectively. In fact, our best model\u2019s capacity to extract meaningful interdependencies among words significantly promotes its classification performance. This analysis contributes to a deeper understanding of the psychological factors and linguistic markers indicative of suicidal thoughts, thereby informing future research and intervention strategies.","doi":"10.1007\/s13748-024-00326-z","cleaned_title":"advanced deep learning and large language models for suicide ideation detection on social media","cleaned_abstract":"recently, suicide ideations represent a worldwide health concern and pose many anticipation challenges. actually, the prevalence of expressing self-destructive thoughts especially on forums and social media requires effective monitoring for suicide prevention, and early intervention. meanwhile, deep learning techniques and large language models (llms) have emerged as promising tools in diverse natural language processing (nlp) tasks, including sentiment analysis and text classification. in this paper, we propose a deep learning model incorporating triple models of word embeddings, as well as various fine-tuned llms, to identify suicidal thoughts in reddit posts. in effect, we implemented a bidirectional long short-term memory (bilstm), and a convolutional neural network (cnn) model to categorize posts associated with non-suicidal and suicidal thoughts. besides, through the combination of word2vec, fasttext and glove embeddings, our models learn intricate patterns and prevalent nuances in suicide-related language. furthermore, we employed a merged version of cnn and bilstm models, entitled c-bilstm, and several llms, including pre-trained bidirectional encoder representations from transformers (bert) models and a generative pre-training transformer (gpt) model. the analysis of all our proposed models shows that our c-bilstm model with triple word embedding and our gpt model got the best performance compared to deep learning and llms baseline models, reaching accuracies of 94.5% and 97.69%, respectively. in fact, our best model\u2019s capacity to extract meaningful interdependencies among words significantly promotes its classification performance. this analysis contributes to a deeper understanding of the psychological factors and linguistic markers indicative of suicidal thoughts, thereby informing future research and intervention strategies.","key_phrases":["suicide ideations","a worldwide health concern","many anticipation challenges","the prevalence","self-destructive thoughts","forums","social media","effective monitoring","suicide prevention","early intervention","deep learning techniques","large language models","llms","promising tools","nlp","sentiment analysis","text classification","this paper","we","a deep learning model","triple models","word embeddings","various fine-tuned llms","suicidal thoughts","reddit posts","effect","we","a bidirectional long short-term memory","bilstm","a convolutional neural network (cnn) model","posts","-suicidal and suicidal thoughts","the combination","word2vec, fasttext and glove embeddings","our models","intricate patterns","prevalent nuances","suicide-related language","we","a merged version","cnn","bilstm models","c-bilstm","several llms","pre-trained bidirectional encoder representations","transformers","(bert) models","a generative pre-training transformer","(gpt) model","the analysis","all our proposed models","our c-bilstm model","triple word","our gpt model","the best performance","deep learning and llms baseline models","accuracies","94.5%","97.69%","fact","our best model\u2019s capacity","meaningful interdependencies","words","its classification performance","this analysis","a deeper understanding","the psychological factors","linguistic markers","suicidal thoughts","future research and intervention strategies","cnn","cnn","gpt","gpt","94.5%","97.69%"]},{"title":"Study on automatic lithology identification based on convolutional neural network and deep transfer learning","authors":["Shiliang Li","Yuelong Dong","Zhanrong Zhang","Chengyuan Lin","Huaji Liu","Yafei Wang","Youyan Bian","Feng Xiong","Guohua Zhang"],"abstract":"Automatic and fast rock classification identification is an important part of geotechnical intelligent survey system. Image based supervised deep learning analysis, especially for convolutional neural networks (CNN), has potential in optimizing lithologic classification and interpretation using borehole core images. However, the accuracy and efficiency of lithology identification models are low at present. In this work, a systematic and enormous rock data framework based on the geological rock classification system is firstly established to provide rock learning datasets. The dataset is composed of approximately 150,000 images of rock samples, which covers igneous rocks, sedimentary rocks, and metamorphic rocks. Secondly, based on CNN-deep transfer learning algorithm, an end-to-end, image-to-label rock lithology identification is established. Finally, the generalization of the proposed model and the field drilling core verification test show that the constructed intelligent rock recognition model has an ability to identify rocks quickly and accurately, and the recognition accuracy of 12 kinds of common engineering rocks is more than 95%. The proposed rock intelligent classification model provides a convenient and fast tool for field geologists and scientific researchers.","doi":"10.1007\/s42452-024-06020-y","cleaned_title":"study on automatic lithology identification based on convolutional neural network and deep transfer learning","cleaned_abstract":"automatic and fast rock classification identification is an important part of geotechnical intelligent survey system. image based supervised deep learning analysis, especially for convolutional neural networks (cnn), has potential in optimizing lithologic classification and interpretation using borehole core images. however, the accuracy and efficiency of lithology identification models are low at present. in this work, a systematic and enormous rock data framework based on the geological rock classification system is firstly established to provide rock learning datasets. the dataset is composed of approximately 150,000 images of rock samples, which covers igneous rocks, sedimentary rocks, and metamorphic rocks. secondly, based on cnn-deep transfer learning algorithm, an end-to-end, image-to-label rock lithology identification is established. finally, the generalization of the proposed model and the field drilling core verification test show that the constructed intelligent rock recognition model has an ability to identify rocks quickly and accurately, and the recognition accuracy of 12 kinds of common engineering rocks is more than 95%. the proposed rock intelligent classification model provides a convenient and fast tool for field geologists and scientific researchers.","key_phrases":["automatic and fast rock classification identification","an important part","geotechnical intelligent survey system","image","deep learning analysis","convolutional neural networks","cnn","potential","lithologic classification","interpretation","borehole core images","the accuracy","efficiency","lithology identification models","present","this work","a systematic and enormous rock data framework","the geological rock classification system","rock learning datasets","the dataset","approximately 150,000 images","rock samples","which","igneous rocks","sedimentary rocks","metamorphic rocks","cnn-deep transfer learning algorithm","an end","end","label","the generalization","the proposed model","the field","the constructed intelligent rock recognition model","an ability","rocks","the recognition accuracy","12 kinds","common engineering rocks","more than 95%","the proposed rock intelligent classification model","a convenient and fast tool","field geologists","scientific researchers","cnn","approximately 150,000","secondly","cnn","12","more than 95%"]},{"title":"Molecular design with automated quantum computing-based deep learning and optimization","authors":["Akshay Ajagekar","Fengqi You"],"abstract":"Computer-aided design of novel molecules and compounds is a challenging task that can be addressed with quantum computing (QC) owing to its notable advances in optimization and machine learning. Here, we use QC-assisted learning and optimization techniques implemented with near-term QC devices for molecular property prediction and generation tasks. The proposed probabilistic energy-based deep learning model trained in a generative manner facilitated by QC yields robust latent representations of molecules, while the proposed data-driven QC-based optimization framework performs guided navigation of the target chemical space by exploiting the structure\u2013property relationships captured by the energy-based model. We demonstrate the viability of the proposed molecular design approach by generating several molecular candidates that satisfy specific property target requirements. The proposed QC-based methods exhibit an improved predictive performance while efficiently generating novel molecules that accurately fulfill target conditions and exemplify the potential of QC for automated molecular design, thus accentuating its utility.","doi":"10.1038\/s41524-023-01099-0","cleaned_title":"molecular design with automated quantum computing-based deep learning and optimization","cleaned_abstract":"computer-aided design of novel molecules and compounds is a challenging task that can be addressed with quantum computing (qc) owing to its notable advances in optimization and machine learning. here, we use qc-assisted learning and optimization techniques implemented with near-term qc devices for molecular property prediction and generation tasks. the proposed probabilistic energy-based deep learning model trained in a generative manner facilitated by qc yields robust latent representations of molecules, while the proposed data-driven qc-based optimization framework performs guided navigation of the target chemical space by exploiting the structure\u2013property relationships captured by the energy-based model. we demonstrate the viability of the proposed molecular design approach by generating several molecular candidates that satisfy specific property target requirements. the proposed qc-based methods exhibit an improved predictive performance while efficiently generating novel molecules that accurately fulfill target conditions and exemplify the potential of qc for automated molecular design, thus accentuating its utility.","key_phrases":["computer-aided design","novel molecules","compounds","a challenging task","that","quantum computing","qc","its notable advances","optimization","machine learning","we","qc-assisted learning and optimization techniques","near-term qc devices","molecular property prediction","generation tasks","the proposed probabilistic energy-based deep learning model","a generative manner","qc yields","robust latent representations","molecules","the proposed data-driven qc-based optimization framework","guided navigation","the target chemical space","the structure","property relationships","the energy-based model","we","the viability","the proposed molecular design approach","several molecular candidates","that","specific property target requirements","the proposed qc-based methods","an improved predictive performance","novel molecules","that","target conditions","the potential","qc","automated molecular design","its utility","quantum"]},{"title":"Automated detection and recognition system for chewable food items using advanced deep learning models","authors":["Yogesh Kumar","Apeksha Koul","Kamini","Marcin Wo\u017aniak","Jana Shafi","Muhammad Fazal Ijaz"],"abstract":"Identifying and recognizing the food on the basis of its eating sounds is a challenging task, as it plays an important role in avoiding allergic foods, providing dietary preferences to people who are restricted to a particular diet, showcasing its cultural significance, etc. In this research paper, the aim is to design a novel methodology that helps to identify food items by analyzing their eating sounds using various deep learning models. To achieve this objective, a system has been proposed that extracts meaningful features from food-eating sounds with the help of signal processing techniques and deep learning models for classifying them into their respective food classes. Initially, 1200 audio files for 20 food items labeled have been collected and visualized to find relationships between the sound files of different food items. Later, to extract meaningful features, various techniques such as spectrograms, spectral rolloff, spectral bandwidth, and mel-frequency cepstral coefficients are used for the cleaning of audio files as well as to capture the unique characteristics of different food items. In the next phase, various deep learning models like GRU, LSTM, InceptionResNetV2, and the customized CNN model have been trained to learn spectral and temporal patterns in audio signals. Besides this, the models have also been hybridized i.e. Bidirectional LSTM\u2009+\u2009GRU and RNN\u2009+\u2009Bidirectional LSTM, and RNN\u2009+\u2009Bidirectional GRU to analyze their performance for the same labeled data in order to associate particular patterns of sound with their corresponding class of food item. During evaluation, the highest accuracy, precision,F1 score, and recall have been obtained by GRU with 99.28%, Bidirectional LSTM\u2009+\u2009GRU with 97.7% as well as 97.3%, and RNN\u2009+\u2009Bidirectional LSTM with 97.45%, respectively. The results of this study demonstrate that deep learning models have the potential to precisely identify foods on the basis of their sound by computing the best outcomes.","doi":"10.1038\/s41598-024-57077-z","cleaned_title":"automated detection and recognition system for chewable food items using advanced deep learning models","cleaned_abstract":"identifying and recognizing the food on the basis of its eating sounds is a challenging task, as it plays an important role in avoiding allergic foods, providing dietary preferences to people who are restricted to a particular diet, showcasing its cultural significance, etc. in this research paper, the aim is to design a novel methodology that helps to identify food items by analyzing their eating sounds using various deep learning models. to achieve this objective, a system has been proposed that extracts meaningful features from food-eating sounds with the help of signal processing techniques and deep learning models for classifying them into their respective food classes. initially, 1200 audio files for 20 food items labeled have been collected and visualized to find relationships between the sound files of different food items. later, to extract meaningful features, various techniques such as spectrograms, spectral rolloff, spectral bandwidth, and mel-frequency cepstral coefficients are used for the cleaning of audio files as well as to capture the unique characteristics of different food items. in the next phase, various deep learning models like gru, lstm, inceptionresnetv2, and the customized cnn model have been trained to learn spectral and temporal patterns in audio signals. besides this, the models have also been hybridized i.e. bidirectional lstm + gru and rnn + bidirectional lstm, and rnn + bidirectional gru to analyze their performance for the same labeled data in order to associate particular patterns of sound with their corresponding class of food item. during evaluation, the highest accuracy, precision,f1 score, and recall have been obtained by gru with 99.28%, bidirectional lstm + gru with 97.7% as well as 97.3%, and rnn + bidirectional lstm with 97.45%, respectively. the results of this study demonstrate that deep learning models have the potential to precisely identify foods on the basis of their sound by computing the best outcomes.","key_phrases":["the food","the basis","its eating sounds","a challenging task","it","an important role","allergic foods","dietary preferences","people","who","a particular diet","its cultural significance","this research paper","the aim","a novel methodology","that","food items","their eating sounds","various deep learning models","this objective","a system","meaningful features","the help","signal processing techniques","deep learning models","them","their respective food classes","1200 audio files","20 food items","relationships","the sound files","different food items","meaningful features","various techniques","spectrograms","spectral rolloff","spectral","mel-frequency cepstral coefficients","the cleaning","audio files","the unique characteristics","different food items","the next phase","various deep learning models","gru","lstm","inceptionresnetv2","the customized cnn model","spectral and temporal patterns","audio signals","this","the models","i.e. bidirectional lstm","bidirectional lstm","bidirectional gru","their performance","the same labeled data","order","particular patterns","sound","their corresponding class","food item","evaluation","the highest accuracy","precision","f1 score","recall","gru","99.28%","bidirectional lstm","97.7%","97.3%","bidirectional lstm","97.45%","the results","this study","deep learning models","the potential","foods","the basis","their sound","the best outcomes","1200","20","mel","inceptionresnetv2","cnn","99.28%","97.7%","97.3%","+ bidirectional lstm","97.45%"]},{"title":"Advances in Deep Learning Techniques for Short-term Energy Load Forecasting Applications: A Review","authors":["Radhika Chandrasekaran","Senthil Kumar Paramasivan"],"abstract":"Today, the majority of the leading power companies place a significant emphasis on forecasting the electricity load in the balance of power and administration. Meanwhile, since electricity is an integral component of every person\u2019s contemporary life, energy load forecasting is necessary to afford the energy demand required. The expansion of the electrical infrastructure is a key factor in increasing sustainable economic growth, and the planning and control of the utility power system rely on accurate load forecasting. Due to uncertainty in energy utilization, forecasting is turning into a complex task, and it makes an impact on applications that include energy scheduling and management, price forecasting, etc. The statistical methods involving time series for regression analysis and machine learning techniques have been used in energy load forecasting extensively over the last few decades to precisely predict future energy demands. However, they have some drawbacks with limited model flexibility, generalization, and overfitting. Deep learning addresses the issues of handling unstructured and unlabeled data, automatic feature learning, non-linear model flexibility, the ability to handle high-dimensional data, and simultaneous computation using GPUs efficiently. This paper investigates factors influencing energy load forecasting, then discusses the most commonly used deep learning approaches in energy load forecasting, as well as evaluation metrics to evaluate the performance of the model, followed by bio-inspired algorithms to optimize the model, and other advanced technologies for energy load forecasting. This study discusses the research findings, challenges, and opportunities in energy load forecasting.","doi":"10.1007\/s11831-024-10155-x","cleaned_title":"advances in deep learning techniques for short-term energy load forecasting applications: a review","cleaned_abstract":"today, the majority of the leading power companies place a significant emphasis on forecasting the electricity load in the balance of power and administration. meanwhile, since electricity is an integral component of every person\u2019s contemporary life, energy load forecasting is necessary to afford the energy demand required. the expansion of the electrical infrastructure is a key factor in increasing sustainable economic growth, and the planning and control of the utility power system rely on accurate load forecasting. due to uncertainty in energy utilization, forecasting is turning into a complex task, and it makes an impact on applications that include energy scheduling and management, price forecasting, etc. the statistical methods involving time series for regression analysis and machine learning techniques have been used in energy load forecasting extensively over the last few decades to precisely predict future energy demands. however, they have some drawbacks with limited model flexibility, generalization, and overfitting. deep learning addresses the issues of handling unstructured and unlabeled data, automatic feature learning, non-linear model flexibility, the ability to handle high-dimensional data, and simultaneous computation using gpus efficiently. this paper investigates factors influencing energy load forecasting, then discusses the most commonly used deep learning approaches in energy load forecasting, as well as evaluation metrics to evaluate the performance of the model, followed by bio-inspired algorithms to optimize the model, and other advanced technologies for energy load forecasting. this study discusses the research findings, challenges, and opportunities in energy load forecasting.","key_phrases":["the majority","the leading power companies","a significant emphasis","the electricity load","the balance","power","administration","electricity","an integral component","every person\u2019s contemporary life","energy load forecasting","the energy demand","the expansion","the electrical infrastructure","a key factor","sustainable economic growth","the planning","control","the utility power system","accurate load forecasting","uncertainty","energy utilization","forecasting","a complex task","it","an impact","applications","that","energy scheduling","management","price forecasting","the statistical methods","time series","regression analysis","machine learning techniques","energy load","the last few decades","future energy demands","they","some drawbacks","limited model flexibility","generalization","deep learning addresses","the issues","unstructured and unlabeled data","automatic feature learning","non-linear model flexibility","the ability","high-dimensional data","simultaneous computation","gpus","this paper investigates","energy load forecasting","the most commonly used deep learning approaches","energy load forecasting","evaluation metrics","the performance","the model","bio-inspired algorithms","the model","other advanced technologies","energy load forecasting","this study","the research findings","challenges","opportunities","energy load forecasting","today","the last few decades"]},{"title":"Optimized COVID-19 detection using sparse deep learning models from multimodal imaging data","authors":["MohammadMahdi Moradi","Alireza Hassanzadeh","Arman Haghanifar","Seok Bum Ko"],"abstract":"This paper explores the global spread of the COVID-19 virus since 2019, impacting 219 countries worldwide. Despite the absence of a definitive cure, the utilization of artificial intelligence (AI) methods for disease diagnosis has demonstrated commendable effectiveness in promptly diagnosing patients and curbing infection transmission. The study introduces a deep learning-based model tailored for COVID-19 detection, leveraging three prevalent medical imaging modalities: computed tomography (CT), chest X-ray (CXR), and Ultrasound. Various deep Transfer Learning Convolutional Neural Network-based (CNN) models have undergone assessment for each imaging modality. For each imaging modality, this study has selected the two most accurate models based on evaluation metrics such as accuracy and loss. Additionally, efforts have been made to prune unnecessary weights from these models to obtain more efficient and sparse models. By fusing these pruned models, enhanced performance has been achieved. The models have undergone rigorous training and testing using publicly available real-world medical datasets, focusing on classifying these datasets into three distinct categories: Normal, COVID-19 Pneumonia, and non-COVID-19 Pneumonia. The primary objective is to develop an optimized and swift model through strategies like Transfer Learning, Ensemble Learning, and reducing network complexity, making it easier for storage and transfer. The results of the trained network on test data exhibit promising outcomes. The accuracy of these models on the CT scan, X-ray, and ultrasound datasets stands at 99.4%, 98.9%, and 99.3%, respectively. Moreover, these models\u2019 sizes have been substantially reduced and optimized by 51.93%, 38.00%, and 69.07%, respectively. This study proposes a Computer-aided-coronavirus-detection system based on three standard medical imaging techniques. The intention is to assist radiologists in accurately and swiftly diagnosing the disease, especially during the screening process, by providing high accuracy and speed in the identification of COVID-19 cases.","doi":"10.1007\/s11042-024-18987-2","cleaned_title":"optimized covid-19 detection using sparse deep learning models from multimodal imaging data","cleaned_abstract":"this paper explores the global spread of the covid-19 virus since 2019, impacting 219 countries worldwide. despite the absence of a definitive cure, the utilization of artificial intelligence (ai) methods for disease diagnosis has demonstrated commendable effectiveness in promptly diagnosing patients and curbing infection transmission. the study introduces a deep learning-based model tailored for covid-19 detection, leveraging three prevalent medical imaging modalities: computed tomography (ct), chest x-ray (cxr), and ultrasound. various deep transfer learning convolutional neural network-based (cnn) models have undergone assessment for each imaging modality. for each imaging modality, this study has selected the two most accurate models based on evaluation metrics such as accuracy and loss. additionally, efforts have been made to prune unnecessary weights from these models to obtain more efficient and sparse models. by fusing these pruned models, enhanced performance has been achieved. the models have undergone rigorous training and testing using publicly available real-world medical datasets, focusing on classifying these datasets into three distinct categories: normal, covid-19 pneumonia, and non-covid-19 pneumonia. the primary objective is to develop an optimized and swift model through strategies like transfer learning, ensemble learning, and reducing network complexity, making it easier for storage and transfer. the results of the trained network on test data exhibit promising outcomes. the accuracy of these models on the ct scan, x-ray, and ultrasound datasets stands at 99.4%, 98.9%, and 99.3%, respectively. moreover, these models\u2019 sizes have been substantially reduced and optimized by 51.93%, 38.00%, and 69.07%, respectively. this study proposes a computer-aided-coronavirus-detection system based on three standard medical imaging techniques. the intention is to assist radiologists in accurately and swiftly diagnosing the disease, especially during the screening process, by providing high accuracy and speed in the identification of covid-19 cases.","key_phrases":["this paper","the global spread","the covid-19 virus","219 countries","the absence","a definitive cure","the utilization","artificial intelligence (ai) methods","disease diagnosis","commendable effectiveness","patients","infection transmission","the study","a deep learning-based model","covid-19 detection","three prevalent medical imaging modalities","computed tomography","ct","chest","-","ray","(cxr","ultrasound","various deep transfer","convolutional neural network-based (cnn) models","assessment","each imaging modality","each imaging modality","this study","the two most accurate models","evaluation metrics","accuracy","loss","efforts","unnecessary weights","these models","more efficient and sparse models","these pruned models","enhanced performance","the models","rigorous training","testing","publicly available real-world medical datasets","these datasets","three distinct categories","the primary objective","an optimized and swift model","strategies","transfer learning","ensemble learning","network complexity","it","storage","the results","the trained network","test data","promising outcomes","the accuracy","these models","the ct scan","x","-","ray","ultrasound datasets","99.4%","98.9%","99.3%","these models\u2019 sizes","51.93%","38.00%","69.07%","this study","a computer-aided-coronavirus-detection system","three standard medical imaging techniques","the intention","radiologists","the disease","the screening process","high accuracy","speed","the identification","covid-19 cases","covid-19","2019","219","covid-19","three","cnn","two","three","covid-19","non-covid-19","99.4%","98.9%","99.3%","51.93%","38.00%","69.07%","three","covid-19"]},{"title":"Automated classification of polyps using deep learning architectures and few-shot learning","authors":["Adrian Krenzer","Stefan Heil","Daniel Fitting","Safa Matti","Wolfram G. Zoller","Alexander Hann","Frank Puppe"],"abstract":"BackgroundColorectal cancer is a leading cause of cancer-related deaths worldwide. The best method to prevent CRC is a colonoscopy. However, not all colon polyps have the risk of becoming cancerous. Therefore, polyps are classified using different classification systems. After the classification, further treatment and procedures are based on the classification of the polyp. Nevertheless, classification is not easy. Therefore, we suggest two novel automated classifications system assisting gastroenterologists in classifying polyps based on the NICE\u00a0and Paris\u00a0classification.MethodsWe build two classification systems. One is classifying polyps based on their shape (Paris). The other classifies polyps based on their texture and surface patterns (NICE). A two-step process for the Paris\u00a0classification is introduced: First, detecting and cropping the polyp on the image, and secondly, classifying the polyp based on the cropped area with a transformer network. For the NICE\u00a0classification, we design a few-shot learning algorithm based on the Deep Metric Learning approach. The algorithm creates an embedding space for polyps, which allows classification from a few examples to account for the data scarcity of NICE annotated images in our database.ResultsFor the Paris classification, we achieve an accuracy of 89.35 %, surpassing all papers in the literature and establishing a new state-of-the-art and baseline accuracy for other publications on a public data set. For the NICE classification, we achieve a competitive accuracy of 81.13 % and demonstrate thereby the viability of the few-shot learning paradigm in polyp classification in data-scarce environments. Additionally, we show different ablations of the algorithms. Finally, we further elaborate on the explainability of the system by showing heat maps of the neural network explaining neural activations.ConclusionOverall we introduce two polyp classification systems to assist gastroenterologists. We achieve state-of-the-art performance in the Paris classification and demonstrate the viability of the few-shot learning paradigm in the NICE classification, addressing the prevalent data scarcity issues faced in medical machine learning.","doi":"10.1186\/s12880-023-01007-4","cleaned_title":"automated classification of polyps using deep learning architectures and few-shot learning","cleaned_abstract":"backgroundcolorectal cancer is a leading cause of cancer-related deaths worldwide. the best method to prevent crc is a colonoscopy. however, not all colon polyps have the risk of becoming cancerous. therefore, polyps are classified using different classification systems. after the classification, further treatment and procedures are based on the classification of the polyp. nevertheless, classification is not easy. therefore, we suggest two novel automated classifications system assisting gastroenterologists in classifying polyps based on the nice and paris classification.methodswe build two classification systems. one is classifying polyps based on their shape (paris). the other classifies polyps based on their texture and surface patterns (nice). a two-step process for the paris classification is introduced: first, detecting and cropping the polyp on the image, and secondly, classifying the polyp based on the cropped area with a transformer network. for the nice classification, we design a few-shot learning algorithm based on the deep metric learning approach. the algorithm creates an embedding space for polyps, which allows classification from a few examples to account for the data scarcity of nice annotated images in our database.resultsfor the paris classification, we achieve an accuracy of 89.35 %, surpassing all papers in the literature and establishing a new state-of-the-art and baseline accuracy for other publications on a public data set. for the nice classification, we achieve a competitive accuracy of 81.13 % and demonstrate thereby the viability of the few-shot learning paradigm in polyp classification in data-scarce environments. additionally, we show different ablations of the algorithms. finally, we further elaborate on the explainability of the system by showing heat maps of the neural network explaining neural activations.conclusionoverall we introduce two polyp classification systems to assist gastroenterologists. we achieve state-of-the-art performance in the paris classification and demonstrate the viability of the few-shot learning paradigm in the nice classification, addressing the prevalent data scarcity issues faced in medical machine learning.","key_phrases":["backgroundcolorectal cancer","a leading cause","cancer-related deaths","the best method","crc","a colonoscopy","not all colon polyps","the risk","polyps","different classification systems","the classification","further treatment","procedures","the classification","the polyp","classification","we","two novel automated classifications system","gastroenterologists","polyps","the nice and paris classification.methodswe","two classification systems","polyps","their shape","paris","the other classifies polyps","their texture and surface patterns","a two-step process","the paris classification","the polyp","the image","the polyp","the cropped area","a transformer network","the nice classification","we","a few-shot learning algorithm","the deep metric learning approach","the algorithm","an embedding space","polyps","which","classification","a few examples","the data scarcity","nice annotated images","our database.resultsfor","the paris classification","we","an accuracy","89.35 %","all papers","the literature","the-art","baseline","other publications","a public data set","the nice classification","we","a competitive accuracy","81.13 %","thereby the viability","the few-shot learning paradigm","polyp classification","data-scarce environments","we","different ablations","the algorithms","we","the explainability","the system","heat maps","the neural network","neural activations.conclusionoverall","we","two polyp classification systems","gastroenterologists","we","the-art","the paris classification","the viability","the few-shot learning paradigm","the nice classification","the prevalent data scarcity issues","medical machine learning","two","paris","two","paris","two","paris","first","secondly","paris","89.35 %","81.13 %","two","paris"]},{"title":"Deep learning prediction of survival in patients with heart failure using chest radiographs","authors":["Han Jia","Shengen Liao","Xiaomei Zhu","Wangyan Liu","Yi Xu","Rongjun Ge","Yinsu Zhu"],"abstract":"Heart failure (HF) is associated with high rates of morbidity and mortality. The value of deep learning survival prediction models using chest radiographs in patients with heart failure is currently unclear. The aim of our study\u00a0is\u00a0to develop and validate a deep learning survival prediction model using chest X-ray (DLSPCXR) in patients with HF. The study retrospectively enrolled a cohort of 353 patients with HF who underwent chest X-ray (CXR) at our institution between March 2012 and March 2017. The dataset was randomly divided into training (n\u2009=\u2009247) and validation (n\u2009=\u2009106) datasets. Univariate and multivariate Cox analysis were conducted on the training dataset to develop clinical and imaging survival prediction models. The DLSPCXR was trained and the selected clinical parameters were incorporated into DLSPCXR to establish a new model called DLSPinteg. Discrimination performance was evaluated using the time-dependent area under the receiver operating characteristic curves (TD AUC) at 1, 3, and 5-years survival. Delong\u2019s test was employed for the comparison of differences between two AUCs of different models. The risk-discrimination capability of the optimal model was evaluated by the Kaplan\u2013Meier curve. In multivariable Cox analysis, older age, higher N-terminal pro-B-type natriuretic peptide (NT-ProBNP), systolic pulmonary artery pressure (sPAP)\u2009>\u200950\u00a0mmHg, New York Heart Association (NYHA) functional class III\u2013IV and cardiothoracic ratio (CTR)\u2009\u2265\u20090.62 in CXR were independent predictors of poor prognosis in patients with HF. Based on the receiver operating characteristic (ROC) curve analysis, DLSPCXR had better performance at predicting 5-year survival than the imaging Cox model in the validation cohort (AUC: 0.757 vs. 0.561, P\u2009=\u20090.01). DLSPinteg as the optimal model outperforms the clinical Cox model (AUC: 0.826 vs. 0.633, P\u2009=\u20090.03), imaging Cox model (AUC: 0.826 vs. 0.555, P\u2009<\u20090.001), and DLSPCXR (AUC: 0.826 vs. 0.767, P\u2009=\u20090.06). Deep learning models using chest radiographs can predict survival in patients with heart failure with acceptable accuracy.","doi":"10.1007\/s10554-024-03177-w","cleaned_title":"deep learning prediction of survival in patients with heart failure using chest radiographs","cleaned_abstract":"heart failure (hf) is associated with high rates of morbidity and mortality. the value of deep learning survival prediction models using chest radiographs in patients with heart failure is currently unclear. the aim of our study is to develop and validate a deep learning survival prediction model using chest x-ray (dlspcxr) in patients with hf. the study retrospectively enrolled a cohort of 353 patients with hf who underwent chest x-ray (cxr) at our institution between march 2012 and march 2017. the dataset was randomly divided into training (n = 247) and validation (n = 106) datasets. univariate and multivariate cox analysis were conducted on the training dataset to develop clinical and imaging survival prediction models. the dlspcxr was trained and the selected clinical parameters were incorporated into dlspcxr to establish a new model called dlspinteg. discrimination performance was evaluated using the time-dependent area under the receiver operating characteristic curves (td auc) at 1, 3, and 5-years survival. delong\u2019s test was employed for the comparison of differences between two aucs of different models. the risk-discrimination capability of the optimal model was evaluated by the kaplan\u2013meier curve. in multivariable cox analysis, older age, higher n-terminal pro-b-type natriuretic peptide (nt-probnp), systolic pulmonary artery pressure (spap) > 50 mmhg, new york heart association (nyha) functional class iii\u2013iv and cardiothoracic ratio (ctr) \u2265 0.62 in cxr were independent predictors of poor prognosis in patients with hf. based on the receiver operating characteristic (roc) curve analysis, dlspcxr had better performance at predicting 5-year survival than the imaging cox model in the validation cohort (auc: 0.757 vs. 0.561, p = 0.01). dlspinteg as the optimal model outperforms the clinical cox model (auc: 0.826 vs. 0.633, p = 0.03), imaging cox model (auc: 0.826 vs. 0.555, p < 0.001), and dlspcxr (auc: 0.826 vs. 0.767, p = 0.06). deep learning models using chest radiographs can predict survival in patients with heart failure with acceptable accuracy.","key_phrases":["heart failure","hf","high rates","morbidity","mortality","the value","deep learning survival prediction models","chest radiographs","patients","heart failure","the aim","our study","a deep learning survival prediction model","chest x","-","ray","dlspcxr","patients","hf","the study","a cohort","353 patients","hf","who","chest","x","-","ray","(cxr","our institution","march","march","the dataset","training","validation","= 106) datasets","univariate and multivariate cox analysis","the training dataset","clinical and imaging survival prediction models","the dlspcxr","the selected clinical parameters","dlspcxr","a new model","discrimination performance","the time-dependent area","the receiver operating characteristic curves","td auc","5-years survival","delong\u2019s test","the comparison","differences","two aucs","different models","the risk-discrimination capability","the optimal model","the kaplan\u2013meier curve","multivariable cox analysis","higher n-terminal pro-b-type natriuretic peptide","nt-probnp","systolic pulmonary artery pressure","spap","50 mmhg","new york heart association","nyha) functional class iii","iv","cardiothoracic ratio","ctr","\u2265","cxr","independent predictors","poor prognosis","patients","hf","the receiver operating characteristic (roc) curve analysis","dlspcxr","better performance","5-year survival","the imaging cox model","the validation cohort","auc","the optimal model","the clinical cox model","auc","cox model","auc","auc","deep learning models","chest radiographs","survival","patients","heart failure","acceptable accuracy","353","march 2012 and","march 2017","247","106","1","3","5-years","delong","between two","50","new york","ctr","\u2265 0.62","roc","dlspcxr","5-year","0.757","0.561","0.01","0.826","0.633","0.03","0.826","0.555","p < 0.001","0.826","0.767","0.06"]},{"title":"Leveraging distant supervision and deep learning for twitter sentiment and emotion classification","authors":["Muhamet Kastrati","Zenun Kastrati","Ali Shariq Imran","Marenglen Biba"],"abstract":"Nowadays, various applications across industries, healthcare, and security have begun adopting automatic sentiment analysis and emotion detection in short texts, such as posts from social media. Twitter stands out as one of the most popular online social media platforms due to its easy, unique, and advanced accessibility using the API. On the other hand, supervised learning is the most widely used paradigm for tasks involving sentiment polarity and fine-grained emotion detection in short and informal texts, such as Twitter posts. However, supervised learning models are data-hungry and heavily reliant on abundant labeled data, which remains a challenge. This study aims to address this challenge by creating a large-scale real-world dataset of 17.5 million tweets. A distant supervision approach relying on emojis available in tweets is applied to label tweets corresponding to Ekman\u2019s six basic emotions. Additionally, we conducted a series of experiments using various conventional machine learning models and deep learning, including transformer-based models, on our dataset to establish baseline results. The experimental results and an extensive ablation analysis on the dataset showed that BiLSTM with FastText and an attention mechanism outperforms other models in both classification tasks, achieving an F1-score of 70.92% for sentiment classification and 54.85% for emotion detection.","doi":"10.1007\/s10844-024-00845-0","cleaned_title":"leveraging distant supervision and deep learning for twitter sentiment and emotion classification","cleaned_abstract":"nowadays, various applications across industries, healthcare, and security have begun adopting automatic sentiment analysis and emotion detection in short texts, such as posts from social media. twitter stands out as one of the most popular online social media platforms due to its easy, unique, and advanced accessibility using the api. on the other hand, supervised learning is the most widely used paradigm for tasks involving sentiment polarity and fine-grained emotion detection in short and informal texts, such as twitter posts. however, supervised learning models are data-hungry and heavily reliant on abundant labeled data, which remains a challenge. this study aims to address this challenge by creating a large-scale real-world dataset of 17.5 million tweets. a distant supervision approach relying on emojis available in tweets is applied to label tweets corresponding to ekman\u2019s six basic emotions. additionally, we conducted a series of experiments using various conventional machine learning models and deep learning, including transformer-based models, on our dataset to establish baseline results. the experimental results and an extensive ablation analysis on the dataset showed that bilstm with fasttext and an attention mechanism outperforms other models in both classification tasks, achieving an f1-score of 70.92% for sentiment classification and 54.85% for emotion detection.","key_phrases":["various applications","industries","healthcare","security","automatic sentiment analysis","emotion detection","short texts","posts","social media","twitter","the most popular online social media platforms","its easy, unique, and advanced accessibility","the api","the other hand","supervised learning","the most widely used paradigm","tasks","sentiment polarity","fine-grained emotion detection","short and informal texts","twitter posts","supervised learning models","abundant labeled data","which","a challenge","this study","this challenge","a large-scale real-world dataset","17.5 million tweets","a distant supervision approach","emojis","tweets","label tweets","ekman\u2019s six basic emotions","we","a series","experiments","various conventional machine learning models","deep learning","transformer-based models","our dataset","baseline results","the experimental results","an extensive ablation analysis","the dataset","that bilstm","fasttext","an attention mechanism","other models","both classification tasks","an f1-score","70.92%","sentiment classification","54.85%","emotion detection","17.5 million","six","70.92%","54.85%"]},{"title":"Octorotor flight control system design with stochastic optimal tuning, deep learning and differential morphing","authors":["Oguz Kose"],"abstract":"In this paper, simultaneous longitudinal and lateral flight control is investigated for an octorotor by using stochastic optimal tuning and deep learning under differential morphing. Octorotor models for differential morphing were drawn in SOLIDWORKS drawing program. Arm lengths are randomly estimated in the algorithm. Moments of inertia changing according to morphing ratios are estimated with deep neural network. In addition, the proportional\u2013integral\u2013derivative controller coefficients required for both longitudinal and lateral flight according to the morphing ratios are estimated by simultaneous perturbation stochastic approximation. Considering the design performance criteria, 49.95% improvement was achieved in the total cost. The estimation of unknown parameters by optimization method and deep learning was tested in simulations, and the octorotor successfully followed the given reference angle.","doi":"10.1007\/s40430-024-04972-1","cleaned_title":"octorotor flight control system design with stochastic optimal tuning, deep learning and differential morphing","cleaned_abstract":"in this paper, simultaneous longitudinal and lateral flight control is investigated for an octorotor by using stochastic optimal tuning and deep learning under differential morphing. octorotor models for differential morphing were drawn in solidworks drawing program. arm lengths are randomly estimated in the algorithm. moments of inertia changing according to morphing ratios are estimated with deep neural network. in addition, the proportional\u2013integral\u2013derivative controller coefficients required for both longitudinal and lateral flight according to the morphing ratios are estimated by simultaneous perturbation stochastic approximation. considering the design performance criteria, 49.95% improvement was achieved in the total cost. the estimation of unknown parameters by optimization method and deep learning was tested in simulations, and the octorotor successfully followed the given reference angle.","key_phrases":["this paper","simultaneous longitudinal and lateral flight control","an octorotor","stochastic optimal tuning","deep learning","differential morphing","octorotor models","differential morphing","solidworks drawing program","arm lengths","the algorithm","moments","inertia","morphing ratios","deep neural network","addition","the proportional\u2013integral\u2013derivative controller coefficients","both longitudinal and lateral flight","the morphing ratios","simultaneous perturbation stochastic approximation","the design performance criteria","49.95% improvement","the total cost","the estimation","unknown parameters","optimization method","deep learning","simulations","the octorotor","the given reference angle","49.95%"]},{"title":"Advances in Deep Learning Models for Resolving Medical Image Segmentation Data Scarcity Problem: A Topical Review","authors":["Ashwini Kumar Upadhyay","Ashish Kumar Bhandari"],"abstract":"Deep learning (DL) methods have recently become state-of-the-art in most automated medical image segmentation tasks. Some of the biggest challenges in this field are related to datasets. This paper aims to review the recent developments in deep learning architectures and approaches that aim to resolve dataset-related challenges faced in DL-based medical image segmentation. We have studied architectural developments in deep learning models and their recent applications in medical image segmentation tasks. Popular U-Net-based models are tested for segmentation performance comparison on a Coronavirus disease 2019 (Covid-19) lung infection Computed Tomography segmentation dataset. The comparison results prove the effectiveness of the original U-Net architecture, even in present-day medical image segmentation tasks. To overcome major dataset-related challenges such as labeled data scarcity, high annotation time and cost, distribution shifts, low-quality of images, and generalizability issues; we have studied recent developments in deep learning approaches like active learning, data augmentation, domain adaptation, and self- and semi-supervised learning, that aim to provide innovative solutions for those challenges. With rapid developments in the field, approaches like data augmentation, domain adaptation, and semi-supervised learning have become some of the hot areas of research, aiming for more efficient use of datasets, better segmentation prediction, and model generalizability.","doi":"10.1007\/s11831-023-10028-9","cleaned_title":"advances in deep learning models for resolving medical image segmentation data scarcity problem: a topical review","cleaned_abstract":"deep learning (dl) methods have recently become state-of-the-art in most automated medical image segmentation tasks. some of the biggest challenges in this field are related to datasets. this paper aims to review the recent developments in deep learning architectures and approaches that aim to resolve dataset-related challenges faced in dl-based medical image segmentation. we have studied architectural developments in deep learning models and their recent applications in medical image segmentation tasks. popular u-net-based models are tested for segmentation performance comparison on a coronavirus disease 2019 (covid-19) lung infection computed tomography segmentation dataset. the comparison results prove the effectiveness of the original u-net architecture, even in present-day medical image segmentation tasks. to overcome major dataset-related challenges such as labeled data scarcity, high annotation time and cost, distribution shifts, low-quality of images, and generalizability issues; we have studied recent developments in deep learning approaches like active learning, data augmentation, domain adaptation, and self- and semi-supervised learning, that aim to provide innovative solutions for those challenges. with rapid developments in the field, approaches like data augmentation, domain adaptation, and semi-supervised learning have become some of the hot areas of research, aiming for more efficient use of datasets, better segmentation prediction, and model generalizability.","key_phrases":["deep learning (dl) methods","state","the-art","most automated medical image segmentation tasks","some","the biggest challenges","this field","datasets","this paper","the recent developments","deep learning architectures","approaches","that","dataset-related challenges","dl-based medical image segmentation","we","architectural developments","deep learning models","their recent applications","medical image segmentation tasks","popular u-net-based models","segmentation performance comparison","a coronavirus disease","covid-19) lung infection","tomography segmentation","the comparison results","the effectiveness","the original u-net architecture","present-day medical image segmentation tasks","major dataset-related challenges","labeled data scarcity","high annotation time","cost","distribution shifts","low-quality","images","generalizability issues","we","recent developments","deep learning approaches","active learning","data augmentation","domain adaptation","self-","semi-supervised learning","that","innovative solutions","those challenges","rapid developments","the field","approaches","data augmentation","domain adaptation","semi-supervised learning","some","the hot areas","research","more efficient use","datasets","better segmentation prediction","model generalizability","2019","covid-19","present-day"]},{"title":"Advancing differential diagnosis: a comprehensive review of deep learning approaches for differentiating tuberculosis, pneumonia, and COVID-19","authors":["Kajal Kansal","Tej Bahadur Chandra","Akansha Singh"],"abstract":"In the realm of medical diagnostics, particularly in differential diagnosis, where differentiating between illnesses or ailments with comparable symptoms is essential, deep learning has gained importance. Recent developments in deep learning have demonstrated considerable promise for revolutionizing medical diagnostics by using the ability of artificial intelligence (AI) to accurately interpret radiological images. We examine the most cutting-edge deep learning techniques currently being utilized for the differential diagnosis of tuberculosis, pneumonia, and COVID-19 in this in-depth review. The study presents an in-depth critical review of several SOTA (state-of-the-art) studies used for differential diagnosis of different respiratory abnormalities like TB, Pneumonia, and COVID-19. In addition, an overview of various approaches, datasets employed in each method, various diagnosis tests, used assessment measures, and obtained performance is summarized and comprehensively compared to assist future research. We suggest a pathway for future research and development of deep learning solutions for differential diagnosis by critically analyzing the current literature and outlining the limitations and potential in this sector.","doi":"10.1007\/s11042-024-19350-1","cleaned_title":"advancing differential diagnosis: a comprehensive review of deep learning approaches for differentiating tuberculosis, pneumonia, and covid-19","cleaned_abstract":"in the realm of medical diagnostics, particularly in differential diagnosis, where differentiating between illnesses or ailments with comparable symptoms is essential, deep learning has gained importance. recent developments in deep learning have demonstrated considerable promise for revolutionizing medical diagnostics by using the ability of artificial intelligence (ai) to accurately interpret radiological images. we examine the most cutting-edge deep learning techniques currently being utilized for the differential diagnosis of tuberculosis, pneumonia, and covid-19 in this in-depth review. the study presents an in-depth critical review of several sota (state-of-the-art) studies used for differential diagnosis of different respiratory abnormalities like tb, pneumonia, and covid-19. in addition, an overview of various approaches, datasets employed in each method, various diagnosis tests, used assessment measures, and obtained performance is summarized and comprehensively compared to assist future research. we suggest a pathway for future research and development of deep learning solutions for differential diagnosis by critically analyzing the current literature and outlining the limitations and potential in this sector.","key_phrases":["the realm","medical diagnostics","differential diagnosis","illnesses","ailments","comparable symptoms","deep learning","importance","recent developments","deep learning","considerable promise","medical diagnostics","the ability","artificial intelligence","radiological images","we","the most cutting-edge deep learning techniques","the differential diagnosis","tuberculosis","pneumonia","covid-19","-depth","the study","an in-depth critical review","the-art","differential diagnosis","different respiratory abnormalities","tb","pneumonia","covid-19","addition","an overview","various approaches","datasets","each method","various diagnosis tests","assessment measures","performance","future research","we","a pathway","future research","development","deep learning solutions","differential diagnosis","the current literature","the limitations","potential","this sector","covid-19","covid-19"]},{"title":"An efficient intrusive deep reinforcement learning framework for OpenFOAM","authors":["Saeed Salehi"],"abstract":"Recent advancements in artificial intelligence and deep learning offer tremendous opportunities to tackle high-dimensional and challenging problems. Particularly, deep reinforcement learning (DRL) has been shown to be able to address optimal decision-making problems and control complex dynamical systems. DRL has received increased attention in the realm of computational fluid dynamics (CFD) due to its demonstrated ability to optimize complex flow control strategies. However, DRL algorithms often suffer from low sampling efficiency and require numerous interactions between the agent and the environment, necessitating frequent data exchanges. One significant bottleneck in coupled DRL\u2013CFD algorithms is the extensive data communication between DRL and CFD codes. Non-intrusive algorithms where the DRL agent treats the CFD environment as a black box may come with the deficiency of increased computational cost due to overhead associated with the information exchange between the two DRL and CFD modules. In this article, a TensorFlow-based intrusive DRL\u2013CFD framework is introduced where the agent model is integrated within the open-source CFD solver OpenFOAM. The integration eliminates the need for any external information exchange during DRL episodes. The framework is parallelized using the message passing interface to manage parallel environments for computationally intensive CFD cases through distributed computing. The performance and effectiveness of the framework are verified by controlling the vortex shedding behind two and three-dimensional cylinders, achieved as a result of minimizing drag and lift forces through an active flow control mechanism. The simulation results indicate that the trained controller can stabilize the flow and effectively mitigate the vortex shedding.","doi":"10.1007\/s11012-024-01830-1","cleaned_title":"an efficient intrusive deep reinforcement learning framework for openfoam","cleaned_abstract":"recent advancements in artificial intelligence and deep learning offer tremendous opportunities to tackle high-dimensional and challenging problems. particularly, deep reinforcement learning (drl) has been shown to be able to address optimal decision-making problems and control complex dynamical systems. drl has received increased attention in the realm of computational fluid dynamics (cfd) due to its demonstrated ability to optimize complex flow control strategies. however, drl algorithms often suffer from low sampling efficiency and require numerous interactions between the agent and the environment, necessitating frequent data exchanges. one significant bottleneck in coupled drl\u2013cfd algorithms is the extensive data communication between drl and cfd codes. non-intrusive algorithms where the drl agent treats the cfd environment as a black box may come with the deficiency of increased computational cost due to overhead associated with the information exchange between the two drl and cfd modules. in this article, a tensorflow-based intrusive drl\u2013cfd framework is introduced where the agent model is integrated within the open-source cfd solver openfoam. the integration eliminates the need for any external information exchange during drl episodes. the framework is parallelized using the message passing interface to manage parallel environments for computationally intensive cfd cases through distributed computing. the performance and effectiveness of the framework are verified by controlling the vortex shedding behind two and three-dimensional cylinders, achieved as a result of minimizing drag and lift forces through an active flow control mechanism. the simulation results indicate that the trained controller can stabilize the flow and effectively mitigate the vortex shedding.","key_phrases":["recent advancements","artificial intelligence","deep learning","tremendous opportunities","high-dimensional and challenging problems","deep reinforcement learning","drl","optimal decision-making problems","complex dynamical systems","drl","increased attention","the realm","computational fluid dynamics","cfd","its demonstrated ability","complex flow control strategies","drl algorithms","low sampling efficiency","numerous interactions","the agent","the environment","frequent data exchanges","one significant bottleneck","coupled drl","cfd algorithms","the extensive data communication","drl and cfd codes","non-intrusive algorithms","the drl agent","the cfd environment","a black box","the deficiency","increased computational cost","overhead","the information exchange","the two drl and cfd modules","this article","a tensorflow-based intrusive drl","cfd framework","the agent model","the open-source cfd","solver openfoam","the integration","the need","any external information exchange","drl episodes","the framework","the message","interface","parallel environments","computationally intensive cfd cases","distributed computing","the performance","effectiveness","the framework","the vortex","two and three-dimensional cylinders","a result","drag","forces","an active flow control mechanism","the simulation results","the trained controller","the flow","the vortex shedding","drl","one","drl","two","cfd modules","two"]},{"title":"Heart failure classification using deep learning to extract spatiotemporal features from ECG","authors":["Chang-Jiang Zhang","Yuan-Lu","Fu-Qin Tang","Hai-Peng Cai","Yin-Fen Qian","Chao-Wang"],"abstract":"BackgroundHeart failure is a syndrome with complex clinical manifestations. Due to increasing population aging, heart failure has become a major medical problem worldwide. In this study, we used the MIMIC-III public database to extract the temporal and spatial characteristics of electrocardiogram (ECG) signals from patients with heart failure.MethodsWe developed a NYHA functional classification model for heart failure based on a deep learning method. We introduced an integrating attention mechanism based on the CNN-LSTM-SE model, segmenting the ECG signal into 2 to 20\u2009s long segments. Ablation experiments showed that the 12\u2009s ECG signal segments could be used with the proposed deep learning model for superior classification of heart failure.ResultsThe accuracy, positive predictive value, sensitivity, and specificity of the NYHA functional classification method were 99.09, 98.9855, 99.033, and 99.649%, respectively.ConclusionsThe comprehensive performance of this model exceeds similar methods and can be used to assist in clinical medical diagnoses.","doi":"10.1186\/s12911-024-02415-4","cleaned_title":"heart failure classification using deep learning to extract spatiotemporal features from ecg","cleaned_abstract":"backgroundheart failure is a syndrome with complex clinical manifestations. due to increasing population aging, heart failure has become a major medical problem worldwide. in this study, we used the mimic-iii public database to extract the temporal and spatial characteristics of electrocardiogram (ecg) signals from patients with heart failure.methodswe developed a nyha functional classification model for heart failure based on a deep learning method. we introduced an integrating attention mechanism based on the cnn-lstm-se model, segmenting the ecg signal into 2 to 20 s long segments. ablation experiments showed that the 12 s ecg signal segments could be used with the proposed deep learning model for superior classification of heart failure.resultsthe accuracy, positive predictive value, sensitivity, and specificity of the nyha functional classification method were 99.09, 98.9855, 99.033, and 99.649%, respectively.conclusionsthe comprehensive performance of this model exceeds similar methods and can be used to assist in clinical medical diagnoses.","key_phrases":["backgroundheart failure","a syndrome","complex clinical manifestations","increasing population aging","heart failure","a major medical problem","this study","we","the mimic-iii public database","the temporal and spatial characteristics","electrocardiogram","ecg","patients","a nyha functional classification model","heart failure","a deep learning method","we","an integrating attention mechanism","the cnn-lstm-se model","the ecg signal","2 to 20 s long segments","ablation experiments","the 12 s ecg signal segments","the proposed deep learning model","superior classification","heart","failure.resultsthe accuracy","positive predictive value","sensitivity","specificity","the nyha functional classification method","99.649%","respectively.conclusionsthe comprehensive performance","this model","similar methods","clinical medical diagnoses","cnn","2","12","99.09","98.9855","99.033","99.649%"]},{"title":"Drug target prediction through deep learning functional representation of gene signatures","authors":["Hao Chen","Frederick J. King","Bin Zhou","Yu Wang","Carter J. Canedy","Joel Hayashi","Yang Zhong","Max W. Chang","Lars Pache","Julian L. Wong","Yong Jia","John Joslin","Tao Jiang","Christopher Benner","Sumit K. Chanda","Yingyao Zhou"],"abstract":"Many machine learning applications in bioinformatics currently rely on matching gene identities when analyzing input gene signatures and fail to take advantage of preexisting knowledge about gene functions. To further enable comparative analysis of OMICS datasets, including target deconvolution and mechanism of action studies, we develop an approach that represents gene signatures projected onto their biological functions, instead of their identities, similar to how the word2vec technique works in natural language processing. We develop the Functional Representation of Gene Signatures (FRoGS) approach by training a deep learning model and demonstrate that its application to the Broad Institute\u2019s L1000 datasets results in more effective compound-target predictions than models based on gene identities alone. By integrating additional pharmacological activity data sources, FRoGS significantly increases the number of high-quality compound-target predictions relative to existing approaches, many of which are supported by in silico and\/or experimental evidence. These results underscore the general utility of FRoGS in machine learning-based bioinformatics applications. Prediction networks pre-equipped with the knowledge of gene functions may help uncover new relationships among gene signatures acquired by large-scale OMICs studies on compounds, cell types, disease models, and patient cohorts.","doi":"10.1038\/s41467-024-46089-y","cleaned_title":"drug target prediction through deep learning functional representation of gene signatures","cleaned_abstract":"many machine learning applications in bioinformatics currently rely on matching gene identities when analyzing input gene signatures and fail to take advantage of preexisting knowledge about gene functions. to further enable comparative analysis of omics datasets, including target deconvolution and mechanism of action studies, we develop an approach that represents gene signatures projected onto their biological functions, instead of their identities, similar to how the word2vec technique works in natural language processing. we develop the functional representation of gene signatures (frogs) approach by training a deep learning model and demonstrate that its application to the broad institute\u2019s l1000 datasets results in more effective compound-target predictions than models based on gene identities alone. by integrating additional pharmacological activity data sources, frogs significantly increases the number of high-quality compound-target predictions relative to existing approaches, many of which are supported by in silico and\/or experimental evidence. these results underscore the general utility of frogs in machine learning-based bioinformatics applications. prediction networks pre-equipped with the knowledge of gene functions may help uncover new relationships among gene signatures acquired by large-scale omics studies on compounds, cell types, disease models, and patient cohorts.","key_phrases":["many machine learning applications","bioinformatics","gene identities","input gene signatures","advantage","knowledge","gene functions","comparative analysis","omics datasets","target deconvolution","mechanism","action studies","we","an approach","that","gene signatures","their biological functions","their identities","the word2vec technique","natural language processing","we","the functional representation","gene signatures","frogs","approach","a deep learning model","the broad institute\u2019s l1000 datasets results","more effective compound-target predictions","models","gene identities","additional pharmacological activity data sources","frogs","the number","high-quality compound-target predictions","existing approaches","which","silico","experimental evidence","these results","the general utility","frogs","machine learning-based bioinformatics applications","prediction networks","the knowledge","gene functions","new relationships","gene signatures","large-scale omics studies","compounds","cell types","disease models","patient cohorts"]},{"title":"An Outlook for Deep Learning in Ecosystem Science","authors":["George L. W. Perry","Rupert Seidl","Andr\u00e9 M. Bellv\u00e9","Werner Rammer"],"abstract":"Rapid advances in hardware and software, accompanied by public- and private-sector investment, have led to a new generation of data-driven computational tools. Recently, there has been a particular focus on deep learning\u2014a class of machine learning algorithms that uses deep neural networks to identify patterns in large and heterogeneous datasets. These developments have been accompanied by both hype and scepticism by ecologists and others. This review describes the context in which deep learning methods have emerged, the deep learning methods most relevant to ecosystem ecologists, and some of the problem domains they have been applied to. Deep learning methods have high predictive performance in a range of ecological contexts, leveraging the large data resources now available. Furthermore, deep learning tools offer ecosystem ecologists new ways to learn about ecosystem dynamics. In particular, recent advances in interpretable machine learning and in developing hybrid approaches combining deep learning and mechanistic models provide a bridge between pure prediction and causal explanation. We conclude by looking at the opportunities that deep learning tools offer ecosystem ecologists and assess the challenges in interpretability that deep learning applications pose.","doi":"10.1007\/s10021-022-00789-y","cleaned_title":"an outlook for deep learning in ecosystem science","cleaned_abstract":"rapid advances in hardware and software, accompanied by public- and private-sector investment, have led to a new generation of data-driven computational tools. recently, there has been a particular focus on deep learning\u2014a class of machine learning algorithms that uses deep neural networks to identify patterns in large and heterogeneous datasets. these developments have been accompanied by both hype and scepticism by ecologists and others. this review describes the context in which deep learning methods have emerged, the deep learning methods most relevant to ecosystem ecologists, and some of the problem domains they have been applied to. deep learning methods have high predictive performance in a range of ecological contexts, leveraging the large data resources now available. furthermore, deep learning tools offer ecosystem ecologists new ways to learn about ecosystem dynamics. in particular, recent advances in interpretable machine learning and in developing hybrid approaches combining deep learning and mechanistic models provide a bridge between pure prediction and causal explanation. we conclude by looking at the opportunities that deep learning tools offer ecosystem ecologists and assess the challenges in interpretability that deep learning applications pose.","key_phrases":["rapid advances","hardware","software","public-","private-sector investment","a new generation","data-driven computational tools","a particular focus","deep learning","a class","machine learning algorithms","that","deep neural networks","patterns","large and heterogeneous datasets","these developments","both hype","scepticism","ecologists","others","this review","the context","which","deep learning methods","the deep learning methods","ecosystem ecologists","some","the problem domains","they","deep learning methods","high predictive performance","a range","ecological contexts","the large data resources","deep learning tools","ecosystem","new ways","ecosystem dynamics","recent advances","interpretable machine learning","hybrid approaches","deep learning","mechanistic models","a bridge","pure prediction","causal explanation","we","the opportunities","deep learning tools","ecosystem ecologists","the challenges","interpretability","deep learning applications"]},{"title":"Deep learning-based automated steel surface defect segmentation: a comparative experimental study","authors":["Dejene M. Sime","Guotai Wang","Zhi Zeng","Bei Peng"],"abstract":"The use of machine vision and deep learning for intelligent industrial inspection has become increasingly important in automating the production processes. Despite the fact that machine vision approaches are used for industrial inspection, deep learning-based defect segmentation has not been widely studied. While state-of-the-art segmentation methods are often tuned for a specific purpose, extending them to unknown sets or other datasets, such as defect segmentation datasets, require further analysis. In addition, recent contributions and improvements in image segmentation methods have not been extensively investigated for defect segmentation. To address these problems, we conducted a comparative experimental study on several recent state-of-the-art deep learning-based segmentation methods for steel surface defect segmentation and evaluated them on the basis of segmentation performance, processing time, and computational complexity using two public datasets, NEU-Seg and Severstal Steel Defect Detection (SSDD). In addition we proposed and trained a hybrid transformer-based encoder with CNN-based decoder head and achieved state-of-the-art results, a Dice score of 95.22% (NEU-Seg) and 95.55% (SSDD).","doi":"10.1007\/s11042-023-15307-y","cleaned_title":"deep learning-based automated steel surface defect segmentation: a comparative experimental study","cleaned_abstract":"the use of machine vision and deep learning for intelligent industrial inspection has become increasingly important in automating the production processes. despite the fact that machine vision approaches are used for industrial inspection, deep learning-based defect segmentation has not been widely studied. while state-of-the-art segmentation methods are often tuned for a specific purpose, extending them to unknown sets or other datasets, such as defect segmentation datasets, require further analysis. in addition, recent contributions and improvements in image segmentation methods have not been extensively investigated for defect segmentation. to address these problems, we conducted a comparative experimental study on several recent state-of-the-art deep learning-based segmentation methods for steel surface defect segmentation and evaluated them on the basis of segmentation performance, processing time, and computational complexity using two public datasets, neu-seg and severstal steel defect detection (ssdd). in addition we proposed and trained a hybrid transformer-based encoder with cnn-based decoder head and achieved state-of-the-art results, a dice score of 95.22% (neu-seg) and 95.55% (ssdd).","key_phrases":["the use","machine vision","deep learning","intelligent industrial inspection","the production processes","the fact","machine vision approaches","industrial inspection","deep learning-based defect segmentation","the-art","a specific purpose","them","unknown sets","other datasets","defect segmentation datasets","further analysis","addition","recent contributions","improvements","image segmentation methods","defect segmentation","these problems","we","a comparative experimental study","the-art","steel surface defect segmentation","them","the basis","segmentation performance","processing time","computational complexity","two public datasets","neu-seg and severstal steel defect detection","ssdd","addition","we","a hybrid transformer-based encoder","cnn-based decoder head","the-art","a dice score","95.22%","neu-seg","95.55%","ssdd","two","cnn","95.22%","neu","95.55%"]},{"title":"Research on Real-time Detection of Stacked Objects Based on Deep Learning","authors":["Kaiguo Geng","Jinwei Qiao","Na Liu","Zhi Yang","Rongmin Zhang","Huiling Li"],"abstract":"Deep Learning has garnered significant attention in the field of object detection and is widely used in both industry and everyday life. The objective of this study is to investigate the applicability and targeted improvements of Deep Learning-based object detection in complex stacked environments. We analyzed the limitations in practical applications under such conditions, pinpointed the specific problems, and proposed corresponding improvement strategies. First, the study provided an overview of recent advancements in mainstream one-stage object detection algorithms, which included Anchor-based, Anchor-free, and Transformer-based architectures. The high real-time performance of these algorithms holds particular significance in practical engineering applications. It then looked at relevant technologies in three emerging research areas: Parts Recognition, Intelligent Driving, and Agricultural Picking. The study summarized existing limitations in real-time object detection within complex stacked environments and provided a comprehensive analysis of prevalent improvement strategies such as multi-level feature fusion, knowledge distillation, and hyperparameter optimization. Finally, after analyzing the performance of recent advanced one-stage algorithms on official datasets, this paper conducted empirical tests on a self-constructed industrial stacked dataset with algorithms of different structure and analyzed the experimental results in detail. A comprehensive analysis shows that Deep Learning-based object detection algorithms offer extensive applicability in complex stacked environments. In addressing diverse target sizes, overlapping occlusions, real-time constraints, and the need for lightweight solutions in complex stacked environments, each improvement strategy has its own advantages and limitations. Selecting and integrating appropriate enhancement strategies is critical and typically requires holistic evaluation, tailored to specific application contexts and challenges.","doi":"10.1007\/s10846-023-02009-8","cleaned_title":"research on real-time detection of stacked objects based on deep learning","cleaned_abstract":"deep learning has garnered significant attention in the field of object detection and is widely used in both industry and everyday life. the objective of this study is to investigate the applicability and targeted improvements of deep learning-based object detection in complex stacked environments. we analyzed the limitations in practical applications under such conditions, pinpointed the specific problems, and proposed corresponding improvement strategies. first, the study provided an overview of recent advancements in mainstream one-stage object detection algorithms, which included anchor-based, anchor-free, and transformer-based architectures. the high real-time performance of these algorithms holds particular significance in practical engineering applications. it then looked at relevant technologies in three emerging research areas: parts recognition, intelligent driving, and agricultural picking. the study summarized existing limitations in real-time object detection within complex stacked environments and provided a comprehensive analysis of prevalent improvement strategies such as multi-level feature fusion, knowledge distillation, and hyperparameter optimization. finally, after analyzing the performance of recent advanced one-stage algorithms on official datasets, this paper conducted empirical tests on a self-constructed industrial stacked dataset with algorithms of different structure and analyzed the experimental results in detail. a comprehensive analysis shows that deep learning-based object detection algorithms offer extensive applicability in complex stacked environments. in addressing diverse target sizes, overlapping occlusions, real-time constraints, and the need for lightweight solutions in complex stacked environments, each improvement strategy has its own advantages and limitations. selecting and integrating appropriate enhancement strategies is critical and typically requires holistic evaluation, tailored to specific application contexts and challenges.","key_phrases":["deep learning","significant attention","the field","object detection","both industry","everyday life","the objective","this study","the applicability","targeted improvements","deep learning-based object detection","complex stacked environments","we","the limitations","practical applications","such conditions","the specific problems","corresponding improvement strategies","the study","an overview","recent advancements","mainstream one-stage object detection algorithms","which","anchor-based, anchor-free, and transformer-based architectures","the high real-time performance","these algorithms","particular significance","practical engineering applications","it","relevant technologies","three emerging research areas","parts recognition","intelligent driving","agricultural picking","the study","existing limitations","real-time object detection","complex stacked environments","a comprehensive analysis","prevalent improvement strategies","multi-level feature fusion","knowledge distillation","hyperparameter optimization","the performance","recent advanced one-stage algorithms","official datasets","this paper","empirical tests","a self-constructed industrial stacked dataset","algorithms","different structure","the experimental results","detail","a comprehensive analysis","deep learning-based object detection algorithms","extensive applicability","complex stacked environments","diverse target sizes","occlusions","real-time constraints","lightweight solutions","complex stacked environments","each improvement strategy","its own advantages","limitations","appropriate enhancement strategies","holistic evaluation","specific application contexts","challenges","first","one","three","one"]},{"title":"Metasurface-empowered snapshot hyperspectral imaging with convex\/deep (CODE) small-data learning theory","authors":["Chia-Hsiang Lin","Shih-Hsiu Huang","Ting-Hsuan Lin","Pin Chieh Wu"],"abstract":"Hyperspectral imaging is vital for material identification but traditional systems are bulky, hindering the development of compact systems. While previous metasurfaces address volume issues, the requirements of complicated fabrication processes and significant footprint still limit their applications. This work reports a compact snapshot hyperspectral imager by incorporating the meta-optics with a small-data convex\/deep (CODE) deep learning theory. Our snapshot hyperspectral imager comprises only one single multi-wavelength metasurface chip working in the visible window (500-650 nm), significantly reducing the device area. To demonstrate the high performance of our hyperspectral imager, a 4-band multispectral imaging dataset is used as the input. Through the CODE-driven imaging system, it efficiently generates an 18-band hyperspectral data cube with high fidelity using only 18 training data points. We expect the elegant integration of multi-resonant metasurfaces with small-data learning theory will enable low-profile advanced instruments for fundamental science studies and real-world applications.","doi":"10.1038\/s41467-023-42381-5","cleaned_title":"metasurface-empowered snapshot hyperspectral imaging with convex\/deep (code) small-data learning theory","cleaned_abstract":"hyperspectral imaging is vital for material identification but traditional systems are bulky, hindering the development of compact systems. while previous metasurfaces address volume issues, the requirements of complicated fabrication processes and significant footprint still limit their applications. this work reports a compact snapshot hyperspectral imager by incorporating the meta-optics with a small-data convex\/deep (code) deep learning theory. our snapshot hyperspectral imager comprises only one single multi-wavelength metasurface chip working in the visible window (500-650 nm), significantly reducing the device area. to demonstrate the high performance of our hyperspectral imager, a 4-band multispectral imaging dataset is used as the input. through the code-driven imaging system, it efficiently generates an 18-band hyperspectral data cube with high fidelity using only 18 training data points. we expect the elegant integration of multi-resonant metasurfaces with small-data learning theory will enable low-profile advanced instruments for fundamental science studies and real-world applications.","key_phrases":["hyperspectral imaging","material identification","traditional systems","the development","compact systems","previous metasurfaces address volume issues","the requirements","complicated fabrication processes","significant footprint","their applications","this work","a compact snapshot hyperspectral","the meta-optics","a small-data convex\/deep (code) deep learning theory","only one single multi-wavelength metasurface chip","the visible window","500-650 nm","the device area","the high performance","our hyperspectral imager","a 4-band multispectral imaging dataset","the input","the code-driven imaging system","it","an 18-band hyperspectral data cube","high fidelity","only 18 training data points","we","the elegant integration","multi-resonant metasurfaces","small-data learning theory","low-profile advanced instruments","fundamental science studies","real-world applications","only one","500-650","4","18","only 18"]},{"title":"Performance evaluation of deep learning approaches for predicting mechanical fields in composites","authors":["Marwa Yacouti","Maryam Shakiba"],"abstract":"This paper presents a rigorous and critical methodology for evaluating the performance of deep learning (DL) techniques in predicting mechanical responses within the microstructural representation of composites. In the past few years, deep learning has emerged as a powerful tool and an efficient surrogate for finite element analysis in computational mechanics. This research addresses questions regarding the suitability of common error metrics for evaluating the accuracy of DL techniques in predicting full-field mechanical responses of composites. Through comparative analysis, we evaluate the performance and identify the limitations of two DL frameworks in predicting the linear von Mises stress distribution within the microstructure of the selected fiber-reinforced composite. The first DL method is based on the residual network, while the second utilizes U-Net architecture. We use several evaluation metrics, including different types of error and statistical measures and examine their suitability. Additionally, this study also investigates the influence of the size of the training and validation dataset, ranging from 50 to 2000 samples, on the predictive performance of the employed ResNet and U-Net based approaches.","doi":"10.1007\/s00366-024-01966-4","cleaned_title":"performance evaluation of deep learning approaches for predicting mechanical fields in composites","cleaned_abstract":"this paper presents a rigorous and critical methodology for evaluating the performance of deep learning (dl) techniques in predicting mechanical responses within the microstructural representation of composites. in the past few years, deep learning has emerged as a powerful tool and an efficient surrogate for finite element analysis in computational mechanics. this research addresses questions regarding the suitability of common error metrics for evaluating the accuracy of dl techniques in predicting full-field mechanical responses of composites. through comparative analysis, we evaluate the performance and identify the limitations of two dl frameworks in predicting the linear von mises stress distribution within the microstructure of the selected fiber-reinforced composite. the first dl method is based on the residual network, while the second utilizes u-net architecture. we use several evaluation metrics, including different types of error and statistical measures and examine their suitability. additionally, this study also investigates the influence of the size of the training and validation dataset, ranging from 50 to 2000 samples, on the predictive performance of the employed resnet and u-net based approaches.","key_phrases":["this paper","a rigorous and critical methodology","the performance","deep learning","(dl) techniques","mechanical responses","the microstructural representation","composites","the past few years","deep learning","a powerful tool","an efficient surrogate","finite element analysis","computational mechanics","questions","the suitability","common error metrics","the accuracy","dl techniques","full-field mechanical responses","composites","comparative analysis","we","the performance","the limitations","two dl frameworks","the linear von mises","distribution","the microstructure","the selected fiber-reinforced composite","the first dl method","the residual network","u-net architecture","we","several evaluation metrics","different types","error","statistical measures","their suitability","this study","the influence","the size","the training","validation dataset","50 to 2000 samples","the predictive performance","the employed resnet and u-net based approaches","the past few years","two","linear","first","second","50 to 2000"]},{"title":"Application of Deep Learning in the Classification of Maritime Safety Information","authors":["Pengbo Sun","Yi Zuo","Xinyu Li","Yudi Wang"],"abstract":"Maritime safety information (MSI) refers to urgent information concerning navigation warnings, weather warnings, weather forecasts, and other safety-related information for navigation. MSI is primarily disseminated through two systems: the international NAVTEX (Navigational Telex) system and the Enhanced Group Calling (EGC) system. NAVTEX receivers on ships can automatically receive MSI, enhancing safety of life and property at sea and reducing the workload for communication personnel. Due to the narrowband direct printing telegraphy (NBDP) technology used in the broadcast of MSI, the information must be concise, it makes the information difficult for navigators to interpret. To address this challenge, various machine learning solutions have been proposed. Among these, deep neural networks have shown superior performance in MSI classification. This study provides a more detailed analysis of deep neural networks applied to MSI classification. It utilizes collected MSI datasets, including thousands of navigation warnings, and compares the performances of different models based on accuracy, precision, recall, and F1-score.","doi":"10.1007\/s12626-024-00167-1","cleaned_title":"application of deep learning in the classification of maritime safety information","cleaned_abstract":"maritime safety information (msi) refers to urgent information concerning navigation warnings, weather warnings, weather forecasts, and other safety-related information for navigation. msi is primarily disseminated through two systems: the international navtex (navigational telex) system and the enhanced group calling (egc) system. navtex receivers on ships can automatically receive msi, enhancing safety of life and property at sea and reducing the workload for communication personnel. due to the narrowband direct printing telegraphy (nbdp) technology used in the broadcast of msi, the information must be concise, it makes the information difficult for navigators to interpret. to address this challenge, various machine learning solutions have been proposed. among these, deep neural networks have shown superior performance in msi classification. this study provides a more detailed analysis of deep neural networks applied to msi classification. it utilizes collected msi datasets, including thousands of navigation warnings, and compares the performances of different models based on accuracy, precision, recall, and f1-score.","key_phrases":["maritime safety information","msi","urgent information","navigation warnings","weather warnings","weather forecasts","other safety-related information","navigation","msi","two systems","the international navtex","navigational telex) system","the enhanced group","(egc) system","navtex receivers","ships","msi","safety","life","property","sea","the workload","communication personnel","the narrowband direct printing telegraphy (nbdp) technology","the broadcast","msi","the information","it","the information","navigators","this challenge","various machine learning solutions","these","deep neural networks","superior performance","msi classification","this study","a more detailed analysis","deep neural networks","msi classification","it","msi datasets","thousands","navigation warnings","the performances","different models","accuracy","precision","recall","f1-score","two","thousands"]},{"title":"Characterization of fault-karst reservoirs based on deep learning and attribute fusion","authors":["Zhipeng Gui","Junhua Zhang","Yintao Zhang","Chong Sun"],"abstract":"The identification of fault-karst reservoir is crucial for the exploration and development of fault-controlled oil and gas reservoirs. Traditional methods primarily rely on well logging and seismic attribute analysis for karst cave identification. However, these methods often lack the resolution needed to meet practical demands. Deep learning methods offer promising solutions by effectively overcoming the complex response characteristics of seismic wave fields, owing to their high learning capabilities. Therefore, this research proposes a method for fault-karst reservoir identification. Initially, a comparative analysis between the improved U-Net++ network and traditional deep convolutional networks is conducted to select appropriate training parameters for separate training of karst caves and faults. Subsequently, the trained models are applied to actual seismic data to predict karst caves and faults within the research area, followed by attribute fusion to acquire data on fault-karst reservoirs. The results indicate that: (1) The proposed method effectively identifies karst caves and faults, outperforming traditional seismic attribute and coherence methods in terms of identification accuracy, and slightly surpassing U-Net and FCN; (2) The fusion of predicted karst caves and faults yields clear delineation of the relationship between top karst caves and bottom fractures within the research area. In summary, the proposed method for fault-karst reservoirs identification and characterization provides valuable insights for the exploration and development of fault-controlled oil and gas reservoirs in the region.","doi":"10.1007\/s11600-024-01420-5","cleaned_title":"characterization of fault-karst reservoirs based on deep learning and attribute fusion","cleaned_abstract":"the identification of fault-karst reservoir is crucial for the exploration and development of fault-controlled oil and gas reservoirs. traditional methods primarily rely on well logging and seismic attribute analysis for karst cave identification. however, these methods often lack the resolution needed to meet practical demands. deep learning methods offer promising solutions by effectively overcoming the complex response characteristics of seismic wave fields, owing to their high learning capabilities. therefore, this research proposes a method for fault-karst reservoir identification. initially, a comparative analysis between the improved u-net++ network and traditional deep convolutional networks is conducted to select appropriate training parameters for separate training of karst caves and faults. subsequently, the trained models are applied to actual seismic data to predict karst caves and faults within the research area, followed by attribute fusion to acquire data on fault-karst reservoirs. the results indicate that: (1) the proposed method effectively identifies karst caves and faults, outperforming traditional seismic attribute and coherence methods in terms of identification accuracy, and slightly surpassing u-net and fcn; (2) the fusion of predicted karst caves and faults yields clear delineation of the relationship between top karst caves and bottom fractures within the research area. in summary, the proposed method for fault-karst reservoirs identification and characterization provides valuable insights for the exploration and development of fault-controlled oil and gas reservoirs in the region.","key_phrases":["the identification","fault-karst reservoir","the exploration","development","fault-controlled oil and gas reservoirs","traditional methods","well logging and seismic attribute analysis","karst cave identification","these methods","the resolution","practical demands","deep learning methods","promising solutions","the complex response characteristics","seismic wave fields","their high learning capabilities","this research","a method","fault-karst reservoir identification","a comparative analysis","the improved u-net++ network","traditional deep convolutional networks","appropriate training parameters","separate training","karst caves","faults","the trained models","actual seismic data","karst caves","faults","the research area","attribute fusion","data","fault-karst reservoirs","the results","the proposed method","karst caves","faults","traditional seismic attribute and coherence methods","terms","identification accuracy","u","-","net","fcn","(2) the fusion","predicted karst caves","faults","clear delineation","the relationship","top karst caves","bottom fractures","the research area","summary","the proposed method","fault-karst reservoirs identification","characterization","valuable insights","the exploration","development","fault-controlled oil and gas reservoirs","the region","1","2"]},{"title":"Integrated deep learning with explainable artificial intelligence for enhanced landslide management","authors":["Saeed Alqadhi","Javed Mallick","Meshel Alkahtani"],"abstract":"Landslides pose significant threats to mountainous regions, causing widespread damage to both property and human lives. This study seeks to enhance landslide prediction in the Aqabat Al-Sulbat Asir region of Saudi Arabia by integrating deep neural networks (DNNs), 1D convolutional neural networks (CNNs), and a combined DNN and CNN ensemble (DCN) with explainable artificial intelligence (XAI) techniques. These XAI techniques enhance the interpretability of these complex deep learning models, thereby facilitating better decision-making strategies. Furthermore, the DNN model is employed to incorporate game theory principles, assessing the individual impact of variables on landslide prediction. Our findings indicate high and very high landslide susceptibility zones covering 35.1\u201341.32 km2 and 15.14\u201316.2 km2, respectively. The DCN model boasts the highest area under the curve (AUC) at 0.97, followed by CNN (0.94) and DNN (0.9), showcasing DCN's superiority. XAI analysis exposes significant residuals in CNN's posterior despite its high AUC. Notably, precipitation, slope, soil texture, and line density emerge as pivotal parameters for accurate landslide prediction. Game theory results highlight line density's preeminence, trailed by topographic wetness index, curvature, and slope in landslide occurrence. By incorporating deep learning models, XAI, and game theory, this study presents a holistic approach to landslide management. This comprehensive framework equips authorities and stakeholders with valuable tools for informed decision-making in landslide-prone areas, delivering accurate predictions and insights into crucial parameters.","doi":"10.1007\/s11069-023-06260-y","cleaned_title":"integrated deep learning with explainable artificial intelligence for enhanced landslide management","cleaned_abstract":"landslides pose significant threats to mountainous regions, causing widespread damage to both property and human lives. this study seeks to enhance landslide prediction in the aqabat al-sulbat asir region of saudi arabia by integrating deep neural networks (dnns), 1d convolutional neural networks (cnns), and a combined dnn and cnn ensemble (dcn) with explainable artificial intelligence (xai) techniques. these xai techniques enhance the interpretability of these complex deep learning models, thereby facilitating better decision-making strategies. furthermore, the dnn model is employed to incorporate game theory principles, assessing the individual impact of variables on landslide prediction. our findings indicate high and very high landslide susceptibility zones covering 35.1\u201341.32 km2 and 15.14\u201316.2 km2, respectively. the dcn model boasts the highest area under the curve (auc) at 0.97, followed by cnn (0.94) and dnn (0.9), showcasing dcn's superiority. xai analysis exposes significant residuals in cnn's posterior despite its high auc. notably, precipitation, slope, soil texture, and line density emerge as pivotal parameters for accurate landslide prediction. game theory results highlight line density's preeminence, trailed by topographic wetness index, curvature, and slope in landslide occurrence. by incorporating deep learning models, xai, and game theory, this study presents a holistic approach to landslide management. this comprehensive framework equips authorities and stakeholders with valuable tools for informed decision-making in landslide-prone areas, delivering accurate predictions and insights into crucial parameters.","key_phrases":["landslides","significant threats","mountainous regions","widespread damage","both property","human lives","this study","landslide prediction","the aqabat al-sulbat asir region","saudi arabia","deep neural networks","dnns","1d convolutional neural networks","cnns","a combined dnn","dcn","explainable artificial intelligence (xai) techniques","these xai techniques","the interpretability","these complex deep learning models","better decision-making strategies","the dnn model","game theory principles","the individual impact","variables","landslide prediction","our findings","high and very high landslide susceptibility zones","35.1\u201341.32 km2","15.14\u201316.2 km2","the dcn model","the highest area","the curve","auc","cnn","dnn","dcn's superiority","xai analysis","significant residuals","cnn's posterior","its high auc","precipitation","slope","soil texture","line density","pivotal parameters","accurate landslide prediction","game theory results","line density's preeminence","topographic wetness index","curvature","slope","landslide occurrence","deep learning models","this study","a holistic approach","landslide management","this comprehensive framework","authorities","stakeholders","valuable tools","informed decision-making","landslide-prone areas","accurate predictions","insights","crucial parameters","al-sulbat","saudi arabia","1d","cnn","0.97","cnn","0.94","0.9","cnn"]},{"title":"Deep learning-based multi-parametric magnetic resonance imaging (mp-MRI) nomogram for predicting Ki-67 expression in rectal cancer","authors":["Sikai Wu","Neng Wang","Weiqun Ao","Jinwen Hu","Wenjie Xu","Guoqun Mao"],"abstract":"PurposeTo explore the value of deep learning-based multi-parametric magnetic resonance imaging (mp-MRI) nomogram in predicting the Ki-67 expression in rectal cancer.MethodsThe data of 491 patients with rectal cancer from two centers were retrospectively analyzed and divided into training, internal validation, and external validation sets. They were categorized into high- and low-expression group based on postoperative pathological Ki-67 expression. Each patient\u2019s mp-MRI data were analyzed to extract and select the most relevant features of deep learning, and a deep learning model was constructed. Independent predictive risk factors were identified and incorporated into a clinical model, and the clinical and deep learning models were combined to obtain a nomogram for the prediction of Ki-67 expression. The performance characteristics of the DL-model, clinical model, and nomogram were assessed using ROCs, calibration curve, decision curve, and clinical impact curve analysis.ResultsThe strongest deep learning features were extracted and screened from mp-MRI data. Two independent predictive factors, namely Magnetic Resonance Imaging T (mrT) staging and differentiation degree, were identified through clinical feature selection. Three models were constructed: a deep learning (DL)-model, a clinical model, and a nomogram. The AUCs of clinical model in the training, internal validation, and external validation set were 0.69, 0.78, and 0.67, respectively. The AUCs of the deep model and nomogram ranged from 0.88 to 0.98. The prediction performance of the deep learning model and nomogram was significantly better than the clinical model (P\u2009<\u20090.001).ConclusionThe nomogram based on deep learning can help clinicians accurately and conveniently predict the expression status of Ki-67 in rectal cancer.Graphical abstract","doi":"10.1007\/s00261-024-04232-9","cleaned_title":"deep learning-based multi-parametric magnetic resonance imaging (mp-mri) nomogram for predicting ki-67 expression in rectal cancer","cleaned_abstract":"purposeto explore the value of deep learning-based multi-parametric magnetic resonance imaging (mp-mri) nomogram in predicting the ki-67 expression in rectal cancer.methodsthe data of 491 patients with rectal cancer from two centers were retrospectively analyzed and divided into training, internal validation, and external validation sets. they were categorized into high- and low-expression group based on postoperative pathological ki-67 expression. each patient\u2019s mp-mri data were analyzed to extract and select the most relevant features of deep learning, and a deep learning model was constructed. independent predictive risk factors were identified and incorporated into a clinical model, and the clinical and deep learning models were combined to obtain a nomogram for the prediction of ki-67 expression. the performance characteristics of the dl-model, clinical model, and nomogram were assessed using rocs, calibration curve, decision curve, and clinical impact curve analysis.resultsthe strongest deep learning features were extracted and screened from mp-mri data. two independent predictive factors, namely magnetic resonance imaging t (mrt) staging and differentiation degree, were identified through clinical feature selection. three models were constructed: a deep learning (dl)-model, a clinical model, and a nomogram. the aucs of clinical model in the training, internal validation, and external validation set were 0.69, 0.78, and 0.67, respectively. the aucs of the deep model and nomogram ranged from 0.88 to 0.98. the prediction performance of the deep learning model and nomogram was significantly better than the clinical model (p < 0.001).conclusionthe nomogram based on deep learning can help clinicians accurately and conveniently predict the expression status of ki-67 in rectal cancer.graphical abstract","key_phrases":["purposeto","the value","deep learning-based multi-parametric magnetic resonance imaging (mp-mri) nomogram","the ki-67 expression","rectal cancer.methodsthe data","491 patients","rectal cancer","two centers","training","internal validation","external validation sets","they","high- and low-expression group","postoperative pathological ki-67 expression","each patient\u2019s mp-mri data","the most relevant features","deep learning","a deep learning model","independent predictive risk factors","a clinical model","the clinical and deep learning models","a nomogram","the prediction","ki-67 expression","the performance characteristics","the dl-model, clinical model","nomogram","rocs","calibration curve","decision curve","clinical impact","analysis.resultsthe strongest deep learning features","mp-mri data","two independent predictive factors","mrt","clinical feature selection","three models","a deep learning","dl)-model","a clinical model","a nomogram","the aucs","clinical model","the training","internal validation","external validation set","the aucs","the deep model","nomogram","the prediction performance","the deep learning model","nomogram","the clinical model","(p < 0.001).conclusionthe nomogram","deep learning","clinicians","the expression status","ki-67","rectal cancer.graphical abstract","ki-67","491","two","ki-67","ki-67","two","three","0.69","0.78","0.67","0.88","0.98","ki-67"]},{"title":"Fast high-frequency porosity characterization from computer tomography images and deep learning","authors":["Manuel R. V. Avila","Julio de C. V. Fernandes","Carlos E. M. dos Anjos","Adna G. P. Vasconcelos","Igor R. Cartucho","Felipe B. F. Guimaraes","Rodrigo Surmas","Alexandre G. Evsukoff"],"abstract":"This work presents a methodology for combining deep learning with whole-core computed tomography (CT) imaging and laboratory data for fast and high-frequency estimation of porosity logs. The proposed method trains a convolutional neural network (CNN) model to predict laboratory-scale porosities by using core-scale CT images as input. Despite the different scales, the model works well to predict the porosities measured at laboratory scales using the textures present in core-scale CT images. The main contribution of this work is the data-driven methodology for fast and high-frequency characterization of porosity at the core scale. The proposed methodology can provide a much more efficient porosity characterization than the traditional workflow based on the qualitative and subjective evaluation of experts. The method was evaluated in a well-based cross-validation process in 26 wells from 3 different Brazilian pre-salt fields, using two distinct CNN architectures of varying complexity. The results show that the models can make good predictions for new wells with acceptable margins of error. The proposed method can be easily integrated into digital rock workflows as it is based on data available for traditional petrophysical analyses and deep learning models available in open source.","doi":"10.1007\/s13137-024-00249-w","cleaned_title":"fast high-frequency porosity characterization from computer tomography images and deep learning","cleaned_abstract":"this work presents a methodology for combining deep learning with whole-core computed tomography (ct) imaging and laboratory data for fast and high-frequency estimation of porosity logs. the proposed method trains a convolutional neural network (cnn) model to predict laboratory-scale porosities by using core-scale ct images as input. despite the different scales, the model works well to predict the porosities measured at laboratory scales using the textures present in core-scale ct images. the main contribution of this work is the data-driven methodology for fast and high-frequency characterization of porosity at the core scale. the proposed methodology can provide a much more efficient porosity characterization than the traditional workflow based on the qualitative and subjective evaluation of experts. the method was evaluated in a well-based cross-validation process in 26 wells from 3 different brazilian pre-salt fields, using two distinct cnn architectures of varying complexity. the results show that the models can make good predictions for new wells with acceptable margins of error. the proposed method can be easily integrated into digital rock workflows as it is based on data available for traditional petrophysical analyses and deep learning models available in open source.","key_phrases":["this work","a methodology","deep learning","whole-core computed tomography (ct) imaging and laboratory data","fast and high-frequency estimation","porosity logs","the proposed method","a convolutional neural network (cnn) model","laboratory-scale porosities","core-scale ct images","input","the different scales","the model","the porosities","laboratory scales","the textures","core-scale ct images","the main contribution","this work","the data-driven methodology","fast and high-frequency characterization","porosity","the core scale","the proposed methodology","a much more efficient porosity characterization","the traditional workflow","the qualitative and subjective evaluation","experts","the method","a well-based cross-validation process","26 wells","3 different brazilian pre-salt fields","two distinct cnn architectures","varying complexity","the results","the models","good predictions","new wells","acceptable margins","error","the proposed method","digital rock workflows","it","data","traditional petrophysical analyses","deep learning models","open source","cnn","26","3","two","cnn"]}]