id
stringlengths
9
10
submitter
stringlengths
1
64
authors
stringlengths
4
20.7k
title
stringlengths
4
246
comments
stringlengths
1
523
journal-ref
stringlengths
4
404
doi
stringlengths
11
153
report-no
stringlengths
2
254
categories
stringlengths
5
98
license
stringclasses
9 values
orig_abstract
stringlengths
14
3.35k
versions
listlengths
1
60
update_date
stringlengths
10
10
authors_parsed
listlengths
1
1.35k
abstract
stringlengths
11
3.34k
2207.08000
Ruizhi Shao
Ruizhi Shao, Zerong Zheng, Hongwen Zhang, Jingxiang Sun, Yebin Liu
DiffuStereo: High Quality Human Reconstruction via Diffusion-based Stereo Using Sparse Cameras
Accepted by ECCV2022
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We propose DiffuStereo, a novel system using only sparse cameras (8 in this work) for high-quality 3D human reconstruction. At its core is a novel diffusion-based stereo module, which introduces diffusion models, a type of powerful generative models, into the iterative stereo matching network. To this end, we design a new diffusion kernel and additional stereo constraints to facilitate stereo matching and depth estimation in the network. We further present a multi-level stereo network architecture to handle high-resolution (up to 4k) inputs without requiring unaffordable memory footprint. Given a set of sparse-view color images of a human, the proposed multi-level diffusion-based stereo network can produce highly accurate depth maps, which are then converted into a high-quality 3D human model through an efficient multi-view fusion strategy. Overall, our method enables automatic reconstruction of human models with quality on par to high-end dense-view camera rigs, and this is achieved using a much more light-weight hardware setup. Experiments show that our method outperforms state-of-the-art methods by a large margin both qualitatively and quantitatively.
[ { "created": "Sat, 16 Jul 2022 19:08:18 GMT", "version": "v1" }, { "created": "Wed, 20 Jul 2022 08:12:00 GMT", "version": "v2" } ]
2022-07-21
[ [ "Shao", "Ruizhi", "" ], [ "Zheng", "Zerong", "" ], [ "Zhang", "Hongwen", "" ], [ "Sun", "Jingxiang", "" ], [ "Liu", "Yebin", "" ] ]
We propose DiffuStereo, a novel system using only sparse cameras (8 in this work) for high-quality 3D human reconstruction. At its core is a novel diffusion-based stereo module, which introduces diffusion models, a type of powerful generative models, into the iterative stereo matching network. To this end, we design a new diffusion kernel and additional stereo constraints to facilitate stereo matching and depth estimation in the network. We further present a multi-level stereo network architecture to handle high-resolution (up to 4k) inputs without requiring unaffordable memory footprint. Given a set of sparse-view color images of a human, the proposed multi-level diffusion-based stereo network can produce highly accurate depth maps, which are then converted into a high-quality 3D human model through an efficient multi-view fusion strategy. Overall, our method enables automatic reconstruction of human models with quality on par to high-end dense-view camera rigs, and this is achieved using a much more light-weight hardware setup. Experiments show that our method outperforms state-of-the-art methods by a large margin both qualitatively and quantitatively.
2403.04135
Yui Uehara
Yui Uehara
Unsupervised Learning of Harmonic Analysis Based on Neural HSMM with Code Quality Templates
20 pages, 5 figures, the original edition of this paper will be published in the ICNMC2024 Proceedings and this arXiv publication is a copy
null
null
null
cs.AI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
This paper presents a method of unsupervised learning of harmonic analysis based on a hidden semi-Markov model (HSMM). We introduce the chord quality templates, which specify the probability of pitch class emissions given a root note and a chord quality. Other probability distributions that comprise the HSMM are automatically learned via unsupervised learning, which has been a challenge in existing research. The results of the harmonic analysis of the proposed model were evaluated using existing labeled data. While our proposed method has yet to perform as well as existing models that used supervised learning and complex rule design, it has the advantage of not requiring expensive labeled data or rule elaboration. Furthermore, we also show how to recognize the tonic without prior knowledge, based on the transition probabilities of the Markov model.
[ { "created": "Thu, 7 Mar 2024 01:29:48 GMT", "version": "v1" } ]
2024-03-08
[ [ "Uehara", "Yui", "" ] ]
This paper presents a method of unsupervised learning of harmonic analysis based on a hidden semi-Markov model (HSMM). We introduce the chord quality templates, which specify the probability of pitch class emissions given a root note and a chord quality. Other probability distributions that comprise the HSMM are automatically learned via unsupervised learning, which has been a challenge in existing research. The results of the harmonic analysis of the proposed model were evaluated using existing labeled data. While our proposed method has yet to perform as well as existing models that used supervised learning and complex rule design, it has the advantage of not requiring expensive labeled data or rule elaboration. Furthermore, we also show how to recognize the tonic without prior knowledge, based on the transition probabilities of the Markov model.
2401.15544
Ramy Harik
Ramy Harik, Fadi El Kalach, Jad Samaha, Devon Clark, Drew Sander, Philip Samaha, Liam Burns, Ibrahim Yousif, Victor Gadow, Theodros Tarekegne, Nitol Saha
Analog and Multi-modal Manufacturing Datasets Acquired on the Future Factories Platform
11 pages, datasets for Future Factories
null
null
null
cs.LG
http://creativecommons.org/licenses/by-sa/4.0/
Two industry-grade datasets are presented in this paper that were collected at the Future Factories Lab at the University of South Carolina on December 11th and 12th of 2023. These datasets are generated by a manufacturing assembly line that utilizes industrial standards with respect to actuators, control mechanisms, and transducers. The two datasets were both generated simultaneously by operating the assembly line for 30 consecutive hours (with minor filtering) and collecting data from sensors equipped throughout the system. During operation, defects were also introduced into the assembly operation by manually removing parts needed for the final assembly. The datasets generated include a time series analog dataset and the other is a time series multi-modal dataset which includes images of the system alongside the analog data. These datasets were generated with the objective of providing tools to further the research towards enhancing intelligence in manufacturing. Real manufacturing datasets can be scarce let alone datasets with anomalies or defects. As such these datasets hope to address this gap and provide researchers with a foundation to build and train Artificial Intelligence models applicable for the manufacturing industry. Finally, these datasets are the first iteration of published data from the future Factories lab and can be further adjusted to fit more researchers needs moving forward.
[ { "created": "Sun, 28 Jan 2024 02:26:58 GMT", "version": "v1" } ]
2024-01-30
[ [ "Harik", "Ramy", "" ], [ "Kalach", "Fadi El", "" ], [ "Samaha", "Jad", "" ], [ "Clark", "Devon", "" ], [ "Sander", "Drew", "" ], [ "Samaha", "Philip", "" ], [ "Burns", "Liam", "" ], [ "Yousif", "Ibrahim", "" ], [ "Gadow", "Victor", "" ], [ "Tarekegne", "Theodros", "" ], [ "Saha", "Nitol", "" ] ]
Two industry-grade datasets are presented in this paper that were collected at the Future Factories Lab at the University of South Carolina on December 11th and 12th of 2023. These datasets are generated by a manufacturing assembly line that utilizes industrial standards with respect to actuators, control mechanisms, and transducers. The two datasets were both generated simultaneously by operating the assembly line for 30 consecutive hours (with minor filtering) and collecting data from sensors equipped throughout the system. During operation, defects were also introduced into the assembly operation by manually removing parts needed for the final assembly. The datasets generated include a time series analog dataset and the other is a time series multi-modal dataset which includes images of the system alongside the analog data. These datasets were generated with the objective of providing tools to further the research towards enhancing intelligence in manufacturing. Real manufacturing datasets can be scarce let alone datasets with anomalies or defects. As such these datasets hope to address this gap and provide researchers with a foundation to build and train Artificial Intelligence models applicable for the manufacturing industry. Finally, these datasets are the first iteration of published data from the future Factories lab and can be further adjusted to fit more researchers needs moving forward.
2107.07728
Pascal Pfeiffer
Christof Henkel, Pascal Pfeiffer and Philipp Singer
Recognizing bird species in diverse soundscapes under weak supervision
All authors contributed equally, 8 pages, 4 figures, submitted to CEUR-WS
null
null
null
cs.SD cs.LG eess.AS
http://creativecommons.org/licenses/by/4.0/
We present a robust classification approach for avian vocalization in complex and diverse soundscapes, achieving second place in the BirdCLEF2021 challenge. We illustrate how to make full use of pre-trained convolutional neural networks, by using an efficient modeling and training routine supplemented by novel augmentation methods. Thereby, we improve the generalization of weakly labeled crowd-sourced data to productive data collected by autonomous recording units. As such, we illustrate how to progress towards an accurate automated assessment of avian population which would enable global biodiversity monitoring at scale, impossible by manual annotation.
[ { "created": "Fri, 16 Jul 2021 06:54:38 GMT", "version": "v1" } ]
2021-07-19
[ [ "Henkel", "Christof", "" ], [ "Pfeiffer", "Pascal", "" ], [ "Singer", "Philipp", "" ] ]
We present a robust classification approach for avian vocalization in complex and diverse soundscapes, achieving second place in the BirdCLEF2021 challenge. We illustrate how to make full use of pre-trained convolutional neural networks, by using an efficient modeling and training routine supplemented by novel augmentation methods. Thereby, we improve the generalization of weakly labeled crowd-sourced data to productive data collected by autonomous recording units. As such, we illustrate how to progress towards an accurate automated assessment of avian population which would enable global biodiversity monitoring at scale, impossible by manual annotation.
0711.3176
Seyed Abolfazl Motahari
Abolfazl S. Motahari and Amir K. Khandani
To Decode the Interference or To Consider it as Noise
submitted to IEEE Transactions on Information Theory
null
null
null
cs.IT math.IT
null
We address single-user data transmission over a channel where the received signal incurs interference from a finite number of users (interfering users) that use single codebooks for transmitting their own messages. The receiver, however, is allowed to decode interfering users' messages. This means the signal transmitted from any interfering user is either decoded or considered as noise at the receiver side. We propose the following method to obtain an achievable rate for this channel. Assuming its own data is decoded successfully, the receiver partitions the set of interfering users into two disjoint subsets, namely the set of decodable users and the set of non-decodable users. Then the transmitter's rate is chosen such that the intended signal can be jointly decoded with the set of decodable users. To show the strength of this method, we prove that for the additive Gaussian channel with Gaussian interfering users, the Gaussian distribution is optimal and the achievable rate is the capacity of this channel. To obtain the maximum achievable rate, one needs to find the maximum decodable subset of interfering users. Due to the large number of possible choices, having efficient algorithms that find the set of decodable users with maximum cardinality is desired. To this end, we propose an algorithm that enables the receiver to accomplish this task in polynomial time.
[ { "created": "Tue, 20 Nov 2007 17:34:30 GMT", "version": "v1" } ]
2007-11-21
[ [ "Motahari", "Abolfazl S.", "" ], [ "Khandani", "Amir K.", "" ] ]
We address single-user data transmission over a channel where the received signal incurs interference from a finite number of users (interfering users) that use single codebooks for transmitting their own messages. The receiver, however, is allowed to decode interfering users' messages. This means the signal transmitted from any interfering user is either decoded or considered as noise at the receiver side. We propose the following method to obtain an achievable rate for this channel. Assuming its own data is decoded successfully, the receiver partitions the set of interfering users into two disjoint subsets, namely the set of decodable users and the set of non-decodable users. Then the transmitter's rate is chosen such that the intended signal can be jointly decoded with the set of decodable users. To show the strength of this method, we prove that for the additive Gaussian channel with Gaussian interfering users, the Gaussian distribution is optimal and the achievable rate is the capacity of this channel. To obtain the maximum achievable rate, one needs to find the maximum decodable subset of interfering users. Due to the large number of possible choices, having efficient algorithms that find the set of decodable users with maximum cardinality is desired. To this end, we propose an algorithm that enables the receiver to accomplish this task in polynomial time.
2107.13118
Qiaoyong Zhong
Jinlei Hou, Yingying Zhang, Qiaoyong Zhong, Di Xie, Shiliang Pu, Hong Zhou
Divide-and-Assemble: Learning Block-wise Memory for Unsupervised Anomaly Detection
accepted by ICCV 2021
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Reconstruction-based methods play an important role in unsupervised anomaly detection in images. Ideally, we expect a perfect reconstruction for normal samples and poor reconstruction for abnormal samples. Since the generalizability of deep neural networks is difficult to control, existing models such as autoencoder do not work well. In this work, we interpret the reconstruction of an image as a divide-and-assemble procedure. Surprisingly, by varying the granularity of division on feature maps, we are able to modulate the reconstruction capability of the model for both normal and abnormal samples. That is, finer granularity leads to better reconstruction, while coarser granularity leads to poorer reconstruction. With proper granularity, the gap between the reconstruction error of normal and abnormal samples can be maximized. The divide-and-assemble framework is implemented by embedding a novel multi-scale block-wise memory module into an autoencoder network. Besides, we introduce adversarial learning and explore the semantic latent representation of the discriminator, which improves the detection of subtle anomaly. We achieve state-of-the-art performance on the challenging MVTec AD dataset. Remarkably, we improve the vanilla autoencoder model by 10.1% in terms of the AUROC score.
[ { "created": "Wed, 28 Jul 2021 01:14:32 GMT", "version": "v1" } ]
2021-07-29
[ [ "Hou", "Jinlei", "" ], [ "Zhang", "Yingying", "" ], [ "Zhong", "Qiaoyong", "" ], [ "Xie", "Di", "" ], [ "Pu", "Shiliang", "" ], [ "Zhou", "Hong", "" ] ]
Reconstruction-based methods play an important role in unsupervised anomaly detection in images. Ideally, we expect a perfect reconstruction for normal samples and poor reconstruction for abnormal samples. Since the generalizability of deep neural networks is difficult to control, existing models such as autoencoder do not work well. In this work, we interpret the reconstruction of an image as a divide-and-assemble procedure. Surprisingly, by varying the granularity of division on feature maps, we are able to modulate the reconstruction capability of the model for both normal and abnormal samples. That is, finer granularity leads to better reconstruction, while coarser granularity leads to poorer reconstruction. With proper granularity, the gap between the reconstruction error of normal and abnormal samples can be maximized. The divide-and-assemble framework is implemented by embedding a novel multi-scale block-wise memory module into an autoencoder network. Besides, we introduce adversarial learning and explore the semantic latent representation of the discriminator, which improves the detection of subtle anomaly. We achieve state-of-the-art performance on the challenging MVTec AD dataset. Remarkably, we improve the vanilla autoencoder model by 10.1% in terms of the AUROC score.
2404.01615
Melanie McGrath
Melanie J. McGrath (CSIRO), Andreas Duenser (CSIRO), Justine Lacey (CSIRO), Cecile Paris (CSIRO)
Collaborative human-AI trust (CHAI-T): A process framework for active management of trust in human-AI collaboration
36 pages, 2 figures
null
null
null
cs.HC
http://creativecommons.org/licenses/by/4.0/
Collaborative human-AI (HAI) teaming combines the unique skills and capabilities of humans and machines in sustained teaming interactions leveraging the strengths of each. In tasks involving regular exposure to novelty and uncertainty, collaboration between adaptive, creative humans and powerful, precise artificial intelligence (AI) promises new solutions and efficiencies. User trust is essential to creating and maintaining these collaborative relationships. Established models of trust in traditional forms of AI typically recognize the contribution of three primary categories of trust antecedents: characteristics of the human user, characteristics of the technology, and environmental factors. The emergence of HAI teams, however, requires an understanding of human trust that accounts for the specificity of task contexts and goals, integrates processes of interaction, and captures how trust evolves in a teaming environment over time. Drawing on both the psychological and computer science literature, the process framework of trust in collaborative HAI teams (CHAI-T) presented in this paper adopts the tripartite structure of antecedents established by earlier models, while incorporating team processes and performance phases to capture the dynamism inherent to trust in teaming contexts. These features enable active management of trust in collaborative AI systems, with practical implications for the design and deployment of collaborative HAI teams.
[ { "created": "Tue, 2 Apr 2024 03:39:06 GMT", "version": "v1" } ]
2024-04-03
[ [ "McGrath", "Melanie J.", "", "CSIRO" ], [ "Duenser", "Andreas", "", "CSIRO" ], [ "Lacey", "Justine", "", "CSIRO" ], [ "Paris", "Cecile", "", "CSIRO" ] ]
Collaborative human-AI (HAI) teaming combines the unique skills and capabilities of humans and machines in sustained teaming interactions leveraging the strengths of each. In tasks involving regular exposure to novelty and uncertainty, collaboration between adaptive, creative humans and powerful, precise artificial intelligence (AI) promises new solutions and efficiencies. User trust is essential to creating and maintaining these collaborative relationships. Established models of trust in traditional forms of AI typically recognize the contribution of three primary categories of trust antecedents: characteristics of the human user, characteristics of the technology, and environmental factors. The emergence of HAI teams, however, requires an understanding of human trust that accounts for the specificity of task contexts and goals, integrates processes of interaction, and captures how trust evolves in a teaming environment over time. Drawing on both the psychological and computer science literature, the process framework of trust in collaborative HAI teams (CHAI-T) presented in this paper adopts the tripartite structure of antecedents established by earlier models, while incorporating team processes and performance phases to capture the dynamism inherent to trust in teaming contexts. These features enable active management of trust in collaborative AI systems, with practical implications for the design and deployment of collaborative HAI teams.
2206.12766
Md Jobair Hossain Faruk
Md Jobair Hossain Faruk, Hossain Shahriar, Maria Valero, Sweta Sneha, Sheikh I. Ahamed Mohammad Rahman
Towards Blockchain-Based Secure Data Management for Remote Patient Monitoring
null
2021 IEEE International Conference on Digital Health (ICDH)
10.1109/ICDH52753.2021.00054
null
cs.DB cs.CR
http://creativecommons.org/licenses/by/4.0/
Traditional data collection, storage and processing of Electronic Health Records (EHR) utilize centralized techniques that pose several risks of single point of failure and lean the systems to a number of internal and external data breaches that compromise their reliability and availability. Blockchain is an emerging distributed technology that can solve these issues due to its immutability and architectural nature that prevent records manipulation or alterations. In this paper, we discuss the progress and opportunities of remote patient monitoring using futuristic blockchain technologies and its two primary frameworks: Ethereum and Hyperledger Fabric. We also discuss the possible blockchain use cases in software engineering for systematic, disciplined, and quantifiable application development. The study extends by introducing a system architecture for EHR data management using Ethereum as a model. We discuss the challenges and limitations along with the initial evaluation results of the proposed system and draw future research directions in this promising area.
[ { "created": "Sun, 26 Jun 2022 02:20:38 GMT", "version": "v1" } ]
2022-06-28
[ [ "Faruk", "Md Jobair Hossain", "" ], [ "Shahriar", "Hossain", "" ], [ "Valero", "Maria", "" ], [ "Sneha", "Sweta", "" ], [ "Rahman", "Sheikh I. Ahamed Mohammad", "" ] ]
Traditional data collection, storage and processing of Electronic Health Records (EHR) utilize centralized techniques that pose several risks of single point of failure and lean the systems to a number of internal and external data breaches that compromise their reliability and availability. Blockchain is an emerging distributed technology that can solve these issues due to its immutability and architectural nature that prevent records manipulation or alterations. In this paper, we discuss the progress and opportunities of remote patient monitoring using futuristic blockchain technologies and its two primary frameworks: Ethereum and Hyperledger Fabric. We also discuss the possible blockchain use cases in software engineering for systematic, disciplined, and quantifiable application development. The study extends by introducing a system architecture for EHR data management using Ethereum as a model. We discuss the challenges and limitations along with the initial evaluation results of the proposed system and draw future research directions in this promising area.
2006.16411
Ali Hadian
Ali Hadian, Ankit Kumar, Thomas Heinis
Hands-off Model Integration in Spatial Index Structures
Proceedings of the 2nd International Workshop on Applied AI for Database Systems and Applications (2020)
null
null
null
cs.DB cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Spatial indexes are crucial for the analysis of the increasing amounts of spatial data, for example generated through IoT applications. The plethora of indexes that has been developed in recent decades has primarily been optimised for disk. With increasing amounts of memory even on commodity machines, however, moving them to main memory is an option. Doing so opens up the opportunity to use additional optimizations that are only amenable to main memory. In this paper we thus explore the opportunity to use light-weight machine learning models to accelerate queries on spatial indexes. We do so by exploring the potential of using interpolation and similar techniques on the R-tree, arguably the most broadly used spatial index. As we show in our experimental analysis, the query execution time can be reduced by up to 60% while simultaneously shrinking the index's memory footprint by over 90%
[ { "created": "Mon, 29 Jun 2020 22:05:28 GMT", "version": "v1" }, { "created": "Sun, 9 Aug 2020 19:43:38 GMT", "version": "v2" } ]
2020-08-11
[ [ "Hadian", "Ali", "" ], [ "Kumar", "Ankit", "" ], [ "Heinis", "Thomas", "" ] ]
Spatial indexes are crucial for the analysis of the increasing amounts of spatial data, for example generated through IoT applications. The plethora of indexes that has been developed in recent decades has primarily been optimised for disk. With increasing amounts of memory even on commodity machines, however, moving them to main memory is an option. Doing so opens up the opportunity to use additional optimizations that are only amenable to main memory. In this paper we thus explore the opportunity to use light-weight machine learning models to accelerate queries on spatial indexes. We do so by exploring the potential of using interpolation and similar techniques on the R-tree, arguably the most broadly used spatial index. As we show in our experimental analysis, the query execution time can be reduced by up to 60% while simultaneously shrinking the index's memory footprint by over 90%
1811.02068
Yuqi Zhou
Yuqi Zhou, Jorge Cisneros-Saldana, Le Xie
False Analog Data Injection Attack Towards Topology Errors: Formulation and Feasibility Analysis
5 pages, 7 figures, Proc. of 2018 IEEE Power and Energy Society General Meeting
null
10.1109/PESGM.2018.8586585
null
cs.SY math.NA math.OC
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In this paper, we propose a class of false analog data injection attack that can misguide the system as if topology errors had occurred. By utilizing the measurement redundancy with respect to the state variables, the adversary who knows the system configuration is shown to be capable of computing the corresponding measurement value with the intentionally misguided topology. The attack is designed such that the state as well as residue distribution after state estimation will converge to those in the system with a topology error. It is shown that the attack can be launched even if the attacker is constrained to some specific meters. The attack is detrimental to the system since manipulation of analog data will lead to a forged digital topology status, and the state after the error is identified and modified will be significantly biased with the intended wrong topology. The feasibility of the proposed attack is demonstrated with an IEEE 14-bus system.
[ { "created": "Mon, 5 Nov 2018 22:33:46 GMT", "version": "v1" }, { "created": "Thu, 8 Nov 2018 00:05:45 GMT", "version": "v2" } ]
2019-07-11
[ [ "Zhou", "Yuqi", "" ], [ "Cisneros-Saldana", "Jorge", "" ], [ "Xie", "Le", "" ] ]
In this paper, we propose a class of false analog data injection attack that can misguide the system as if topology errors had occurred. By utilizing the measurement redundancy with respect to the state variables, the adversary who knows the system configuration is shown to be capable of computing the corresponding measurement value with the intentionally misguided topology. The attack is designed such that the state as well as residue distribution after state estimation will converge to those in the system with a topology error. It is shown that the attack can be launched even if the attacker is constrained to some specific meters. The attack is detrimental to the system since manipulation of analog data will lead to a forged digital topology status, and the state after the error is identified and modified will be significantly biased with the intended wrong topology. The feasibility of the proposed attack is demonstrated with an IEEE 14-bus system.
2402.05680
Antti Kuusisto
Reijo Jaakkola, Tomi Janhunen, Antti Kuusisto, Masood Feyzbakhsh Rankooh, Miikka Vilander
Interpretable classifiers for tabular data via discretization and feature selection
Changes in relation to version 1: more thorough and detailed experiments, general corrections and refinements
null
null
null
cs.LG cs.AI cs.LO
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We introduce a method for computing immediately human interpretable yet accurate classifiers from tabular data. The classifiers obtained are short Boolean formulas, computed via first discretizing the original data and then using feature selection coupled with a very fast algorithm for producing the best possible Boolean classifier for the setting. We demonstrate the approach via 13 experiments, obtaining results with accuracies comparable to ones obtained via random forests, XGBoost, and existing results for the same datasets in the literature. In most cases, the accuracy of our method is in fact similar to that of the reference methods, even though the main objective of our study is the immediate interpretability of our classifiers. We also prove a new result on the probability that the classifier we obtain from real-life data corresponds to the ideally best classifier with respect to the background distribution the data comes from.
[ { "created": "Thu, 8 Feb 2024 13:58:16 GMT", "version": "v1" }, { "created": "Thu, 30 May 2024 14:12:54 GMT", "version": "v2" } ]
2024-05-31
[ [ "Jaakkola", "Reijo", "" ], [ "Janhunen", "Tomi", "" ], [ "Kuusisto", "Antti", "" ], [ "Rankooh", "Masood Feyzbakhsh", "" ], [ "Vilander", "Miikka", "" ] ]
We introduce a method for computing immediately human interpretable yet accurate classifiers from tabular data. The classifiers obtained are short Boolean formulas, computed via first discretizing the original data and then using feature selection coupled with a very fast algorithm for producing the best possible Boolean classifier for the setting. We demonstrate the approach via 13 experiments, obtaining results with accuracies comparable to ones obtained via random forests, XGBoost, and existing results for the same datasets in the literature. In most cases, the accuracy of our method is in fact similar to that of the reference methods, even though the main objective of our study is the immediate interpretability of our classifiers. We also prove a new result on the probability that the classifier we obtain from real-life data corresponds to the ideally best classifier with respect to the background distribution the data comes from.
1408.4443
Daphney-Stavroula Zois
Daphney-Stavroula Zois, Urbashi Mitra
Controlled Sensing: A Myopic Fisher Information Sensor Selection Algorithm
6 pages, 3 figures, accepted in Globecom 2014
null
null
null
cs.SY
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
This paper considers the problem of state tracking with observation control for a particular class of dynamical systems. The system state evolution is described by a discrete-time, finite-state Markov chain, while the measurement process is characterized by a controlled multi-variate Gaussian observation model. The computational complexity of the optimal control strategy proposed in our prior work proves to be prohibitive. A suboptimal, lower complexity algorithm based on the Fisher information measure is proposed. Toward this end, the preceding measure is generalized to account for multi-valued discrete parameters and control inputs. A closed-form formula for our system model is also derived. Numerical simulations are provided for a physical activity tracking application showing the near-optimal performance of the proposed algorithm.
[ { "created": "Tue, 19 Aug 2014 19:50:26 GMT", "version": "v1" } ]
2014-08-20
[ [ "Zois", "Daphney-Stavroula", "" ], [ "Mitra", "Urbashi", "" ] ]
This paper considers the problem of state tracking with observation control for a particular class of dynamical systems. The system state evolution is described by a discrete-time, finite-state Markov chain, while the measurement process is characterized by a controlled multi-variate Gaussian observation model. The computational complexity of the optimal control strategy proposed in our prior work proves to be prohibitive. A suboptimal, lower complexity algorithm based on the Fisher information measure is proposed. Toward this end, the preceding measure is generalized to account for multi-valued discrete parameters and control inputs. A closed-form formula for our system model is also derived. Numerical simulations are provided for a physical activity tracking application showing the near-optimal performance of the proposed algorithm.
2008.11906
Oliver Biggar
Oliver Biggar (1), Mohammad Zamani (1), Iman Shames (2) ((1) Defence Science and Technology Group, Australia, (2) The Australian National University, Australia)
A principled analysis of Behavior Trees and their generalisations
13 pages, 11 figures. The content of the previous version is now split between this and arXiv:2104.07919, which have both been significantly updated
null
null
null
cs.AI cs.RO
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
As complex autonomous robotic systems become more widespread, the need for transparent and reusable Artificial Intelligence (AI) designs becomes more apparent. In this paper we analyse how the principles behind Behavior Trees (BTs), an increasingly popular tree-structured control architecture, are applicable to these goals. Using structured programming as a guide, we analyse the BT principles of reactiveness and modularity in a formal framework of action selection. Proceeding from these principles, we review a number of challenging use cases of BTs in the literature, and show that reasoning via these principles leads to compatible solutions. Extending these arguments, we introduce a new class of control architectures we call generalised BTs or $k$-BTs and show how they can extend the applicability of BTs to some of the aforementioned challenging BT use cases while preserving the BT principles.
[ { "created": "Thu, 27 Aug 2020 04:09:31 GMT", "version": "v1" }, { "created": "Tue, 25 May 2021 05:42:14 GMT", "version": "v2" } ]
2021-05-26
[ [ "Biggar", "Oliver", "" ], [ "Zamani", "Mohammad", "" ], [ "Shames", "Iman", "" ] ]
As complex autonomous robotic systems become more widespread, the need for transparent and reusable Artificial Intelligence (AI) designs becomes more apparent. In this paper we analyse how the principles behind Behavior Trees (BTs), an increasingly popular tree-structured control architecture, are applicable to these goals. Using structured programming as a guide, we analyse the BT principles of reactiveness and modularity in a formal framework of action selection. Proceeding from these principles, we review a number of challenging use cases of BTs in the literature, and show that reasoning via these principles leads to compatible solutions. Extending these arguments, we introduce a new class of control architectures we call generalised BTs or $k$-BTs and show how they can extend the applicability of BTs to some of the aforementioned challenging BT use cases while preserving the BT principles.
2108.09836
Jan Kaiser
Jan Kaiser and Supriyo Datta
Probabilistic computing with p-bits
null
Appl. Phys. Lett. 119, 150503 (2021)
10.1063/5.0067927
null
cs.ET cond-mat.dis-nn quant-ph
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Digital computers store information in the form of bits that can take on one of two values 0 and 1, while quantum computers are based on qubits that are described by a complex wavefunction, whose squared magnitude gives the probability of measuring either 0 or 1. Here, we make the case for a probabilistic computer based on p-bits, which take on values 0 and 1 with controlled probabilities and can be implemented with specialized compact energy-efficient hardware. We propose a generic architecture for such p-computers and emulate systems with thousands of p-bits to show that they can significantly accelerate randomized algorithms used in a wide variety of applications including but not limited to Bayesian networks, optimization, Ising models, and quantum Monte Carlo.
[ { "created": "Sun, 22 Aug 2021 20:50:01 GMT", "version": "v1" }, { "created": "Tue, 12 Oct 2021 21:15:49 GMT", "version": "v2" } ]
2021-10-14
[ [ "Kaiser", "Jan", "" ], [ "Datta", "Supriyo", "" ] ]
Digital computers store information in the form of bits that can take on one of two values 0 and 1, while quantum computers are based on qubits that are described by a complex wavefunction, whose squared magnitude gives the probability of measuring either 0 or 1. Here, we make the case for a probabilistic computer based on p-bits, which take on values 0 and 1 with controlled probabilities and can be implemented with specialized compact energy-efficient hardware. We propose a generic architecture for such p-computers and emulate systems with thousands of p-bits to show that they can significantly accelerate randomized algorithms used in a wide variety of applications including but not limited to Bayesian networks, optimization, Ising models, and quantum Monte Carlo.
1901.04670
Xuefeng Peng
Xuefeng Peng, Yi Ding, David Wihl, Omer Gottesman, Matthieu Komorowski, Li-wei H. Lehman, Andrew Ross, Aldo Faisal, Finale Doshi-Velez
Improving Sepsis Treatment Strategies by Combining Deep and Kernel-Based Reinforcement Learning
AMIA 2018 Annual Symposium
null
null
null
cs.LG cs.AI stat.ML
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Sepsis is the leading cause of mortality in the ICU. It is challenging to manage because individual patients respond differently to treatment. Thus, tailoring treatment to the individual patient is essential for the best outcomes. In this paper, we take steps toward this goal by applying a mixture-of-experts framework to personalize sepsis treatment. The mixture model selectively alternates between neighbor-based (kernel) and deep reinforcement learning (DRL) experts depending on patient's current history. On a large retrospective cohort, this mixture-based approach outperforms physician, kernel only, and DRL-only experts.
[ { "created": "Tue, 15 Jan 2019 05:40:27 GMT", "version": "v1" } ]
2019-01-16
[ [ "Peng", "Xuefeng", "" ], [ "Ding", "Yi", "" ], [ "Wihl", "David", "" ], [ "Gottesman", "Omer", "" ], [ "Komorowski", "Matthieu", "" ], [ "Lehman", "Li-wei H.", "" ], [ "Ross", "Andrew", "" ], [ "Faisal", "Aldo", "" ], [ "Doshi-Velez", "Finale", "" ] ]
Sepsis is the leading cause of mortality in the ICU. It is challenging to manage because individual patients respond differently to treatment. Thus, tailoring treatment to the individual patient is essential for the best outcomes. In this paper, we take steps toward this goal by applying a mixture-of-experts framework to personalize sepsis treatment. The mixture model selectively alternates between neighbor-based (kernel) and deep reinforcement learning (DRL) experts depending on patient's current history. On a large retrospective cohort, this mixture-based approach outperforms physician, kernel only, and DRL-only experts.
1704.04293
Mohammad Amin
Mohammad Amin and Marta Molinas
Model Predictive Control of Voltage Source Converter in a HVDC System
null
null
null
null
cs.SY
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Model Predictive Control (MPC) method is a class of advanced control techniques most widely applied in industry. The major advantages of the MPC are its straightforward procedure which can be applied for both linear and nonlinear system. This paper proposes the use of MPC for voltage source converter (VSC) in a high voltage direct current (HVDC) system. A MPC controller is modeled based on the state-space model of a single VSC-HVDC station including the dynamics of the main ac grid. A full scale nonlinear switching model of point-to-point connected VSC-based HVDC system is developed in Matlab/Simulink association with SimPower system to demonstrate the application of the proposed controller.
[ { "created": "Thu, 13 Apr 2017 22:35:54 GMT", "version": "v1" } ]
2017-04-17
[ [ "Amin", "Mohammad", "" ], [ "Molinas", "Marta", "" ] ]
Model Predictive Control (MPC) method is a class of advanced control techniques most widely applied in industry. The major advantages of the MPC are its straightforward procedure which can be applied for both linear and nonlinear system. This paper proposes the use of MPC for voltage source converter (VSC) in a high voltage direct current (HVDC) system. A MPC controller is modeled based on the state-space model of a single VSC-HVDC station including the dynamics of the main ac grid. A full scale nonlinear switching model of point-to-point connected VSC-based HVDC system is developed in Matlab/Simulink association with SimPower system to demonstrate the application of the proposed controller.
1909.05964
Ashish Tiwari
Sumit Gulwani and Kunal Pathak and Arjun Radhakrishna and Ashish Tiwari and Abhishek Udupa
Quantitative Programming by Examples
null
null
null
null
cs.PL cs.AI cs.SE
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Programming-by-Example (PBE) systems synthesize an intended program in some (relatively constrained) domain-specific language from a small number of input-output examples provided by the user. In this paper, we motivate and define the problem of quantitative PBE (qPBE) that relates to synthesizing an intended program over an underlying (real world) programming language that also minimizes a given quantitative cost function. We present a modular approach for solving qPBE that consists of three phases: intent disambiguation, global search, and local search. On two concrete objectives, namely program performance and size, our qPBE procedure achieves $1.53 X$ and $1.26 X$ improvement respectively over the baseline FlashFill PBE system, averaged over $701$ benchmarks. Our detailed experiments validate the design of our procedure and show the value of combining global and local search for qPBE.
[ { "created": "Thu, 12 Sep 2019 21:55:00 GMT", "version": "v1" } ]
2019-09-16
[ [ "Gulwani", "Sumit", "" ], [ "Pathak", "Kunal", "" ], [ "Radhakrishna", "Arjun", "" ], [ "Tiwari", "Ashish", "" ], [ "Udupa", "Abhishek", "" ] ]
Programming-by-Example (PBE) systems synthesize an intended program in some (relatively constrained) domain-specific language from a small number of input-output examples provided by the user. In this paper, we motivate and define the problem of quantitative PBE (qPBE) that relates to synthesizing an intended program over an underlying (real world) programming language that also minimizes a given quantitative cost function. We present a modular approach for solving qPBE that consists of three phases: intent disambiguation, global search, and local search. On two concrete objectives, namely program performance and size, our qPBE procedure achieves $1.53 X$ and $1.26 X$ improvement respectively over the baseline FlashFill PBE system, averaged over $701$ benchmarks. Our detailed experiments validate the design of our procedure and show the value of combining global and local search for qPBE.
2401.08868
Manuel Tran
Manuel Tran and Amal Lahiani and Yashin Dicente Cid and Melanie Boxberg and Peter Lienemann and Christian Matek and Sophia J. Wagner and Fabian J. Theis and Eldad Klaiman and Tingying Peng
B-Cos Aligned Transformers Learn Human-Interpretable Features
Accepted at MICCAI 2023 (oral). Camera-ready available at https://doi.org/10.1007/978-3-031-43993-3_50
null
null
null
cs.CV
http://creativecommons.org/licenses/by/4.0/
Vision Transformers (ViTs) and Swin Transformers (Swin) are currently state-of-the-art in computational pathology. However, domain experts are still reluctant to use these models due to their lack of interpretability. This is not surprising, as critical decisions need to be transparent and understandable. The most common approach to understanding transformers is to visualize their attention. However, attention maps of ViTs are often fragmented, leading to unsatisfactory explanations. Here, we introduce a novel architecture called the B-cos Vision Transformer (BvT) that is designed to be more interpretable. It replaces all linear transformations with the B-cos transform to promote weight-input alignment. In a blinded study, medical experts clearly ranked BvTs above ViTs, suggesting that our network is better at capturing biomedically relevant structures. This is also true for the B-cos Swin Transformer (Bwin). Compared to the Swin Transformer, it even improves the F1-score by up to 4.7% on two public datasets.
[ { "created": "Tue, 16 Jan 2024 22:46:29 GMT", "version": "v1" }, { "created": "Thu, 18 Jan 2024 07:14:00 GMT", "version": "v2" } ]
2024-01-19
[ [ "Tran", "Manuel", "" ], [ "Lahiani", "Amal", "" ], [ "Cid", "Yashin Dicente", "" ], [ "Boxberg", "Melanie", "" ], [ "Lienemann", "Peter", "" ], [ "Matek", "Christian", "" ], [ "Wagner", "Sophia J.", "" ], [ "Theis", "Fabian J.", "" ], [ "Klaiman", "Eldad", "" ], [ "Peng", "Tingying", "" ] ]
Vision Transformers (ViTs) and Swin Transformers (Swin) are currently state-of-the-art in computational pathology. However, domain experts are still reluctant to use these models due to their lack of interpretability. This is not surprising, as critical decisions need to be transparent and understandable. The most common approach to understanding transformers is to visualize their attention. However, attention maps of ViTs are often fragmented, leading to unsatisfactory explanations. Here, we introduce a novel architecture called the B-cos Vision Transformer (BvT) that is designed to be more interpretable. It replaces all linear transformations with the B-cos transform to promote weight-input alignment. In a blinded study, medical experts clearly ranked BvTs above ViTs, suggesting that our network is better at capturing biomedically relevant structures. This is also true for the B-cos Swin Transformer (Bwin). Compared to the Swin Transformer, it even improves the F1-score by up to 4.7% on two public datasets.
1012.0634
Radwa El Shawi
Radwa El Shawi, Joachim Gudmundsson, and Christos Levcopoulos
Quickest Path Queries on Transportation Network
16 pages, 7 figures
null
null
null
cs.CG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
This paper considers the problem of finding a quickest path between two points in the Euclidean plane in the presence of a transportation network. A transportation network consists of a planar network where each road (edge) has an individual speed. A traveller may enter and exit the network at any point on the roads. Along any road the traveller moves with a fixed speed depending on the road, and outside the network the traveller moves at unit speed in any direction. We give an exact algorithm for the basic version of the problem: given a transportation network of total complexity n in the Euclidean plane, a source point s and a destination point t, and the quickest path between s and t. We also show how the transportation network can be preprocessed in time O(n^2 log n) into a data structure of size O(n^2) such that (1 + \epsilon)-approximate cheapest path cost queries between any two points in the plane can be answered in time O(1\epsilon^4 log n).
[ { "created": "Fri, 3 Dec 2010 04:03:46 GMT", "version": "v1" }, { "created": "Tue, 11 Jan 2011 10:40:59 GMT", "version": "v2" }, { "created": "Sun, 29 May 2011 15:13:53 GMT", "version": "v3" } ]
2015-03-17
[ [ "Shawi", "Radwa El", "" ], [ "Gudmundsson", "Joachim", "" ], [ "Levcopoulos", "Christos", "" ] ]
This paper considers the problem of finding a quickest path between two points in the Euclidean plane in the presence of a transportation network. A transportation network consists of a planar network where each road (edge) has an individual speed. A traveller may enter and exit the network at any point on the roads. Along any road the traveller moves with a fixed speed depending on the road, and outside the network the traveller moves at unit speed in any direction. We give an exact algorithm for the basic version of the problem: given a transportation network of total complexity n in the Euclidean plane, a source point s and a destination point t, and the quickest path between s and t. We also show how the transportation network can be preprocessed in time O(n^2 log n) into a data structure of size O(n^2) such that (1 + \epsilon)-approximate cheapest path cost queries between any two points in the plane can be answered in time O(1\epsilon^4 log n).
1203.4063
Jesper Nederlof
Petteri Kaski, Mikko Koivisto, Jesper Nederlof
Homomorphic Hashing for Sparse Coefficient Extraction
null
null
null
null
cs.CC
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We study classes of Dynamic Programming (DP) algorithms which, due to their algebraic definitions, are closely related to coefficient extraction methods. DP algorithms can easily be modified to exploit sparseness in the DP table through memorization. Coefficient extraction techniques on the other hand are both space-efficient and parallelisable, but no tools have been available to exploit sparseness. We investigate the systematic use of homomorphic hash functions to combine the best of these methods and obtain improved space-efficient algorithms for problems including LINEAR SAT, SET PARTITION, and SUBSET SUM. Our algorithms run in time proportional to the number of nonzero entries of the last segment of the DP table, which presents a strict improvement over sparse DP. The last property also gives an improved algorithm for CNF SAT with sparse projections.
[ { "created": "Mon, 19 Mar 2012 09:30:28 GMT", "version": "v1" } ]
2012-03-20
[ [ "Kaski", "Petteri", "" ], [ "Koivisto", "Mikko", "" ], [ "Nederlof", "Jesper", "" ] ]
We study classes of Dynamic Programming (DP) algorithms which, due to their algebraic definitions, are closely related to coefficient extraction methods. DP algorithms can easily be modified to exploit sparseness in the DP table through memorization. Coefficient extraction techniques on the other hand are both space-efficient and parallelisable, but no tools have been available to exploit sparseness. We investigate the systematic use of homomorphic hash functions to combine the best of these methods and obtain improved space-efficient algorithms for problems including LINEAR SAT, SET PARTITION, and SUBSET SUM. Our algorithms run in time proportional to the number of nonzero entries of the last segment of the DP table, which presents a strict improvement over sparse DP. The last property also gives an improved algorithm for CNF SAT with sparse projections.
2105.05560
Yuanjie Li
Yuanjie Li, Hewu Li, Lixin Liu, Wei Liu, Jiayi Liu, Jianping Wu, Qian Wu, Jun Liu, Zeqi Lai, Guojie Fan
Fractal Rosette: A Stable Space-Ground Network Structure in Mega-Constellation
null
null
null
null
cs.NI
http://creativecommons.org/licenses/by-nc-nd/4.0/
We present F-Rosette, a stable space-ground network structure for low-earth orbit (LEO) satellite mega-constellations at scale. Due to the dynamic many-to-many space-ground mapping in high mobility, existing LEO mega-constellations with IP protocol stack suffer from frequent user IP address changes (every 133~510s per user) and network routing re-convergence (<20% network usability). To provably stabilize the space-ground network under high mobility and many-to-many dynamics, F-Rosette adopts a recursive structure over the Rosette constellation, derives a hierarchical and time-invariant network address space from a new geographical coordinate system, and ensures efficient and stable routing via geographical-to-topological routing embedding without re-convergence. Our hardware-in-the-loop, trace-driven emulations validate F-Rosette's stability, near-optimal routing (<1.4% additional delays), and marginal overhead (<1% CPU, <2MB memory) for resource-constrained satellites.
[ { "created": "Wed, 12 May 2021 10:18:57 GMT", "version": "v1" } ]
2021-05-13
[ [ "Li", "Yuanjie", "" ], [ "Li", "Hewu", "" ], [ "Liu", "Lixin", "" ], [ "Liu", "Wei", "" ], [ "Liu", "Jiayi", "" ], [ "Wu", "Jianping", "" ], [ "Wu", "Qian", "" ], [ "Liu", "Jun", "" ], [ "Lai", "Zeqi", "" ], [ "Fan", "Guojie", "" ] ]
We present F-Rosette, a stable space-ground network structure for low-earth orbit (LEO) satellite mega-constellations at scale. Due to the dynamic many-to-many space-ground mapping in high mobility, existing LEO mega-constellations with IP protocol stack suffer from frequent user IP address changes (every 133~510s per user) and network routing re-convergence (<20% network usability). To provably stabilize the space-ground network under high mobility and many-to-many dynamics, F-Rosette adopts a recursive structure over the Rosette constellation, derives a hierarchical and time-invariant network address space from a new geographical coordinate system, and ensures efficient and stable routing via geographical-to-topological routing embedding without re-convergence. Our hardware-in-the-loop, trace-driven emulations validate F-Rosette's stability, near-optimal routing (<1.4% additional delays), and marginal overhead (<1% CPU, <2MB memory) for resource-constrained satellites.
2107.12190
Lalli Myllyaho
Lalli Myllyaho, Mikko Raatikainen, Tomi M\"annist\"o, Tommi Mikkonen and Jukka K. Nurminen
Systematic Literature Review of Validation Methods for AI Systems
25 pages, 6 figures, 12 tables. The manuscript has been accepted to the Journal of Systems and Software
null
10.1016/j.jss.2021.111050
null
cs.SE
http://creativecommons.org/licenses/by-nc-nd/4.0/
Context: Artificial intelligence (AI) has made its way into everyday activities, particularly through new techniques such as machine learning (ML). These techniques are implementable with little domain knowledge. This, combined with the difficulty of testing AI systems with traditional methods, has made system trustworthiness a pressing issue. Objective: This paper studies the methods used to validate practical AI systems reported in the literature. Our goal is to classify and describe the methods that are used in realistic settings to ensure the dependability of AI systems. Method: A systematic literature review resulted in 90 papers. Systems presented in the papers were analysed based on their domain, task, complexity, and applied validation methods. Results: The validation methods were synthesized into a taxonomy consisting of trial, simulation, model-centred validation, and expert opinion. Failure monitors, safety channels, redundancy, voting, and input and output restrictions are methods used to continuously validate the systems after deployment. Conclusions: Our results clarify existing strategies applied to validation. They form a basis for the synthesization, assessment, and refinement of AI system validation in research and guidelines for validating individual systems in practice. While various validation strategies have all been relatively widely applied, only few studies report on continuous validation. Keywords: artificial intelligence, machine learning, validation, testing, V&V, systematic literature review.
[ { "created": "Mon, 26 Jul 2021 12:54:32 GMT", "version": "v1" } ]
2021-09-17
[ [ "Myllyaho", "Lalli", "" ], [ "Raatikainen", "Mikko", "" ], [ "Männistö", "Tomi", "" ], [ "Mikkonen", "Tommi", "" ], [ "Nurminen", "Jukka K.", "" ] ]
Context: Artificial intelligence (AI) has made its way into everyday activities, particularly through new techniques such as machine learning (ML). These techniques are implementable with little domain knowledge. This, combined with the difficulty of testing AI systems with traditional methods, has made system trustworthiness a pressing issue. Objective: This paper studies the methods used to validate practical AI systems reported in the literature. Our goal is to classify and describe the methods that are used in realistic settings to ensure the dependability of AI systems. Method: A systematic literature review resulted in 90 papers. Systems presented in the papers were analysed based on their domain, task, complexity, and applied validation methods. Results: The validation methods were synthesized into a taxonomy consisting of trial, simulation, model-centred validation, and expert opinion. Failure monitors, safety channels, redundancy, voting, and input and output restrictions are methods used to continuously validate the systems after deployment. Conclusions: Our results clarify existing strategies applied to validation. They form a basis for the synthesization, assessment, and refinement of AI system validation in research and guidelines for validating individual systems in practice. While various validation strategies have all been relatively widely applied, only few studies report on continuous validation. Keywords: artificial intelligence, machine learning, validation, testing, V&V, systematic literature review.
2209.03473
Alexandre Duval
Alexandre Duval, Fragkiskos Malliaros
Higher-order Clustering and Pooling for Graph Neural Networks
CIKM 2022
null
10.1145/3511808.3557353
null
cs.LG cs.AI
http://creativecommons.org/licenses/by/4.0/
Graph Neural Networks achieve state-of-the-art performance on a plethora of graph classification tasks, especially due to pooling operators, which aggregate learned node embeddings hierarchically into a final graph representation. However, they are not only questioned by recent work showing on par performance with random pooling, but also ignore completely higher-order connectivity patterns. To tackle this issue, we propose HoscPool, a clustering-based graph pooling operator that captures higher-order information hierarchically, leading to richer graph representations. In fact, we learn a probabilistic cluster assignment matrix end-to-end by minimising relaxed formulations of motif spectral clustering in our objective function, and we then extend it to a pooling operator. We evaluate HoscPool on graph classification tasks and its clustering component on graphs with ground-truth community structure, achieving best performance. Lastly, we provide a deep empirical analysis of pooling operators' inner functioning.
[ { "created": "Fri, 2 Sep 2022 09:17:10 GMT", "version": "v1" } ]
2022-09-09
[ [ "Duval", "Alexandre", "" ], [ "Malliaros", "Fragkiskos", "" ] ]
Graph Neural Networks achieve state-of-the-art performance on a plethora of graph classification tasks, especially due to pooling operators, which aggregate learned node embeddings hierarchically into a final graph representation. However, they are not only questioned by recent work showing on par performance with random pooling, but also ignore completely higher-order connectivity patterns. To tackle this issue, we propose HoscPool, a clustering-based graph pooling operator that captures higher-order information hierarchically, leading to richer graph representations. In fact, we learn a probabilistic cluster assignment matrix end-to-end by minimising relaxed formulations of motif spectral clustering in our objective function, and we then extend it to a pooling operator. We evaluate HoscPool on graph classification tasks and its clustering component on graphs with ground-truth community structure, achieving best performance. Lastly, we provide a deep empirical analysis of pooling operators' inner functioning.
1212.3638
Derrick Wing Kwan Ng Dr.
Derrick Wing Kwan Ng, Ernest S. Lo, and Robert Schober
Energy-Efficient Resource Allocation in Multiuser OFDM Systems with Wireless Information and Power Transfer
6 pages. The paper has been accepted for publication at the IEEE Wireless Communications and Networking Conference (WCNC) 2013, Shanghai, China, Apr. 2013
null
10.1109/WCNC.2013.6555184
null
cs.IT math.IT
http://creativecommons.org/licenses/by-nc-sa/3.0/
In this paper, we study the resource allocation algorithm design for multiuser orthogonal frequency division multiplexing (OFDM) downlink systems with simultaneous wireless information and power transfer. The algorithm design is formulated as a non-convex optimization problem for maximizing the energy efficiency of data transmission (bit/Joule delivered to the users). In particular, the problem formulation takes into account the minimum required system data rate, heterogeneous minimum required power transfers to the users, and the circuit power consumption. Subsequently, by exploiting the method of time-sharing and the properties of nonlinear fractional programming, the considered non-convex optimization problem is solved using an efficient iterative resource allocation algorithm. For each iteration, the optimal power allocation and user selection solution are derived based on Lagrange dual decomposition. Simulation results illustrate that the proposed iterative resource allocation algorithm achieves the maximum energy efficiency of the system and reveal how energy efficiency, system capacity, and wireless power transfer benefit from the presence of multiple users in the system.
[ { "created": "Fri, 14 Dec 2012 23:42:59 GMT", "version": "v1" }, { "created": "Mon, 31 Dec 2012 21:01:01 GMT", "version": "v2" } ]
2016-11-17
[ [ "Ng", "Derrick Wing Kwan", "" ], [ "Lo", "Ernest S.", "" ], [ "Schober", "Robert", "" ] ]
In this paper, we study the resource allocation algorithm design for multiuser orthogonal frequency division multiplexing (OFDM) downlink systems with simultaneous wireless information and power transfer. The algorithm design is formulated as a non-convex optimization problem for maximizing the energy efficiency of data transmission (bit/Joule delivered to the users). In particular, the problem formulation takes into account the minimum required system data rate, heterogeneous minimum required power transfers to the users, and the circuit power consumption. Subsequently, by exploiting the method of time-sharing and the properties of nonlinear fractional programming, the considered non-convex optimization problem is solved using an efficient iterative resource allocation algorithm. For each iteration, the optimal power allocation and user selection solution are derived based on Lagrange dual decomposition. Simulation results illustrate that the proposed iterative resource allocation algorithm achieves the maximum energy efficiency of the system and reveal how energy efficiency, system capacity, and wireless power transfer benefit from the presence of multiple users in the system.
1909.12938
Akhil Gupta
Akhil Gupta
Time Series Modeling for Dream Team in Fantasy Premier League
International Conference on Sports Engineering (ICSE'17)
null
null
null
cs.CY cs.LG
http://creativecommons.org/licenses/by-nc-sa/4.0/
The performance of football players in English Premier League varies largely from season to season and for different teams. It is evident that a method capable of forecasting and analyzing the future of these players on-field antics shall assist the management to a great extent. In a simulated environment like the Fantasy Premier League, enthusiasts from all over the world participate and manage the players catalogue for the entire season. Due to the dynamic nature of points system, there is no known approach for the formulation of a dream team. This study aims to tackle this problem by using a hybrid of Autoregressive Integrated Moving Average (ARIMA) and Recurrent Neural Networks (RNNs) for time series prediction of player points and subsequent maximization of total points using Linear Programming (LPP). Given the player points for the past three seasons, the predictions have been made for the current season by modeling differently for ARIMA and RNN, and then creating an ensemble of the same. Prior to that, proper data preprocessing techniques were deployed to enhance the efficacy of the prepared model. Constraints on the type of players like goalkeepers, defenders, midfielders and forwards along with the total budget were effectively optimized using LPP approach. The validation of the proposed team was done with the performance in upcoming season, where the players outperform as expected, and helped in strengthening the feasibility of the solution. Likewise, the proposed approach can be extended to English Premier League by official managers on-field.
[ { "created": "Thu, 19 Sep 2019 20:20:04 GMT", "version": "v1" } ]
2019-10-01
[ [ "Gupta", "Akhil", "" ] ]
The performance of football players in English Premier League varies largely from season to season and for different teams. It is evident that a method capable of forecasting and analyzing the future of these players on-field antics shall assist the management to a great extent. In a simulated environment like the Fantasy Premier League, enthusiasts from all over the world participate and manage the players catalogue for the entire season. Due to the dynamic nature of points system, there is no known approach for the formulation of a dream team. This study aims to tackle this problem by using a hybrid of Autoregressive Integrated Moving Average (ARIMA) and Recurrent Neural Networks (RNNs) for time series prediction of player points and subsequent maximization of total points using Linear Programming (LPP). Given the player points for the past three seasons, the predictions have been made for the current season by modeling differently for ARIMA and RNN, and then creating an ensemble of the same. Prior to that, proper data preprocessing techniques were deployed to enhance the efficacy of the prepared model. Constraints on the type of players like goalkeepers, defenders, midfielders and forwards along with the total budget were effectively optimized using LPP approach. The validation of the proposed team was done with the performance in upcoming season, where the players outperform as expected, and helped in strengthening the feasibility of the solution. Likewise, the proposed approach can be extended to English Premier League by official managers on-field.
2207.08857
Md Nafee Al Islam
Md Nafee Al Islam, Yihong Ma, Pedro Alarcon Granadeno, Nitesh Chawla, Jane Cleland-Huang
RESAM: Requirements Elicitation and Specification for Deep-Learning Anomaly Models with Applications to UAV Flight Controllers
null
null
null
null
cs.SE cs.AI cs.MA
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
CyberPhysical systems (CPS) must be closely monitored to identify and potentially mitigate emergent problems that arise during their routine operations. However, the multivariate time-series data which they typically produce can be complex to understand and analyze. While formal product documentation often provides example data plots with diagnostic suggestions, the sheer diversity of attributes, critical thresholds, and data interactions can be overwhelming to non-experts who subsequently seek help from discussion forums to interpret their data logs. Deep learning models, such as Long Short-term memory (LSTM) networks can be used to automate these tasks and to provide clear explanations of diverse anomalies detected in real-time multivariate data-streams. In this paper we present RESAM, a requirements process that integrates knowledge from domain experts, discussion forums, and formal product documentation, to discover and specify requirements and design definitions in the form of time-series attributes that contribute to the construction of effective deep learning anomaly detectors. We present a case-study based on a flight control system for small Uncrewed Aerial Systems and demonstrate that its use guides the construction of effective anomaly detection models whilst also providing underlying support for explainability. RESAM is relevant to domains in which open or closed online forums provide discussion support for log analysis.
[ { "created": "Mon, 18 Jul 2022 18:09:59 GMT", "version": "v1" } ]
2022-07-20
[ [ "Islam", "Md Nafee Al", "" ], [ "Ma", "Yihong", "" ], [ "Granadeno", "Pedro Alarcon", "" ], [ "Chawla", "Nitesh", "" ], [ "Cleland-Huang", "Jane", "" ] ]
CyberPhysical systems (CPS) must be closely monitored to identify and potentially mitigate emergent problems that arise during their routine operations. However, the multivariate time-series data which they typically produce can be complex to understand and analyze. While formal product documentation often provides example data plots with diagnostic suggestions, the sheer diversity of attributes, critical thresholds, and data interactions can be overwhelming to non-experts who subsequently seek help from discussion forums to interpret their data logs. Deep learning models, such as Long Short-term memory (LSTM) networks can be used to automate these tasks and to provide clear explanations of diverse anomalies detected in real-time multivariate data-streams. In this paper we present RESAM, a requirements process that integrates knowledge from domain experts, discussion forums, and formal product documentation, to discover and specify requirements and design definitions in the form of time-series attributes that contribute to the construction of effective deep learning anomaly detectors. We present a case-study based on a flight control system for small Uncrewed Aerial Systems and demonstrate that its use guides the construction of effective anomaly detection models whilst also providing underlying support for explainability. RESAM is relevant to domains in which open or closed online forums provide discussion support for log analysis.
2401.13980
Qianqian Yang
Weixuan Chen, Shuo Shao, Qianqian Yang, Zhaoyang Zhang, Ping Zhang
A Nearly Information Theoretically Secure Approach for Semantic Communications over Wiretap Channel
13 pages, 16 figures
null
null
null
cs.IT eess.IV math.IT
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
This paper addresses the challenge of achieving information-theoretic security in semantic communication (SeCom) over a wiretap channel, where a legitimate receiver coexists with an eavesdropper experiencing a poorer channel condition. Despite previous efforts to secure SeCom against eavesdroppers, achieving information-theoretic security in such schemes remains an open issue. In this work, we propose a secure digital SeCom approach based on superposition codes, aiming to attain nearly information-theoretic security. Our proposed method involves associating semantic information with satellite constellation points within a double-layered constellation map, where cloud center constellation points are randomly selected. By carefully allocating power between these two layers of constellation, we ensure that the symbol error probability (SEP) of the eavesdropper decoding satellite constellation points is nearly equivalent to random guessing, while maintaining a low SEP for the legitimate receiver to successfully decode the semantic information. Simulation results showcase that the Peak Signal-to-Noise Ratio (PSNR) and Mean Squared Error (MSE) for the eavesdropper's reconstructed data, using our proposed method, can range from decoding Gaussian-distributed random noise to approaching the variance of the data. This validates the ability of our method to achieve nearly information-theoretic security, demonstrating superior data security compared to benchmark methods.
[ { "created": "Thu, 25 Jan 2024 06:46:13 GMT", "version": "v1" } ]
2024-01-26
[ [ "Chen", "Weixuan", "" ], [ "Shao", "Shuo", "" ], [ "Yang", "Qianqian", "" ], [ "Zhang", "Zhaoyang", "" ], [ "Zhang", "Ping", "" ] ]
This paper addresses the challenge of achieving information-theoretic security in semantic communication (SeCom) over a wiretap channel, where a legitimate receiver coexists with an eavesdropper experiencing a poorer channel condition. Despite previous efforts to secure SeCom against eavesdroppers, achieving information-theoretic security in such schemes remains an open issue. In this work, we propose a secure digital SeCom approach based on superposition codes, aiming to attain nearly information-theoretic security. Our proposed method involves associating semantic information with satellite constellation points within a double-layered constellation map, where cloud center constellation points are randomly selected. By carefully allocating power between these two layers of constellation, we ensure that the symbol error probability (SEP) of the eavesdropper decoding satellite constellation points is nearly equivalent to random guessing, while maintaining a low SEP for the legitimate receiver to successfully decode the semantic information. Simulation results showcase that the Peak Signal-to-Noise Ratio (PSNR) and Mean Squared Error (MSE) for the eavesdropper's reconstructed data, using our proposed method, can range from decoding Gaussian-distributed random noise to approaching the variance of the data. This validates the ability of our method to achieve nearly information-theoretic security, demonstrating superior data security compared to benchmark methods.
1512.07901
Marco Bressan
Marco Bressan, Enoch Peserico, Luca Pretto
Simple set cardinality estimation through random sampling
3 pages
null
null
null
cs.DM
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We present a simple algorithm that estimates the cardinality $n$ of a set $V$ when allowed to sample elements of $V$ uniformly and independently at random. Our algorithm with probability $(1-\delta)$ returns a $(1\pm\epsilon)-$approximation of $n$ drawing $O\big(\sqrt{n} \cdot \epsilon^{-1}\sqrt{\log(\delta^{-1})}\big)$ samples (for $\epsilon^{-1}\sqrt{\log(\delta^{-1})} = O(\sqrt{n})$).
[ { "created": "Thu, 24 Dec 2015 20:42:10 GMT", "version": "v1" }, { "created": "Mon, 18 Jan 2016 15:25:18 GMT", "version": "v2" }, { "created": "Wed, 11 Apr 2018 21:09:20 GMT", "version": "v3" } ]
2018-04-13
[ [ "Bressan", "Marco", "" ], [ "Peserico", "Enoch", "" ], [ "Pretto", "Luca", "" ] ]
We present a simple algorithm that estimates the cardinality $n$ of a set $V$ when allowed to sample elements of $V$ uniformly and independently at random. Our algorithm with probability $(1-\delta)$ returns a $(1\pm\epsilon)-$approximation of $n$ drawing $O\big(\sqrt{n} \cdot \epsilon^{-1}\sqrt{\log(\delta^{-1})}\big)$ samples (for $\epsilon^{-1}\sqrt{\log(\delta^{-1})} = O(\sqrt{n})$).
1808.09796
Bingjie Xu
Bingjie Xu, Junnan Li, Yongkang Wong, Mohan S. Kankanhalli, and Qi Zhao
Interact as You Intend: Intention-Driven Human-Object Interaction Detection
null
null
null
null
cs.CV
http://creativecommons.org/licenses/by/4.0/
The recent advances in instance-level detection tasks lay strong foundation for genuine comprehension of the visual scenes. However, the ability to fully comprehend a social scene is still in its preliminary stage. In this work, we focus on detecting human-object interactions (HOIs) in social scene images, which is demanding in terms of research and increasingly useful for practical applications. To undertake social tasks interacting with objects, humans direct their attention and move their body based on their intention. Based on this observation, we provide a unique computational perspective to explore human intention in HOI detection. Specifically, the proposed human intention-driven HOI detection (iHOI) framework models human pose with the relative distances from body joints to the object instances. It also utilizes human gaze to guide the attended contextual regions in a weakly-supervised setting. In addition, we propose a hard negative sampling strategy to address the problem of mis-grouping. We perform extensive experiments on two benchmark datasets, namely V-COCO and HICO-DET. The efficacy of each proposed component has also been validated.
[ { "created": "Wed, 29 Aug 2018 13:25:50 GMT", "version": "v1" }, { "created": "Sun, 22 Sep 2019 11:45:38 GMT", "version": "v2" } ]
2019-09-24
[ [ "Xu", "Bingjie", "" ], [ "Li", "Junnan", "" ], [ "Wong", "Yongkang", "" ], [ "Kankanhalli", "Mohan S.", "" ], [ "Zhao", "Qi", "" ] ]
The recent advances in instance-level detection tasks lay strong foundation for genuine comprehension of the visual scenes. However, the ability to fully comprehend a social scene is still in its preliminary stage. In this work, we focus on detecting human-object interactions (HOIs) in social scene images, which is demanding in terms of research and increasingly useful for practical applications. To undertake social tasks interacting with objects, humans direct their attention and move their body based on their intention. Based on this observation, we provide a unique computational perspective to explore human intention in HOI detection. Specifically, the proposed human intention-driven HOI detection (iHOI) framework models human pose with the relative distances from body joints to the object instances. It also utilizes human gaze to guide the attended contextual regions in a weakly-supervised setting. In addition, we propose a hard negative sampling strategy to address the problem of mis-grouping. We perform extensive experiments on two benchmark datasets, namely V-COCO and HICO-DET. The efficacy of each proposed component has also been validated.
2101.03230
Junjie Zhong
Junjie Zhong and Hiromitsu Hattori
Generation of Traffic Flows in Multi-Agent Traffic Simulation with Agent Behavior Model based on Deep Reinforcement Learning
Experiment data maybe wrong due to the method " Repeated and Partial Training". This method may not converge to the optimal policy
null
null
null
cs.MA cs.SY eess.SY
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In multi-agent based traffic simulation, agents are always supposed to move following existing instructions, and mechanically and unnaturally imitate human behavior. The human drivers perform acceleration or deceleration irregularly all the time, which seems unnecessary in some conditions. For letting agents in traffic simulation behave more like humans and recognize other agents' behavior in complex conditions, we propose a unified mechanism for agents learn to decide various accelerations by using deep reinforcement learning based on a combination of regenerated visual images revealing some notable features, and numerical vectors containing some important data such as instantaneous speed. By handling batches of sequential data, agents are enabled to recognize surrounding agents' behavior and decide their own acceleration. In addition, we can generate a traffic flow behaving diversely to simulate the real traffic flow by using an architecture of fully decentralized training and fully centralized execution without violating Markov assumptions.
[ { "created": "Sat, 26 Dec 2020 15:13:06 GMT", "version": "v1" }, { "created": "Mon, 25 Jan 2021 05:00:00 GMT", "version": "v2" } ]
2021-01-26
[ [ "Zhong", "Junjie", "" ], [ "Hattori", "Hiromitsu", "" ] ]
In multi-agent based traffic simulation, agents are always supposed to move following existing instructions, and mechanically and unnaturally imitate human behavior. The human drivers perform acceleration or deceleration irregularly all the time, which seems unnecessary in some conditions. For letting agents in traffic simulation behave more like humans and recognize other agents' behavior in complex conditions, we propose a unified mechanism for agents learn to decide various accelerations by using deep reinforcement learning based on a combination of regenerated visual images revealing some notable features, and numerical vectors containing some important data such as instantaneous speed. By handling batches of sequential data, agents are enabled to recognize surrounding agents' behavior and decide their own acceleration. In addition, we can generate a traffic flow behaving diversely to simulate the real traffic flow by using an architecture of fully decentralized training and fully centralized execution without violating Markov assumptions.
1505.04785
Emanuel Diamant
Emanuel Diamant
Advances in Bioinformatics and Computational Biology: Don't take them too seriously anyway
The paper was submitted to the BIOCOMP'15 conference (Las Vegas, Nevada, USA, July 27-30, 2015) and was accepted as a poster presentation. arXiv admin note: text overlap with arXiv:1505.04578
null
null
null
cs.CE
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In the last few decades or so, we witness a paradigm shift in our nature studies - from a data-processing based computational approach to an information-processing based cognitive approach. The process is restricted and often misguided by the lack of a clear understanding about what information is and how it should be treated in research applications (in general) and in biological studies (in particular). The paper intend to provide some remedies for this bizarre situation.
[ { "created": "Mon, 18 May 2015 10:18:20 GMT", "version": "v1" } ]
2015-05-20
[ [ "Diamant", "Emanuel", "" ] ]
In the last few decades or so, we witness a paradigm shift in our nature studies - from a data-processing based computational approach to an information-processing based cognitive approach. The process is restricted and often misguided by the lack of a clear understanding about what information is and how it should be treated in research applications (in general) and in biological studies (in particular). The paper intend to provide some remedies for this bizarre situation.
1303.3469
Hassan Bashir A
Hassan A. Bashir and Richard S. Neville
Hybrid Evolutionary Computation for Continuous Optimization
Companion Publications for this Technical Memorandum, available at IEEE Xplore, are: [1] H. A. Bashir and R. S. Neville, "Convergence measurement in evolutionary computation using Price's theorem," IEEE (CEC), 2012. [2] H. A. Bashir and R. S. Neville, "A hybrid evolutionary computation algorithm for global optimization," IEEE (CEC), 2012
null
null
Technical Memorandum 2011-v.01
cs.NE
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Hybrid optimization algorithms have gained popularity as it has become apparent there cannot be a universal optimization strategy which is globally more beneficial than any other. Despite their popularity, hybridization frameworks require more detailed categorization regarding: the nature of the problem domain, the constituent algorithms, the coupling schema and the intended area of application. This report proposes a hybrid algorithm for solving small to large-scale continuous global optimization problems. It comprises evolutionary computation (EC) algorithms and a sequential quadratic programming (SQP) algorithm; combined in a collaborative portfolio. The SQP is a gradient based local search method. To optimize the individual contributions of the EC and SQP algorithms for the overall success of the proposed hybrid system, improvements were made in key features of these algorithms. The report proposes enhancements in: i) the evolutionary algorithm, ii) a new convergence detection mechanism was proposed; and iii) in the methods for evaluating the search directions and step sizes for the SQP local search algorithm. The proposed hybrid design aim was to ensure that the two algorithms complement each other by exploring and exploiting the problem search space. Preliminary results justify that an adept hybridization of evolutionary algorithms with a suitable local search method, could yield a robust and efficient means of solving wide range of global optimization problems. Finally, a discussion of the outcomes of the initial investigation and a review of the associated challenges and inherent limitations of the proposed method is presented to complete the investigation. The report highlights extensive research, particularly, some potential case studies and application areas.
[ { "created": "Thu, 14 Mar 2013 14:59:32 GMT", "version": "v1" } ]
2013-03-15
[ [ "Bashir", "Hassan A.", "" ], [ "Neville", "Richard S.", "" ] ]
Hybrid optimization algorithms have gained popularity as it has become apparent there cannot be a universal optimization strategy which is globally more beneficial than any other. Despite their popularity, hybridization frameworks require more detailed categorization regarding: the nature of the problem domain, the constituent algorithms, the coupling schema and the intended area of application. This report proposes a hybrid algorithm for solving small to large-scale continuous global optimization problems. It comprises evolutionary computation (EC) algorithms and a sequential quadratic programming (SQP) algorithm; combined in a collaborative portfolio. The SQP is a gradient based local search method. To optimize the individual contributions of the EC and SQP algorithms for the overall success of the proposed hybrid system, improvements were made in key features of these algorithms. The report proposes enhancements in: i) the evolutionary algorithm, ii) a new convergence detection mechanism was proposed; and iii) in the methods for evaluating the search directions and step sizes for the SQP local search algorithm. The proposed hybrid design aim was to ensure that the two algorithms complement each other by exploring and exploiting the problem search space. Preliminary results justify that an adept hybridization of evolutionary algorithms with a suitable local search method, could yield a robust and efficient means of solving wide range of global optimization problems. Finally, a discussion of the outcomes of the initial investigation and a review of the associated challenges and inherent limitations of the proposed method is presented to complete the investigation. The report highlights extensive research, particularly, some potential case studies and application areas.
2012.12394
Stefano Giovanni Rizzo
Stefano Giovanni Rizzo, Linsey Pang, Yixian Chen, Sanjay Chawla
Probabilistic Outlier Detection and Generation
null
null
null
null
cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
A new method for outlier detection and generation is introduced by lifting data into the space of probability distributions which are not analytically expressible, but from which samples can be drawn using a neural generator. Given a mixture of unknown latent inlier and outlier distributions, a Wasserstein double autoencoder is used to both detect and generate inliers and outliers. The proposed method, named WALDO (Wasserstein Autoencoder for Learning the Distribution of Outliers), is evaluated on classical data sets including MNIST, CIFAR10 and KDD99 for detection accuracy and robustness. We give an example of outlier detection on a real retail sales data set and an example of outlier generation for simulating intrusion attacks. However we foresee many application scenarios where WALDO can be used. To the best of our knowledge this is the first work that studies both outlier detection and generation together.
[ { "created": "Tue, 22 Dec 2020 22:42:56 GMT", "version": "v1" } ]
2020-12-24
[ [ "Rizzo", "Stefano Giovanni", "" ], [ "Pang", "Linsey", "" ], [ "Chen", "Yixian", "" ], [ "Chawla", "Sanjay", "" ] ]
A new method for outlier detection and generation is introduced by lifting data into the space of probability distributions which are not analytically expressible, but from which samples can be drawn using a neural generator. Given a mixture of unknown latent inlier and outlier distributions, a Wasserstein double autoencoder is used to both detect and generate inliers and outliers. The proposed method, named WALDO (Wasserstein Autoencoder for Learning the Distribution of Outliers), is evaluated on classical data sets including MNIST, CIFAR10 and KDD99 for detection accuracy and robustness. We give an example of outlier detection on a real retail sales data set and an example of outlier generation for simulating intrusion attacks. However we foresee many application scenarios where WALDO can be used. To the best of our knowledge this is the first work that studies both outlier detection and generation together.
2407.20535
Cynthia Steinhardt
Cynthia R. Steinhardt, Menoua Keshishian, Nima Mesgarani, Kim Stachenfeld
DeepSpeech models show Human-like Performance and Processing of Cochlear Implant Inputs
NEURIPS preprint
null
null
null
cs.NE cs.SD eess.AS
http://creativecommons.org/licenses/by-nc-nd/4.0/
Cochlear implants(CIs) are arguably the most successful neural implant, having restored hearing to over one million people worldwide. While CI research has focused on modeling the cochlear activations in response to low-level acoustic features, we hypothesize that the success of these implants is due in large part to the role of the upstream network in extracting useful features from a degraded signal and learned statistics of language to resolve the signal. In this work, we use the deep neural network (DNN) DeepSpeech2, as a paradigm to investigate how natural input and cochlear implant-based inputs are processed over time. We generate naturalistic and cochlear implant-like inputs from spoken sentences and test the similarity of model performance to human performance on analogous phoneme recognition tests. Our model reproduces error patterns in reaction time and phoneme confusion patterns under noise conditions in normal hearing and CI participant studies. We then use interpretability techniques to determine where and when confusions arise when processing naturalistic and CI-like inputs. We find that dynamics over time in each layer are affected by context as well as input type. Dynamics of all phonemes diverge during confusion and comprehension within the same time window, which is temporally shifted backward in each layer of the network. There is a modulation of this signal during processing of CI which resembles changes in human EEG signals in the auditory stream. This reduction likely relates to the reduction of encoded phoneme identity. These findings suggest that we have a viable model in which to explore the loss of speech-related information in time and that we can use it to find population-level encoding signals to target when optimizing cochlear implant inputs to improve encoding of essential speech-related information and improve perception.
[ { "created": "Tue, 30 Jul 2024 04:32:27 GMT", "version": "v1" } ]
2024-07-31
[ [ "Steinhardt", "Cynthia R.", "" ], [ "Keshishian", "Menoua", "" ], [ "Mesgarani", "Nima", "" ], [ "Stachenfeld", "Kim", "" ] ]
Cochlear implants(CIs) are arguably the most successful neural implant, having restored hearing to over one million people worldwide. While CI research has focused on modeling the cochlear activations in response to low-level acoustic features, we hypothesize that the success of these implants is due in large part to the role of the upstream network in extracting useful features from a degraded signal and learned statistics of language to resolve the signal. In this work, we use the deep neural network (DNN) DeepSpeech2, as a paradigm to investigate how natural input and cochlear implant-based inputs are processed over time. We generate naturalistic and cochlear implant-like inputs from spoken sentences and test the similarity of model performance to human performance on analogous phoneme recognition tests. Our model reproduces error patterns in reaction time and phoneme confusion patterns under noise conditions in normal hearing and CI participant studies. We then use interpretability techniques to determine where and when confusions arise when processing naturalistic and CI-like inputs. We find that dynamics over time in each layer are affected by context as well as input type. Dynamics of all phonemes diverge during confusion and comprehension within the same time window, which is temporally shifted backward in each layer of the network. There is a modulation of this signal during processing of CI which resembles changes in human EEG signals in the auditory stream. This reduction likely relates to the reduction of encoded phoneme identity. These findings suggest that we have a viable model in which to explore the loss of speech-related information in time and that we can use it to find population-level encoding signals to target when optimizing cochlear implant inputs to improve encoding of essential speech-related information and improve perception.
1302.7080
Wesam Elshamy
Hassan M Emara, Wesam Elshamy, Ahmed Bahgat
Parameter Identification of Induction Motor Using Modified Particle Swarm Optimization Algorithm
IEEE International Symposium on Industrial Electronics Jul 2008, Cambridge, UK
null
10.1109/ISIE.2008.4677254
null
cs.NE
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
This paper presents a new technique for induction motor parameter identification. The proposed technique is based on a simple startup test using a standard V/F inverter. The recorded startup currents are compared to that obtained by simulation of an induction motor model. A Modified PSO optimization is used to find out the best model parameter that minimizes the sum square error between the measured and the simulated currents. The performance of the modified PSO is compared with other optimization methods including line search, conventional PSO and Genetic Algorithms. Simulation results demonstrate the ability of the proposed technique to capture the true values of the machine parameters and the superiority of the results obtained using the modified PSO over other optimization techniques.
[ { "created": "Thu, 28 Feb 2013 04:41:53 GMT", "version": "v1" } ]
2016-11-17
[ [ "Emara", "Hassan M", "" ], [ "Elshamy", "Wesam", "" ], [ "Bahgat", "Ahmed", "" ] ]
This paper presents a new technique for induction motor parameter identification. The proposed technique is based on a simple startup test using a standard V/F inverter. The recorded startup currents are compared to that obtained by simulation of an induction motor model. A Modified PSO optimization is used to find out the best model parameter that minimizes the sum square error between the measured and the simulated currents. The performance of the modified PSO is compared with other optimization methods including line search, conventional PSO and Genetic Algorithms. Simulation results demonstrate the ability of the proposed technique to capture the true values of the machine parameters and the superiority of the results obtained using the modified PSO over other optimization techniques.
2010.14916
Yoann Dieudonn\'e
S\'ebastien Bouchard, Yoann Dieudonn\'e, Arnaud Labourel, Andrzej Pelc
Almost-Optimal Deterministic Treasure Hunt in Arbitrary Graphs
null
null
null
null
cs.DS cs.CC
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
A mobile agent navigating along edges of a simple connected graph, either finite or countably infinite, has to find an inert target (treasure) hidden in one of the nodes. This task is known as treasure hunt. The agent has no a priori knowledge of the graph, of the location of the treasure or of the initial distance to it. The cost of a treasure hunt algorithm is the worst-case number of edge traversals performed by the agent until finding the treasure. Awerbuch, Betke, Rivest and Singh [3] considered graph exploration and treasure hunt for finite graphs in a restricted model where the agent has a fuel tank that can be replenished only at the starting node $s$. The size of the tank is $B=2(1+\alpha)r$, for some positive real constant $\alpha$, where $r$, called the radius of the graph, is the maximum distance from $s$ to any other node. The tank of size $B$ allows the agent to make at most $\lfloor B\rfloor$ edge traversals between two consecutive visits at node $s$. Let $e(d)$ be the number of edges whose at least one extremity is at distance less than $d$ from $s$. Awerbuch, Betke, Rivest and Singh [3] conjectured that it is impossible to find a treasure hidden in a node at distance at most $d$ at cost nearly linear in $e(d)$. We first design a deterministic treasure hunt algorithm working in the model without any restrictions on the moves of the agent at cost $\mathcal{O}(e(d) \log d)$, and then show how to modify this algorithm to work in the model from [3] with the same complexity. Thus we refute the above twenty-year-old conjecture. We observe that no treasure hunt algorithm can beat cost $\Theta(e(d))$ for all graphs and thus our algorithms are also almost optimal.
[ { "created": "Wed, 28 Oct 2020 12:25:23 GMT", "version": "v1" }, { "created": "Tue, 3 Nov 2020 12:17:11 GMT", "version": "v2" }, { "created": "Wed, 4 Nov 2020 14:32:43 GMT", "version": "v3" }, { "created": "Thu, 11 Feb 2021 12:43:49 GMT", "version": "v4" }, { "created": "Sat, 13 Feb 2021 10:27:18 GMT", "version": "v5" } ]
2021-02-16
[ [ "Bouchard", "Sébastien", "" ], [ "Dieudonné", "Yoann", "" ], [ "Labourel", "Arnaud", "" ], [ "Pelc", "Andrzej", "" ] ]
A mobile agent navigating along edges of a simple connected graph, either finite or countably infinite, has to find an inert target (treasure) hidden in one of the nodes. This task is known as treasure hunt. The agent has no a priori knowledge of the graph, of the location of the treasure or of the initial distance to it. The cost of a treasure hunt algorithm is the worst-case number of edge traversals performed by the agent until finding the treasure. Awerbuch, Betke, Rivest and Singh [3] considered graph exploration and treasure hunt for finite graphs in a restricted model where the agent has a fuel tank that can be replenished only at the starting node $s$. The size of the tank is $B=2(1+\alpha)r$, for some positive real constant $\alpha$, where $r$, called the radius of the graph, is the maximum distance from $s$ to any other node. The tank of size $B$ allows the agent to make at most $\lfloor B\rfloor$ edge traversals between two consecutive visits at node $s$. Let $e(d)$ be the number of edges whose at least one extremity is at distance less than $d$ from $s$. Awerbuch, Betke, Rivest and Singh [3] conjectured that it is impossible to find a treasure hidden in a node at distance at most $d$ at cost nearly linear in $e(d)$. We first design a deterministic treasure hunt algorithm working in the model without any restrictions on the moves of the agent at cost $\mathcal{O}(e(d) \log d)$, and then show how to modify this algorithm to work in the model from [3] with the same complexity. Thus we refute the above twenty-year-old conjecture. We observe that no treasure hunt algorithm can beat cost $\Theta(e(d))$ for all graphs and thus our algorithms are also almost optimal.
2201.02374
Haisen Zhao
Fanchao Zhong and Yonglai Xu and Haisen Zhao and Lin Lu
As-Continuous-As-Possible Extrusion Fabrication of Surface Models
16 pages, 23 figures
null
null
null
cs.GR
http://creativecommons.org/licenses/by/4.0/
We propose a novel computational framework for optimizing the toolpath continuity in fabricating surface models on an extrusion-based 3D printer. Toolpath continuity has been a critical issue for extrusion-based fabrications that affects both quality and efficiency. Transfer moves cause non-smoothor bumpy surfaces and get worse for materials with large inertia like clay. For surface models, the effects of continuity are even more severe, in terms of surface quality and model stability. In this paper, we introduce an original criterion "one-path-patch" (OPP), for representing a shell surface patch that can be traversed in one path considering fabrication constraints. We study the properties of an OPP and the merging operations for OPPs, and propose a bottom-up OPP merging procedure for decomposing the given shell surface into a minimal number of OPPs and generating the "as-continuous-as-possible" (ACAP) toolpath. Furthermore, we customize the path planning algorithm with a curved layer printing scheme, which reduces the staircase defect and improves the toolpath continuity via possibly connecting multiple segments. We evaluate the ACAP algorithm for both ceramic and thermoplastic materials, and results demonstrate that it improves the fabrication of surface models in both surface quality and efficiency.
[ { "created": "Fri, 7 Jan 2022 09:18:59 GMT", "version": "v1" }, { "created": "Mon, 10 Jan 2022 20:22:50 GMT", "version": "v2" }, { "created": "Sat, 28 May 2022 08:39:43 GMT", "version": "v3" } ]
2022-05-31
[ [ "Zhong", "Fanchao", "" ], [ "Xu", "Yonglai", "" ], [ "Zhao", "Haisen", "" ], [ "Lu", "Lin", "" ] ]
We propose a novel computational framework for optimizing the toolpath continuity in fabricating surface models on an extrusion-based 3D printer. Toolpath continuity has been a critical issue for extrusion-based fabrications that affects both quality and efficiency. Transfer moves cause non-smoothor bumpy surfaces and get worse for materials with large inertia like clay. For surface models, the effects of continuity are even more severe, in terms of surface quality and model stability. In this paper, we introduce an original criterion "one-path-patch" (OPP), for representing a shell surface patch that can be traversed in one path considering fabrication constraints. We study the properties of an OPP and the merging operations for OPPs, and propose a bottom-up OPP merging procedure for decomposing the given shell surface into a minimal number of OPPs and generating the "as-continuous-as-possible" (ACAP) toolpath. Furthermore, we customize the path planning algorithm with a curved layer printing scheme, which reduces the staircase defect and improves the toolpath continuity via possibly connecting multiple segments. We evaluate the ACAP algorithm for both ceramic and thermoplastic materials, and results demonstrate that it improves the fabrication of surface models in both surface quality and efficiency.
2110.10548
Ningning Xie
Ningning Xie, Tamara Norman, Dominik Grewe, Dimitrios Vytiniotis
Synthesizing Optimal Parallelism Placement and Reduction Strategies on Hierarchical Systems for Deep Learning
null
null
null
null
cs.PL cs.DC cs.LG
http://creativecommons.org/licenses/by/4.0/
We present a novel characterization of the mapping of multiple parallelism forms (e.g. data and model parallelism) onto hierarchical accelerator systems that is hierarchy-aware and greatly reduces the space of software-to-hardware mapping. We experimentally verify the substantial effect of these mappings on all-reduce performance (up to 448x). We offer a novel syntax-guided program synthesis framework that is able to decompose reductions over one or more parallelism axes to sequences of collectives in a hierarchy- and mapping-aware way. For 69% of parallelism placements and user requested reductions, our framework synthesizes programs that outperform the default all-reduce implementation when evaluated on different GPU hierarchies (max 2.04x, average 1.27x). We complement our synthesis tool with a simulator exceeding 90% top-10 accuracy, which therefore reduces the need for massive evaluations of synthesis results to determine a small set of optimal programs and mappings.
[ { "created": "Wed, 20 Oct 2021 13:05:49 GMT", "version": "v1" }, { "created": "Tue, 16 Nov 2021 12:54:39 GMT", "version": "v2" } ]
2021-11-17
[ [ "Xie", "Ningning", "" ], [ "Norman", "Tamara", "" ], [ "Grewe", "Dominik", "" ], [ "Vytiniotis", "Dimitrios", "" ] ]
We present a novel characterization of the mapping of multiple parallelism forms (e.g. data and model parallelism) onto hierarchical accelerator systems that is hierarchy-aware and greatly reduces the space of software-to-hardware mapping. We experimentally verify the substantial effect of these mappings on all-reduce performance (up to 448x). We offer a novel syntax-guided program synthesis framework that is able to decompose reductions over one or more parallelism axes to sequences of collectives in a hierarchy- and mapping-aware way. For 69% of parallelism placements and user requested reductions, our framework synthesizes programs that outperform the default all-reduce implementation when evaluated on different GPU hierarchies (max 2.04x, average 1.27x). We complement our synthesis tool with a simulator exceeding 90% top-10 accuracy, which therefore reduces the need for massive evaluations of synthesis results to determine a small set of optimal programs and mappings.
2312.02207
Xiaojun Jia
Xiaojun Jia, Jindong Gu, Yihao Huang, Simeng Qin, Qing Guo, Yang Liu, Xiaochun Cao
TranSegPGD: Improving Transferability of Adversarial Examples on Semantic Segmentation
null
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Transferability of adversarial examples on image classification has been systematically explored, which generates adversarial examples in black-box mode. However, the transferability of adversarial examples on semantic segmentation has been largely overlooked. In this paper, we propose an effective two-stage adversarial attack strategy to improve the transferability of adversarial examples on semantic segmentation, dubbed TranSegPGD. Specifically, at the first stage, every pixel in an input image is divided into different branches based on its adversarial property. Different branches are assigned different weights for optimization to improve the adversarial performance of all pixels.We assign high weights to the loss of the hard-to-attack pixels to misclassify all pixels. At the second stage, the pixels are divided into different branches based on their transferable property which is dependent on Kullback-Leibler divergence. Different branches are assigned different weights for optimization to improve the transferability of the adversarial examples. We assign high weights to the loss of the high-transferability pixels to improve the transferability of adversarial examples. Extensive experiments with various segmentation models are conducted on PASCAL VOC 2012 and Cityscapes datasets to demonstrate the effectiveness of the proposed method. The proposed adversarial attack method can achieve state-of-the-art performance.
[ { "created": "Sun, 3 Dec 2023 00:48:33 GMT", "version": "v1" } ]
2023-12-06
[ [ "Jia", "Xiaojun", "" ], [ "Gu", "Jindong", "" ], [ "Huang", "Yihao", "" ], [ "Qin", "Simeng", "" ], [ "Guo", "Qing", "" ], [ "Liu", "Yang", "" ], [ "Cao", "Xiaochun", "" ] ]
Transferability of adversarial examples on image classification has been systematically explored, which generates adversarial examples in black-box mode. However, the transferability of adversarial examples on semantic segmentation has been largely overlooked. In this paper, we propose an effective two-stage adversarial attack strategy to improve the transferability of adversarial examples on semantic segmentation, dubbed TranSegPGD. Specifically, at the first stage, every pixel in an input image is divided into different branches based on its adversarial property. Different branches are assigned different weights for optimization to improve the adversarial performance of all pixels.We assign high weights to the loss of the hard-to-attack pixels to misclassify all pixels. At the second stage, the pixels are divided into different branches based on their transferable property which is dependent on Kullback-Leibler divergence. Different branches are assigned different weights for optimization to improve the transferability of the adversarial examples. We assign high weights to the loss of the high-transferability pixels to improve the transferability of adversarial examples. Extensive experiments with various segmentation models are conducted on PASCAL VOC 2012 and Cityscapes datasets to demonstrate the effectiveness of the proposed method. The proposed adversarial attack method can achieve state-of-the-art performance.
2204.04421
Ronghao Dang
Ronghao Dang, Zhuofan Shi, Liuyi Wang, Zongtao He, Chengju Liu, Qijun Chen
Unbiased Directed Object Attention Graph for Object Navigation
13 pages, accepted by ACM Mutimedia 2022
null
null
null
cs.CV cs.AI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Object navigation tasks require agents to locate specific objects in unknown environments based on visual information. Previously, graph convolutions were used to implicitly explore the relationships between objects. However, due to differences in visibility among objects, it is easy to generate biases in object attention. Thus, in this paper, we propose a directed object attention (DOA) graph to guide the agent in explicitly learning the attention relationships between objects, thereby reducing the object attention bias. In particular, we use the DOA graph to perform unbiased adaptive object attention (UAOA) on the object features and unbiased adaptive image attention (UAIA) on the raw images, respectively. To distinguish features in different branches, a concise adaptive branch energy distribution (ABED) method is proposed. We assess our methods on the AI2-Thor dataset. Compared with the state-of-the-art (SOTA) method, our method reports 7.4%, 8.1% and 17.6% increase in success rate (SR), success weighted by path length (SPL) and success weighted by action efficiency (SAE), respectively.
[ { "created": "Sat, 9 Apr 2022 08:13:05 GMT", "version": "v1" }, { "created": "Fri, 8 Jul 2022 01:41:38 GMT", "version": "v2" } ]
2022-07-11
[ [ "Dang", "Ronghao", "" ], [ "Shi", "Zhuofan", "" ], [ "Wang", "Liuyi", "" ], [ "He", "Zongtao", "" ], [ "Liu", "Chengju", "" ], [ "Chen", "Qijun", "" ] ]
Object navigation tasks require agents to locate specific objects in unknown environments based on visual information. Previously, graph convolutions were used to implicitly explore the relationships between objects. However, due to differences in visibility among objects, it is easy to generate biases in object attention. Thus, in this paper, we propose a directed object attention (DOA) graph to guide the agent in explicitly learning the attention relationships between objects, thereby reducing the object attention bias. In particular, we use the DOA graph to perform unbiased adaptive object attention (UAOA) on the object features and unbiased adaptive image attention (UAIA) on the raw images, respectively. To distinguish features in different branches, a concise adaptive branch energy distribution (ABED) method is proposed. We assess our methods on the AI2-Thor dataset. Compared with the state-of-the-art (SOTA) method, our method reports 7.4%, 8.1% and 17.6% increase in success rate (SR), success weighted by path length (SPL) and success weighted by action efficiency (SAE), respectively.
2205.11908
Siddhartha Siddhartha
Siddhartha
An interpretation of the final fully connected layer
null
null
null
null
cs.LG cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In recent years neural networks have achieved state-of-the-art accuracy for various tasks but the the interpretation of the generated outputs still remains difficult. In this work we attempt to provide a method to understand the learnt weights in the final fully connected layer in image classification models. We motivate our method by drawing a connection between the policy gradient objective in RL and supervised learning objective. We suggest that the commonly used cross entropy based supervised learning objective can be regarded as a special case of the policy gradient objective. Using this insight we propose a method to find the most discriminative and confusing parts of an image. Our method does not make any prior assumption about neural network achitecture and has low computational cost. We apply our method on publicly available pre-trained models and report the generated results.
[ { "created": "Tue, 24 May 2022 09:05:19 GMT", "version": "v1" } ]
2022-05-25
[ [ "Siddhartha", "", "" ] ]
In recent years neural networks have achieved state-of-the-art accuracy for various tasks but the the interpretation of the generated outputs still remains difficult. In this work we attempt to provide a method to understand the learnt weights in the final fully connected layer in image classification models. We motivate our method by drawing a connection between the policy gradient objective in RL and supervised learning objective. We suggest that the commonly used cross entropy based supervised learning objective can be regarded as a special case of the policy gradient objective. Using this insight we propose a method to find the most discriminative and confusing parts of an image. Our method does not make any prior assumption about neural network achitecture and has low computational cost. We apply our method on publicly available pre-trained models and report the generated results.
2311.11707
Dimitri Watel
Dominique Barth (DAVID, UVSQ), Thierry Mautor (DAVID, UVSQ), Dimitri Watel (SAMOVAR, ENSIIE), Marc-Antoine Weisser (LISN, GALaC)
Configuring an heterogeneous smartgrid network: complexity and approximations for tree topologies
Journal of Global Optimization, 2023
null
10.1007/s10898-023-01338-0
null
cs.CC
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We address the problem of configuring a power distribution network with reliability and resilience objectives by satisfying the demands of the consumers and saturating each production source as little as possible. We consider power distribution networks containing source nodes producing electricity, nodes representing electricity consumers and switches between them. Configuring this network consists in deciding the orientation of the links between the nodes of the network. The electric flow is a direct consequence of the chosen configuration and can be computed in polynomial time. It is valid if it satisfies the demand of each consumer and capacity constraints on the network. In such a case, we study the problem of determining a feasible solution that balances the loads of the sources, that is their production rates. We use three metrics to measure the quality of a solution: minimizing the maximum load, maximizing the minimum load and minimizing the difference of the maximum and the minimum loads. This defines optimization problems called respectively min-M, max-m and min-R. In the case where the graph of the network is a tree, it is known that the problem of building a valid configuration is polynomial. We show the three optimization variants have distinct properties regarding the theoretical complexity and the approximability. Particularly, we show that min-M is polynomial, that max-m is NP-Hard but belongs to the class FPTAS and that min-R is NP-Hard, cannot 1 be approximated to within any exponential relative ratio but, for any $\epsilon$ > 0, there exists an algorithm for which the value of the returned solution equals the value of an optimal solution shifted by at most $\epsilon$.
[ { "created": "Mon, 20 Nov 2023 12:22:26 GMT", "version": "v1" } ]
2023-11-21
[ [ "Barth", "Dominique", "", "DAVID, UVSQ" ], [ "Mautor", "Thierry", "", "DAVID, UVSQ" ], [ "Watel", "Dimitri", "", "SAMOVAR, ENSIIE" ], [ "Weisser", "Marc-Antoine", "", "LISN, GALaC" ] ]
We address the problem of configuring a power distribution network with reliability and resilience objectives by satisfying the demands of the consumers and saturating each production source as little as possible. We consider power distribution networks containing source nodes producing electricity, nodes representing electricity consumers and switches between them. Configuring this network consists in deciding the orientation of the links between the nodes of the network. The electric flow is a direct consequence of the chosen configuration and can be computed in polynomial time. It is valid if it satisfies the demand of each consumer and capacity constraints on the network. In such a case, we study the problem of determining a feasible solution that balances the loads of the sources, that is their production rates. We use three metrics to measure the quality of a solution: minimizing the maximum load, maximizing the minimum load and minimizing the difference of the maximum and the minimum loads. This defines optimization problems called respectively min-M, max-m and min-R. In the case where the graph of the network is a tree, it is known that the problem of building a valid configuration is polynomial. We show the three optimization variants have distinct properties regarding the theoretical complexity and the approximability. Particularly, we show that min-M is polynomial, that max-m is NP-Hard but belongs to the class FPTAS and that min-R is NP-Hard, cannot 1 be approximated to within any exponential relative ratio but, for any $\epsilon$ > 0, there exists an algorithm for which the value of the returned solution equals the value of an optimal solution shifted by at most $\epsilon$.
2301.10608
Arlindo Oliveira L
Tiago Oliveira, Tiago Marques, Arlindo L. Oliveira
Connecting metrics for shape-texture knowledge in computer vision
7 pages, 3 figures
null
null
null
cs.CV cs.LG
http://creativecommons.org/licenses/by/4.0/
Modern artificial neural networks, including convolutional neural networks and vision transformers, have mastered several computer vision tasks, including object recognition. However, there are many significant differences between the behavior and robustness of these systems and of the human visual system. Deep neural networks remain brittle and susceptible to many changes in the image that do not cause humans to misclassify images. Part of this different behavior may be explained by the type of features humans and deep neural networks use in vision tasks. Humans tend to classify objects according to their shape while deep neural networks seem to rely mostly on texture. Exploring this question is relevant, since it may lead to better performing neural network architectures and to a better understanding of the workings of the vision system of primates. In this work, we advance the state of the art in our understanding of this phenomenon, by extending previous analyses to a much larger set of deep neural network architectures. We found that the performance of models in image classification tasks is highly correlated with their shape bias measured at the output and penultimate layer. Furthermore, our results showed that the number of neurons that represent shape and texture are strongly anti-correlated, thus providing evidence that there is competition between these two types of features. Finally, we observed that while in general there is a correlation between performance and shape bias, there are significant variations between architecture families.
[ { "created": "Wed, 25 Jan 2023 14:37:42 GMT", "version": "v1" } ]
2023-01-26
[ [ "Oliveira", "Tiago", "" ], [ "Marques", "Tiago", "" ], [ "Oliveira", "Arlindo L.", "" ] ]
Modern artificial neural networks, including convolutional neural networks and vision transformers, have mastered several computer vision tasks, including object recognition. However, there are many significant differences between the behavior and robustness of these systems and of the human visual system. Deep neural networks remain brittle and susceptible to many changes in the image that do not cause humans to misclassify images. Part of this different behavior may be explained by the type of features humans and deep neural networks use in vision tasks. Humans tend to classify objects according to their shape while deep neural networks seem to rely mostly on texture. Exploring this question is relevant, since it may lead to better performing neural network architectures and to a better understanding of the workings of the vision system of primates. In this work, we advance the state of the art in our understanding of this phenomenon, by extending previous analyses to a much larger set of deep neural network architectures. We found that the performance of models in image classification tasks is highly correlated with their shape bias measured at the output and penultimate layer. Furthermore, our results showed that the number of neurons that represent shape and texture are strongly anti-correlated, thus providing evidence that there is competition between these two types of features. Finally, we observed that while in general there is a correlation between performance and shape bias, there are significant variations between architecture families.
2302.06354
Gal Kaplun
Gal Kaplun, Andrey Gurevich, Tal Swisa, Mazor David, Shai Shalev-Shwartz and Eran Malach
Less is More: Selective Layer Finetuning with SubTuning
null
null
null
null
cs.LG cs.AI
http://creativecommons.org/licenses/by-nc-nd/4.0/
Finetuning a pretrained model has become a standard approach for training neural networks on novel tasks, resulting in fast convergence and improved performance. In this work, we study an alternative finetuning method, where instead of finetuning all the weights of the network, we only train a carefully chosen subset of layers, keeping the rest of the weights frozen at their initial (pretrained) values. We demonstrate that \emph{subset finetuning} (or SubTuning) often achieves accuracy comparable to full finetuning of the model, and even surpasses the performance of full finetuning when training data is scarce. Therefore, SubTuning allows deploying new tasks at minimal computational cost, while enjoying the benefits of finetuning the entire model. This yields a simple and effective method for multi-task learning, where different tasks do not interfere with one another, and yet share most of the resources at inference time. We demonstrate the efficiency of SubTuning across multiple tasks, using different network architectures and pretraining methods.
[ { "created": "Mon, 13 Feb 2023 13:38:46 GMT", "version": "v1" }, { "created": "Tue, 14 Feb 2023 02:03:11 GMT", "version": "v2" }, { "created": "Sun, 2 Jul 2023 12:28:46 GMT", "version": "v3" } ]
2023-07-04
[ [ "Kaplun", "Gal", "" ], [ "Gurevich", "Andrey", "" ], [ "Swisa", "Tal", "" ], [ "David", "Mazor", "" ], [ "Shalev-Shwartz", "Shai", "" ], [ "Malach", "Eran", "" ] ]
Finetuning a pretrained model has become a standard approach for training neural networks on novel tasks, resulting in fast convergence and improved performance. In this work, we study an alternative finetuning method, where instead of finetuning all the weights of the network, we only train a carefully chosen subset of layers, keeping the rest of the weights frozen at their initial (pretrained) values. We demonstrate that \emph{subset finetuning} (or SubTuning) often achieves accuracy comparable to full finetuning of the model, and even surpasses the performance of full finetuning when training data is scarce. Therefore, SubTuning allows deploying new tasks at minimal computational cost, while enjoying the benefits of finetuning the entire model. This yields a simple and effective method for multi-task learning, where different tasks do not interfere with one another, and yet share most of the resources at inference time. We demonstrate the efficiency of SubTuning across multiple tasks, using different network architectures and pretraining methods.
2304.02444
P\'eter Antal
P\'eter Antal, Tam\'as P\'eni, and Roland T\'oth
Autonomous Hook-Based Grasping and Transportation with Quadcopters
null
null
null
null
cs.RO cs.SY eess.SY
http://creativecommons.org/licenses/by-nc-sa/4.0/
Payload grasping and transportation with quadcopters is an active research area that has rapidly developed over the last decade. To grasp a payload without human interaction, most state-of-the-art approaches apply robotic arms that are attached to the quadcopter body. However, due to the large weight and power consumption of these aerial manipulators, their agility and flight time are limited. This paper proposes a motion control and planning method for transportation with a lightweight, passive manipulator structure that consists of a hook attached to a quadrotor using a 1 DoF revolute joint. To perform payload grasping, transportation, and release, first, time-optimal reference trajectories are designed through specific waypoints to ensure the fast and reliable execution of the tasks. Then, a two-stage motion control approach is developed based on a robust geometric controller for precise and reliable reference tracking and a linear--quadratic payload regulator for rapid setpoint stabilization of the payload swing. Furthermore, stability of the closed-loop system is mathematically proven to give safety guarantee for its operation. The proposed control architecture and design are evaluated in a high-fidelity physical simulator, and also in real flight experiments, using a custom-made quadrotor--hook manipulator platform.
[ { "created": "Wed, 5 Apr 2023 14:02:53 GMT", "version": "v1" }, { "created": "Tue, 26 Mar 2024 08:13:01 GMT", "version": "v2" } ]
2024-03-27
[ [ "Antal", "Péter", "" ], [ "Péni", "Tamás", "" ], [ "Tóth", "Roland", "" ] ]
Payload grasping and transportation with quadcopters is an active research area that has rapidly developed over the last decade. To grasp a payload without human interaction, most state-of-the-art approaches apply robotic arms that are attached to the quadcopter body. However, due to the large weight and power consumption of these aerial manipulators, their agility and flight time are limited. This paper proposes a motion control and planning method for transportation with a lightweight, passive manipulator structure that consists of a hook attached to a quadrotor using a 1 DoF revolute joint. To perform payload grasping, transportation, and release, first, time-optimal reference trajectories are designed through specific waypoints to ensure the fast and reliable execution of the tasks. Then, a two-stage motion control approach is developed based on a robust geometric controller for precise and reliable reference tracking and a linear--quadratic payload regulator for rapid setpoint stabilization of the payload swing. Furthermore, stability of the closed-loop system is mathematically proven to give safety guarantee for its operation. The proposed control architecture and design are evaluated in a high-fidelity physical simulator, and also in real flight experiments, using a custom-made quadrotor--hook manipulator platform.
2103.03434
Ojas Kanhere
O. Kanhere, A. Chopra, A. Thornburg, T. S. Rappaport, and S. S. Ghassemzadeh
Performance Impact Analysis of Beam Switching in Millimeter Wave Vehicular Communications
IEEE 93rd Vehicular Technology Conference (VTC-Spring)
null
null
null
cs.IT math.IT
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Millimeter wave wireless spectrum deployments will allow vehicular communications to share high data rate vehicular sensor data in real-time. The highly directional nature of wireless links in millimeter spectral bands will require continuous channel measurements to ensure the transmitter (TX) and receiver (RX) beams are aligned to provide the best channel. Using real-world vehicular mmWave measurement data at 28 GHz, we determine the optimal beam sweeping period, i.e. the frequency of the channel measurements, to align the RX beams to the best channel directions for maximizing the vehicle-to-infrastructure (V2I) throughput. We show that in a realistic vehicular traffic environment in Austin, TX, for a vehicle traveling at an average speed of 10.5 mph, a beam sweeping period of 300 ms in future V2I communication standards would maximize the V2I throughput, using a system of four RX phased arrays that scanned the channel 360 degrees in the azimuth and 30 degrees above and below the boresight. We also investigate the impact of the number of active RX chains controlling the steerable phased arrays on V2I throughput. Reducing the number of RX chains controlling the phased arrays helps reduce the cost of the vehicular mmWave hardware while multiple RX chains, although more expensive, provide more robustness to beam direction changes at the vehicle, allowing near maximum throughput over a wide range of beam sweep periods. We show that the overhead of utilizing one RX chain instead of four leads to a 10% drop in mean V2I throughput over six non-line-of-sight runs in real traffic conditions, with each run being 10 to 20 seconds long over a distance of 40 to 90 meters.
[ { "created": "Fri, 5 Mar 2021 02:16:01 GMT", "version": "v1" } ]
2021-03-08
[ [ "Kanhere", "O.", "" ], [ "Chopra", "A.", "" ], [ "Thornburg", "A.", "" ], [ "Rappaport", "T. S.", "" ], [ "Ghassemzadeh", "S. S.", "" ] ]
Millimeter wave wireless spectrum deployments will allow vehicular communications to share high data rate vehicular sensor data in real-time. The highly directional nature of wireless links in millimeter spectral bands will require continuous channel measurements to ensure the transmitter (TX) and receiver (RX) beams are aligned to provide the best channel. Using real-world vehicular mmWave measurement data at 28 GHz, we determine the optimal beam sweeping period, i.e. the frequency of the channel measurements, to align the RX beams to the best channel directions for maximizing the vehicle-to-infrastructure (V2I) throughput. We show that in a realistic vehicular traffic environment in Austin, TX, for a vehicle traveling at an average speed of 10.5 mph, a beam sweeping period of 300 ms in future V2I communication standards would maximize the V2I throughput, using a system of four RX phased arrays that scanned the channel 360 degrees in the azimuth and 30 degrees above and below the boresight. We also investigate the impact of the number of active RX chains controlling the steerable phased arrays on V2I throughput. Reducing the number of RX chains controlling the phased arrays helps reduce the cost of the vehicular mmWave hardware while multiple RX chains, although more expensive, provide more robustness to beam direction changes at the vehicle, allowing near maximum throughput over a wide range of beam sweep periods. We show that the overhead of utilizing one RX chain instead of four leads to a 10% drop in mean V2I throughput over six non-line-of-sight runs in real traffic conditions, with each run being 10 to 20 seconds long over a distance of 40 to 90 meters.
2310.13343
Yuxuan Zhao
Xiaoliang Chen, Liangbin Li, Le Chang, Yunhe Huang, Yuxuan Zhao, Yuxiao Zhang, Dinuo Li
Challenges and Contributing Factors in the Utilization of Large Language Models (LLMs)
null
null
null
null
cs.CL cs.AI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
With the development of large language models (LLMs) like the GPT series, their widespread use across various application scenarios presents a myriad of challenges. This review initially explores the issue of domain specificity, where LLMs may struggle to provide precise answers to specialized questions within niche fields. The problem of knowledge forgetting arises as these LLMs might find it hard to balance old and new information. The knowledge repetition phenomenon reveals that sometimes LLMs might deliver overly mechanized responses, lacking depth and originality. Furthermore, knowledge illusion describes situations where LLMs might provide answers that seem insightful but are actually superficial, while knowledge toxicity focuses on harmful or biased information outputs. These challenges underscore problems in the training data and algorithmic design of LLMs. To address these issues, it's suggested to diversify training data, fine-tune models, enhance transparency and interpretability, and incorporate ethics and fairness training. Future technological trends might lean towards iterative methodologies, multimodal learning, model personalization and customization, and real-time learning and feedback mechanisms. In conclusion, future LLMs should prioritize fairness, transparency, and ethics, ensuring they uphold high moral and ethical standards when serving humanity.
[ { "created": "Fri, 20 Oct 2023 08:13:36 GMT", "version": "v1" } ]
2023-10-23
[ [ "Chen", "Xiaoliang", "" ], [ "Li", "Liangbin", "" ], [ "Chang", "Le", "" ], [ "Huang", "Yunhe", "" ], [ "Zhao", "Yuxuan", "" ], [ "Zhang", "Yuxiao", "" ], [ "Li", "Dinuo", "" ] ]
With the development of large language models (LLMs) like the GPT series, their widespread use across various application scenarios presents a myriad of challenges. This review initially explores the issue of domain specificity, where LLMs may struggle to provide precise answers to specialized questions within niche fields. The problem of knowledge forgetting arises as these LLMs might find it hard to balance old and new information. The knowledge repetition phenomenon reveals that sometimes LLMs might deliver overly mechanized responses, lacking depth and originality. Furthermore, knowledge illusion describes situations where LLMs might provide answers that seem insightful but are actually superficial, while knowledge toxicity focuses on harmful or biased information outputs. These challenges underscore problems in the training data and algorithmic design of LLMs. To address these issues, it's suggested to diversify training data, fine-tune models, enhance transparency and interpretability, and incorporate ethics and fairness training. Future technological trends might lean towards iterative methodologies, multimodal learning, model personalization and customization, and real-time learning and feedback mechanisms. In conclusion, future LLMs should prioritize fairness, transparency, and ethics, ensuring they uphold high moral and ethical standards when serving humanity.
2004.03153
Yang Zhang
Yang Zhang, Changhui Hu, Xiaobo Lu
Adaptive Multiscale Illumination-Invariant Feature Representation for Undersampled Face Recognition
null
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
This paper presents an novel illumination-invariant feature representation approach used to eliminate the varying illumination affection in undersampled face recognition. Firstly, a new illumination level classification technique based on Singular Value Decomposition (SVD) is proposed to judge the illumination level of input image. Secondly, we construct the logarithm edgemaps feature (LEF) based on lambertian model and local near neighbor feature of the face image, applying to local region within multiple scales. Then, the illumination level is referenced to construct the high performance LEF as well realize adaptive fusion for multiple scales LEFs for the face image, performing JLEF-feature. In addition, the constrain operation is used to remove the useless high-frequency interference, disentangling useful facial feature edges and constructing AJLEF-face. Finally, the effects of the our methods and other state-of-the-art algorithms including deep learning methods are tested on Extended Yale B, CMU PIE, AR as well as our Self-build Driver database (SDB). The experimental results demonstrate that the JLEF-feature and AJLEF-face outperform other related approaches for undersampled face recognition under varying illumination.
[ { "created": "Tue, 7 Apr 2020 06:48:44 GMT", "version": "v1" } ]
2020-04-08
[ [ "Zhang", "Yang", "" ], [ "Hu", "Changhui", "" ], [ "Lu", "Xiaobo", "" ] ]
This paper presents an novel illumination-invariant feature representation approach used to eliminate the varying illumination affection in undersampled face recognition. Firstly, a new illumination level classification technique based on Singular Value Decomposition (SVD) is proposed to judge the illumination level of input image. Secondly, we construct the logarithm edgemaps feature (LEF) based on lambertian model and local near neighbor feature of the face image, applying to local region within multiple scales. Then, the illumination level is referenced to construct the high performance LEF as well realize adaptive fusion for multiple scales LEFs for the face image, performing JLEF-feature. In addition, the constrain operation is used to remove the useless high-frequency interference, disentangling useful facial feature edges and constructing AJLEF-face. Finally, the effects of the our methods and other state-of-the-art algorithms including deep learning methods are tested on Extended Yale B, CMU PIE, AR as well as our Self-build Driver database (SDB). The experimental results demonstrate that the JLEF-feature and AJLEF-face outperform other related approaches for undersampled face recognition under varying illumination.
1912.07747
William Hsu
Huichen Yang, Carlos A. Aguirre, Maria F. De La Torre, Derek Christensen, Luis Bobadilla, Emily Davich, Jordan Roth, Lei Luo, Yihong Theis, Alice Lam, T. Yong-Jin Han, David Buttler, William H. Hsu
Pipelines for Procedural Information Extraction from Scientific Literature: Towards Recipes using Machine Learning and Data Science
15th International Conference on Document Analysis and Recognition Workshops (ICDARW 2019)
null
10.1109/ICDARW.2019.10037
2019-1
cs.IR cs.CL cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
This paper describes a machine learning and data science pipeline for structured information extraction from documents, implemented as a suite of open-source tools and extensions to existing tools. It centers around a methodology for extracting procedural information in the form of recipes, stepwise procedures for creating an artifact (in this case synthesizing a nanomaterial), from published scientific literature. From our overall goal of producing recipes from free text, we derive the technical objectives of a system consisting of pipeline stages: document acquisition and filtering, payload extraction, recipe step extraction as a relationship extraction task, recipe assembly, and presentation through an information retrieval interface with question answering (QA) functionality. This system meets computational information and knowledge management (CIKM) requirements of metadata-driven payload extraction, named entity extraction, and relationship extraction from text. Functional contributions described in this paper include semi-supervised machine learning methods for PDF filtering and payload extraction tasks, followed by structured extraction and data transformation tasks beginning with section extraction, recipe steps as information tuples, and finally assembled recipes. Measurable objective criteria for extraction quality include precision and recall of recipe steps, ordering constraints, and QA accuracy, precision, and recall. Results, key novel contributions, and significant open problems derived from this work center around the attribution of these holistic quality measures to specific machine learning and inference stages of the pipeline, each with their performance measures. The desired recipes contain identified preconditions, material inputs, and operations, and constitute the overall output generated by our computational information and knowledge management (CIKM) system.
[ { "created": "Mon, 16 Dec 2019 23:04:03 GMT", "version": "v1" } ]
2019-12-18
[ [ "Yang", "Huichen", "" ], [ "Aguirre", "Carlos A.", "" ], [ "De La Torre", "Maria F.", "" ], [ "Christensen", "Derek", "" ], [ "Bobadilla", "Luis", "" ], [ "Davich", "Emily", "" ], [ "Roth", "Jordan", "" ], [ "Luo", "Lei", "" ], [ "Theis", "Yihong", "" ], [ "Lam", "Alice", "" ], [ "Han", "T. Yong-Jin", "" ], [ "Buttler", "David", "" ], [ "Hsu", "William H.", "" ] ]
This paper describes a machine learning and data science pipeline for structured information extraction from documents, implemented as a suite of open-source tools and extensions to existing tools. It centers around a methodology for extracting procedural information in the form of recipes, stepwise procedures for creating an artifact (in this case synthesizing a nanomaterial), from published scientific literature. From our overall goal of producing recipes from free text, we derive the technical objectives of a system consisting of pipeline stages: document acquisition and filtering, payload extraction, recipe step extraction as a relationship extraction task, recipe assembly, and presentation through an information retrieval interface with question answering (QA) functionality. This system meets computational information and knowledge management (CIKM) requirements of metadata-driven payload extraction, named entity extraction, and relationship extraction from text. Functional contributions described in this paper include semi-supervised machine learning methods for PDF filtering and payload extraction tasks, followed by structured extraction and data transformation tasks beginning with section extraction, recipe steps as information tuples, and finally assembled recipes. Measurable objective criteria for extraction quality include precision and recall of recipe steps, ordering constraints, and QA accuracy, precision, and recall. Results, key novel contributions, and significant open problems derived from this work center around the attribution of these holistic quality measures to specific machine learning and inference stages of the pipeline, each with their performance measures. The desired recipes contain identified preconditions, material inputs, and operations, and constitute the overall output generated by our computational information and knowledge management (CIKM) system.
2302.12458
Hoi Man Lam
Hoi Man Lam, W. Jared Walker, Lucas Jonasch, Dimitri Schreiber, and Michael C. Yip
Design and Mechanics of Cable-Driven Rolling Diaphragm Transmission for High-Transparency Robotic Motion
7 pages, 13 figures
null
null
null
cs.RO cs.SY eess.SY
http://creativecommons.org/licenses/by/4.0/
Applications of rolling diaphragm transmissions for medical and teleoperated robotics are of great interest, due to the low friction of rolling diaphragms combined with the power density and stiffness of hydraulic transmissions. However, the stiffness-enabling pressure preloads can form a tradeoff against bearing loading in some rolling diaphragm layouts, and transmission setup can be difficult. Utilization of cable drives compliment the rolling diaphragm transmission's advantages, but maintaining cable tension is crucial for optimal and consistent performance. In this paper, a coaxial opposed rolling diaphragm layout with cable drive and an electronic transmission control system are investigated, with a focus on system reliability and scalability. Mechanical features are proposed which enable force balancing, decoupling of transmission pressure from bearing loads, and maintenance of cable tension. Key considerations and procedures for automation of transmission setup, phasing, and operation are also presented. We also present an analysis of system stiffness to identify key compliance contributors, and conduct experiments to validate prototype design performance.
[ { "created": "Fri, 24 Feb 2023 05:18:00 GMT", "version": "v1" } ]
2023-02-27
[ [ "Lam", "Hoi Man", "" ], [ "Walker", "W. Jared", "" ], [ "Jonasch", "Lucas", "" ], [ "Schreiber", "Dimitri", "" ], [ "Yip", "Michael C.", "" ] ]
Applications of rolling diaphragm transmissions for medical and teleoperated robotics are of great interest, due to the low friction of rolling diaphragms combined with the power density and stiffness of hydraulic transmissions. However, the stiffness-enabling pressure preloads can form a tradeoff against bearing loading in some rolling diaphragm layouts, and transmission setup can be difficult. Utilization of cable drives compliment the rolling diaphragm transmission's advantages, but maintaining cable tension is crucial for optimal and consistent performance. In this paper, a coaxial opposed rolling diaphragm layout with cable drive and an electronic transmission control system are investigated, with a focus on system reliability and scalability. Mechanical features are proposed which enable force balancing, decoupling of transmission pressure from bearing loads, and maintenance of cable tension. Key considerations and procedures for automation of transmission setup, phasing, and operation are also presented. We also present an analysis of system stiffness to identify key compliance contributors, and conduct experiments to validate prototype design performance.
1211.1265
Emmanuel d'Angelo
Emmanuel d'Angelo, Laurent jacques, Alexandre Alahi, Pierre Vandergheynst
From Bits to Images: Inversion of Local Binary Descriptors
null
null
null
null
cs.CV cs.IT math.IT
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Local Binary Descriptors are becoming more and more popular for image matching tasks, especially when going mobile. While they are extensively studied in this context, their ability to carry enough information in order to infer the original image is seldom addressed. In this work, we leverage an inverse problem approach to show that it is possible to directly reconstruct the image content from Local Binary Descriptors. This process relies on very broad assumptions besides the knowledge of the pattern of the descriptor at hand. This generalizes previous results that required either a prior learning database or non-binarized features. Furthermore, our reconstruction scheme reveals differences in the way different Local Binary Descriptors capture and encode image information. Hence, the potential applications of our work are multiple, ranging from privacy issues caused by eavesdropping image keypoints streamed by mobile devices to the design of better descriptors through the visualization and the analysis of their geometric content.
[ { "created": "Tue, 6 Nov 2012 15:32:34 GMT", "version": "v1" } ]
2012-11-07
[ [ "d'Angelo", "Emmanuel", "" ], [ "jacques", "Laurent", "" ], [ "Alahi", "Alexandre", "" ], [ "Vandergheynst", "Pierre", "" ] ]
Local Binary Descriptors are becoming more and more popular for image matching tasks, especially when going mobile. While they are extensively studied in this context, their ability to carry enough information in order to infer the original image is seldom addressed. In this work, we leverage an inverse problem approach to show that it is possible to directly reconstruct the image content from Local Binary Descriptors. This process relies on very broad assumptions besides the knowledge of the pattern of the descriptor at hand. This generalizes previous results that required either a prior learning database or non-binarized features. Furthermore, our reconstruction scheme reveals differences in the way different Local Binary Descriptors capture and encode image information. Hence, the potential applications of our work are multiple, ranging from privacy issues caused by eavesdropping image keypoints streamed by mobile devices to the design of better descriptors through the visualization and the analysis of their geometric content.
1403.2431
Gabriele Fici
Gabriele Fici, Travis Gagie, Juha K\"arkk\"ainen, Dominik Kempa
A Subquadratic Algorithm for Minimum Palindromic Factorization
Accepted for publication in Journal of Discrete Algorithms
null
10.1016/j.jda.2014.08.001
null
cs.DS cs.DM
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We give an $\mathcal{O}(n \log n)$-time, $\mathcal{O}(n)$-space algorithm for factoring a string into the minimum number of palindromic substrings. That is, given a string $S [1..n]$, in $\mathcal{O}(n \log n)$ time our algorithm returns the minimum number of palindromes $S_1,\ldots, S_\ell$ such that $S = S_1 \cdots S_\ell$. We also show that the time complexity is $\mathcal{O}(n)$ on average and $\Omega(n\log n)$ in the worst case. The last result is based on a characterization of the palindromic structure of Zimin words.
[ { "created": "Mon, 10 Mar 2014 22:18:40 GMT", "version": "v1" }, { "created": "Thu, 7 Aug 2014 09:52:23 GMT", "version": "v2" } ]
2020-12-15
[ [ "Fici", "Gabriele", "" ], [ "Gagie", "Travis", "" ], [ "Kärkkäinen", "Juha", "" ], [ "Kempa", "Dominik", "" ] ]
We give an $\mathcal{O}(n \log n)$-time, $\mathcal{O}(n)$-space algorithm for factoring a string into the minimum number of palindromic substrings. That is, given a string $S [1..n]$, in $\mathcal{O}(n \log n)$ time our algorithm returns the minimum number of palindromes $S_1,\ldots, S_\ell$ such that $S = S_1 \cdots S_\ell$. We also show that the time complexity is $\mathcal{O}(n)$ on average and $\Omega(n\log n)$ in the worst case. The last result is based on a characterization of the palindromic structure of Zimin words.
1904.03828
Dacheng Tao
Chen Gong, Dacheng Tao, Xiaojun Chang, Jian Yang
Ensemble Teaching for Hybrid Label Propagation
null
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Label propagation aims to iteratively diffuse the label information from labeled examples to unlabeled examples over a similarity graph. Current label propagation algorithms cannot consistently yield satisfactory performance due to two reasons: one is the instability of single propagation method in dealing with various practical data, and the other one is the improper propagation sequence ignoring the labeling difficulties of different examples. To remedy above defects, this paper proposes a novel propagation algorithm called hybrid diffusion under ensemble teaching (HyDEnT). Specifically, HyDEnT integrates multiple propagation methods as base learners to fully exploit their individual wisdom, which helps HyDEnT to be stable and obtain consistent encouraging results. More importantly, HyDEnT conducts propagation under the guidance of an ensemble of teachers. That is to say, in every propagation round the simplest curriculum examples are wisely designated by a teaching algorithm, so that their labels can be reliably and accurately decided by the learners. To optimally choose these simplest examples, every teacher in the ensemble should comprehensively consider the examples' difficulties from its own viewpoint, as well as the common knowledge shared by all the teachers. This is accomplished by a designed optimization problem, which can be efficiently solved via the block coordinate descent method. Thanks to the efforts of the teachers, all the unlabeled examples are logically propagated from simple to difficult, leading to better propagation quality of HyDEnT than the existing methods.
[ { "created": "Mon, 8 Apr 2019 04:10:40 GMT", "version": "v1" } ]
2019-04-09
[ [ "Gong", "Chen", "" ], [ "Tao", "Dacheng", "" ], [ "Chang", "Xiaojun", "" ], [ "Yang", "Jian", "" ] ]
Label propagation aims to iteratively diffuse the label information from labeled examples to unlabeled examples over a similarity graph. Current label propagation algorithms cannot consistently yield satisfactory performance due to two reasons: one is the instability of single propagation method in dealing with various practical data, and the other one is the improper propagation sequence ignoring the labeling difficulties of different examples. To remedy above defects, this paper proposes a novel propagation algorithm called hybrid diffusion under ensemble teaching (HyDEnT). Specifically, HyDEnT integrates multiple propagation methods as base learners to fully exploit their individual wisdom, which helps HyDEnT to be stable and obtain consistent encouraging results. More importantly, HyDEnT conducts propagation under the guidance of an ensemble of teachers. That is to say, in every propagation round the simplest curriculum examples are wisely designated by a teaching algorithm, so that their labels can be reliably and accurately decided by the learners. To optimally choose these simplest examples, every teacher in the ensemble should comprehensively consider the examples' difficulties from its own viewpoint, as well as the common knowledge shared by all the teachers. This is accomplished by a designed optimization problem, which can be efficiently solved via the block coordinate descent method. Thanks to the efforts of the teachers, all the unlabeled examples are logically propagated from simple to difficult, leading to better propagation quality of HyDEnT than the existing methods.
2306.02275
Yulin He
Yulin He, Wei Chen, Yusong Tan, Siqi Wang
USD: Unknown Sensitive Detector Empowered by Decoupled Objectness and Segment Anything Model
null
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Open World Object Detection (OWOD) is a novel and challenging computer vision task that enables object detection with the ability to detect unknown objects. Existing methods typically estimate the object likelihood with an additional objectness branch, but ignore the conflict in learning objectness and classification boundaries, which oppose each other on the semantic manifold and training objective. To address this issue, we propose a simple yet effective learning strategy, namely Decoupled Objectness Learning (DOL), which divides the learning of these two boundaries into suitable decoder layers. Moreover, detecting unknown objects comprehensively requires a large amount of annotations, but labeling all unknown objects is both difficult and expensive. Therefore, we propose to take advantage of the recent Large Vision Model (LVM), specifically the Segment Anything Model (SAM), to enhance the detection of unknown objects. Nevertheless, the output results of SAM contain noise, including backgrounds and fragments, so we introduce an Auxiliary Supervision Framework (ASF) that uses a pseudo-labeling and a soft-weighting strategies to alleviate the negative impact of noise. Extensive experiments on popular benchmarks, including Pascal VOC and MS COCO, demonstrate the effectiveness of our approach. Our proposed Unknown Sensitive Detector (USD) outperforms the recent state-of-the-art methods in terms of Unknown Recall, achieving significant improvements of 14.3\%, 15.5\%, and 8.9\% on the M-OWODB, and 27.1\%, 29.1\%, and 25.1\% on the S-OWODB.
[ { "created": "Sun, 4 Jun 2023 06:42:09 GMT", "version": "v1" } ]
2023-06-06
[ [ "He", "Yulin", "" ], [ "Chen", "Wei", "" ], [ "Tan", "Yusong", "" ], [ "Wang", "Siqi", "" ] ]
Open World Object Detection (OWOD) is a novel and challenging computer vision task that enables object detection with the ability to detect unknown objects. Existing methods typically estimate the object likelihood with an additional objectness branch, but ignore the conflict in learning objectness and classification boundaries, which oppose each other on the semantic manifold and training objective. To address this issue, we propose a simple yet effective learning strategy, namely Decoupled Objectness Learning (DOL), which divides the learning of these two boundaries into suitable decoder layers. Moreover, detecting unknown objects comprehensively requires a large amount of annotations, but labeling all unknown objects is both difficult and expensive. Therefore, we propose to take advantage of the recent Large Vision Model (LVM), specifically the Segment Anything Model (SAM), to enhance the detection of unknown objects. Nevertheless, the output results of SAM contain noise, including backgrounds and fragments, so we introduce an Auxiliary Supervision Framework (ASF) that uses a pseudo-labeling and a soft-weighting strategies to alleviate the negative impact of noise. Extensive experiments on popular benchmarks, including Pascal VOC and MS COCO, demonstrate the effectiveness of our approach. Our proposed Unknown Sensitive Detector (USD) outperforms the recent state-of-the-art methods in terms of Unknown Recall, achieving significant improvements of 14.3\%, 15.5\%, and 8.9\% on the M-OWODB, and 27.1\%, 29.1\%, and 25.1\% on the S-OWODB.
cs/0408063
Alexander Haubold
Alexander Haubold, John R. Kender
Analysis and Visualization of Index Words from Audio Transcripts of Instructional Videos
2004 IEEE International Workshop on Multimedia Content-based Analysis and Retrieval; 20 pages, 8 figures, 7 tables
null
10.1109/MMSE.2004.27
null
cs.IR cs.MM
null
We introduce new techniques for extracting, analyzing, and visualizing textual contents from instructional videos of low production quality. Using Automatic Speech Recognition, approximate transcripts (H75% Word Error Rate) are obtained from the originally highly compressed videos of university courses, each comprising between 10 to 30 lectures. Text material in the form of books or papers that accompany the course are then used to filter meaningful phrases from the seemingly incoherent transcripts. The resulting index into the transcripts is tied together and visualized in 3 experimental graphs that help in understanding the overall course structure and provide a tool for localizing certain topics for indexing. We specifically discuss a Transcript Index Map, which graphically lays out key phrases for a course, a Textbook Chapter to Transcript Match, and finally a Lecture Transcript Similarity graph, which clusters semantically similar lectures. We test our methods and tools on 7 full courses with 230 hours of video and 273 transcripts. We are able to extract up to 98 unique key terms for a given transcript and up to 347 unique key terms for an entire course. The accuracy of the Textbook Chapter to Transcript Match exceeds 70% on average. The methods used can be applied to genres of video in which there are recurrent thematic words (news, sports, meetings,...)
[ { "created": "Fri, 27 Aug 2004 20:45:32 GMT", "version": "v1" } ]
2016-11-15
[ [ "Haubold", "Alexander", "" ], [ "Kender", "John R.", "" ] ]
We introduce new techniques for extracting, analyzing, and visualizing textual contents from instructional videos of low production quality. Using Automatic Speech Recognition, approximate transcripts (H75% Word Error Rate) are obtained from the originally highly compressed videos of university courses, each comprising between 10 to 30 lectures. Text material in the form of books or papers that accompany the course are then used to filter meaningful phrases from the seemingly incoherent transcripts. The resulting index into the transcripts is tied together and visualized in 3 experimental graphs that help in understanding the overall course structure and provide a tool for localizing certain topics for indexing. We specifically discuss a Transcript Index Map, which graphically lays out key phrases for a course, a Textbook Chapter to Transcript Match, and finally a Lecture Transcript Similarity graph, which clusters semantically similar lectures. We test our methods and tools on 7 full courses with 230 hours of video and 273 transcripts. We are able to extract up to 98 unique key terms for a given transcript and up to 347 unique key terms for an entire course. The accuracy of the Textbook Chapter to Transcript Match exceeds 70% on average. The methods used can be applied to genres of video in which there are recurrent thematic words (news, sports, meetings,...)
2104.01322
Wolfgang Utschick
Wolfgang Utschick, Valentina Rizzello, Michael Joham, Zhengxiang Ma, and Leonard Piazzi
Learning the CSI Recovery in FDD Systems
null
null
null
null
cs.IT math.IT
http://creativecommons.org/licenses/by/4.0/
We propose an innovative machine learning-based technique to address the problem of channel acquisition at the base station in frequency division duplex systems. In this context, the base station reconstructs the full channel state information in the downlink frequency range based on limited downlink channel state information feedback from the mobile terminal. The channel state information recovery is based on a convolutional neural network which is trained exclusively on collected channel state samples acquired in the uplink frequency domain. No acquisition of training samples in the downlink frequency range is required at all. Finally, after a detailed presentation and analysis of the proposed technique and its performance, the "transfer learning'' assumption of the convolutional neural network that is central to the proposed approach is validated with an analysis based on the maximum mean discrepancy metric.
[ { "created": "Sat, 3 Apr 2021 06:35:24 GMT", "version": "v1" }, { "created": "Sun, 3 Oct 2021 10:31:55 GMT", "version": "v2" } ]
2021-10-05
[ [ "Utschick", "Wolfgang", "" ], [ "Rizzello", "Valentina", "" ], [ "Joham", "Michael", "" ], [ "Ma", "Zhengxiang", "" ], [ "Piazzi", "Leonard", "" ] ]
We propose an innovative machine learning-based technique to address the problem of channel acquisition at the base station in frequency division duplex systems. In this context, the base station reconstructs the full channel state information in the downlink frequency range based on limited downlink channel state information feedback from the mobile terminal. The channel state information recovery is based on a convolutional neural network which is trained exclusively on collected channel state samples acquired in the uplink frequency domain. No acquisition of training samples in the downlink frequency range is required at all. Finally, after a detailed presentation and analysis of the proposed technique and its performance, the "transfer learning'' assumption of the convolutional neural network that is central to the proposed approach is validated with an analysis based on the maximum mean discrepancy metric.
1204.4111
Tobias Harks
Tobias Harks and Britta Peis
Resource Buying Games
null
null
null
null
cs.GT
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In resource buying games a set of players jointly buys a subset of a finite resource set E (e.g., machines, edges, or nodes in a digraph). The cost of a resource e depends on the number (or load) of players using e, and has to be paid completely by the players before it becomes available. Each player i needs at least one set of a predefined family S_i in 2^E to be available. Thus, resource buying games can be seen as a variant of congestion games in which the load-dependent costs of the resources can be shared arbitrarily among the players. A strategy of player i in resource buying games is a tuple consisting of one of i's desired configurations S_i together with a payment vector p_i in R^E_+ indicating how much i is willing to contribute towards the purchase of the chosen resources. In this paper, we study the existence and computational complexity of pure Nash equilibria (PNE, for short) of resource buying games. In contrast to classical congestion games for which equilibria are guaranteed to exist, the existence of equilibria in resource buying games strongly depends on the underlying structure of the S_i's and the behavior of the cost functions. We show that for marginally non-increasing cost functions, matroids are exactly the right structure to consider, and that resource buying games with marginally non-decreasing cost functions always admit a PNE.
[ { "created": "Wed, 18 Apr 2012 15:47:25 GMT", "version": "v1" } ]
2012-04-19
[ [ "Harks", "Tobias", "" ], [ "Peis", "Britta", "" ] ]
In resource buying games a set of players jointly buys a subset of a finite resource set E (e.g., machines, edges, or nodes in a digraph). The cost of a resource e depends on the number (or load) of players using e, and has to be paid completely by the players before it becomes available. Each player i needs at least one set of a predefined family S_i in 2^E to be available. Thus, resource buying games can be seen as a variant of congestion games in which the load-dependent costs of the resources can be shared arbitrarily among the players. A strategy of player i in resource buying games is a tuple consisting of one of i's desired configurations S_i together with a payment vector p_i in R^E_+ indicating how much i is willing to contribute towards the purchase of the chosen resources. In this paper, we study the existence and computational complexity of pure Nash equilibria (PNE, for short) of resource buying games. In contrast to classical congestion games for which equilibria are guaranteed to exist, the existence of equilibria in resource buying games strongly depends on the underlying structure of the S_i's and the behavior of the cost functions. We show that for marginally non-increasing cost functions, matroids are exactly the right structure to consider, and that resource buying games with marginally non-decreasing cost functions always admit a PNE.
2304.08464
Yifan Yin
Yifan Yin, Yutai Wang, Yunpu Zhang, Russell H. Taylor, and Balazs P. Vagvolgyi
Applications of Uncalibrated Image Based Visual Servoing in Micro- and Macroscale Robotics
null
null
null
null
cs.RO
http://creativecommons.org/licenses/by-nc-nd/4.0/
We present a robust markerless image based visual servoing method that enables precision robot control without hand-eye and camera calibrations in 1, 3, and 5 degrees-of-freedom. The system uses two cameras for observing the workspace and a combination of classical image processing algorithms and deep learning based methods to detect features on camera images. The only restriction on the placement of the two cameras is that relevant image features must be visible in both views. The system enables precise robot-tool to workspace interactions even when the physical setup is disturbed, for example if cameras are moved or the workspace shifts during manipulation. The usefulness of the visual servoing method is demonstrated and evaluated in two applications: in the calibration of a micro-robotic system that dissects mosquitoes for the automated production of a malaria vaccine, and a macro-scale manipulation system for fastening screws using a UR10 robot. Evaluation results indicate that our image based visual servoing method achieves human-like manipulation accuracy in challenging setups even without camera calibration.
[ { "created": "Mon, 17 Apr 2023 17:41:02 GMT", "version": "v1" } ]
2023-04-18
[ [ "Yin", "Yifan", "" ], [ "Wang", "Yutai", "" ], [ "Zhang", "Yunpu", "" ], [ "Taylor", "Russell H.", "" ], [ "Vagvolgyi", "Balazs P.", "" ] ]
We present a robust markerless image based visual servoing method that enables precision robot control without hand-eye and camera calibrations in 1, 3, and 5 degrees-of-freedom. The system uses two cameras for observing the workspace and a combination of classical image processing algorithms and deep learning based methods to detect features on camera images. The only restriction on the placement of the two cameras is that relevant image features must be visible in both views. The system enables precise robot-tool to workspace interactions even when the physical setup is disturbed, for example if cameras are moved or the workspace shifts during manipulation. The usefulness of the visual servoing method is demonstrated and evaluated in two applications: in the calibration of a micro-robotic system that dissects mosquitoes for the automated production of a malaria vaccine, and a macro-scale manipulation system for fastening screws using a UR10 robot. Evaluation results indicate that our image based visual servoing method achieves human-like manipulation accuracy in challenging setups even without camera calibration.
0705.0817
Andrea Lo Pumo
Andrea Lo Pumo
Quantum Shortest Path Netsukuku
null
null
null
null
cs.NI
null
This document describes the QSPN, the routing discovery algorithm used by Netsukuku. Through a deductive analysis the main proprieties of the QSPN are shown. Moreover, a second version of the algorithm, is presented.
[ { "created": "Sun, 6 May 2007 20:05:44 GMT", "version": "v1" } ]
2007-05-23
[ [ "Pumo", "Andrea Lo", "" ] ]
This document describes the QSPN, the routing discovery algorithm used by Netsukuku. Through a deductive analysis the main proprieties of the QSPN are shown. Moreover, a second version of the algorithm, is presented.
1810.09786
Fabian Falck
Fabian Falck, Sagar Doshi, Nico Smuts, John Lingi, Kim Rants, Petar Kormushev
Human-centered manipulation and navigation with Robot DE NIRO
In Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2018) Workshop "Towards Robots that Exhibit Manipulation Intelligence", Madrid, Spain, Oct. 1, 2018
null
null
null
cs.RO
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Social assistance robots in health and elderly care have the potential to support and ease human lives. Given the macrosocial trends of aging and long-lived populations, robotics-based care research mainly focused on helping the elderly live independently. In this paper, we introduce Robot DE NIRO, a research platform that aims to support the supporter (the caregiver) and also offers direct human-robot interaction for the care recipient. Augmented by several sensors, DE NIRO is capable of complex manipulation tasks. It reliably interacts with humans and can autonomously and swiftly navigate through dynamically changing environments. We describe preliminary experiments in a demonstrative scenario and discuss DE NIRO's design and capabilities. We put particular emphases on safe, human-centered interaction procedures implemented in both hardware and software, including collision avoidance in manipulation and navigation as well as an intuitive perception stack through speech and face recognition.
[ { "created": "Tue, 23 Oct 2018 11:30:33 GMT", "version": "v1" } ]
2018-10-24
[ [ "Falck", "Fabian", "" ], [ "Doshi", "Sagar", "" ], [ "Smuts", "Nico", "" ], [ "Lingi", "John", "" ], [ "Rants", "Kim", "" ], [ "Kormushev", "Petar", "" ] ]
Social assistance robots in health and elderly care have the potential to support and ease human lives. Given the macrosocial trends of aging and long-lived populations, robotics-based care research mainly focused on helping the elderly live independently. In this paper, we introduce Robot DE NIRO, a research platform that aims to support the supporter (the caregiver) and also offers direct human-robot interaction for the care recipient. Augmented by several sensors, DE NIRO is capable of complex manipulation tasks. It reliably interacts with humans and can autonomously and swiftly navigate through dynamically changing environments. We describe preliminary experiments in a demonstrative scenario and discuss DE NIRO's design and capabilities. We put particular emphases on safe, human-centered interaction procedures implemented in both hardware and software, including collision avoidance in manipulation and navigation as well as an intuitive perception stack through speech and face recognition.
1605.02043
Lingda Li
Lingda Li, Ari B. Hayes, Stephen A. Hackler, Eddy Z. Zhang, Mario Szegedy, Shuaiwen Leon Song
A Graph-based Model for GPU Caching Problems
Currently under submission
null
null
null
cs.DC cs.PL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Modeling data sharing in GPU programs is a challenging task because of the massive parallelism and complex data sharing patterns provided by GPU architectures. Better GPU caching efficiency can be achieved through careful task scheduling among different threads. Traditionally, in the field of parallel computing, graph partition models are used to model data communication and guide task scheduling. However, we discover that the previous methods are either inaccurate or expensive when applied to GPU programs. In this paper, we propose a novel task partition model that is accurate and gives rise to the development of fast and high quality task/data reorganization algorithms. We demonstrate the effectiveness of the proposed model by rigorous theoretical analysis of the algorithm bounds and extensive experimental analysis. The experimental results show that it achieves significant performance improvement across a representative set of GPU applications.
[ { "created": "Fri, 6 May 2016 19:12:06 GMT", "version": "v1" } ]
2016-10-04
[ [ "Li", "Lingda", "" ], [ "Hayes", "Ari B.", "" ], [ "Hackler", "Stephen A.", "" ], [ "Zhang", "Eddy Z.", "" ], [ "Szegedy", "Mario", "" ], [ "Song", "Shuaiwen Leon", "" ] ]
Modeling data sharing in GPU programs is a challenging task because of the massive parallelism and complex data sharing patterns provided by GPU architectures. Better GPU caching efficiency can be achieved through careful task scheduling among different threads. Traditionally, in the field of parallel computing, graph partition models are used to model data communication and guide task scheduling. However, we discover that the previous methods are either inaccurate or expensive when applied to GPU programs. In this paper, we propose a novel task partition model that is accurate and gives rise to the development of fast and high quality task/data reorganization algorithms. We demonstrate the effectiveness of the proposed model by rigorous theoretical analysis of the algorithm bounds and extensive experimental analysis. The experimental results show that it achieves significant performance improvement across a representative set of GPU applications.
2308.05368
Jacopo Tagliabue
Jacopo Tagliabue, Ciro Greco, Luca Bigon
Building a serverless Data Lakehouse from spare parts
Paper accepted for the Second International Workshop on Composable Data Management Systems (@ VLDB 2023)
null
null
null
cs.DB cs.DC cs.SE
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The recently proposed Data Lakehouse architecture is built on open file formats, performance, and first-class support for data transformation, BI and data science: while the vision stresses the importance of lowering the barrier for data work, existing implementations often struggle to live up to user expectations. At Bauplan, we decided to build a new serverless platform to fulfill the Lakehouse vision. Since building from scratch is a challenge unfit for a startup, we started by re-using (sometimes unconventionally) existing projects, and then investing in improving the areas that would give us the highest marginal gains for the developer experience. In this work, we review user experience, high-level architecture and tooling decisions, and conclude by sharing plans for future development.
[ { "created": "Thu, 10 Aug 2023 06:24:25 GMT", "version": "v1" } ]
2023-08-11
[ [ "Tagliabue", "Jacopo", "" ], [ "Greco", "Ciro", "" ], [ "Bigon", "Luca", "" ] ]
The recently proposed Data Lakehouse architecture is built on open file formats, performance, and first-class support for data transformation, BI and data science: while the vision stresses the importance of lowering the barrier for data work, existing implementations often struggle to live up to user expectations. At Bauplan, we decided to build a new serverless platform to fulfill the Lakehouse vision. Since building from scratch is a challenge unfit for a startup, we started by re-using (sometimes unconventionally) existing projects, and then investing in improving the areas that would give us the highest marginal gains for the developer experience. In this work, we review user experience, high-level architecture and tooling decisions, and conclude by sharing plans for future development.
2005.04864
Xingyu Chen
Xingyu Chen and Zijie Liu
The Fairness of Leximin in Allocation of Indivisible Chores
null
null
null
null
cs.GT
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The leximin solution -- which selects an allocation that maximizes the minimum utility, then the second minimum utility, and so forth -- is known to provide EFX (envy-free up to any good) fairness guarantee in some contexts when allocating indivisible goods. However, it remains unknown how fair the leximin solution is when used to allocate indivisible chores. In this paper, we demonstrate that the leximin solution can be modified to also provide compelling fairness guarantees for the allocation of indivisible chores. First, we generalize the definition of the leximin solution. Then, we show that the leximin solution finds a PROP1 (proportional up to one good) and PO (Pareto-optimal) allocation for 3 or 4 agents in the context of chores allocation with additive distinct valuations. Additionally, we prove that the leximin solution is EFX for combinations of goods and chores for agents with general but identical valuations.
[ { "created": "Mon, 11 May 2020 05:15:43 GMT", "version": "v1" } ]
2020-05-12
[ [ "Chen", "Xingyu", "" ], [ "Liu", "Zijie", "" ] ]
The leximin solution -- which selects an allocation that maximizes the minimum utility, then the second minimum utility, and so forth -- is known to provide EFX (envy-free up to any good) fairness guarantee in some contexts when allocating indivisible goods. However, it remains unknown how fair the leximin solution is when used to allocate indivisible chores. In this paper, we demonstrate that the leximin solution can be modified to also provide compelling fairness guarantees for the allocation of indivisible chores. First, we generalize the definition of the leximin solution. Then, we show that the leximin solution finds a PROP1 (proportional up to one good) and PO (Pareto-optimal) allocation for 3 or 4 agents in the context of chores allocation with additive distinct valuations. Additionally, we prove that the leximin solution is EFX for combinations of goods and chores for agents with general but identical valuations.
2204.02337
Vignesh Ram Somnath
Vignesh Ram Somnath, Charlotte Bunne, Andreas Krause
Multi-Scale Representation Learning on Proteins
Neural Information Processing Systems 2021
null
null
null
cs.LG cs.AI q-bio.BM
http://creativecommons.org/licenses/by/4.0/
Proteins are fundamental biological entities mediating key roles in cellular function and disease. This paper introduces a multi-scale graph construction of a protein -- HoloProt -- connecting surface to structure and sequence. The surface captures coarser details of the protein, while sequence as primary component and structure -- comprising secondary and tertiary components -- capture finer details. Our graph encoder then learns a multi-scale representation by allowing each level to integrate the encoding from level(s) below with the graph at that level. We test the learned representation on different tasks, (i.) ligand binding affinity (regression), and (ii.) protein function prediction (classification). On the regression task, contrary to previous methods, our model performs consistently and reliably across different dataset splits, outperforming all baselines on most splits. On the classification task, it achieves a performance close to the top-performing model while using 10x fewer parameters. To improve the memory efficiency of our construction, we segment the multiplex protein surface manifold into molecular superpixels and substitute the surface with these superpixels at little to no performance loss.
[ { "created": "Mon, 4 Apr 2022 08:29:17 GMT", "version": "v1" } ]
2022-04-06
[ [ "Somnath", "Vignesh Ram", "" ], [ "Bunne", "Charlotte", "" ], [ "Krause", "Andreas", "" ] ]
Proteins are fundamental biological entities mediating key roles in cellular function and disease. This paper introduces a multi-scale graph construction of a protein -- HoloProt -- connecting surface to structure and sequence. The surface captures coarser details of the protein, while sequence as primary component and structure -- comprising secondary and tertiary components -- capture finer details. Our graph encoder then learns a multi-scale representation by allowing each level to integrate the encoding from level(s) below with the graph at that level. We test the learned representation on different tasks, (i.) ligand binding affinity (regression), and (ii.) protein function prediction (classification). On the regression task, contrary to previous methods, our model performs consistently and reliably across different dataset splits, outperforming all baselines on most splits. On the classification task, it achieves a performance close to the top-performing model while using 10x fewer parameters. To improve the memory efficiency of our construction, we segment the multiplex protein surface manifold into molecular superpixels and substitute the surface with these superpixels at little to no performance loss.
2311.15460
Anantaa Kotal
Anantaa Kotal, Lavanya Elluri, Deepti Gupta, Varun Mandalapu and Anupam Joshi
Privacy-Preserving Data Sharing in Agriculture: Enforcing Policy Rules for Secure and Confidential Data Synthesis
null
null
10.1109/BigData59044.2023.10386276
null
cs.CR cs.AI cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Big Data empowers the farming community with the information needed to optimize resource usage, increase productivity, and enhance the sustainability of agricultural practices. The use of Big Data in farming requires the collection and analysis of data from various sources such as sensors, satellites, and farmer surveys. While Big Data can provide the farming community with valuable insights and improve efficiency, there is significant concern regarding the security of this data as well as the privacy of the participants. Privacy regulations, such as the EU GDPR, the EU Code of Conduct on agricultural data sharing by contractual agreement, and the proposed EU AI law, have been created to address the issue of data privacy and provide specific guidelines on when and how data can be shared between organizations. To make confidential agricultural data widely available for Big Data analysis without violating the privacy of the data subjects, we consider privacy-preserving methods of data sharing in agriculture. Deep learning-based synthetic data generation has been proposed for privacy-preserving data sharing. However, there is a lack of compliance with documented data privacy policies in such privacy-preserving efforts. In this study, we propose a novel framework for enforcing privacy policy rules in privacy-preserving data generation algorithms. We explore several available agricultural codes of conduct, extract knowledge related to the privacy constraints in data, and use the extracted knowledge to define privacy bounds in a privacy-preserving generative model. We use our framework to generate synthetic agricultural data and present experimental results that demonstrate the utility of the synthetic dataset in downstream tasks. We also show that our framework can evade potential threats and secure data based on applicable regulatory policy rules.
[ { "created": "Mon, 27 Nov 2023 00:12:47 GMT", "version": "v1" } ]
2024-01-29
[ [ "Kotal", "Anantaa", "" ], [ "Elluri", "Lavanya", "" ], [ "Gupta", "Deepti", "" ], [ "Mandalapu", "Varun", "" ], [ "Joshi", "Anupam", "" ] ]
Big Data empowers the farming community with the information needed to optimize resource usage, increase productivity, and enhance the sustainability of agricultural practices. The use of Big Data in farming requires the collection and analysis of data from various sources such as sensors, satellites, and farmer surveys. While Big Data can provide the farming community with valuable insights and improve efficiency, there is significant concern regarding the security of this data as well as the privacy of the participants. Privacy regulations, such as the EU GDPR, the EU Code of Conduct on agricultural data sharing by contractual agreement, and the proposed EU AI law, have been created to address the issue of data privacy and provide specific guidelines on when and how data can be shared between organizations. To make confidential agricultural data widely available for Big Data analysis without violating the privacy of the data subjects, we consider privacy-preserving methods of data sharing in agriculture. Deep learning-based synthetic data generation has been proposed for privacy-preserving data sharing. However, there is a lack of compliance with documented data privacy policies in such privacy-preserving efforts. In this study, we propose a novel framework for enforcing privacy policy rules in privacy-preserving data generation algorithms. We explore several available agricultural codes of conduct, extract knowledge related to the privacy constraints in data, and use the extracted knowledge to define privacy bounds in a privacy-preserving generative model. We use our framework to generate synthetic agricultural data and present experimental results that demonstrate the utility of the synthetic dataset in downstream tasks. We also show that our framework can evade potential threats and secure data based on applicable regulatory policy rules.
1611.02776
Daoyuan Jia
Daoyuan Jia, Yongchi Su, Chunping Li
Deep Convolutional Neural Network for 6-DOF Image Localization
will update soon
null
null
null
cs.CV
http://creativecommons.org/licenses/by/4.0/
We present an accurate and robust method for six degree of freedom image localization. There are two key-points of our method, 1. automatic immense photo synthesis and labeling from point cloud model and, 2. pose estimation with deep convolutional neural networks regression. Our model can directly regresses 6-DOF camera poses from images, accurately describing where and how it was captured. We achieved an accuracy within 1 meters and 1 degree on our out-door dataset, which covers about 2 acres on our school campus.
[ { "created": "Tue, 8 Nov 2016 23:59:16 GMT", "version": "v1" } ]
2016-11-10
[ [ "Jia", "Daoyuan", "" ], [ "Su", "Yongchi", "" ], [ "Li", "Chunping", "" ] ]
We present an accurate and robust method for six degree of freedom image localization. There are two key-points of our method, 1. automatic immense photo synthesis and labeling from point cloud model and, 2. pose estimation with deep convolutional neural networks regression. Our model can directly regresses 6-DOF camera poses from images, accurately describing where and how it was captured. We achieved an accuracy within 1 meters and 1 degree on our out-door dataset, which covers about 2 acres on our school campus.
2102.10708
Mahdi Fahmideh
Mahdi Fahmideh, Aakash Ahmed, Ali Behnaz, John Grundy, Willy Susilo
Software Engineering for Internet of Things: The Practitioner's Perspective
null
null
null
null
cs.SE
http://creativecommons.org/licenses/by/4.0/
Internet of Things based systems (IoT systems for short) are becoming increasingly popular across different industrial domains and their development is rapidly increasing to provide value-added services to end-users and citizens. Little research to date uncovers the core development process lifecycle needed for IoT systems, and thus software engineers find themselves unprepared and unfamiliar with this new genre of system development. To ameliorate this gap, we conducted a mixed quantitative and qualitative research study where we derived a conceptual process framework from the extant literature on IoT, that identifies 27 key tasks for incorporating into development processes for IoT systems. The framework was then validated by means of a survey of 127 IoT systems practitioners developers from 35 countries across 6 continents with 15 different industry backgrounds. Our research provides an understanding of the most important development process tasks and informs both software engineering practitioners and researchers of the challenges and recommendations related to the development of next generation of IoT systems.
[ { "created": "Sun, 21 Feb 2021 23:09:32 GMT", "version": "v1" }, { "created": "Wed, 7 Apr 2021 23:42:45 GMT", "version": "v2" }, { "created": "Wed, 5 May 2021 05:55:32 GMT", "version": "v3" } ]
2021-05-06
[ [ "Fahmideh", "Mahdi", "" ], [ "Ahmed", "Aakash", "" ], [ "Behnaz", "Ali", "" ], [ "Grundy", "John", "" ], [ "Susilo", "Willy", "" ] ]
Internet of Things based systems (IoT systems for short) are becoming increasingly popular across different industrial domains and their development is rapidly increasing to provide value-added services to end-users and citizens. Little research to date uncovers the core development process lifecycle needed for IoT systems, and thus software engineers find themselves unprepared and unfamiliar with this new genre of system development. To ameliorate this gap, we conducted a mixed quantitative and qualitative research study where we derived a conceptual process framework from the extant literature on IoT, that identifies 27 key tasks for incorporating into development processes for IoT systems. The framework was then validated by means of a survey of 127 IoT systems practitioners developers from 35 countries across 6 continents with 15 different industry backgrounds. Our research provides an understanding of the most important development process tasks and informs both software engineering practitioners and researchers of the challenges and recommendations related to the development of next generation of IoT systems.
1607.01383
Karim Banawan
Karim Banawan, Sennur Ulukus
MIMO Wiretap Channel under Receiver Side Power Constraints with Applications to Wireless Power Transfer and Cognitive Radio
Submitted to IEEE Transactions on Communications, September 2015. Accepted for publication, July 2016
null
10.1109/TCOMM.2016.2593739
null
cs.IT cs.CR cs.NI math.IT
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We consider the multiple-input multiple-output (MIMO) wiretap channel under a minimum receiver-side power constraint in addition to the usual maximum transmitter-side power constraint. This problem is motivated by energy harvesting communications with wireless energy transfer, where an added goal is to deliver a minimum amount of energy to a receiver in addition to delivering secure data to another receiver. In this paper, we characterize the exact secrecy capacity of the MIMO wiretap channel under transmitter and receiver-side power constraints. We first show that solving this problem is equivalent to solving the secrecy capacity of the wiretap channel under a double-sided correlation matrix constraint on the channel input. We show the converse by extending the channel enhancement technique to our case. We present two achievable schemes that achieve the secrecy capacity: the first achievable scheme uses a Gaussian codebook with a fixed mean, and the second achievable scheme uses artificial noise (or cooperative jamming) together with a Gaussian codebook. The role of the mean or the artificial noise is to enable energy transfer without sacrificing from the secure rate. This is the first instance of a channel model where either the use of a mean signal or the use of channel prefixing via artificial noise is strictly necessary for the MIMO wiretap channel. We then extend our work to consider a maximum receiver-side power constraint. This problem is motivated by cognitive radio applications, where an added goal is to decrease the received signal energy (interference temperature) at a receiver. We further extend our results to: requiring receiver-side power constraints at both receivers; considering secrecy constraints at both receivers to study broadcast channels with confidential messages; and removing the secrecy constraints to study the classical broadcast channel.
[ { "created": "Tue, 5 Jul 2016 19:49:39 GMT", "version": "v1" } ]
2016-11-17
[ [ "Banawan", "Karim", "" ], [ "Ulukus", "Sennur", "" ] ]
We consider the multiple-input multiple-output (MIMO) wiretap channel under a minimum receiver-side power constraint in addition to the usual maximum transmitter-side power constraint. This problem is motivated by energy harvesting communications with wireless energy transfer, where an added goal is to deliver a minimum amount of energy to a receiver in addition to delivering secure data to another receiver. In this paper, we characterize the exact secrecy capacity of the MIMO wiretap channel under transmitter and receiver-side power constraints. We first show that solving this problem is equivalent to solving the secrecy capacity of the wiretap channel under a double-sided correlation matrix constraint on the channel input. We show the converse by extending the channel enhancement technique to our case. We present two achievable schemes that achieve the secrecy capacity: the first achievable scheme uses a Gaussian codebook with a fixed mean, and the second achievable scheme uses artificial noise (or cooperative jamming) together with a Gaussian codebook. The role of the mean or the artificial noise is to enable energy transfer without sacrificing from the secure rate. This is the first instance of a channel model where either the use of a mean signal or the use of channel prefixing via artificial noise is strictly necessary for the MIMO wiretap channel. We then extend our work to consider a maximum receiver-side power constraint. This problem is motivated by cognitive radio applications, where an added goal is to decrease the received signal energy (interference temperature) at a receiver. We further extend our results to: requiring receiver-side power constraints at both receivers; considering secrecy constraints at both receivers to study broadcast channels with confidential messages; and removing the secrecy constraints to study the classical broadcast channel.
2204.00541
Chuhan Wu
Chuhan Wu, Fangzhao Wu, Tao Qi, Yongfeng Huang
FairRank: Fairness-aware Single-tower Ranking Framework for News Recommendation
null
null
null
null
cs.IR
http://creativecommons.org/licenses/by/4.0/
Single-tower models are widely used in the ranking stage of news recommendation to accurately rank candidate news according to their fine-grained relatedness with user interest indicated by user behaviors. However, these models can easily inherit the biases related to users' sensitive attributes (e.g., demographics) encoded in training click data, and may generate recommendation results that are unfair to users with certain attributes. In this paper, we propose FairRank, which is a fairness-aware single-tower ranking framework for news recommendation. Since candidate news selection can be biased, we propose to use a shared candidate-aware user model to match user interest with a real displayed candidate news and a random news, respectively, to learn a candidate-aware user embedding that reflects user interest in candidate news and a candidate-invariant user embedding that indicates intrinsic user interest. We apply adversarial learning to both of them to reduce the biases brought by sensitive user attributes. In addition, we use a KL loss to regularize the attribute labels inferred from the two user embeddings to be similar, which can make the model capture less candidate-aware bias information. Extensive experiments on two datasets show that FairRank can improve the fairness of various single-tower news ranking models with minor performance losses.
[ { "created": "Fri, 1 Apr 2022 16:07:31 GMT", "version": "v1" } ]
2022-04-04
[ [ "Wu", "Chuhan", "" ], [ "Wu", "Fangzhao", "" ], [ "Qi", "Tao", "" ], [ "Huang", "Yongfeng", "" ] ]
Single-tower models are widely used in the ranking stage of news recommendation to accurately rank candidate news according to their fine-grained relatedness with user interest indicated by user behaviors. However, these models can easily inherit the biases related to users' sensitive attributes (e.g., demographics) encoded in training click data, and may generate recommendation results that are unfair to users with certain attributes. In this paper, we propose FairRank, which is a fairness-aware single-tower ranking framework for news recommendation. Since candidate news selection can be biased, we propose to use a shared candidate-aware user model to match user interest with a real displayed candidate news and a random news, respectively, to learn a candidate-aware user embedding that reflects user interest in candidate news and a candidate-invariant user embedding that indicates intrinsic user interest. We apply adversarial learning to both of them to reduce the biases brought by sensitive user attributes. In addition, we use a KL loss to regularize the attribute labels inferred from the two user embeddings to be similar, which can make the model capture less candidate-aware bias information. Extensive experiments on two datasets show that FairRank can improve the fairness of various single-tower news ranking models with minor performance losses.
2402.06044
Hainiu Xu
Hainiu Xu, Runcong Zhao, Lixing Zhu, Jinhua Du, Yulan He
OpenToM: A Comprehensive Benchmark for Evaluating Theory-of-Mind Reasoning Capabilities of Large Language Models
ACL 2024
null
null
null
cs.AI cs.CL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Neural Theory-of-Mind (N-ToM), machine's ability to understand and keep track of the mental states of others, is pivotal in developing socially intelligent agents. However, prevalent N-ToM benchmarks have several shortcomings, including the presence of ambiguous and artificial narratives, absence of personality traits and preferences, a lack of questions addressing characters' psychological mental states, and limited diversity in the questions posed. In response to these issues, we construct OpenToM, a new benchmark for assessing N-ToM with (1) longer and clearer narrative stories, (2) characters with explicit personality traits, (3) actions that are triggered by character intentions, and (4) questions designed to challenge LLMs' capabilities of modeling characters' mental states of both the physical and psychological world. Using OpenToM, we reveal that state-of-the-art LLMs thrive at modeling certain aspects of mental states in the physical world but fall short when tracking characters' mental states in the psychological world.
[ { "created": "Thu, 8 Feb 2024 20:35:06 GMT", "version": "v1" }, { "created": "Wed, 14 Feb 2024 13:23:51 GMT", "version": "v2" }, { "created": "Mon, 3 Jun 2024 10:48:16 GMT", "version": "v3" } ]
2024-06-04
[ [ "Xu", "Hainiu", "" ], [ "Zhao", "Runcong", "" ], [ "Zhu", "Lixing", "" ], [ "Du", "Jinhua", "" ], [ "He", "Yulan", "" ] ]
Neural Theory-of-Mind (N-ToM), machine's ability to understand and keep track of the mental states of others, is pivotal in developing socially intelligent agents. However, prevalent N-ToM benchmarks have several shortcomings, including the presence of ambiguous and artificial narratives, absence of personality traits and preferences, a lack of questions addressing characters' psychological mental states, and limited diversity in the questions posed. In response to these issues, we construct OpenToM, a new benchmark for assessing N-ToM with (1) longer and clearer narrative stories, (2) characters with explicit personality traits, (3) actions that are triggered by character intentions, and (4) questions designed to challenge LLMs' capabilities of modeling characters' mental states of both the physical and psychological world. Using OpenToM, we reveal that state-of-the-art LLMs thrive at modeling certain aspects of mental states in the physical world but fall short when tracking characters' mental states in the psychological world.
cs/0106008
M. H. van Emden
M.H. van Emden
Computing Functional and Relational Box Consistency by Structured Propagation in Atomic Constraint Systems
Presented at the Sixth Annual Workshop of the ERCIM Working Group on Constraints. 12 pages
null
null
Univ. of Victoria Computer Science Dept Technical Report DCS-266-IR
cs.PL cs.AI
null
Box consistency has been observed to yield exponentially better performance than chaotic constraint propagation in the interval constraint system obtained by decomposing the original expression into primitive constraints. The claim was made that the improvement is due to avoiding decomposition. In this paper we argue that the improvement is due to replacing chaotic iteration by a more structured alternative. To this end we distinguish the existing notion of box consistency from relational box consistency. We show that from a computational point of view it is important to maintain the functional structure in constraint systems that are associated with a system of equations. So far, it has only been considered computationally important that constraint propagation be fair. With the additional structure of functional constraint systems, one can define and implement computationally effective, structured, truncated constraint propagations. The existing algorithm for box consistency is one such. Our results suggest that there are others worth investigating.
[ { "created": "Thu, 7 Jun 2001 14:50:40 GMT", "version": "v1" } ]
2007-05-23
[ [ "van Emden", "M. H.", "" ] ]
Box consistency has been observed to yield exponentially better performance than chaotic constraint propagation in the interval constraint system obtained by decomposing the original expression into primitive constraints. The claim was made that the improvement is due to avoiding decomposition. In this paper we argue that the improvement is due to replacing chaotic iteration by a more structured alternative. To this end we distinguish the existing notion of box consistency from relational box consistency. We show that from a computational point of view it is important to maintain the functional structure in constraint systems that are associated with a system of equations. So far, it has only been considered computationally important that constraint propagation be fair. With the additional structure of functional constraint systems, one can define and implement computationally effective, structured, truncated constraint propagations. The existing algorithm for box consistency is one such. Our results suggest that there are others worth investigating.
2307.16045
Ana-Maria Bucur
Ana-Maria Bucur, Andreea Dinc\u{a}, M\u{a}d\u{a}lina Chitez and Roxana Rogobete
Automatic Extraction of the Romanian Academic Word List: Data and Methods
null
null
null
null
cs.CL
http://creativecommons.org/licenses/by-nc-nd/4.0/
This paper presents the methodology and data used for the automatic extraction of the Romanian Academic Word List (Ro-AWL). Academic Word Lists are useful in both L2 and L1 teaching contexts. For the Romanian language, no such resource exists so far. Ro-AWL has been generated by combining methods from corpus and computational linguistics with L2 academic writing approaches. We use two types of data: (a) existing data, such as the Romanian Frequency List based on the ROMBAC corpus, and (b) self-compiled data, such as the expert academic writing corpus EXPRES. For constructing the academic word list, we follow the methodology for building the Academic Vocabulary List for the English language. The distribution of Ro-AWL features (general distribution, POS distribution) into four disciplinary datasets is in line with previous research. Ro-AWL is freely available and can be used for teaching, research and NLP applications.
[ { "created": "Sat, 29 Jul 2023 18:21:38 GMT", "version": "v1" } ]
2023-08-01
[ [ "Bucur", "Ana-Maria", "" ], [ "Dincă", "Andreea", "" ], [ "Chitez", "Mădălina", "" ], [ "Rogobete", "Roxana", "" ] ]
This paper presents the methodology and data used for the automatic extraction of the Romanian Academic Word List (Ro-AWL). Academic Word Lists are useful in both L2 and L1 teaching contexts. For the Romanian language, no such resource exists so far. Ro-AWL has been generated by combining methods from corpus and computational linguistics with L2 academic writing approaches. We use two types of data: (a) existing data, such as the Romanian Frequency List based on the ROMBAC corpus, and (b) self-compiled data, such as the expert academic writing corpus EXPRES. For constructing the academic word list, we follow the methodology for building the Academic Vocabulary List for the English language. The distribution of Ro-AWL features (general distribution, POS distribution) into four disciplinary datasets is in line with previous research. Ro-AWL is freely available and can be used for teaching, research and NLP applications.
1306.5473
Emilio Ferrara
Michael D. Conover, Clayton Davis, Emilio Ferrara, Karissa McKelvey, Filippo Menczer, Alessandro Flammini
The Geospatial Characteristics of a Social Movement Communication Network
Open access available at: http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0064679
PLoS ONE 8(3):e55957 2013
10.1371/journal.pone.0055957
null
cs.CY cs.SI physics.data-an physics.soc-ph
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Social movements rely in large measure on networked communication technologies to organize and disseminate information relating to the movements' objectives. In this work we seek to understand how the goals and needs of a protest movement are reflected in the geographic patterns of its communication network, and how these patterns differ from those of stable political communication. To this end, we examine an online communication network reconstructed from over 600,000 tweets from a thirty-six week period covering the birth and maturation of the American anticapitalist movement, Occupy Wall Street. We find that, compared to a network of stable domestic political communication, the Occupy Wall Street network exhibits higher levels of locality and a hub and spoke structure, in which the majority of non-local attention is allocated to high-profile locations such as New York, California, and Washington D.C. Moreover, we observe that information flows across state boundaries are more likely to contain framing language and references to the media, while communication among individuals in the same state is more likely to reference protest action and specific places and and times. Tying these results to social movement theory, we propose that these features reflect the movement's efforts to mobilize resources at the local level and to develop narrative frames that reinforce collective purpose at the national level.
[ { "created": "Sun, 23 Jun 2013 21:31:16 GMT", "version": "v1" } ]
2013-06-25
[ [ "Conover", "Michael D.", "" ], [ "Davis", "Clayton", "" ], [ "Ferrara", "Emilio", "" ], [ "McKelvey", "Karissa", "" ], [ "Menczer", "Filippo", "" ], [ "Flammini", "Alessandro", "" ] ]
Social movements rely in large measure on networked communication technologies to organize and disseminate information relating to the movements' objectives. In this work we seek to understand how the goals and needs of a protest movement are reflected in the geographic patterns of its communication network, and how these patterns differ from those of stable political communication. To this end, we examine an online communication network reconstructed from over 600,000 tweets from a thirty-six week period covering the birth and maturation of the American anticapitalist movement, Occupy Wall Street. We find that, compared to a network of stable domestic political communication, the Occupy Wall Street network exhibits higher levels of locality and a hub and spoke structure, in which the majority of non-local attention is allocated to high-profile locations such as New York, California, and Washington D.C. Moreover, we observe that information flows across state boundaries are more likely to contain framing language and references to the media, while communication among individuals in the same state is more likely to reference protest action and specific places and and times. Tying these results to social movement theory, we propose that these features reflect the movement's efforts to mobilize resources at the local level and to develop narrative frames that reinforce collective purpose at the national level.
1504.03363
Gabriel Fernando Pivaro
G. F. Pivaro, G. Fraindenraich
Outage Probability for Multi-Hop Full-Duplex Decode and Forward MIMO Relay
null
null
null
null
cs.IT math.IT
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
In this paper, a multi-hop (MH) decode-and-forward (DF) multiple-input multiple-output (MIMO) relay network has been studied. To consider a more realistic scenario, Full-Duplex (FD) operation with Relay Self-Interference (RSI) is employed. Assuming that the MIMO channels are subject to Rayleigh fading, a simple and compact closed-form outage probability expression has been derived. The key assumption to derive this result is that the mutual information of each channel could be well approximated by a Gaussian random variable. In order to obtain the resultant outage probability, a new excellent accurate approximation has been obtained for the sum of Wishart distributed complex random matrices. Numerical Monte Carlo simulations have been performed to validate our result. These simulations have shown that, for low and medium interference regime, FD mode performs better than Half-Duplex (HD) mode. On the other hand, when RSI increases, HD mode can outperforms FD mode.
[ { "created": "Mon, 13 Apr 2015 21:01:15 GMT", "version": "v1" } ]
2015-04-15
[ [ "Pivaro", "G. F.", "" ], [ "Fraindenraich", "G.", "" ] ]
In this paper, a multi-hop (MH) decode-and-forward (DF) multiple-input multiple-output (MIMO) relay network has been studied. To consider a more realistic scenario, Full-Duplex (FD) operation with Relay Self-Interference (RSI) is employed. Assuming that the MIMO channels are subject to Rayleigh fading, a simple and compact closed-form outage probability expression has been derived. The key assumption to derive this result is that the mutual information of each channel could be well approximated by a Gaussian random variable. In order to obtain the resultant outage probability, a new excellent accurate approximation has been obtained for the sum of Wishart distributed complex random matrices. Numerical Monte Carlo simulations have been performed to validate our result. These simulations have shown that, for low and medium interference regime, FD mode performs better than Half-Duplex (HD) mode. On the other hand, when RSI increases, HD mode can outperforms FD mode.
1411.2404
Jelani Nelson
Kasper Green Larsen, Jelani Nelson
The Johnson-Lindenstrauss lemma is optimal for linear dimensionality reduction
null
null
null
null
cs.IT cs.CG cs.DS math.FA math.IT
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
For any $n>1$ and $0<\varepsilon<1/2$, we show the existence of an $n^{O(1)}$-point subset $X$ of $\mathbb{R}^n$ such that any linear map from $(X,\ell_2)$ to $\ell_2^m$ with distortion at most $1+\varepsilon$ must have $m = \Omega(\min\{n, \varepsilon^{-2}\log n\})$. Our lower bound matches the upper bounds provided by the identity matrix and the Johnson-Lindenstrauss lemma, improving the previous lower bound of Alon by a $\log(1/\varepsilon)$ factor.
[ { "created": "Mon, 10 Nov 2014 12:53:41 GMT", "version": "v1" } ]
2014-11-11
[ [ "Larsen", "Kasper Green", "" ], [ "Nelson", "Jelani", "" ] ]
For any $n>1$ and $0<\varepsilon<1/2$, we show the existence of an $n^{O(1)}$-point subset $X$ of $\mathbb{R}^n$ such that any linear map from $(X,\ell_2)$ to $\ell_2^m$ with distortion at most $1+\varepsilon$ must have $m = \Omega(\min\{n, \varepsilon^{-2}\log n\})$. Our lower bound matches the upper bounds provided by the identity matrix and the Johnson-Lindenstrauss lemma, improving the previous lower bound of Alon by a $\log(1/\varepsilon)$ factor.
2404.01196
Sondre Wold
Sondre Wold, Petter M{\ae}hlum, Oddbj{\o}rn Hove
Estimating Lexical Complexity from Document-Level Distributions
LREC-COLING 2024
null
null
null
cs.CL
http://creativecommons.org/licenses/by/4.0/
Existing methods for complexity estimation are typically developed for entire documents. This limitation in scope makes them inapplicable for shorter pieces of text, such as health assessment tools. These typically consist of lists of independent sentences, all of which are too short for existing methods to apply. The choice of wording in these assessment tools is crucial, as both the cognitive capacity and the linguistic competency of the intended patient groups could vary substantially. As a first step towards creating better tools for supporting health practitioners, we develop a two-step approach for estimating lexical complexity that does not rely on any pre-annotated data. We implement our approach for the Norwegian language and verify its effectiveness using statistical testing and a qualitative evaluation of samples from real assessment tools. We also investigate the relationship between our complexity measure and certain features typically associated with complexity in the literature, such as word length, frequency, and the number of syllables.
[ { "created": "Mon, 1 Apr 2024 15:55:18 GMT", "version": "v1" } ]
2024-04-02
[ [ "Wold", "Sondre", "" ], [ "Mæhlum", "Petter", "" ], [ "Hove", "Oddbjørn", "" ] ]
Existing methods for complexity estimation are typically developed for entire documents. This limitation in scope makes them inapplicable for shorter pieces of text, such as health assessment tools. These typically consist of lists of independent sentences, all of which are too short for existing methods to apply. The choice of wording in these assessment tools is crucial, as both the cognitive capacity and the linguistic competency of the intended patient groups could vary substantially. As a first step towards creating better tools for supporting health practitioners, we develop a two-step approach for estimating lexical complexity that does not rely on any pre-annotated data. We implement our approach for the Norwegian language and verify its effectiveness using statistical testing and a qualitative evaluation of samples from real assessment tools. We also investigate the relationship between our complexity measure and certain features typically associated with complexity in the literature, such as word length, frequency, and the number of syllables.
cs/0605062
Al-Mukaddim Khan Pathan
Al-Mukaddim Khan Pathan and Md. Golam Shagadul Amin Talukder
QoSIP: A QoS Aware IP Routing Ptotocol for Multimedia Data
8th International Conference of Advanced Communication Technology (ICACT 2006)
null
null
null
cs.NI
null
Conventional IP routing protocols are not suitable for multimedia applications which have very stringent Quality-of-Service (QoS) demands and they require a connection oriented service. For multimedia applications it is expected that the router should be able to forward the packet according to the demand of the packet and it is necessary to find a path that satisfies the specific demands of a particular application. In order to address these issues, in this paper, we have presented a QoS aware IP routing protocol where a router stores information about the QoS parameters and routes the packet accordingly. Keywords: IP Routing Protocol, Quality of Service (QoS) parameter, QoSIP, Selective Flooding.
[ { "created": "Mon, 15 May 2006 10:39:40 GMT", "version": "v1" } ]
2007-05-23
[ [ "Pathan", "Al-Mukaddim Khan", "" ], [ "Talukder", "Md. Golam Shagadul Amin", "" ] ]
Conventional IP routing protocols are not suitable for multimedia applications which have very stringent Quality-of-Service (QoS) demands and they require a connection oriented service. For multimedia applications it is expected that the router should be able to forward the packet according to the demand of the packet and it is necessary to find a path that satisfies the specific demands of a particular application. In order to address these issues, in this paper, we have presented a QoS aware IP routing protocol where a router stores information about the QoS parameters and routes the packet accordingly. Keywords: IP Routing Protocol, Quality of Service (QoS) parameter, QoSIP, Selective Flooding.
1502.02519
Fabrizio Montesi
Fabrizio Montesi
Kickstarting Choreographic Programming
null
null
null
null
cs.PL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We present an overview of some recent efforts aimed at the development of Choreographic Programming, a programming paradigm for the production of concurrent software that is guaranteed to be correct by construction from global descriptions of communication behaviour.
[ { "created": "Mon, 9 Feb 2015 15:20:03 GMT", "version": "v1" }, { "created": "Tue, 10 Feb 2015 12:03:10 GMT", "version": "v2" } ]
2015-02-11
[ [ "Montesi", "Fabrizio", "" ] ]
We present an overview of some recent efforts aimed at the development of Choreographic Programming, a programming paradigm for the production of concurrent software that is guaranteed to be correct by construction from global descriptions of communication behaviour.
2203.06766
Van Bang Le
Sun-Yuan Hsieh, Hoang-Oanh Le, Van Bang Le, Sheng-Lung Peng
On the $d$-Claw Vertex Deletion Problem
null
null
null
null
cs.DM cs.DS
http://creativecommons.org/licenses/by/4.0/
Let $d$-claw (or $d$-star) stand for $K_{1,d}$, the complete bipartite graph with 1 and $d\ge 1$ vertices on each part. The $d$-claw vertex deletion problem, $d$-CLAW-VD, asks for a given graph $G$ and an integer $k$ if one can delete at most $k$ vertices from $G$ such that the resulting graph has no $d$-claw as an induced subgraph. Thus, 1-CLAW-VD and 2-CLAW-VD are just the famous VERTEX COVER problem and the CLUSTER VERTEX DELETION problem, respectively. In this paper, we strengthen a hardness result in [M. Yannakakis, Node-Deletion Problems on Bipartite Graphs, SIAM J. Comput. (1981)], by showing that CLUSTER VERTEX DELETION remains NP-complete when restricted to bipartite graphs of maximum degree 3. Moreover, for every $d\ge 3$, we show that $d$-CLAW-VD is NP-complete even when restricted to bipartite graphs of maximum degree $d$. These hardness results are optimal with respect to degree constraint. By extending the hardness result in [F. Bonomo-Braberman et al., Linear-Time Algorithms for Eliminating Claws in Graphs, COCOON 2020], we show that, for every $d\ge 3$, $d$-CLAW-VD is NP-complete even when restricted to split graphs without $(d+1)$-claws, and split graphs of diameter 2. On the positive side, we prove that $d$-CLAW-VD is polynomially solvable on what we call $d$-block graphs, a class properly contains all block graphs. This result extends the polynomial-time algorithm in [Y. Cao et al., Vertex deletion problems on chordal graphs, Theor. Comput. Sci. (2018)] for 2-CLAW-VD on block graphs to $d$-CLAW-VD for all $d\ge 2$ and improves the polynomial-time algorithm proposed by F. Bonomo-Brabeman et al. for (unweighted) 3-CLAW-VD on block graphs to 3-block graphs.
[ { "created": "Sun, 13 Mar 2022 21:36:48 GMT", "version": "v1" } ]
2022-03-15
[ [ "Hsieh", "Sun-Yuan", "" ], [ "Le", "Hoang-Oanh", "" ], [ "Le", "Van Bang", "" ], [ "Peng", "Sheng-Lung", "" ] ]
Let $d$-claw (or $d$-star) stand for $K_{1,d}$, the complete bipartite graph with 1 and $d\ge 1$ vertices on each part. The $d$-claw vertex deletion problem, $d$-CLAW-VD, asks for a given graph $G$ and an integer $k$ if one can delete at most $k$ vertices from $G$ such that the resulting graph has no $d$-claw as an induced subgraph. Thus, 1-CLAW-VD and 2-CLAW-VD are just the famous VERTEX COVER problem and the CLUSTER VERTEX DELETION problem, respectively. In this paper, we strengthen a hardness result in [M. Yannakakis, Node-Deletion Problems on Bipartite Graphs, SIAM J. Comput. (1981)], by showing that CLUSTER VERTEX DELETION remains NP-complete when restricted to bipartite graphs of maximum degree 3. Moreover, for every $d\ge 3$, we show that $d$-CLAW-VD is NP-complete even when restricted to bipartite graphs of maximum degree $d$. These hardness results are optimal with respect to degree constraint. By extending the hardness result in [F. Bonomo-Braberman et al., Linear-Time Algorithms for Eliminating Claws in Graphs, COCOON 2020], we show that, for every $d\ge 3$, $d$-CLAW-VD is NP-complete even when restricted to split graphs without $(d+1)$-claws, and split graphs of diameter 2. On the positive side, we prove that $d$-CLAW-VD is polynomially solvable on what we call $d$-block graphs, a class properly contains all block graphs. This result extends the polynomial-time algorithm in [Y. Cao et al., Vertex deletion problems on chordal graphs, Theor. Comput. Sci. (2018)] for 2-CLAW-VD on block graphs to $d$-CLAW-VD for all $d\ge 2$ and improves the polynomial-time algorithm proposed by F. Bonomo-Brabeman et al. for (unweighted) 3-CLAW-VD on block graphs to 3-block graphs.
2304.11241
Pierre Marza
Pierre Marza, Laetitia Matignon, Olivier Simonin, Dhruv Batra, Christian Wolf, Devendra Singh Chaplot
AutoNeRF: Training Implicit Scene Representations with Autonomous Agents
null
null
null
null
cs.CV cs.LG cs.RO
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Implicit representations such as Neural Radiance Fields (NeRF) have been shown to be very effective at novel view synthesis. However, these models typically require manual and careful human data collection for training. In this paper, we present AutoNeRF, a method to collect data required to train NeRFs using autonomous embodied agents. Our method allows an agent to explore an unseen environment efficiently and use the experience to build an implicit map representation autonomously. We compare the impact of different exploration strategies including handcrafted frontier-based exploration, end-to-end and modular approaches composed of trained high-level planners and classical low-level path followers. We train these models with different reward functions tailored to this problem and evaluate the quality of the learned representations on four different downstream tasks: classical viewpoint rendering, map reconstruction, planning, and pose refinement. Empirical results show that NeRFs can be trained on actively collected data using just a single episode of experience in an unseen environment, and can be used for several downstream robotic tasks, and that modular trained exploration models outperform other classical and end-to-end baselines. Finally, we show that AutoNeRF can reconstruct large-scale scenes, and is thus a useful tool to perform scene-specific adaptation as the produced 3D environment models can be loaded into a simulator to fine-tune a policy of interest.
[ { "created": "Fri, 21 Apr 2023 20:22:17 GMT", "version": "v1" }, { "created": "Fri, 22 Dec 2023 13:55:53 GMT", "version": "v2" } ]
2023-12-25
[ [ "Marza", "Pierre", "" ], [ "Matignon", "Laetitia", "" ], [ "Simonin", "Olivier", "" ], [ "Batra", "Dhruv", "" ], [ "Wolf", "Christian", "" ], [ "Chaplot", "Devendra Singh", "" ] ]
Implicit representations such as Neural Radiance Fields (NeRF) have been shown to be very effective at novel view synthesis. However, these models typically require manual and careful human data collection for training. In this paper, we present AutoNeRF, a method to collect data required to train NeRFs using autonomous embodied agents. Our method allows an agent to explore an unseen environment efficiently and use the experience to build an implicit map representation autonomously. We compare the impact of different exploration strategies including handcrafted frontier-based exploration, end-to-end and modular approaches composed of trained high-level planners and classical low-level path followers. We train these models with different reward functions tailored to this problem and evaluate the quality of the learned representations on four different downstream tasks: classical viewpoint rendering, map reconstruction, planning, and pose refinement. Empirical results show that NeRFs can be trained on actively collected data using just a single episode of experience in an unseen environment, and can be used for several downstream robotic tasks, and that modular trained exploration models outperform other classical and end-to-end baselines. Finally, we show that AutoNeRF can reconstruct large-scale scenes, and is thus a useful tool to perform scene-specific adaptation as the produced 3D environment models can be loaded into a simulator to fine-tune a policy of interest.
1302.6641
John Iacono
John Iacono
Why some heaps support constant-amortized-time decrease-key operations, and others do not
null
null
null
null
cs.DS
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
A lower bound is presented which shows that a class of heap algorithms in the pointer model with only heap pointers must spend Omega(log log n / log log log n) amortized time on the decrease-key operation (given O(log n) amortized-time extract-min). Intuitively, this bound shows the key to having O(1)-time decrease-key is the ability to sort O(log n) items in O(log n) time; Fibonacci heaps [M.L. Fredman and R. E. Tarjan. J. ACM 34(3):596-615 (1987)] do this through the use of bucket sort. Our lower bound also holds no matter how much data is augmented; this is in contrast to the lower bound of Fredman [J. ACM 46(4):473-501 (1999)] who showed a tradeoff between the number of augmented bits and the amortized cost of decrease-key. A new heap data structure, the sort heap, is presented. This heap is a simplification of the heap of Elmasry [SODA 2009: 471-476] and shares with it a O(log log n) amortized-time decrease-key, but with a straightforward implementation such that our lower bound holds. Thus a natural model is presented for a pointer-based heap such that the amortized runtime of a self-adjusting structure and amortized lower asymptotic bounds for decrease-key differ by but a O(log log log n) factor.
[ { "created": "Wed, 27 Feb 2013 01:52:21 GMT", "version": "v1" }, { "created": "Wed, 3 Apr 2013 22:12:24 GMT", "version": "v2" }, { "created": "Tue, 16 Jul 2013 18:50:20 GMT", "version": "v3" } ]
2013-07-17
[ [ "Iacono", "John", "" ] ]
A lower bound is presented which shows that a class of heap algorithms in the pointer model with only heap pointers must spend Omega(log log n / log log log n) amortized time on the decrease-key operation (given O(log n) amortized-time extract-min). Intuitively, this bound shows the key to having O(1)-time decrease-key is the ability to sort O(log n) items in O(log n) time; Fibonacci heaps [M.L. Fredman and R. E. Tarjan. J. ACM 34(3):596-615 (1987)] do this through the use of bucket sort. Our lower bound also holds no matter how much data is augmented; this is in contrast to the lower bound of Fredman [J. ACM 46(4):473-501 (1999)] who showed a tradeoff between the number of augmented bits and the amortized cost of decrease-key. A new heap data structure, the sort heap, is presented. This heap is a simplification of the heap of Elmasry [SODA 2009: 471-476] and shares with it a O(log log n) amortized-time decrease-key, but with a straightforward implementation such that our lower bound holds. Thus a natural model is presented for a pointer-based heap such that the amortized runtime of a self-adjusting structure and amortized lower asymptotic bounds for decrease-key differ by but a O(log log log n) factor.
2011.03667
Yunhao Yang
Yunhao Yang, Andrew Whinston
Identifying Mislabeled Images in Supervised Learning Utilizing Autoencoder
UTCS Tech Report: Honors Thesis. 12 pages, 11 figures
Lecture Notes in Networks and Systems vol 359 (2021) 266-282
10.1007/978-3-030-89880-9_21
null
cs.CV cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Supervised learning is based on the assumption that the ground truth in the training data is accurate. However, this may not be guaranteed in real-world settings. Inaccurate training data will result in some unexpected predictions. In image classification, incorrect labels may cause the classification model to be inaccurate as well. In this paper, I am going to apply unsupervised techniques to the training data before training the classification network. A convolutional autoencoder is applied to encode and reconstruct images. The encoder will project the image data on to latent space. In the latent space, image features are preserved in a lower dimension. The assumption is that data samples with similar features are likely to have the same label. Noised samples can be classified in the latent space by the Density-Base Scan (DBSCAN) clustering algorithm. These incorrectly labeled data are visualized as outliers in the latent space. Therefore, the outliers identified by the DBSCAN algorithm can be classified as incorrectly labeled samples. After the outliers are detected, all the outliers are treated as mislabeled data samples and removed from the dataset. Thus the training data can be directly used in training the supervised learning network. The algorithm can detect and remove above 67\% of mislabeled data in the experimental dataset.
[ { "created": "Sat, 7 Nov 2020 03:09:34 GMT", "version": "v1" }, { "created": "Mon, 18 Jan 2021 22:59:44 GMT", "version": "v2" } ]
2022-01-06
[ [ "Yang", "Yunhao", "" ], [ "Whinston", "Andrew", "" ] ]
Supervised learning is based on the assumption that the ground truth in the training data is accurate. However, this may not be guaranteed in real-world settings. Inaccurate training data will result in some unexpected predictions. In image classification, incorrect labels may cause the classification model to be inaccurate as well. In this paper, I am going to apply unsupervised techniques to the training data before training the classification network. A convolutional autoencoder is applied to encode and reconstruct images. The encoder will project the image data on to latent space. In the latent space, image features are preserved in a lower dimension. The assumption is that data samples with similar features are likely to have the same label. Noised samples can be classified in the latent space by the Density-Base Scan (DBSCAN) clustering algorithm. These incorrectly labeled data are visualized as outliers in the latent space. Therefore, the outliers identified by the DBSCAN algorithm can be classified as incorrectly labeled samples. After the outliers are detected, all the outliers are treated as mislabeled data samples and removed from the dataset. Thus the training data can be directly used in training the supervised learning network. The algorithm can detect and remove above 67\% of mislabeled data in the experimental dataset.
1503.05187
Khaled Fawagreh
Khaled Fawagreh, Mohamad Medhat Gaber, Eyad Elyan
An Outlier Detection-based Tree Selection Approach to Extreme Pruning of Random Forests
21 pages, 4 Figures. arXiv admin note: substantial text overlap with arXiv:1503.04996
null
null
null
cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Random Forest (RF) is an ensemble classification technique that was developed by Breiman over a decade ago. Compared with other ensemble techniques, it has proved its accuracy and superiority. Many researchers, however, believe that there is still room for enhancing and improving its performance in terms of predictive accuracy. This explains why, over the past decade, there have been many extensions of RF where each extension employed a variety of techniques and strategies to improve certain aspect(s) of RF. Since it has been proven empirically that ensembles tend to yield better results when there is a significant diversity among the constituent models, the objective of this paper is twofolds. First, it investigates how an unsupervised learning technique, namely, Local Outlier Factor (LOF) can be used to identify diverse trees in the RF. Second, trees with the highest LOF scores are then used to produce an extension of RF termed LOFB-DRF that is much smaller in size than RF, and yet performs at least as good as RF, but mostly exhibits higher performance in terms of accuracy. The latter refers to a known technique called ensemble pruning. Experimental results on 10 real datasets prove the superiority of our proposed extension over the traditional RF. Unprecedented pruning levels reaching 99% have been achieved at the time of boosting the predictive accuracy of the ensemble. The notably high pruning level makes the technique a good candidate for real-time applications.
[ { "created": "Tue, 17 Mar 2015 11:05:31 GMT", "version": "v1" } ]
2015-03-19
[ [ "Fawagreh", "Khaled", "" ], [ "Gaber", "Mohamad Medhat", "" ], [ "Elyan", "Eyad", "" ] ]
Random Forest (RF) is an ensemble classification technique that was developed by Breiman over a decade ago. Compared with other ensemble techniques, it has proved its accuracy and superiority. Many researchers, however, believe that there is still room for enhancing and improving its performance in terms of predictive accuracy. This explains why, over the past decade, there have been many extensions of RF where each extension employed a variety of techniques and strategies to improve certain aspect(s) of RF. Since it has been proven empirically that ensembles tend to yield better results when there is a significant diversity among the constituent models, the objective of this paper is twofolds. First, it investigates how an unsupervised learning technique, namely, Local Outlier Factor (LOF) can be used to identify diverse trees in the RF. Second, trees with the highest LOF scores are then used to produce an extension of RF termed LOFB-DRF that is much smaller in size than RF, and yet performs at least as good as RF, but mostly exhibits higher performance in terms of accuracy. The latter refers to a known technique called ensemble pruning. Experimental results on 10 real datasets prove the superiority of our proposed extension over the traditional RF. Unprecedented pruning levels reaching 99% have been achieved at the time of boosting the predictive accuracy of the ensemble. The notably high pruning level makes the technique a good candidate for real-time applications.
1205.6423
Plawan Kumar Rath
Plawan Kumar Rath, G. N. Anil
Proposed Challenges And Areas of Concern in Operating System Research and Development
5 pages; International Journal for Computer Science Issues(IJCSI), Volume 9, Issue 2, March 2012
null
null
null
cs.OS
http://creativecommons.org/licenses/publicdomain/
Computers are a very important part of our lives and the major reason why they have been such a success is because of the excellent graphical operating systems that run on these powerful machines. As the computer hardware is becoming more and more powerful, it is also vital to keep the software updated in order to utilize the hardware of the system efficiently and make it faster and smarter. This paper highlights some core issues that if dealt with in the operating system level would make use of the full potential of the computer hardware and provide an excellent user experience.
[ { "created": "Tue, 29 May 2012 17:10:29 GMT", "version": "v1" } ]
2012-05-30
[ [ "Rath", "Plawan Kumar", "" ], [ "Anil", "G. N.", "" ] ]
Computers are a very important part of our lives and the major reason why they have been such a success is because of the excellent graphical operating systems that run on these powerful machines. As the computer hardware is becoming more and more powerful, it is also vital to keep the software updated in order to utilize the hardware of the system efficiently and make it faster and smarter. This paper highlights some core issues that if dealt with in the operating system level would make use of the full potential of the computer hardware and provide an excellent user experience.
2104.14579
Matteo Zecchin
Matteo Zecchin, Mahdi Boloursaz Mashhadi, Mikolaj Jankowski, Deniz Gunduz, Marios Kountouris, David Gesbert
LIDAR and Position-Aided mmWave Beam Selection with Non-local CNNs and Curriculum Training
Submitted for publication
null
null
null
cs.IT cs.LG eess.SP math.IT
http://creativecommons.org/licenses/by-nc-nd/4.0/
Efficient millimeter wave (mmWave) beam selection in vehicle-to-infrastructure (V2I) communication is a crucial yet challenging task due to the narrow mmWave beamwidth and high user mobility. To reduce the search overhead of iterative beam discovery procedures, contextual information from light detection and ranging (LIDAR) sensors mounted on vehicles has been leveraged by data-driven methods to produce useful side information. In this paper, we propose a lightweight neural network (NN) architecture along with the corresponding LIDAR preprocessing, which significantly outperforms previous works. Our solution comprises multiple novelties that improve both the convergence speed and the final accuracy of the model. In particular, we define a novel loss function inspired by the knowledge distillation idea, introduce a curriculum training approach exploiting line-of-sight (LOS)/non-line-of-sight (NLOS) information, and we propose a non-local attention module to improve the performance for the more challenging NLOS cases. Simulation results on benchmark datasets show that, utilizing solely LIDAR data and the receiver position, our NN-based beam selection scheme can achieve 79.9% throughput of an exhaustive beam sweeping approach without any beam search overhead and 95% by searching among as few as 6 beams. In a typical mmWave V2I scenario, our proposed method considerably reduces the beam search time required to achieve a desired throughput, in comparison with the inverse fingerprinting and hierarchical beam selection schemes.
[ { "created": "Thu, 29 Apr 2021 18:07:31 GMT", "version": "v1" }, { "created": "Mon, 3 May 2021 12:02:06 GMT", "version": "v2" }, { "created": "Wed, 17 Nov 2021 10:16:51 GMT", "version": "v3" } ]
2021-11-18
[ [ "Zecchin", "Matteo", "" ], [ "Mashhadi", "Mahdi Boloursaz", "" ], [ "Jankowski", "Mikolaj", "" ], [ "Gunduz", "Deniz", "" ], [ "Kountouris", "Marios", "" ], [ "Gesbert", "David", "" ] ]
Efficient millimeter wave (mmWave) beam selection in vehicle-to-infrastructure (V2I) communication is a crucial yet challenging task due to the narrow mmWave beamwidth and high user mobility. To reduce the search overhead of iterative beam discovery procedures, contextual information from light detection and ranging (LIDAR) sensors mounted on vehicles has been leveraged by data-driven methods to produce useful side information. In this paper, we propose a lightweight neural network (NN) architecture along with the corresponding LIDAR preprocessing, which significantly outperforms previous works. Our solution comprises multiple novelties that improve both the convergence speed and the final accuracy of the model. In particular, we define a novel loss function inspired by the knowledge distillation idea, introduce a curriculum training approach exploiting line-of-sight (LOS)/non-line-of-sight (NLOS) information, and we propose a non-local attention module to improve the performance for the more challenging NLOS cases. Simulation results on benchmark datasets show that, utilizing solely LIDAR data and the receiver position, our NN-based beam selection scheme can achieve 79.9% throughput of an exhaustive beam sweeping approach without any beam search overhead and 95% by searching among as few as 6 beams. In a typical mmWave V2I scenario, our proposed method considerably reduces the beam search time required to achieve a desired throughput, in comparison with the inverse fingerprinting and hierarchical beam selection schemes.
2212.01005
Zhiying Xu
Zhiying Xu, Hongding Peng, Wei Wang
AGO: Boosting Mobile AI Inference Performance by Removing Constraints on Graph Optimization
null
null
null
null
cs.LG cs.CL cs.DC
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Traditional deep learning compilers rely on heuristics for subgraph generation, which impose extra constraints on graph optimization, e.g., each subgraph can only contain at most one complex operator. In this paper, we propose AGO, a framework for graph optimization with arbitrary structures to boost the inference performance of deep models by removing such constraints. To create new optimization opportunities for complicated subgraphs, we propose intensive operator fusion, which can effectively stitch multiple complex operators together for better performance. Further, we design a graph partitioning scheme that allows an arbitrary structure for each subgraph while guaranteeing the acyclic property among all generated subgraphs. Additionally, to enable efficient performance tuning on complicated subgraphs, we devise a novel divide-and-conquer tuning mechanism to orchestrate different system components. Through extensive experiments on various neural networks and mobile devices, we show that our system can improve the inference performance by up to 3.3x when compared with state-of-the-art deep compilers.
[ { "created": "Fri, 2 Dec 2022 07:16:49 GMT", "version": "v1" } ]
2022-12-05
[ [ "Xu", "Zhiying", "" ], [ "Peng", "Hongding", "" ], [ "Wang", "Wei", "" ] ]
Traditional deep learning compilers rely on heuristics for subgraph generation, which impose extra constraints on graph optimization, e.g., each subgraph can only contain at most one complex operator. In this paper, we propose AGO, a framework for graph optimization with arbitrary structures to boost the inference performance of deep models by removing such constraints. To create new optimization opportunities for complicated subgraphs, we propose intensive operator fusion, which can effectively stitch multiple complex operators together for better performance. Further, we design a graph partitioning scheme that allows an arbitrary structure for each subgraph while guaranteeing the acyclic property among all generated subgraphs. Additionally, to enable efficient performance tuning on complicated subgraphs, we devise a novel divide-and-conquer tuning mechanism to orchestrate different system components. Through extensive experiments on various neural networks and mobile devices, we show that our system can improve the inference performance by up to 3.3x when compared with state-of-the-art deep compilers.
2405.01159
Aleksei Dorkin
Aleksei Dorkin and Kairit Sirts
TartuNLP at EvaLatin 2024: Emotion Polarity Detection
Accepted to The Third Workshop on Language Technologies for Historical and Ancient Languages (LT4HALA 2024)
null
null
null
cs.CL
http://creativecommons.org/licenses/by/4.0/
This paper presents the TartuNLP team submission to EvaLatin 2024 shared task of the emotion polarity detection for historical Latin texts. Our system relies on two distinct approaches to annotating training data for supervised learning: 1) creating heuristics-based labels by adopting the polarity lexicon provided by the organizers and 2) generating labels with GPT4. We employed parameter efficient fine-tuning using the adapters framework and experimented with both monolingual and cross-lingual knowledge transfer for training language and task adapters. Our submission with the LLM-generated labels achieved the overall first place in the emotion polarity detection task. Our results show that LLM-based annotations show promising results on texts in Latin.
[ { "created": "Thu, 2 May 2024 10:28:52 GMT", "version": "v1" } ]
2024-05-03
[ [ "Dorkin", "Aleksei", "" ], [ "Sirts", "Kairit", "" ] ]
This paper presents the TartuNLP team submission to EvaLatin 2024 shared task of the emotion polarity detection for historical Latin texts. Our system relies on two distinct approaches to annotating training data for supervised learning: 1) creating heuristics-based labels by adopting the polarity lexicon provided by the organizers and 2) generating labels with GPT4. We employed parameter efficient fine-tuning using the adapters framework and experimented with both monolingual and cross-lingual knowledge transfer for training language and task adapters. Our submission with the LLM-generated labels achieved the overall first place in the emotion polarity detection task. Our results show that LLM-based annotations show promising results on texts in Latin.
1804.04438
Ari Morcos
Avraham Ruderman, Neil C. Rabinowitz, Ari S. Morcos, Daniel Zoran
Pooling is neither necessary nor sufficient for appropriate deformation stability in CNNs
NIPS 2018 submission
null
null
null
cs.CV cs.LG stat.ML
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Many of our core assumptions about how neural networks operate remain empirically untested. One common assumption is that convolutional neural networks need to be stable to small translations and deformations to solve image recognition tasks. For many years, this stability was baked into CNN architectures by incorporating interleaved pooling layers. Recently, however, interleaved pooling has largely been abandoned. This raises a number of questions: Are our intuitions about deformation stability right at all? Is it important? Is pooling necessary for deformation invariance? If not, how is deformation invariance achieved in its absence? In this work, we rigorously test these questions, and find that deformation stability in convolutional networks is more nuanced than it first appears: (1) Deformation invariance is not a binary property, but rather that different tasks require different degrees of deformation stability at different layers. (2) Deformation stability is not a fixed property of a network and is heavily adjusted over the course of training, largely through the smoothness of the convolutional filters. (3) Interleaved pooling layers are neither necessary nor sufficient for achieving the optimal form of deformation stability for natural image classification. (4) Pooling confers too much deformation stability for image classification at initialization, and during training, networks have to learn to counteract this inductive bias. Together, these findings provide new insights into the role of interleaved pooling and deformation invariance in CNNs, and demonstrate the importance of rigorous empirical testing of even our most basic assumptions about the working of neural networks.
[ { "created": "Thu, 12 Apr 2018 11:44:05 GMT", "version": "v1" }, { "created": "Fri, 25 May 2018 13:03:50 GMT", "version": "v2" } ]
2018-05-28
[ [ "Ruderman", "Avraham", "" ], [ "Rabinowitz", "Neil C.", "" ], [ "Morcos", "Ari S.", "" ], [ "Zoran", "Daniel", "" ] ]
Many of our core assumptions about how neural networks operate remain empirically untested. One common assumption is that convolutional neural networks need to be stable to small translations and deformations to solve image recognition tasks. For many years, this stability was baked into CNN architectures by incorporating interleaved pooling layers. Recently, however, interleaved pooling has largely been abandoned. This raises a number of questions: Are our intuitions about deformation stability right at all? Is it important? Is pooling necessary for deformation invariance? If not, how is deformation invariance achieved in its absence? In this work, we rigorously test these questions, and find that deformation stability in convolutional networks is more nuanced than it first appears: (1) Deformation invariance is not a binary property, but rather that different tasks require different degrees of deformation stability at different layers. (2) Deformation stability is not a fixed property of a network and is heavily adjusted over the course of training, largely through the smoothness of the convolutional filters. (3) Interleaved pooling layers are neither necessary nor sufficient for achieving the optimal form of deformation stability for natural image classification. (4) Pooling confers too much deformation stability for image classification at initialization, and during training, networks have to learn to counteract this inductive bias. Together, these findings provide new insights into the role of interleaved pooling and deformation invariance in CNNs, and demonstrate the importance of rigorous empirical testing of even our most basic assumptions about the working of neural networks.
1407.3193
Awais Mansoor
Awais Mansoor, Ulas Bagci, Daniel J. Mollura
Optimally Stabilized PET Image Denoising Using Trilateral Filtering
8 pages, 3 figures; to appear in the Lecture Notes in Computer Science (MICCAI 2014)
null
null
null
cs.CV
http://creativecommons.org/licenses/publicdomain/
Low-resolution and signal-dependent noise distribution in positron emission tomography (PET) images makes denoising process an inevitable step prior to qualitative and quantitative image analysis tasks. Conventional PET denoising methods either over-smooth small-sized structures due to resolution limitation or make incorrect assumptions about the noise characteristics. Therefore, clinically important quantitative information may be corrupted. To address these challenges, we introduced a novel approach to remove signal-dependent noise in the PET images where the noise distribution was considered as Poisson-Gaussian mixed. Meanwhile, the generalized Anscombe's transformation (GAT) was used to stabilize varying nature of the PET noise. Other than noise stabilization, it is also desirable for the noise removal filter to preserve the boundaries of the structures while smoothing the noisy regions. Indeed, it is important to avoid significant loss of quantitative information such as standard uptake value (SUV)-based metrics as well as metabolic lesion volume. To satisfy all these properties, we extended bilateral filtering method into trilateral filtering through multiscaling and optimal Gaussianization process. The proposed method was tested on more than 50 PET-CT images from various patients having different cancers and achieved the superior performance compared to the widely used denoising techniques in the literature.
[ { "created": "Fri, 11 Jul 2014 15:08:18 GMT", "version": "v1" } ]
2014-07-14
[ [ "Mansoor", "Awais", "" ], [ "Bagci", "Ulas", "" ], [ "Mollura", "Daniel J.", "" ] ]
Low-resolution and signal-dependent noise distribution in positron emission tomography (PET) images makes denoising process an inevitable step prior to qualitative and quantitative image analysis tasks. Conventional PET denoising methods either over-smooth small-sized structures due to resolution limitation or make incorrect assumptions about the noise characteristics. Therefore, clinically important quantitative information may be corrupted. To address these challenges, we introduced a novel approach to remove signal-dependent noise in the PET images where the noise distribution was considered as Poisson-Gaussian mixed. Meanwhile, the generalized Anscombe's transformation (GAT) was used to stabilize varying nature of the PET noise. Other than noise stabilization, it is also desirable for the noise removal filter to preserve the boundaries of the structures while smoothing the noisy regions. Indeed, it is important to avoid significant loss of quantitative information such as standard uptake value (SUV)-based metrics as well as metabolic lesion volume. To satisfy all these properties, we extended bilateral filtering method into trilateral filtering through multiscaling and optimal Gaussianization process. The proposed method was tested on more than 50 PET-CT images from various patients having different cancers and achieved the superior performance compared to the widely used denoising techniques in the literature.
2306.10300
Subhashis Das
Subhashis Das, Debashis Naskar and Sayon Roy
Reorganizing Educational Institutional Domain using Faceted Ontological Principles
26 pages, 12 figures, KNOWLEDGE ORGANIZATION Journal Paper
KO KNOWLEDGE ORGANIZATION, 49(1), 6-21 (2022)
10.5771/0943-7444-2022-1
null
cs.AI cs.DL
http://creativecommons.org/licenses/by/4.0/
The purpose of this work is to find out how different library classification systems and linguistic ontologies arrange a particular domain of interest and what are the limitations for information retrieval. We use knowledge representation techniques and languages for construction of a domain specific ontology. This ontology would help not only in problem solving, but it would demonstrate the ease with which complex queries can be handled using principles of domain ontology, thereby facilitating better information retrieval.
[ { "created": "Sat, 17 Jun 2023 09:06:07 GMT", "version": "v1" } ]
2023-06-21
[ [ "Das", "Subhashis", "" ], [ "Naskar", "Debashis", "" ], [ "Roy", "Sayon", "" ] ]
The purpose of this work is to find out how different library classification systems and linguistic ontologies arrange a particular domain of interest and what are the limitations for information retrieval. We use knowledge representation techniques and languages for construction of a domain specific ontology. This ontology would help not only in problem solving, but it would demonstrate the ease with which complex queries can be handled using principles of domain ontology, thereby facilitating better information retrieval.
0901.3950
Moshe Mishali
Moshe Mishali, Yonina C. Eldar and Joel A. Tropp
Efficient Sampling of Sparse Wideband Analog Signals
13 pages, 5 figs, conference paper (see ref. below)
Proc. of IEEEI, 25th convention, pp. 290-294, Dec. 2008
null
CCIT Report #705, Oct. 2008, EE Dept., Technion Israel
cs.IT math.IT
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Periodic nonuniform sampling is a known method to sample spectrally sparse signals below the Nyquist rate. This strategy relies on the implicit assumption that the individual samplers are exposed to the entire frequency range. This assumption becomes impractical for wideband sparse signals. The current paper proposes an alternative sampling stage that does not require a full-band front end. Instead, signals are captured with an analog front end that consists of a bank of multipliers and lowpass filters whose cutoff is much lower than the Nyquist rate. The problem of recovering the original signal from the low-rate samples can be studied within the framework of compressive sampling. An appropriate parameter selection ensures that the samples uniquely determine the analog input. Moreover, the analog input can be stably reconstructed with digital algorithms. Numerical experiments support the theoretical analysis.
[ { "created": "Mon, 26 Jan 2009 07:00:52 GMT", "version": "v1" } ]
2009-01-27
[ [ "Mishali", "Moshe", "" ], [ "Eldar", "Yonina C.", "" ], [ "Tropp", "Joel A.", "" ] ]
Periodic nonuniform sampling is a known method to sample spectrally sparse signals below the Nyquist rate. This strategy relies on the implicit assumption that the individual samplers are exposed to the entire frequency range. This assumption becomes impractical for wideband sparse signals. The current paper proposes an alternative sampling stage that does not require a full-band front end. Instead, signals are captured with an analog front end that consists of a bank of multipliers and lowpass filters whose cutoff is much lower than the Nyquist rate. The problem of recovering the original signal from the low-rate samples can be studied within the framework of compressive sampling. An appropriate parameter selection ensures that the samples uniquely determine the analog input. Moreover, the analog input can be stably reconstructed with digital algorithms. Numerical experiments support the theoretical analysis.
2109.07923
Peisen Yao
Peisen Yao and Jinguo Zhou and Xiao Xiao and Qingkai Shi and Rongxin Wu and Charles Zhang
Efficient Path-Sensitive Data-Dependence Analysis
null
null
null
null
cs.PL cs.SE
http://creativecommons.org/licenses/by/4.0/
This paper presents a scalable path- and context-sensitive data-dependence analysis. The key is to address the aliasing-path-explosion problem via a sparse, demand-driven, and fused approach that piggybacks the computation of pointer information with the resolution of data dependence. Specifically, our approach decomposes the computational efforts of disjunctive reasoning into 1) a context- and semi-path-sensitive analysis that concisely summarizes data dependence as the symbolic and storeless value-flow graphs, and 2) a demand-driven phase that resolves transitive data dependence over the graphs. We have applied the approach to two clients, namely thin slicing and value flow analysis. Using a suite of 16 programs ranging from 13 KLoC to 8 MLoC, we compare our techniques against a diverse group of state-of-the-art analyses, illustrating significant precision and scalability advantages of our approach.
[ { "created": "Thu, 16 Sep 2021 12:17:05 GMT", "version": "v1" } ]
2021-09-20
[ [ "Yao", "Peisen", "" ], [ "Zhou", "Jinguo", "" ], [ "Xiao", "Xiao", "" ], [ "Shi", "Qingkai", "" ], [ "Wu", "Rongxin", "" ], [ "Zhang", "Charles", "" ] ]
This paper presents a scalable path- and context-sensitive data-dependence analysis. The key is to address the aliasing-path-explosion problem via a sparse, demand-driven, and fused approach that piggybacks the computation of pointer information with the resolution of data dependence. Specifically, our approach decomposes the computational efforts of disjunctive reasoning into 1) a context- and semi-path-sensitive analysis that concisely summarizes data dependence as the symbolic and storeless value-flow graphs, and 2) a demand-driven phase that resolves transitive data dependence over the graphs. We have applied the approach to two clients, namely thin slicing and value flow analysis. Using a suite of 16 programs ranging from 13 KLoC to 8 MLoC, we compare our techniques against a diverse group of state-of-the-art analyses, illustrating significant precision and scalability advantages of our approach.
2110.14271
Sigal Oren
Sigal Oren and Oren Roth
Mechanisms for Trading Durable Goods
WINE'21
null
null
null
cs.GT
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We consider trading indivisible and easily transferable \emph{durable goods}, which are goods that an agent can receive, use, and trade again for a different good. This is often the case with books that can be read and later exchanged for unread ones. Other examples of such easily transferable durable goods include puzzles, video games and baby clothes. We introduce a model for the exchange of easily transferable durable goods. In our model, each agent owns a set of items and demands a different set of items. An agent is interested in receiving as many items as possible from his demand set. We consider mechanisms that exchange items in cycles in which each participating agent receives an item that he demands and gives an item that he owns. We aim to develop mechanisms that have the following properties: they are \emph{efficient}, in the sense that they maximize the total number of items that agents receive from their demand set, they are \emph{strategyproof} (i.e., it is in the agents' best interest to report their preferences truthfully) and they run in \emph{polynomial time}. One challenge in developing mechanisms for our setting is that the supply and demand sets of the agents are updated after a trade cycle is executed. This makes constructing strategyproof mechanisms in our model significantly different from previous works, both technically and conceptually and requires developing new tools and techniques. We prove that simultaneously satisfying all desired properties is impossible and thus focus on studying the tradeoffs between these properties. To this end, we provide both approximation algorithms and impossibility results.
[ { "created": "Wed, 27 Oct 2021 08:51:23 GMT", "version": "v1" } ]
2021-10-28
[ [ "Oren", "Sigal", "" ], [ "Roth", "Oren", "" ] ]
We consider trading indivisible and easily transferable \emph{durable goods}, which are goods that an agent can receive, use, and trade again for a different good. This is often the case with books that can be read and later exchanged for unread ones. Other examples of such easily transferable durable goods include puzzles, video games and baby clothes. We introduce a model for the exchange of easily transferable durable goods. In our model, each agent owns a set of items and demands a different set of items. An agent is interested in receiving as many items as possible from his demand set. We consider mechanisms that exchange items in cycles in which each participating agent receives an item that he demands and gives an item that he owns. We aim to develop mechanisms that have the following properties: they are \emph{efficient}, in the sense that they maximize the total number of items that agents receive from their demand set, they are \emph{strategyproof} (i.e., it is in the agents' best interest to report their preferences truthfully) and they run in \emph{polynomial time}. One challenge in developing mechanisms for our setting is that the supply and demand sets of the agents are updated after a trade cycle is executed. This makes constructing strategyproof mechanisms in our model significantly different from previous works, both technically and conceptually and requires developing new tools and techniques. We prove that simultaneously satisfying all desired properties is impossible and thus focus on studying the tradeoffs between these properties. To this end, we provide both approximation algorithms and impossibility results.
2208.09815
Pengqian Yu
Xinhan Di, Pengqian Yu
LWA-HAND: Lightweight Attention Hand for Interacting Hand Reconstruction
Accepted by ECCV 2022 Computer Vision for Metaverse Workshop (16 pages, 6 figures, 1 table)
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Recent years have witnessed great success for hand reconstruction in real-time applications such as visual reality and augmented reality while interacting with two-hand reconstruction through efficient transformers is left unexplored. In this paper, we propose a method called lightweight attention hand (LWA-HAND) to reconstruct hands in low flops from a single RGB image. To solve the occlusion and interaction problem in efficient attention architectures, we propose three mobile attention modules in this paper. The first module is a lightweight feature attention module that extracts both local occlusion representation and global image patch representation in a coarse-to-fine manner. The second module is a cross image and graph bridge module which fuses image context and hand vertex. The third module is a lightweight cross-attention mechanism that uses element-wise operation for the cross-attention of two hands in linear complexity. The resulting model achieves comparable performance on the InterHand2.6M benchmark in comparison with the state-of-the-art models. Simultaneously, it reduces the flops to $0.47GFlops$ while the state-of-the-art models have heavy computations between $10GFlops$ and $20GFlops$.
[ { "created": "Sun, 21 Aug 2022 06:25:56 GMT", "version": "v1" }, { "created": "Tue, 23 Aug 2022 03:54:47 GMT", "version": "v2" }, { "created": "Sat, 27 Aug 2022 13:06:34 GMT", "version": "v3" } ]
2022-08-30
[ [ "Di", "Xinhan", "" ], [ "Yu", "Pengqian", "" ] ]
Recent years have witnessed great success for hand reconstruction in real-time applications such as visual reality and augmented reality while interacting with two-hand reconstruction through efficient transformers is left unexplored. In this paper, we propose a method called lightweight attention hand (LWA-HAND) to reconstruct hands in low flops from a single RGB image. To solve the occlusion and interaction problem in efficient attention architectures, we propose three mobile attention modules in this paper. The first module is a lightweight feature attention module that extracts both local occlusion representation and global image patch representation in a coarse-to-fine manner. The second module is a cross image and graph bridge module which fuses image context and hand vertex. The third module is a lightweight cross-attention mechanism that uses element-wise operation for the cross-attention of two hands in linear complexity. The resulting model achieves comparable performance on the InterHand2.6M benchmark in comparison with the state-of-the-art models. Simultaneously, it reduces the flops to $0.47GFlops$ while the state-of-the-art models have heavy computations between $10GFlops$ and $20GFlops$.
2101.10620
Liang Lin
Liang Lin and Yiming Gao and Ke Gong and Meng Wang and Xiaodan Liang
Graphonomy: Universal Image Parsing via Graph Reasoning and Transfer
To appear in IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE (T-PAMI) 2021. We propose a graph reasoning and transfer learning framework, which incorporates human knowledge and label taxonomy into the intermediate graph representation learning beyond local convolutions. arXiv admin note: substantial text overlap with arXiv:1904.04536
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Prior highly-tuned image parsing models are usually studied in a certain domain with a specific set of semantic labels and can hardly be adapted into other scenarios (e.g., sharing discrepant label granularity) without extensive re-training. Learning a single universal parsing model by unifying label annotations from different domains or at various levels of granularity is a crucial but rarely addressed topic. This poses many fundamental learning challenges, e.g., discovering underlying semantic structures among different label granularity or mining label correlation across relevant tasks. To address these challenges, we propose a graph reasoning and transfer learning framework, named "Graphonomy", which incorporates human knowledge and label taxonomy into the intermediate graph representation learning beyond local convolutions. In particular, Graphonomy learns the global and structured semantic coherency in multiple domains via semantic-aware graph reasoning and transfer, enforcing the mutual benefits of the parsing across domains (e.g., different datasets or co-related tasks). The Graphonomy includes two iterated modules: Intra-Graph Reasoning and Inter-Graph Transfer modules. The former extracts the semantic graph in each domain to improve the feature representation learning by propagating information with the graph; the latter exploits the dependencies among the graphs from different domains for bidirectional knowledge transfer. We apply Graphonomy to two relevant but different image understanding research topics: human parsing and panoptic segmentation, and show Graphonomy can handle both of them well via a standard pipeline against current state-of-the-art approaches. Moreover, some extra benefit of our framework is demonstrated, e.g., generating the human parsing at various levels of granularity by unifying annotations across different datasets.
[ { "created": "Tue, 26 Jan 2021 08:19:03 GMT", "version": "v1" } ]
2021-01-27
[ [ "Lin", "Liang", "" ], [ "Gao", "Yiming", "" ], [ "Gong", "Ke", "" ], [ "Wang", "Meng", "" ], [ "Liang", "Xiaodan", "" ] ]
Prior highly-tuned image parsing models are usually studied in a certain domain with a specific set of semantic labels and can hardly be adapted into other scenarios (e.g., sharing discrepant label granularity) without extensive re-training. Learning a single universal parsing model by unifying label annotations from different domains or at various levels of granularity is a crucial but rarely addressed topic. This poses many fundamental learning challenges, e.g., discovering underlying semantic structures among different label granularity or mining label correlation across relevant tasks. To address these challenges, we propose a graph reasoning and transfer learning framework, named "Graphonomy", which incorporates human knowledge and label taxonomy into the intermediate graph representation learning beyond local convolutions. In particular, Graphonomy learns the global and structured semantic coherency in multiple domains via semantic-aware graph reasoning and transfer, enforcing the mutual benefits of the parsing across domains (e.g., different datasets or co-related tasks). The Graphonomy includes two iterated modules: Intra-Graph Reasoning and Inter-Graph Transfer modules. The former extracts the semantic graph in each domain to improve the feature representation learning by propagating information with the graph; the latter exploits the dependencies among the graphs from different domains for bidirectional knowledge transfer. We apply Graphonomy to two relevant but different image understanding research topics: human parsing and panoptic segmentation, and show Graphonomy can handle both of them well via a standard pipeline against current state-of-the-art approaches. Moreover, some extra benefit of our framework is demonstrated, e.g., generating the human parsing at various levels of granularity by unifying annotations across different datasets.
0909.2622
Jiangyuan Li
Jiangyuan Li and Athina Petropulu
Transmitter Optimization for Achieving Secrecy Capacity in Gaussian MIMO Wiretap Channels
29 pages, 10 figures
null
null
null
cs.IT math.IT
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We consider a Gaussian multiple-input multiple-output (MIMO) wiretap channel model, where there exists a transmitter, a legitimate receiver and an eavesdropper, each node equipped with multiple antennas. We study the problem of finding the optimal input covariance matrix that achieves secrecy capacity subject to a power constraint, which leads to a non-convex optimization problem that is in general difficult to solve. Existing results for this problem address the case in which the transmitter and the legitimate receiver have two antennas each and the eavesdropper has one antenna. For the general cases, it has been shown that the optimal input covariance matrix has low rank when the difference between the Grams of the eavesdropper and the legitimate receiver channel matrices is indefinite or semi-definite, while it may have low rank or full rank when the difference is positive definite. In this paper, the aforementioned non-convex optimization problem is investigated. In particular, for the multiple-input single-output (MISO) wiretap channel, the optimal input covariance matrix is obtained in closed form. For general cases, we derive the necessary conditions for the optimal input covariance matrix consisting of a set of equations. For the case in which the transmitter has two antennas, the derived necessary conditions can result in a closed form solution; For the case in which the difference between the Grams is indefinite and has all negative eigenvalues except one positive eigenvalue, the optimal input covariance matrix has rank one and can be obtained in closed form; For other cases, the solution is proved to be a fixed point of a mapping from a convex set to itself and an iterative procedure is provided to search for it. Numerical results are presented to illustrate the proposed theoretical findings.
[ { "created": "Mon, 14 Sep 2009 19:07:49 GMT", "version": "v1" } ]
2009-09-15
[ [ "Li", "Jiangyuan", "" ], [ "Petropulu", "Athina", "" ] ]
We consider a Gaussian multiple-input multiple-output (MIMO) wiretap channel model, where there exists a transmitter, a legitimate receiver and an eavesdropper, each node equipped with multiple antennas. We study the problem of finding the optimal input covariance matrix that achieves secrecy capacity subject to a power constraint, which leads to a non-convex optimization problem that is in general difficult to solve. Existing results for this problem address the case in which the transmitter and the legitimate receiver have two antennas each and the eavesdropper has one antenna. For the general cases, it has been shown that the optimal input covariance matrix has low rank when the difference between the Grams of the eavesdropper and the legitimate receiver channel matrices is indefinite or semi-definite, while it may have low rank or full rank when the difference is positive definite. In this paper, the aforementioned non-convex optimization problem is investigated. In particular, for the multiple-input single-output (MISO) wiretap channel, the optimal input covariance matrix is obtained in closed form. For general cases, we derive the necessary conditions for the optimal input covariance matrix consisting of a set of equations. For the case in which the transmitter has two antennas, the derived necessary conditions can result in a closed form solution; For the case in which the difference between the Grams is indefinite and has all negative eigenvalues except one positive eigenvalue, the optimal input covariance matrix has rank one and can be obtained in closed form; For other cases, the solution is proved to be a fixed point of a mapping from a convex set to itself and an iterative procedure is provided to search for it. Numerical results are presented to illustrate the proposed theoretical findings.
2402.06331
Joanna Komorniczak
Joanna Komorniczak and Pawel Ksieniewicz
Taking Class Imbalance Into Account in Open Set Recognition Evaluation
null
null
null
null
cs.LG cs.CV
http://creativecommons.org/licenses/by-nc-sa/4.0/
In recent years Deep Neural Network-based systems are not only increasing in popularity but also receive growing user trust. However, due to the closed-world assumption of such systems, they cannot recognize samples from unknown classes and often induce an incorrect label with high confidence. Presented work looks at the evaluation of methods for Open Set Recognition, focusing on the impact of class imbalance, especially in the dichotomy between known and unknown samples. As an outcome of problem analysis, we present a set of guidelines for evaluation of methods in this field.
[ { "created": "Fri, 9 Feb 2024 11:15:49 GMT", "version": "v1" } ]
2024-02-12
[ [ "Komorniczak", "Joanna", "" ], [ "Ksieniewicz", "Pawel", "" ] ]
In recent years Deep Neural Network-based systems are not only increasing in popularity but also receive growing user trust. However, due to the closed-world assumption of such systems, they cannot recognize samples from unknown classes and often induce an incorrect label with high confidence. Presented work looks at the evaluation of methods for Open Set Recognition, focusing on the impact of class imbalance, especially in the dichotomy between known and unknown samples. As an outcome of problem analysis, we present a set of guidelines for evaluation of methods in this field.
2403.04384
Tomasz Winiarski
Tomasz Winiarski, Daniel Gie{\l}dowski, Jan Kaniuka, Jakub Ostrysz, Jakub Sadowski
HeROS: a miniaturised platform for research and development on Heterogeneous RObotic Systems
null
null
null
null
cs.RO
http://creativecommons.org/licenses/by/4.0/
Tests and prototyping are vital in the research and development of robotic systems. Work with target hardware is problematic. Hence, in the article, a low-cost, miniaturised physical platform is presented to deal with experiments on heterogeneous robotic systems. The platform comprises a physical board with tiles of the standardised base, diverse mobile robots, and manipulation robots. The number of exemplary applications validates the usefulness of the solution.
[ { "created": "Thu, 7 Mar 2024 10:23:39 GMT", "version": "v1" } ]
2024-03-08
[ [ "Winiarski", "Tomasz", "" ], [ "Giełdowski", "Daniel", "" ], [ "Kaniuka", "Jan", "" ], [ "Ostrysz", "Jakub", "" ], [ "Sadowski", "Jakub", "" ] ]
Tests and prototyping are vital in the research and development of robotic systems. Work with target hardware is problematic. Hence, in the article, a low-cost, miniaturised physical platform is presented to deal with experiments on heterogeneous robotic systems. The platform comprises a physical board with tiles of the standardised base, diverse mobile robots, and manipulation robots. The number of exemplary applications validates the usefulness of the solution.
2403.04712
Harry Zhang Mr.
Sangli Teng, Harry Zhang, David Jin, Ashkan Jasour, Maani Ghaffari, Luca Carlone
GMKF: Generalized Moment Kalman Filter for Polynomial Systems with Arbitrary Noise
null
null
null
null
cs.RO cs.SY eess.SY
http://creativecommons.org/licenses/by/4.0/
This paper develops a new filtering approach for state estimation in polynomial systems corrupted by arbitrary noise, which commonly arise in robotics. We first consider a batch setup where we perform state estimation using all data collected from the initial to the current time. We formulate the batch state estimation problem as a Polynomial Optimization Problem (POP) and relax the assumption of Gaussian noise by specifying a finite number of moments of the noise. We solve the resulting POP using a moment relaxation and prove that under suitable conditions on the rank of the relaxation, (i) we can extract a provably optimal estimate from the moment relaxation, and (ii) we can obtain a belief representation from the dual (sum-of-squares) relaxation. We then turn our attention to the filtering setup and apply similar insights to develop a GMKF for recursive state estimation in polynomial systems with arbitrary noise. The GMKF formulates the prediction and update steps as POPs and solves them using moment relaxations, carrying over a possibly non-Gaussian belief. In the linear-Gaussian case, GMKF reduces to the standard Kalman Filter. We demonstrate that GMKF performs well under highly non-Gaussian noise and outperforms common alternatives, including the Extended and Unscented Kalman Filter, and their variants on matrix Lie group.
[ { "created": "Thu, 7 Mar 2024 18:07:41 GMT", "version": "v1" }, { "created": "Fri, 8 Mar 2024 05:08:07 GMT", "version": "v2" } ]
2024-03-11
[ [ "Teng", "Sangli", "" ], [ "Zhang", "Harry", "" ], [ "Jin", "David", "" ], [ "Jasour", "Ashkan", "" ], [ "Ghaffari", "Maani", "" ], [ "Carlone", "Luca", "" ] ]
This paper develops a new filtering approach for state estimation in polynomial systems corrupted by arbitrary noise, which commonly arise in robotics. We first consider a batch setup where we perform state estimation using all data collected from the initial to the current time. We formulate the batch state estimation problem as a Polynomial Optimization Problem (POP) and relax the assumption of Gaussian noise by specifying a finite number of moments of the noise. We solve the resulting POP using a moment relaxation and prove that under suitable conditions on the rank of the relaxation, (i) we can extract a provably optimal estimate from the moment relaxation, and (ii) we can obtain a belief representation from the dual (sum-of-squares) relaxation. We then turn our attention to the filtering setup and apply similar insights to develop a GMKF for recursive state estimation in polynomial systems with arbitrary noise. The GMKF formulates the prediction and update steps as POPs and solves them using moment relaxations, carrying over a possibly non-Gaussian belief. In the linear-Gaussian case, GMKF reduces to the standard Kalman Filter. We demonstrate that GMKF performs well under highly non-Gaussian noise and outperforms common alternatives, including the Extended and Unscented Kalman Filter, and their variants on matrix Lie group.
2008.03325
Nathaniel Grammel
Brian Brubach, Nathaniel Grammel, David G. Harris, Aravind Srinivasan, Leonidas Tsepenekas, Anil Vullikanti
Stochastic Optimization and Learning for Two-Stage Supplier Problems
null
null
null
null
cs.DS
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The main focus of this paper is radius-based (supplier) clustering in the two-stage stochastic setting with recourse, where the inherent stochasticity of the model comes in the form of a budget constraint. In addition to the standard (homogeneous) setting where all clients must be within a distance $R$ of the nearest facility, we provide results for the more general problem where the radius demands may be inhomogeneous (i.e., different for each client). We also explore a number of variants where additional constraints are imposed on the first-stage decisions, specifically matroid and multi-knapsack constraints, and provide results for these settings. We derive results for the most general distributional setting, where there is only black-box access to the underlying distribution. To accomplish this, we first develop algorithms for the polynomial scenarios setting; we then employ a novel scenario-discarding variant of the standard Sample Average Approximation (SAA) method, which crucially exploits properties of the restricted-case algorithms. We note that the scenario-discarding modification to the SAA method is necessary in order to optimize over the radius.
[ { "created": "Fri, 7 Aug 2020 18:18:29 GMT", "version": "v1" }, { "created": "Tue, 16 Feb 2021 16:50:30 GMT", "version": "v2" }, { "created": "Thu, 6 May 2021 15:10:17 GMT", "version": "v3" }, { "created": "Sun, 5 Jun 2022 23:35:10 GMT", "version": "v4" }, { "created": "Wed, 17 Aug 2022 05:51:48 GMT", "version": "v5" }, { "created": "Sat, 15 Jul 2023 16:02:59 GMT", "version": "v6" }, { "created": "Sun, 7 Apr 2024 13:18:54 GMT", "version": "v7" } ]
2024-04-09
[ [ "Brubach", "Brian", "" ], [ "Grammel", "Nathaniel", "" ], [ "Harris", "David G.", "" ], [ "Srinivasan", "Aravind", "" ], [ "Tsepenekas", "Leonidas", "" ], [ "Vullikanti", "Anil", "" ] ]
The main focus of this paper is radius-based (supplier) clustering in the two-stage stochastic setting with recourse, where the inherent stochasticity of the model comes in the form of a budget constraint. In addition to the standard (homogeneous) setting where all clients must be within a distance $R$ of the nearest facility, we provide results for the more general problem where the radius demands may be inhomogeneous (i.e., different for each client). We also explore a number of variants where additional constraints are imposed on the first-stage decisions, specifically matroid and multi-knapsack constraints, and provide results for these settings. We derive results for the most general distributional setting, where there is only black-box access to the underlying distribution. To accomplish this, we first develop algorithms for the polynomial scenarios setting; we then employ a novel scenario-discarding variant of the standard Sample Average Approximation (SAA) method, which crucially exploits properties of the restricted-case algorithms. We note that the scenario-discarding modification to the SAA method is necessary in order to optimize over the radius.