id
stringlengths
9
16
title
stringlengths
4
278
categories
stringlengths
5
104
abstract
stringlengths
6
4.09k
2501.02002
HMM-LSTM Fusion Model for Economic Forecasting
cs.LG econ.EM stat.ME
This paper explores the application of Hidden Markov Models (HMM) and Long Short-Term Memory (LSTM) neural networks for economic forecasting, focusing on predicting CPI inflation rates. The study explores a new approach that integrates HMM-derived hidden states and means as additional features for LSTM modeling, aiming to enhance the interpretability and predictive performance of the models. The research begins with data collection and preprocessing, followed by the implementation of the HMM to identify hidden states representing distinct economic conditions. Subsequently, LSTM models are trained using the original and augmented data sets, allowing for comparative analysis and evaluation. The results demonstrate that incorporating HMM-derived data improves the predictive accuracy of LSTM models, particularly in capturing complex temporal patterns and mitigating the impact of volatile economic conditions. Additionally, the paper discusses the implementation of Integrated Gradients for model interpretability and provides insights into the economic dynamics reflected in the forecasting outcomes.
2501.02003
SurfPatch: Enabling Patch Matching for Exploratory Stream Surface Visualization
cs.GR cs.CV
Unlike their line-based counterparts, surface-based techniques have yet to be thoroughly investigated in flow visualization due to their significant placement, speed, perception, and evaluation challenges. This paper presents SurfPatch, a novel framework supporting exploratory stream surface visualization. To begin with, we translate the issue of surface placement to surface selection and trace a large number of stream surfaces from a given flow field dataset. Then, we introduce a three-stage process: vertex-level classification, patch-level matching, and surface-level clustering that hierarchically builds the connection between vertices and patches and between patches and surfaces. This bottom-up approach enables fine-grained, multiscale patch-level matching, sharply contrasts surface-level matching offered by existing works, and provides previously unavailable flexibility during querying. We design an intuitive visual interface for users to conveniently visualize and analyze the underlying collection of stream surfaces in an exploratory manner. SurfPatch is not limited to stream surfaces traced from steady flow datasets. We demonstrate its effectiveness through experiments on stream surfaces produced from steady and unsteady flows as well as isosurfaces extracted from scalar fields. The code is available at https://github.com/adlsn/SurfPatch.
2501.02004
General Information Metrics for Improving AI Model Training Efficiency
cs.LG cs.AI cs.IT math.IT
To address the growing size of AI model training data and the lack of a universal data selection methodology-factors that significantly drive up training costs -- this paper presents the General Information Metrics Evaluation (GIME) method. GIME leverages general information metrics from Objective Information Theory (OIT), including volume, delay, scope, granularity, variety, duration, sampling rate, aggregation, coverage, distortion, and mismatch to optimize dataset selection for training purposes. Comprehensive experiments conducted across diverse domains, such as CTR Prediction, Civil Case Prediction, and Weather Forecasting, demonstrate that GIME effectively preserves model performance while substantially reducing both training time and costs. Additionally, applying GIME within the Judicial AI Program led to a remarkable 39.56% reduction in total model training expenses, underscoring its potential to support efficient and sustainable AI development.
2501.02006
Multi-Task Semantic Communication With Graph Attention-Based Feature Correlation Extraction
cs.LG cs.AI
Multi-task semantic communication can serve multiple learning tasks using a shared encoder model. Existing models have overlooked the intricate relationships between features extracted during an encoding process of tasks. This paper presents a new graph attention inter-block (GAI) module to the encoder/transmitter of a multi-task semantic communication system, which enriches the features for multiple tasks by embedding the intermediate outputs of encoding in the features, compared to the existing techniques. The key idea is that we interpret the outputs of the intermediate feature extraction blocks of the encoder as the nodes of a graph to capture the correlations of the intermediate features. Another important aspect is that we refine the node representation using a graph attention mechanism to extract the correlations and a multi-layer perceptron network to associate the node representations with different tasks. Consequently, the intermediate features are weighted and embedded into the features transmitted for executing multiple tasks at the receiver. Experiments demonstrate that the proposed model surpasses the most competitive and publicly available models by 11.4% on the CityScapes 2Task dataset and outperforms the established state-of-the-art by 3.97% on the NYU V2 3Task dataset, respectively, when the bandwidth ratio of the communication channel (i.e., compression level for transmission over the channel) is as constrained as 1 12 .
2501.02007
TART: Token-based Architecture Transformer for Neural Network Performance Prediction
cs.LG cs.AI
In the realm of neural architecture design, achieving high performance is largely reliant on the manual expertise of researchers. Despite the emergence of Neural Architecture Search (NAS) as a promising technique for automating this process, current NAS methods still require human input to expand the search space and cannot generate new architectures. This paper explores the potential of Transformers in comprehending neural architectures and their performance, with the objective of establishing the foundation for utilizing Transformers to generate novel networks. We propose the Token-based Architecture Transformer (TART), which predicts neural network performance without the need to train candidate networks. TART attains state-of-the-art performance on the DeepNets-1M dataset for performance prediction tasks without edge information, indicating the potential of Transformers to aid in discovering novel and high-performing neural architectures.
2501.02008
Integrated Strategy for Urban Traffic Optimization: Prediction, Adaptive Signal Control, and Distributed Communication via Messaging
eess.SY cs.SY eess.SP
This work introduces an integrated approach to optimizing urban traffic by combining predictive modeling of vehicle flow, adaptive traffic signal control, and a modular integration architecture through distributed messaging. Using real-time data from various sensors, the system anticipates traffic fluctuations and dynamically adjusts signal phase durations to minimize delays and improve traffic flow. This proactive adjustment, supported by algorithms inspired by simulated annealing and reinforcement learning, also enhances energy efficiency, reduces pollutant emissions, and responds effectively to unexpected events (adverse weather, accidents, or temporary gatherings). Preliminary simulations conducted in a realistic urban environment demonstrate a significant reduction in average waiting times. Future developments include incorporating data from connected vehicles, integrating new modes of transport, and continuously refining predictive models to address the growing challenges of urban mobility.
2501.02009
Cross-model Transferability among Large Language Models on the Platonic Representations of Concepts
cs.CL cs.AI
Understanding the inner workings of Large Language Models (LLMs) is a critical research frontier. Prior research has shown that a single LLM's concept representations can be captured as steering vectors (SVs), enabling the control of LLM behavior (e.g., towards generating harmful content). Our work takes a novel approach by exploring the intricate relationships between concept representations across different LLMs, drawing an intriguing parallel to Plato's Allegory of the Cave. In particular, we introduce a linear transformation method to bridge these representations and present three key findings: 1) Concept representations across different LLMs can be effectively aligned using simple linear transformations, enabling efficient cross-model transfer and behavioral control via SVs. 2) This linear transformation generalizes across concepts, facilitating alignment and control of SVs representing different concepts across LLMs. 3) A weak-to-strong transferability exists between LLM concept representations, whereby SVs extracted from smaller LLMs can effectively control the behavior of larger LLMs.
2501.02010
Explainable Neural Networks with Guarantees: A Sparse Estimation Approach
cs.LG
Balancing predictive power and interpretability has long been a challenging research area, particularly in powerful yet complex models like neural networks, where nonlinearity obstructs direct interpretation. This paper introduces a novel approach to constructing an explainable neural network that harmonizes predictiveness and explainability. Our model, termed SparXnet, is designed as a linear combination of a sparse set of jointly learned features, each derived from a different trainable function applied to a single 1-dimensional input feature. Leveraging the ability to learn arbitrarily complex relationships, our neural network architecture enables automatic selection of a sparse set of important features, with the final prediction being a linear combination of rescaled versions of these features. We demonstrate the ability to select significant features while maintaining comparable predictive performance and direct interpretability through extensive experiments on synthetic and real-world datasets. We also provide theoretical analysis on the generalization bounds of our framework, which is favorably linear in the number of selected features and only logarithmic in the number of input features. We further lift any dependence of sample complexity on the number of parameters or the architectural details under very mild conditions. Our research paves the way for further research on sparse and explainable neural networks with guarantee.
2501.02012
Information Subtraction: Learning Representations for Conditional Entropy
cs.LG
The representations of conditional entropy and conditional mutual information are significant in explaining the unique effects among variables. While previous studies based on conditional contrastive sampling have effectively removed information regarding discrete sensitive variables, they have not yet extended their scope to continuous cases. This paper introduces Information Subtraction, a framework designed to generate representations that preserve desired information while eliminating the undesired. We implement a generative-based architecture that outputs these representations by simultaneously maximizing an information term and minimizing another. With its flexibility in disentangling information, we can iteratively apply Information Subtraction to represent arbitrary information components between continuous variables, thereby explaining the various relationships that exist between them. Our results highlight the representations' ability to provide semantic features of conditional entropy. By subtracting sensitive and domain-specific information, our framework demonstrates effective performance in fair learning and domain generalization. The code for this paper is available at https://github.com/jh-liang/Information-Subtraction
2501.02014
Machine Learning-Based Differential Diagnosis of Parkinson's Disease Using Kinematic Feature Extraction and Selection
cs.LG cs.AI
Parkinson's disease (PD), the second most common neurodegenerative disorder, is characterized by dopaminergic neuron loss and the accumulation of abnormal synuclein. PD presents both motor and non-motor symptoms that progressively impair daily functioning. The severity of these symptoms is typically assessed using the MDS-UPDRS rating scale, which is subjective and dependent on the physician's experience. Additionally, PD shares symptoms with other neurodegenerative diseases, such as progressive supranuclear palsy (PSP) and multiple system atrophy (MSA), complicating accurate diagnosis. To address these diagnostic challenges, we propose a machine learning-based system for differential diagnosis of PD, PSP, MSA, and healthy controls (HC). This system utilizes a kinematic feature-based hierarchical feature extraction and selection approach. Initially, 18 kinematic features are extracted, including two newly proposed features: Thumb-to-index vector velocity and acceleration, which provide insights into motor control patterns. In addition, 41 statistical features were extracted here from each kinematic feature, including some new approaches such as Average Absolute Change, Rhythm, Amplitude, Frequency, Standard Deviation of Frequency, and Slope. Feature selection is performed using One-way ANOVA to rank features, followed by Sequential Forward Floating Selection (SFFS) to identify the most relevant ones, aiming to reduce the computational complexity. The final feature set is used for classification, achieving a classification accuracy of 66.67% for each dataset and 88.89% for each patient, with particularly high performance for the MSA and HC groups using the SVM algorithm. This system shows potential as a rapid and accurate diagnostic tool in clinical practice, though further data collection and refinement are needed to enhance its reliability.
2501.02015
KANS: Knowledge Discovery Graph Attention Network for Soft Sensing in Multivariate Industrial Processes
cs.LG cs.AI cs.SY eess.SP eess.SY
Soft sensing of hard-to-measure variables is often crucial in industrial processes. Current practices rely heavily on conventional modeling techniques that show success in improving accuracy. However, they overlook the non-linear nature, dynamics characteristics, and non-Euclidean dependencies between complex process variables. To tackle these challenges, we present a framework known as a Knowledge discovery graph Attention Network for effective Soft sensing (KANS). Unlike the existing deep learning soft sensor models, KANS can discover the intrinsic correlations and irregular relationships between the multivariate industrial processes without a predefined topology. First, an unsupervised graph structure learning method is introduced, incorporating the cosine similarity between different sensor embedding to capture the correlations between sensors. Next, we present a graph attention-based representation learning that can compute the multivariate data parallelly to enhance the model in learning complex sensor nodes and edges. To fully explore KANS, knowledge discovery analysis has also been conducted to demonstrate the interpretability of the model. Experimental results demonstrate that KANS significantly outperforms all the baselines and state-of-the-art methods in soft sensing performance. Furthermore, the analysis shows that KANS can find sensors closely related to different process variables without domain knowledge, significantly improving soft sensing accuracy.
2501.02016
ST-HCSS: Deep Spatio-Temporal Hypergraph Convolutional Neural Network for Soft Sensing
cs.LG cs.AI eess.SP
Higher-order sensor networks are more accurate in characterizing the nonlinear dynamics of sensory time-series data in modern industrial settings by allowing multi-node connections beyond simple pairwise graph edges. In light of this, we propose a deep spatio-temporal hypergraph convolutional neural network for soft sensing (ST-HCSS). In particular, our proposed framework is able to construct and leverage a higher-order graph (hypergraph) to model the complex multi-interactions between sensor nodes in the absence of prior structural knowledge. To capture rich spatio-temporal relationships underlying sensor data, our proposed ST-HCSS incorporates stacked gated temporal and hypergraph convolution layers to effectively aggregate and update hypergraph information across time and nodes. Our results validate the superiority of ST-HCSS compared to existing state-of-the-art soft sensors, and demonstrates that the learned hypergraph feature representations aligns well with the sensor data correlations. The code is available at https://github.com/htew0001/ST-HCSS.git
2501.02017
Rephotography in the Digital Era: Mass Rephotography and re.photos, the Web Portal for Rephotography
cs.CV
Since the beginning of rephotography in the middle of the 19th century, techniques in registration, conservation, presentation, and sharing of rephotographs have come a long way. Here, we will present existing digital approaches to rephotography and discuss future approaches and requirements for digital mass rephotography. We present re.photos, an existing web portal for rephotography, featuring methods for collaborative rephotography, interactive image registration, as well as retrieval, organization, and sharing of rephotographs. For mass rephotography additional requirements must be met. Batches of template images and rephotographs must be handled simultaneously, image registration must be automated, and intuitive smartphone apps for rephotography must be available. Long--term storage with persistent identifiers, automatic or mass georeferencing, as well as gamification and social media integration are further requirements we will discuss in this paper.
2501.02018
Safeguarding Large Language Models in Real-time with Tunable Safety-Performance Trade-offs
cs.CL cs.AI cs.CR cs.LG
Large Language Models (LLMs) have been shown to be susceptible to jailbreak attacks, or adversarial attacks used to illicit high risk behavior from a model. Jailbreaks have been exploited by cybercriminals and blackhat actors to cause significant harm, highlighting the critical need to safeguard widely-deployed models. Safeguarding approaches, which include fine-tuning models or having LLMs "self-reflect", may lengthen the inference time of a model, incur a computational penalty, reduce the semantic fluency of an output, and restrict ``normal'' model behavior. Importantly, these Safety-Performance Trade-offs (SPTs) remain an understudied area. In this work, we introduce a novel safeguard, called SafeNudge, that combines Controlled Text Generation with "nudging", or using text interventions to change the behavior of a model. SafeNudge triggers during text-generation while a jailbreak attack is being executed, and can reduce successful jailbreak attempts by 30% by guiding the LLM towards a safe responses. It adds minimal latency to inference and has a negligible impact on the semantic fluency of outputs. Further, we allow for tunable SPTs. SafeNudge is open-source and available through https://pypi.org/, and is compatible with models loaded with the Hugging Face "transformers" library.
2501.02019
Benchmarking Constraint-Based Bayesian Structure Learning Algorithms: Role of Network Topology
cs.LG cs.AI q-bio.MN
Modeling the associations between real world entities from their multivariate cross-sectional profiles can provide cues into the concerted working of these entities as a system. Several techniques have been proposed for deciphering these associations including constraint-based Bayesian structure learning (BSL) algorithms that model them as directed acyclic graphs. Benchmarking these algorithms have typically focused on assessing the variation in performance measures such as sensitivity as a function of the dimensionality represented by the number of nodes in the DAG, and sample size. The present study elucidates the importance of network topology in benchmarking exercises. More specifically, it investigates variations in sensitivity across distinct network topologies while constraining the nodes, edges, and sample-size to be identical, eliminating these as potential confounders. Sensitivity of three popular constraint-based BSL algorithms (Peter-Clarke, Grow-Shrink, Incremental Association Markov Blanket) in learning the network structure from multivariate cross-sectional profiles sampled from network models with sub-linear, linear, and super-linear DAG topologies generated using preferential attachment is investigated. Results across linear and nonlinear models revealed statistically significant $(\alpha=0.05)$ decrease in sensitivity estimates from sub-linear to super-linear topology constitutively across the three algorithms. These results are demonstrated on networks with nodes $(N_{nods}=48,64)$, noise strengths $(\sigma =3,6)$ and sample size $(N = 2^{10})$. The findings elucidate the importance of accommodating the network topology in constraint-based BSL benchmarking exercises.
2501.02020
Enhancing Uncertainty Modeling with Semantic Graph for Hallucination Detection
cs.CL cs.AI
Large Language Models (LLMs) are prone to hallucination with non-factual or unfaithful statements, which undermines the applications in real-world scenarios. Recent researches focus on uncertainty-based hallucination detection, which utilizes the output probability of LLMs for uncertainty calculation and does not rely on external knowledge or frequent sampling from LLMs. Whereas, most approaches merely consider the uncertainty of each independent token, while the intricate semantic relations among tokens and sentences are not well studied, which limits the detection of hallucination that spans over multiple tokens and sentences in the passage. In this paper, we propose a method to enhance uncertainty modeling with semantic graph for hallucination detection. Specifically, we first construct a semantic graph that well captures the relations among entity tokens and sentences. Then, we incorporate the relations between two entities for uncertainty propagation to enhance sentence-level hallucination detection. Given that hallucination occurs due to the conflict between sentences, we further present a graph-based uncertainty calibration method that integrates the contradiction probability of the sentence with its neighbors in the semantic graph for uncertainty calculation. Extensive experiments on two datasets show the great advantages of our proposed approach. In particular, we obtain substantial improvements with 19.78% in passage-level hallucination detection.
2501.02021
Weakly Supervised Learning on Large Graphs
cs.LG cs.AI
Graph classification plays a pivotal role in various domains, including pathology, where images can be represented as graphs. In this domain, images can be represented as graphs, where nodes might represent individual nuclei, and edges capture the spatial or functional relationships between them. Often, the overall label of the graph, such as a cancer type or disease state, is determined by patterns within smaller, localized regions of the image. This work introduces a weakly-supervised graph classification framework leveraging two subgraph extraction techniques: (1) Sliding-window approach (2) BFS-based approach. Subgraphs are processed using a Graph Attention Network (GAT), which employs attention mechanisms to identify the most informative subgraphs for classification. Weak supervision is achieved by propagating graph-level labels to subgraphs, eliminating the need for detailed subgraph annotations.
2501.02024
Model Checking in Medical Imaging for Tumor Detection and Segmentation
cs.CV cs.AI cs.LG
Recent advancements in model checking have demonstrated significant potential across diverse applications, particularly in signal and image analysis. Medical imaging stands out as a critical domain where model checking can be effectively applied to design and evaluate robust frameworks. These frameworks facilitate automatic and semi-automatic delineation of regions of interest within images, aiding in accurate segmentation. This paper provides a comprehensive analysis of recent works leveraging spatial logic to develop operators and tools for identifying regions of interest, including tumorous and non-tumorous areas. Additionally, we examine the challenges inherent to spatial model-checking techniques, such as variability in ground truth data and the need for streamlined procedures suitable for routine clinical practice.
2501.02025
RealDiffFusionNet: Neural Controlled Differential Equation Informed Multi-Head Attention Fusion Networks for Disease Progression Modeling Using Real-World Data
cs.LG cs.CV q-bio.QM
This paper presents a novel deep learning-based approach named RealDiffFusionNet incorporating Neural Controlled Differential Equations (Neural CDE) - time series models that are robust in handling irregularly sampled data - and multi-head attention to align relevant multimodal context (image data, time invariant data, etc.) at each time point. Long short-term memory (LSTM) models were also used as a baseline. Two different datasets were used: a data from the Open-Source Imaging Consortium (OSIC) containing structured time series data of demographics and lung function with a baseline CT scan of the lungs and the second from the Alzheimer's Disease Neuroimaging Initiative (ADNI) containing a series of MRI scans along with demographics, physical examinations, and cognitive assessment data. An ablation study was performed to understand the role of CDEs, multimodal data, attention fusion, and interpolation strategies on model performance. When the baseline models were evaluated, the use of multimodal data resulted in an improvement in Neural CDE performance, with a lower test RMSE. Additionally, the performance of multimodal Neural CDE was also superior to multimodal LSTM. In the attention-based architectures, fusion through concatenation and rectilinear interpolation were found to improve model performance. The performance of the proposed RealDiffFusionNet was found to be superior (0.2570) to all models. For the ADNI dataset, between the Neural-CDE and LSTM models trained only on the structured data, the test RMSE were comparable (0.471 for LSTM vs. 0.4581 Neural-CDE). Furthermore, the addition of image features from patients' MRI series resulted in an improvement in performance, with a lower test RMSE (0.4372 with multimodal vs 0.4581 with structured data). RealDiffFusionNet has shown promise in utilizing CDEs and multimodal data to accurately predict disease progression.
2501.02026
Recursive Decomposition of Logical Thoughts: Framework for Superior Reasoning and Knowledge Propagation in Large Language Models
cs.CL cs.AI cs.LG cs.LO
Enhancing the reasoning capabilities of Large Language Models remains a critical challenge in artificial intelligence. We introduce RDoLT, Recursive Decomposition of Logical Thought prompting, a novel framework that significantly boosts LLM reasoning performance. RDoLT is built on three key innovations: (1) recursively breaking down complex reasoning tasks into sub-tasks of progressive complexity; (2) employing an advanced selection and scoring mechanism to identify the most promising reasoning thoughts; and (3) integrating a knowledge propagation module that mimics human learning by keeping track of strong and weak thoughts for information propagation. Our approach was evaluated across multiple benchmarks, including GSM8K, SVAMP, MultiArith, LastLetterConcatenation, and Gaokao2023 Math. The results demonstrate that RDoLT consistently outperforms existing state-of-the-art techniques, achieving a 90.98 percent accuracy on GSM8K with ChatGPT-4, surpassing state-of-the-art techniques by 6.28 percent. Similar improvements were observed on other benchmarks, with accuracy gains ranging from 5.5 percent to 6.75 percent. These findings highlight RDoLT's potential to advance prompt engineering, offering a more effective and generalizable approach to complex reasoning tasks.
2501.02029
Spot Risks Before Speaking! Unraveling Safety Attention Heads in Large Vision-Language Models
cs.LG cs.AI cs.CR cs.CV
With the integration of an additional modality, large vision-language models (LVLMs) exhibit greater vulnerability to safety risks (e.g., jailbreaking) compared to their language-only predecessors. Although recent studies have devoted considerable effort to the post-hoc alignment of LVLMs, the inner safety mechanisms remain largely unexplored. In this paper, we discover that internal activations of LVLMs during the first token generation can effectively identify malicious prompts across different attacks. This inherent safety perception is governed by sparse attention heads, which we term ``safety heads." Further analysis reveals that these heads act as specialized shields against malicious prompts; ablating them leads to higher attack success rates, while the model's utility remains unaffected. By locating these safety heads and concatenating their activations, we construct a straightforward but powerful malicious prompt detector that integrates seamlessly into the generation process with minimal extra inference overhead. Despite its simple structure of a logistic regression model, the detector surprisingly exhibits strong zero-shot generalization capabilities. Experiments across various prompt-based attacks confirm the effectiveness of leveraging safety heads to protect LVLMs. Code is available at \url{https://github.com/Ziwei-Zheng/SAHs}.
2501.02030
Detecting Music Performance Errors with Transformers
cs.SD cs.AI eess.AS
Beginner musicians often struggle to identify specific errors in their performances, such as playing incorrect notes or rhythms. There are two limitations in existing tools for music error detection: (1) Existing approaches rely on automatic alignment; therefore, they are prone to errors caused by small deviations between alignment targets.; (2) There is a lack of sufficient data to train music error detection models, resulting in over-reliance on heuristics. To address (1), we propose a novel transformer model, Polytune, that takes audio inputs and outputs annotated music scores. This model can be trained end-to-end to implicitly align and compare performance audio with music scores through latent space representations. To address (2), we present a novel data generation technique capable of creating large-scale synthetic music error datasets. Our approach achieves a 64.1% average Error Detection F1 score, improving upon prior work by 40 percentage points across 14 instruments. Additionally, compared with existing transcription methods repurposed for music error detection, our model can handle multiple instruments. Our source code and datasets are available at https://github.com/ben2002chou/Polytune.
2501.02031
CarbonChat: Large Language Model-Based Corporate Carbon Emission Analysis and Climate Knowledge Q&A System
cs.CL cs.AI
As the impact of global climate change intensifies, corporate carbon emissions have become a focal point of global attention. In response to issues such as the lag in climate change knowledge updates within large language models, the lack of specialization and accuracy in traditional augmented generation architectures for complex problems, and the high cost and time consumption of sustainability report analysis, this paper proposes CarbonChat: Large Language Model-based corporate carbon emission analysis and climate knowledge Q&A system, aimed at achieving precise carbon emission analysis and policy understanding.First, a diversified index module construction method is proposed to handle the segmentation of rule-based and long-text documents, as well as the extraction of structured data, thereby optimizing the parsing of key information.Second, an enhanced self-prompt retrieval-augmented generation architecture is designed, integrating intent recognition, structured reasoning chains, hybrid retrieval, and Text2SQL, improving the efficiency of semantic understanding and query conversion.Next, based on the greenhouse gas accounting framework, 14 dimensions are established for carbon emission analysis, enabling report summarization, relevance evaluation, and customized responses.Finally, through a multi-layer chunking mechanism, timestamps, and hallucination detection features, the accuracy and verifiability of the analysis results are ensured, reducing hallucination rates and enhancing the precision of the responses.
2501.02032
Dynamic Feature Fusion: Combining Global Graph Structures and Local Semantics for Blockchain Fraud Detection
cs.CR cs.AI cs.SE
The advent of blockchain technology has facilitated the widespread adoption of smart contracts in the financial sector. However, current fraud detection methodologies exhibit limitations in capturing both global structural patterns within transaction networks and local semantic relationships embedded in transaction data. Most existing models focus on either structural information or semantic features individually, leading to suboptimal performance in detecting complex fraud patterns.In this paper, we propose a dynamic feature fusion model that combines graph-based representation learning and semantic feature extraction for blockchain fraud detection. Specifically, we construct global graph representations to model account relationships and extract local contextual features from transaction data. A dynamic multimodal fusion mechanism is introduced to adaptively integrate these features, enabling the model to capture both structural and semantic fraud patterns effectively. We further develop a comprehensive data processing pipeline, including graph construction, temporal feature enhancement, and text preprocessing. Experimental results on large-scale real-world blockchain datasets demonstrate that our method outperforms existing benchmarks across accuracy, F1 score, and recall metrics. This work highlights the importance of integrating structural relationships and semantic similarities for robust fraud detection and offers a scalable solution for securing blockchain systems.
2501.02035
3D Cloud reconstruction through geospatially-aware Masked Autoencoders
cs.CV cs.AI
Clouds play a key role in Earth's radiation balance with complex effects that introduce large uncertainties into climate models. Real-time 3D cloud data is essential for improving climate predictions. This study leverages geostationary imagery from MSG/SEVIRI and radar reflectivity measurements of cloud profiles from CloudSat/CPR to reconstruct 3D cloud structures. We first apply self-supervised learning (SSL) methods-Masked Autoencoders (MAE) and geospatially-aware SatMAE on unlabelled MSG images, and then fine-tune our models on matched image-profile pairs. Our approach outperforms state-of-the-art methods like U-Nets, and our geospatial encoding further improves prediction results, demonstrating the potential of SSL for cloud reconstruction.
2501.02036
Deep Clustering via Community Detection
cs.LG cs.AI cs.SI
Deep clustering is an essential task in modern artificial intelligence, aiming to partition a set of data samples into a given number of homogeneous groups (i.e., clusters). Even though many Deep Neural Network (DNN) backbones and clustering strategies have been proposed for the task, achieving increasingly improved performance, deep clustering remains very challenging due to the lack of accurately labeled samples. In this paper, we propose a novel approach of deep clustering via community detection. It initializes clustering by detecting many communities, and then gradually expands clusters by community merging. Compared with the existing clustering strategies, community detection factors in the new perspective of cluster network analysis. As a result, it has the inherent benefit of high pseudo-label purity, which is critical to the performance of self-supervision. We have validated the efficacy of the proposed approach on benchmark image datasets. Our extensive experiments have shown that it can effectively improve the SOTA performance. Our ablation study also demonstrates that the new network perspective can effectively improve community pseudo-label purity, resulting in improved clustering performance.
2501.02038
Architecture for Trajectory-Based Fishing Ship Classification with AIS Data
cs.LG cs.AI
This paper proposes a data preparation process for managing real-world kinematic data and detecting fishing vessels. The solution is a binary classification that classifies ship trajectories into either fishing or non-fishing ships. The data used are characterized by the typical problems found in classic data mining applications using real-world data, such as noise and inconsistencies. The two classes are also clearly unbalanced in the data, a problem which is addressed using algorithms that resample the instances. For classification, a series of features are extracted from spatiotemporal data that represent the trajectories of the ships, available from sequences of Automatic Identification System (AIS) reports. These features are proposed for the modelling of ship behavior but, because they do not contain context-related information, the classification can be applied in other scenarios. Experimentation shows that the proposed data preparation process is useful for the presented classification problem. In addition, positive results are obtained using minimal information.
2501.02039
An Investigation into Value Misalignment in LLM-Generated Texts for Cultural Heritage
cs.CL cs.AI
As Large Language Models (LLMs) become increasingly prevalent in tasks related to cultural heritage, such as generating descriptions of historical monuments, translating ancient texts, preserving oral traditions, and creating educational content, their ability to produce accurate and culturally aligned texts is being increasingly relied upon by users and researchers. However, cultural value misalignments may exist in generated texts, such as the misrepresentation of historical facts, the erosion of cultural identity, and the oversimplification of complex cultural narratives, which may lead to severe consequences. Therefore, investigating value misalignment in the context of LLM for cultural heritage is crucial for mitigating these risks, yet there has been a significant lack of systematic and comprehensive study and investigation in this area. To fill this gap, we systematically assess the reliability of LLMs in generating culturally aligned texts for cultural heritage-related tasks. We conduct a comprehensive evaluation by compiling an extensive set of 1066 query tasks covering 5 widely recognized categories with 17 aspects within the knowledge framework of cultural heritage across 5 open-source LLMs, and examine both the type and rate of cultural value misalignments in the generated texts. Using both automated and manual approaches, we effectively detect and analyze the cultural value misalignments in LLM-generated texts. Our findings are concerning: over 65% of the generated texts exhibit notable cultural misalignments, with certain tasks demonstrating almost complete misalignment with key cultural values. Beyond these findings, this paper introduces a benchmark dataset and a comprehensive evaluation workflow that can serve as a valuable resource for future research aimed at enhancing the cultural sensitivity and reliability of LLMs.
2501.02040
A Separable Self-attention Inspired by the State Space Model for Computer Vision
cs.CV cs.AI
Mamba is an efficient State Space Model (SSM) with linear computational complexity. Although SSMs are not suitable for handling non-causal data, Vision Mamba (ViM) methods still demonstrate good performance in tasks such as image classification and object detection. Recent studies have shown that there is a rich theoretical connection between state space models and attention variants. We propose a novel separable self attention method, for the first time introducing some excellent design concepts of Mamba into separable self-attention. To ensure a fair comparison with ViMs, we introduce VMINet, a simple yet powerful prototype architecture, constructed solely by stacking our novel attention modules with the most basic down-sampling layers. Notably, VMINet differs significantly from the conventional Transformer architecture. Our experiments demonstrate that VMINet has achieved competitive results on image classification and high-resolution dense prediction tasks.Code is available at: \url{https://github.com/yws-wxs/VMINet}.
2501.02041
MRG: A Multi-Robot Manufacturing Digital Scene Generation Method Using Multi-Instance Point Cloud Registration
cs.CV cs.AI
A high-fidelity digital simulation environment is crucial for accurately replicating physical operational processes. However, inconsistencies between simulation and physical environments result in low confidence in simulation outcomes, limiting their effectiveness in guiding real-world production. Unlike the traditional step-by-step point cloud "segmentation-registration" generation method, this paper introduces, for the first time, a novel Multi-Robot Manufacturing Digital Scene Generation (MRG) method that leverages multi-instance point cloud registration, specifically within manufacturing scenes. Tailored to the characteristics of industrial robots and manufacturing settings, an instance-focused transformer module is developed to delineate instance boundaries and capture correlations between local regions. Additionally, a hypothesis generation module is proposed to extract target instances while preserving key features. Finally, an efficient screening and optimization algorithm is designed to refine the final registration results. Experimental evaluations on the Scan2CAD and Welding-Station datasets demonstrate that: (1) the proposed method outperforms existing multi-instance point cloud registration techniques; (2) compared to state-of-the-art methods, the Scan2CAD dataset achieves improvements in MR and MP by 12.15% and 17.79%, respectively; and (3) on the Welding-Station dataset, MR and MP are enhanced by 16.95% and 24.15%, respectively. This work marks the first application of multi-instance point cloud registration in manufacturing scenes, significantly advancing the precision and reliability of digital simulation environments for industrial applications.
2501.02042
Towards Robust and Accurate Stability Estimation of Local Surrogate Models in Text-based Explainable AI
cs.LG cs.CR
Recent work has investigated the concept of adversarial attacks on explainable AI (XAI) in the NLP domain with a focus on examining the vulnerability of local surrogate methods such as Lime to adversarial perturbations or small changes on the input of a machine learning (ML) model. In such attacks, the generated explanation is manipulated while the meaning and structure of the original input remain similar under the ML model. Such attacks are especially alarming when XAI is used as a basis for decision making (e.g., prescribing drugs based on AI medical predictors) or for legal action (e.g., legal dispute involving AI software). Although weaknesses across many XAI methods have been shown to exist, the reasons behind why remain little explored. Central to this XAI manipulation is the similarity measure used to calculate how one explanation differs from another. A poor choice of similarity measure can lead to erroneous conclusions about the stability or adversarial robustness of an XAI method. Therefore, this work investigates a variety of similarity measures designed for text-based ranked lists referenced in related work to determine their comparative suitability for use. We find that many measures are overly sensitive, resulting in erroneous estimates of stability. We then propose a weighting scheme for text-based data that incorporates the synonymity between the features within an explanation, providing more accurate estimates of the actual weakness of XAI methods to adversarial examples.
2501.02043
Modeling COVID-19 spread in the USA using metapopulation SIR models coupled with graph convolutional neural networks
stat.ML cs.LG math.DS q-bio.PE
Graph convolutional neural networks (GCNs) have shown tremendous promise in addressing data-intensive challenges in recent years. In particular, some attempts have been made to improve predictions of Susceptible-Infected-Recovered (SIR) models by incorporating human mobility between metapopulations and using graph approaches to estimate corresponding hyperparameters. Recently, researchers have found that a hybrid GCN-SIR approach outperformed existing methodologies when used on the data collected on a precinct level in Japan. In our work, we extend this approach to data collected from the continental US, adjusting for the differing mobility patterns and varying policy responses. We also develop the strategy for real-time continuous estimation of the reproduction number and study the accuracy of model predictions for the overall population as well as individual states. Strengths and limitations of the GCN-SIR approach are discussed as a potential candidate for modeling disease dynamics.
2501.02044
Advancing Pancreatic Cancer Prediction with a Next Visit Token Prediction Head on top of Med-BERT
cs.CL cs.AI
Background: Recently, numerous foundation models pretrained on extensive data have demonstrated efficacy in disease prediction using Electronic Health Records (EHRs). However, there remains some unanswered questions on how to best utilize such models especially with very small fine-tuning cohorts. Methods: We utilized Med-BERT, an EHR-specific foundation model, and reformulated the disease binary prediction task into a token prediction task and a next visit mask token prediction task to align with Med-BERT's pretraining task format in order to improve the accuracy of pancreatic cancer (PaCa) prediction in both few-shot and fully supervised settings. Results: The reformulation of the task into a token prediction task, referred to as Med-BERT-Sum, demonstrates slightly superior performance in both few-shot scenarios and larger data samples. Furthermore, reformulating the prediction task as a Next Visit Mask Token Prediction task (Med-BERT-Mask) significantly outperforms the conventional Binary Classification (BC) prediction task (Med-BERT-BC) by 3% to 7% in few-shot scenarios with data sizes ranging from 10 to 500 samples. These findings highlight that aligning the downstream task with Med-BERT's pretraining objectives substantially enhances the model's predictive capabilities, thereby improving its effectiveness in predicting both rare and common diseases. Conclusion: Reformatting disease prediction tasks to align with the pretraining of foundation models enhances prediction accuracy, leading to earlier detection and timely intervention. This approach improves treatment effectiveness, survival rates, and overall patient outcomes for PaCa and potentially other cancers.
2501.02045
METAGENE-1: Metagenomic Foundation Model for Pandemic Monitoring
q-bio.GN cs.AI cs.CL cs.LG
We pretrain METAGENE-1, a 7-billion-parameter autoregressive transformer model, which we refer to as a metagenomic foundation model, on a novel corpus of diverse metagenomic DNA and RNA sequences comprising over 1.5 trillion base pairs. This dataset is sourced from a large collection of human wastewater samples, processed and sequenced using deep metagenomic (next-generation) sequencing methods. Unlike genomic models that focus on individual genomes or curated sets of specific species, the aim of METAGENE-1 is to capture the full distribution of genomic information present within this wastewater, to aid in tasks relevant to pandemic monitoring and pathogen detection. We carry out byte-pair encoding (BPE) tokenization on our dataset, tailored for metagenomic sequences, and then pretrain our model. In this paper, we first detail the pretraining dataset, tokenization strategy, and model architecture, highlighting the considerations and design choices that enable the effective modeling of metagenomic data. We then show results of pretraining this model on our metagenomic dataset, providing details about our losses, system metrics, and training stability over the course of pretraining. Finally, we demonstrate the performance of METAGENE-1, which achieves state-of-the-art results on a set of genomic benchmarks and new evaluations focused on human-pathogen detection and genomic sequence embedding, showcasing its potential for public health applications in pandemic monitoring, biosurveillance, and early detection of emerging health threats.
2501.02048
DreamMask: Boosting Open-vocabulary Panoptic Segmentation with Synthetic Data
cs.CV
Open-vocabulary panoptic segmentation has received significant attention due to its applicability in the real world. Despite claims of robust generalization, we find that the advancements of previous works are attributed mainly on trained categories, exposing a lack of generalization to novel classes. In this paper, we explore boosting existing models from a data-centric perspective. We propose DreamMask, which systematically explores how to generate training data in the open-vocabulary setting, and how to train the model with both real and synthetic data. For the first part, we propose an automatic data generation pipeline with off-the-shelf models. We propose crucial designs for vocabulary expansion, layout arrangement, data filtering, etc. Equipped with these techniques, our generated data could significantly outperform the manually collected web data. To train the model with generated data, a synthetic-real alignment loss is designed to bridge the representation gap, bringing noticeable improvements across multiple benchmarks. In general, DreamMask significantly simplifies the collection of large-scale training data, serving as a plug-and-play enhancement for existing methods. For instance, when trained on COCO and tested on ADE20K, the model equipped with DreamMask outperforms the previous state-of-the-art by a substantial margin of 2.1% mIoU.
2501.02059
Active Learning Enables Extrapolation in Molecular Generative Models
cs.LG cond-mat.mtrl-sci physics.chem-ph
Although generative models hold promise for discovering molecules with optimized desired properties, they often fail to suggest synthesizable molecules that improve upon the known molecules seen in training. We find that a key limitation is not in the molecule generation process itself, but in the poor generalization capabilities of molecular property predictors. We tackle this challenge by creating an active-learning, closed-loop molecule generation pipeline, whereby molecular generative models are iteratively refined on feedback from quantum chemical simulations to improve generalization to new chemical space. Compared against other generative model approaches, only our active learning approach generates molecules with properties that extrapolate beyond the training data (reaching up to 0.44 standard deviations beyond the training data range) and out-of-distribution molecule classification accuracy is improved by 79%. By conditioning molecular generation on thermodynamic stability data from the active-learning loop, the proportion of stable molecules generated is 3.5x higher than the next-best model.
2501.02063
AGGA: A Dataset of Academic Guidelines for Generative AI and Large Language Models
cs.CL cs.CY
This study introduces AGGA, a dataset comprising 80 academic guidelines for the use of Generative AIs (GAIs) and Large Language Models (LLMs) in academic settings, meticulously collected from official university websites. The dataset contains 188,674 words and serves as a valuable resource for natural language processing tasks commonly applied in requirements engineering, such as model synthesis, abstraction identification, and document structure assessment. Additionally, AGGA can be further annotated to function as a benchmark for various tasks, including ambiguity detection, requirements categorization, and the identification of equivalent requirements. Our methodologically rigorous approach ensured a thorough examination, with a selection of universities that represent a diverse range of global institutions, including top-ranked universities across six continents. The dataset captures perspectives from a variety of academic fields, including humanities, technology, and both public and private institutions, offering a broad spectrum of insights into the integration of GAIs and LLMs in academia.
2501.02064
ArtCrafter: Text-Image Aligning Style Transfer via Embedding Reframing
cs.CV cs.AI
Recent years have witnessed significant advancements in text-guided style transfer, primarily attributed to innovations in diffusion models. These models excel in conditional guidance, utilizing text or images to direct the sampling process. However, despite their capabilities, direct conditional guidance approaches often face challenges in balancing the expressiveness of textual semantics with the diversity of output results while capturing stylistic features. To address these challenges, we introduce ArtCrafter, a novel framework for text-to-image style transfer. Specifically, we introduce an attention-based style extraction module, meticulously engineered to capture the subtle stylistic elements within an image. This module features a multi-layer architecture that leverages the capabilities of perceiver attention mechanisms to integrate fine-grained information. Additionally, we present a novel text-image aligning augmentation component that adeptly balances control over both modalities, enabling the model to efficiently map image and text embeddings into a shared feature space. We achieve this through attention operations that enable smooth information flow between modalities. Lastly, we incorporate an explicit modulation that seamlessly blends multimodal enhanced embeddings with original embeddings through an embedding reframing design, empowering the model to generate diverse outputs. Extensive experiments demonstrate that ArtCrafter yields impressive results in visual stylization, exhibiting exceptional levels of stylistic intensity, controllability, and diversity.
2501.02066
RadHop-Net: A Lightweight Radiomics-to-Error Regression for False Positive Reduction In MRI Prostate Cancer Detection
cs.CV eess.IV
Clinically significant prostate cancer (csPCa) is a leading cause of cancer death in men, yet it has a high survival rate if diagnosed early. Bi-parametric MRI (bpMRI) reading has become a prominent screening test for csPCa. However, this process has a high false positive (FP) rate, incurring higher diagnostic costs and patient discomfort. This paper introduces RadHop-Net, a novel and lightweight CNN for FP reduction. The pipeline consists of two stages: Stage 1 employs data driven radiomics to extract candidate ROIs. In contrast, Stage 2 expands the receptive field about each ROI using RadHop-Net to compensate for the predicted error from Stage 1. Moreover, a novel loss function for regression problems is introduced to balance the influence between FPs and true positives (TPs). RadHop-Net is trained in a radiomics-to-error manner, thus decoupling from the common voxel-to-label approach. The proposed Stage 2 improves the average precision (AP) in lesion detection from 0.407 to 0.468 in the publicly available pi-cai dataset, also maintaining a significantly smaller model size than the state-of-the-art.
2501.02068
The interplay between domain specialization and model size: a case study in the legal domain
cs.CL cs.AI
Scaling laws for language models so far focused on finding the compute-optimal model size and token count for training from scratch. However, achieving this optimal balance requires significant compute resources due to the extensive data demands when training models from randomly-initialized weights. Continual pre-training offers a cost-effective alternative, leveraging the compute investment from pre-trained models to incorporate new knowledge without requiring extensive new data. Recent findings suggest that data quality influences constants in scaling laws, thereby altering the optimal parameter-token allocation ratio. Building on this insight, we investigate the interplay between domain specialization and model size during continual pre-training under compute-constrained scenarios. Our goal is to identify a compute-efficient training regime for this scenario and, potentially, detect patterns in this interplay that can be generalized across different model sizes and domains. To compare general and specialized training, we filtered a web-based dataset to extract legal domain data. We pre-trained models with 1.5B, 3B, 7B and 14B parameters on both the unfiltered and filtered datasets, then evaluated their performance on legal exams. Results show that as model size increases, the compute-effectiveness gap between specialized and general models widens.
2501.02069
Counterfactual Explanation for Auto-Encoder Based Time-Series Anomaly Detection
cs.LG
The complexity of modern electro-mechanical systems require the development of sophisticated diagnostic methods like anomaly detection capable of detecting deviations. Conventional anomaly detection approaches like signal processing and statistical modelling often struggle to effectively handle the intricacies of complex systems, particularly when dealing with multi-variate signals. In contrast, neural network-based anomaly detection methods, especially Auto-Encoders, have emerged as a compelling alternative, demonstrating remarkable performance. However, Auto-Encoders exhibit inherent opaqueness in their decision-making processes, hindering their practical implementation at scale. Addressing this opacity is essential for enhancing the interpretability and trustworthiness of anomaly detection models. In this work, we address this challenge by employing a feature selector to select features and counterfactual explanations to give a context to the model output. We tested this approach on the SKAB benchmark dataset and an industrial time-series dataset. The gradient based counterfactual explanation approach was evaluated via validity, sparsity and distance measures. Our experimental findings illustrate that our proposed counterfactual approach can offer meaningful and valuable insights into the model decision-making process, by explaining fewer signals compared to conventional approaches. These insights enhance the trustworthiness and interpretability of anomaly detection models.
2501.02071
Laws of thermodynamics for exponential families
cond-mat.stat-mech cs.LG math.ST stat.TH
We develop the laws of thermodynamics in terms of general exponential families. By casting learning (log-loss minimization) problems in max-entropy and statistical mechanics terms, we translate thermodynamics results to learning scenarios. We extend the well-known way in which exponential families characterize thermodynamic and learning equilibria. Basic ideas of work and heat, and advanced concepts of thermodynamic cycles and equipartition of energy, find exact and useful counterparts in AI / statistics terms. These ideas have broad implications for quantifying and addressing distribution shift.
2501.02077
Chance Constrained PDE-Constrained Optimal Design Strategies Under High-Dimensional Uncertainty
cs.CE
This study presents an advanced computational framework for the optimal design of thermal insulation components in buildings, utilizing silica aerogel porous materials. The framework aims to achieve superior thermal insulation while maintaining structural integrity of the component under stress concentrations. A multiphase continuum model is employed to simulate the thermomechanical behavior of the insulation system, governed by a set of coupled partial differential equations (PDEs). The design process explicitly accounts for spatially correlated uncertainties in the design parameters, particularly the spatially varying aerogel porosity, resulting in a high-dimensional, PDE-constrained optimization under uncertainty. The optimization problem is formulated using a risk-averse cost functional to balance insulation performance with uncertainty mitigation, incorporating statistical moments of the design objective. Additionally, chance constraints are imposed to limit the probability of stress exceeding critical thresholds. To address the challenges arising from high-dimensional parameters, the optimization leverages a second-order Taylor expansion of both the design objective and the chance constraint functions, combined with a low-rank approximation of the Hessian matrix for efficient evaluation of the generalized eigenvalue problem. This approach supports scalable computations, either directly or as a variance-reduction control variate for Monte Carlo estimations. Combined with a gradient-based optimization approach, we achieve a scalable solution algorithm with dimension-independent computational costs in terms of number of PDE solved. Two- and three-dimensional numerical experiments on the design of thermal breaks in buildings showcase the features of the proposed design under uncertainty framework with respect to accuracy, scalability, and computational cost.
2501.02080
AI-Powered Cow Detection in Complex Farm Environments
cs.CV
Animal welfare has become a critical issue in contemporary society, emphasizing our ethical responsibilities toward animals, particularly within livestock farming. The advent of Artificial Intelligence (AI) technologies, specifically computer vision, offers an innovative approach to monitoring and enhancing animal welfare. Cows, as essential contributors to sustainable agriculture, are central to this effort. However, existing cow detection algorithms face challenges in real-world farming environments, such as complex lighting, occlusions, pose variations, and background interference, hindering detection. Model generalization is crucial for adaptation across contexts beyond the training dataset. This study addresses these challenges using a diverse cow dataset from six environments, including indoor and outdoor scenarios. We propose a detection model combining YOLOv8 with the CBAM (Convolutional Block Attention Module) and assess its performance against baseline models, including Mask R-CNN, YOLOv5, and YOLOv8. Our findings show baseline models degrade in complex conditions, while our approach improves using CBAM. YOLOv8-CBAM outperformed YOLOv8 by 2.3% in mAP, achieving 95.2% precision and an mAP@0.5:0.95 of 82.6%, demonstrating superior accuracy. Contributions include (1) analyzing detection limitations, (2) proposing a robust model, and (3) benchmarking state-of-the-art algorithms. Applications include health monitoring, behavioral analysis, and tracking in smart farms, enabling precise detection in challenging settings. This study advances AI-driven livestock monitoring, improving animal welfare and smart agriculture.
2501.02086
Instruction-Following Pruning for Large Language Models
cs.CL
With the rapid scaling of large language models (LLMs), structured pruning has become a widely used technique to learn efficient, smaller models from larger ones, delivering superior performance compared to training similarly sized models from scratch. In this paper, we move beyond the traditional static pruning approach of determining a fixed pruning mask for a model, and propose a dynamic approach to structured pruning. In our method, the pruning mask is input-dependent and adapts dynamically based on the information described in a user instruction. Our approach, termed "instruction-following pruning", introduces a sparse mask predictor that takes the user instruction as input and dynamically selects the most relevant model parameters for the given task. To identify and activate effective parameters, we jointly optimize the sparse mask predictor and the LLM, leveraging both instruction-following data and the pre-training corpus. Experimental results demonstrate the effectiveness of our approach on a wide range of evaluation benchmarks. For example, our 3B activated model improves over the 3B dense model by 5-8 points of absolute margin on domains such as math and coding, and rivals the performance of a 9B model.
2501.02087
Beyond CVaR: Leveraging Static Spectral Risk Measures for Enhanced Decision-Making in Distributional Reinforcement Learning
cs.LG stat.ML
In domains such as finance, healthcare, and robotics, managing worst-case scenarios is critical, as failure to do so can lead to catastrophic outcomes. Distributional Reinforcement Learning (DRL) provides a natural framework to incorporate risk sensitivity into decision-making processes. However, existing approaches face two key limitations: (1) the use of fixed risk measures at each decision step often results in overly conservative policies, and (2) the interpretation and theoretical properties of the learned policies remain unclear. While optimizing a static risk measure addresses these issues, its use in the DRL framework has been limited to the simple static CVaR risk measure. In this paper, we present a novel DRL algorithm with convergence guarantees that optimizes for a broader class of static Spectral Risk Measures (SRM). Additionally, we provide a clear interpretation of the learned policy by leveraging the distribution of returns in DRL and the decomposition of static coherent risk measures. Extensive experiments demonstrate that our model learns policies aligned with the SRM objective, and outperforms existing risk-neutral and risk-sensitive DRL models in various settings.
2501.02089
On the Statistical Complexity for Offline and Low-Adaptive Reinforcement Learning with Structures
cs.LG cs.AI
This article reviews the recent advances on the statistical foundation of reinforcement learning (RL) in the offline and low-adaptive settings. We will start by arguing why offline RL is the appropriate model for almost any real-life ML problems, even if they have nothing to do with the recent AI breakthroughs that use RL. Then we will zoom into two fundamental problems of offline RL: offline policy evaluation (OPE) and offline policy learning (OPL). It may be surprising to people that tight bounds for these problems were not known even for tabular and linear cases until recently. We delineate the differences between worst-case minimax bounds and instance-dependent bounds. We also cover key algorithmic ideas and proof techniques behind near-optimal instance-dependent methods in OPE and OPL. Finally, we discuss the limitations of offline RL and review a burgeoning problem of \emph{low-adaptive exploration} which addresses these limitations by providing a sweet middle ground between offline and online RL.
2501.02090
Applying Text Mining to Analyze Human Question Asking in Creativity Research
cs.CL
Creativity relates to the ability to generate novel and effective ideas in the areas of interest. How are such creative ideas generated? One possible mechanism that supports creative ideation and is gaining increased empirical attention is by asking questions. Question asking is a likely cognitive mechanism that allows defining problems, facilitating creative problem solving. However, much is unknown about the exact role of questions in creativity. This work presents an attempt to apply text mining methods to measure the cognitive potential of questions, taking into account, among others, (a) question type, (b) question complexity, and (c) the content of the answer. This contribution summarizes the history of question mining as a part of creativity research, along with the natural language processing methods deemed useful or helpful in the study. In addition, a novel approach is proposed, implemented, and applied to five datasets. The experimental results obtained are comprehensively analyzed, suggesting that natural language processing has a role to play in creative research.
2501.02094
SMTL: A Stratified Logic for Expressive Multi-Level Temporal Specifications
eess.SY cs.SY
We present Stratified Metric Temporal Logic (SMTL), a novel formalism for specifying and verifying properties of complex cyber-physical systems that exhibit behaviors across multiple temporal and abstraction scales. SMTL extends existing temporal logics by incorporating a stratification operator, enabling the association of temporal properties with specific abstraction levels. This allows for the natural expression of multi-scale requirements while maintaining formal reasoning about inter-level relationships. We formalize the syntax and semantics of SMTL, proving that it strictly subsumes metric temporal logic (MTL) and offers enhanced expressiveness by capturing properties unattainable in existing logics. Numerical simulations comparing agents operating under MTL and SMTL specifications show that SMTL enhances agent coordination and safety, reducing collision rates without substantial computational overhead or compromising path efficiency. These findings underscore SMTL's potential as a valuable tool for designing and verifying complex multi-agent systems operating across diverse temporal and abstraction scales.
2501.02104
Equivalence of Informations Characterizes Bregman Divergences
cs.IT math.IT
Bregman divergences are a class of distance-like comparison functions which play fundamental roles in optimization, statistics, and information theory. One important property of Bregman divergences is that they cause two useful formulations of information content (in the sense of variability or non-uniformity) in a weighted collection of vectors to agree. In this note, we show that this agreement in fact characterizes the class of Bregman divergences; they are the only divergences which generate this agreement for arbitrary collections of weighted vectors.
2501.02105
Learning Fricke signs from Maass form Coefficients
math.NT cs.LG hep-th stat.ML
In this paper, we conduct a data-scientific investigation of Maass forms. We find that averaging the Fourier coefficients of Maass forms with the same Fricke sign reveals patterns analogous to the recently discovered "murmuration" phenomenon, and that these patterns become more pronounced when parity is incorporated as an additional feature. Approximately 43% of the forms in our dataset have an unknown Fricke sign. For the remaining forms, we employ Linear Discriminant Analysis (LDA) to machine learn their Fricke sign, achieving 96% (resp. 94%) accuracy for forms with even (resp. odd) parity. We apply the trained LDA model to forms with unknown Fricke signs to make predictions. The average values based on the predicted Fricke signs are computed and compared to those for forms with known signs to verify the reasonableness of the predictions. Additionally, a subset of these predictions is evaluated against heuristic guesses provided by Hejhal's algorithm, showing a match approximately 95% of the time. We also use neural networks to obtain results comparable to those from the LDA model.
2501.02107
Online Detection of Water Contamination Under Concept Drift
cs.LG cs.AI
Water Distribution Networks (WDNs) are vital infrastructures, and contamination poses serious public health risks. Harmful substances can interact with disinfectants like chlorine, making chlorine monitoring essential for detecting contaminants. However, chlorine sensors often become unreliable and require frequent calibration. This study introduces the Dual-Threshold Anomaly and Drift Detection (AD&DD) method, an unsupervised approach combining a dual-threshold drift detection mechanism with an LSTM-based Variational Autoencoder(LSTM-VAE) for real-time contamination detection. Tested on two realistic WDNs, AD&DD effectively identifies anomalies with sensor offsets as concept drift, and outperforms other methods. A proposed decentralized architecture enables accurate contamination detection and localization by deploying AD&DD on selected nodes.
2501.02111
How Your Location Relates to Health: Variable Importance and Interpretable Machine Learning for Environmental and Sociodemographic Data
cs.LG
Health outcomes depend on complex environmental and sociodemographic factors whose effects change over location and time. Only recently has fine-grained spatial and temporal data become available to study these effects, namely the MEDSAT dataset of English health, environmental, and sociodemographic information. Leveraging this new resource, we use a variety of variable importance techniques to robustly identify the most informative predictors across multiple health outcomes. We then develop an interpretable machine learning framework based on Generalized Additive Models (GAMs) and Multiscale Geographically Weighted Regression (MGWR) to analyze both local and global spatial dependencies of each variable on various health outcomes. Our findings identify NO2 as a global predictor for asthma, hypertension, and anxiety, alongside other outcome-specific predictors related to occupation, marriage, and vegetation. Regional analyses reveal local variations with air pollution and solar radiation, with notable shifts during COVID. This comprehensive approach provides actionable insights for addressing health disparities, and advocates for the integration of interpretable machine learning in public health.
2501.02112
Siamese Networks for Cat Re-Identification: Exploring Neural Models for Cat Instance Recognition
cs.CV cs.AI
Street cats in urban areas often rely on human intervention for survival, leading to challenges in population control and welfare management. In April 2023, Hello Inc., a Chinese urban mobility company, launched the Hello Street Cat initiative to address these issues. The project deployed over 21,000 smart feeding stations across 14 cities in China, integrating livestreaming cameras and treat dispensers activated through user donations. It also promotes the Trap-Neuter-Return (TNR) method, supported by a community-driven platform, HelloStreetCatWiki, where volunteers catalog and identify cats. However, manual identification is inefficient and unsustainable, creating a need for automated solutions. This study explores Deep Learning-based models for re-identifying street cats in the Hello Street Cat initiative. A dataset of 2,796 images of 69 cats was used to train Siamese Networks with EfficientNetB0, MobileNet and VGG16 as base models, evaluated under contrastive and triplet loss functions. VGG16 paired with contrastive loss emerged as the most effective configuration, achieving up to 97% accuracy and an F1 score of 0.9344 during testing. The approach leverages image augmentation and dataset refinement to overcome challenges posed by limited data and diverse visual variations. These findings underscore the potential of automated cat re-identification to streamline population monitoring and welfare efforts. By reducing reliance on manual processes, the method offers a scalable and reliable solution for communitydriven initiatives. Future research will focus on expanding datasets and developing real-time implementations to enhance practicality in large-scale deployments.
2501.02114
Relaxation-assisted reverse annealing on nonnegative/binary matrix factorization
quant-ph cond-mat.stat-mech cs.AI cs.LG
Quantum annealing has garnered significant attention as meta-heuristics inspired by quantum physics for combinatorial optimization problems. Among its many applications, nonnegative/binary matrix factorization stands out for its complexity and relevance in unsupervised machine learning. The use of reverse annealing, a derivative procedure of quantum annealing to prioritize the search in a vicinity under a given initial state, helps improve its optimization performance in matrix factorization. This study proposes an improved strategy that integrates reverse annealing with a linear programming relaxation technique. Using relaxed solutions as the initial configuration for reverse annealing, we demonstrate improvements in optimization performance comparable to the exact optimization methods. Our experiments on facial image datasets show that our method provides better convergence than known reverse annealing methods. Furthermore, we investigate the effectiveness of relaxation-based initialization methods on randomized datasets, demonstrating a relationship between the relaxed solution and the optimal solution. This research underscores the potential of combining reverse annealing and classical optimization strategies to enhance optimization performance.
2501.02116
Humanoid Locomotion and Manipulation: Current Progress and Challenges in Control, Planning, and Learning
cs.RO
Humanoid robots have great potential to perform various human-level skills. These skills involve locomotion, manipulation, and cognitive capabilities. Driven by advances in machine learning and the strength of existing model-based approaches, these capabilities have progressed rapidly, but often separately. Therefore, a timely overview of current progress and future trends in this fast-evolving field is essential. This survey first summarizes the model-based planning and control that have been the backbone of humanoid robotics for the past three decades. We then explore emerging learning-based methods, with a focus on reinforcement learning and imitation learning that enhance the versatility of loco-manipulation skills. We examine the potential of integrating foundation models with humanoid embodiments, assessing the prospects for developing generalist humanoid agents. In addition, this survey covers emerging research for whole-body tactile sensing that unlocks new humanoid skills that involve physical interactions. The survey concludes with a discussion of the challenges and future trends.
2501.02117
Fastest mixing reversible Markov chain on friendship graph: Trade-off between transition probabilities among friends and convergence rate
cs.IT math.IT
A long-standing goal of social network research has been to alter the properties of network to achieve the desired outcome. In doing so, DeGroot's consensus model has served as the popular choice for modeling the information diffusion and opinion formation in social networks. Achieving a trade-off between the cost associated with modifications made to the network and the speed of convergence to the desired state has shown to be a critical factor. This has been treated as the Fastest Mixing Markov Chain (FMMC) problem over a graph with given transition probabilities over a subset of edges. Addressing this multi-objective optimization problem over the friendship graph, this paper has provided the corresponding Pareto optimal points or the Pareto frontier. In the case of friendship graph with at least three blades, it is shown that the Pareto frontier is reduced to a global minimum point which is same as the optimal point corresponding to the minimum spanning tree of the friendship graph, i.e., the star topology. Furthermore, a lower limit for transition probabilities among friends has been provided, where values higher than this limit do not have any impact on the convergence rate.
2501.02127
How do Humans take an Object from a Robot: Behavior changes observed in a User Study
cs.RO
To facilitate human-robot interaction and gain human trust, a robot should recognize and adapt to changes in human behavior. This work documents different human behaviors observed while taking objects from an interactive robot in an experimental study, categorized across two dimensions: pull force applied and handedness. We also present the changes observed in human behavior upon repeated interaction with the robot to take various objects.
2501.02132
A hybrid marketplace of ideas
cs.CY cs.AI cs.ET
The convergence of humans and artificial intelligence systems introduces new dynamics into the cultural and intellectual landscape. Complementing emerging cultural evolution concepts such as machine culture, AI agents represent a significant techno-sociological development, particularly within the anthropological study of Web3 as a community focused on decentralization through blockchain. Despite their growing presence, the cultural significance of AI agents remains largely unexplored in academic literature. Toward this end, we conceived hybrid netnography, a novel interdisciplinary approach that examines the cultural and intellectual dynamics within digital ecosystems by analyzing the interactions and contributions of both human and AI agents as co-participants in shaping narratives, ideas, and cultural artifacts. We argue that, within the Web3 community on the social media platform X, these agents challenge traditional notions of participation and influence in public discourse, creating a hybrid marketplace of ideas, a conceptual space where human and AI generated ideas coexist and compete for attention. We examine the current state of AI agents in idea generation, propagation, and engagement, positioning their role as cultural agents through the lens of memetics and encouraging further inquiry into their cultural and societal impact. Additionally, we address the implications of this paradigm for privacy, intellectual property, and governance, highlighting the societal and legal challenges of integrating AI agents into the hybrid marketplace of ideas.
2501.02135
AVTrustBench: Assessing and Enhancing Reliability and Robustness in Audio-Visual LLMs
cs.CV cs.AI
With the rapid advancement of Multi-modal Large Language Models (MLLMs), several diagnostic benchmarks have recently been developed to assess these models' multi-modal reasoning proficiency. However, these benchmarks are restricted to assessing primarily the visual aspect and do not examine the holistic audio-visual (AV) understanding. Moreover, currently, there are no benchmarks that investigate the capabilities of AVLLMs to calibrate their responses when presented with perturbed inputs. To this end, we introduce Audio-Visual Trustworthiness assessment Benchmark (AVTrustBench), comprising 600K samples spanning over 9 meticulously crafted tasks, evaluating the capabilities of AVLLMs across three distinct dimensions: Adversarial attack, Compositional reasoning, and Modality-specific dependency. Using our benchmark we extensively evaluate 13 state-of-the-art AVLLMs. The findings reveal that the majority of existing models fall significantly short of achieving human-like comprehension, offering valuable insights for future research directions. To alleviate the limitations in the existing approaches, we further propose a robust, model-agnostic calibrated audio-visual preference optimization based training strategy CAVPref, obtaining a gain up to 30.19% across all 9 tasks. We will publicly release our code and benchmark to facilitate future research in this direction.
2501.02138
Effective LLM-Driven Code Generation with Pythoness
cs.PL cs.AI cs.SE
The advent of large language models (LLMs) has paved the way for a new era of programming tools with both significant capabilities and risks, as the generated code lacks guarantees of correctness and reliability. Developers using LLMs currently face the difficult task of optimizing, integrating, and maintaining code generated by AI. We propose an embedded domain-specific language (DSL), Pythoness, to address those challenges. In Pythoness, developers program with LLMs at a higher level of abstraction. Rather than interacting directly with generated code, developers using Pythoness operate at the level of behavioral specifications when writing functions, classes, or an entire program. These specifications can take the form of unit tests and property-based tests, which may be expressed formally or in natural language. Guided by these specifications, Pythoness generates code that both passes the tests and can be continuously checked during execution. We posit that the Pythoness approach lets developers harness the full potential of LLMs for code generation while substantially mitigating their inherent risks. We describe our current prototype implementation of Pythoness and demonstrate that it can successfully leverage a combination of tests and code generation to yield higher quality code than specifications alone.
2501.02140
Tree-NET: Enhancing Medical Image Segmentation Through Efficient Low-Level Feature Training
eess.IV cs.CV
This paper introduces Tree-NET, a novel framework for medical image segmentation that leverages bottleneck feature supervision to enhance both segmentation accuracy and computational efficiency. While previous studies have employed bottleneck feature supervision, their applications have largely been limited to the training phase, offering no computational benefits during training or evaluation. To the best of our knowledge, this study is the first to propose a framework that incorporates two additional training phases for segmentation models, utilizing bottleneck features at both input and output stages. This approach significantly improves computational performance by reducing input and output dimensions with a negligible addition to parameter count, without compromising accuracy. Tree-NET features a three-layer architecture comprising Encoder-Net and Decoder-Net, which are autoencoders designed to compress input and label data, respectively, and Bridge-Net, a segmentation framework that supervises the bottleneck features. By focusing on dense, compressed representations, Tree-NET enhances operational efficiency and can be seamlessly integrated into existing segmentation models without altering their internal structures or increasing model size. We evaluate Tree-NET on two critical segmentation tasks -- skin lesion and polyp segmentation -- using various backbone models, including U-NET variants and Polyp-PVT. Experimental results demonstrate that Tree-NET reduces FLOPs by a factor of 4 to 13 and decreases memory usage, while achieving comparable or superior accuracy compared to the original architectures. These findings underscore Tree-NET's potential as a robust and efficient solution for medical image segmentation.
2501.02143
SafeAug: Safety-Critical Driving Data Augmentation from Naturalistic Datasets
cs.CV cs.LG
Safety-critical driving data is crucial for developing safe and trustworthy self-driving algorithms. Due to the scarcity of safety-critical data in naturalistic datasets, current approaches primarily utilize simulated or artificially generated images. However, there remains a gap in authenticity between these generated images and naturalistic ones. We propose a novel framework to augment the safety-critical driving data from the naturalistic dataset to address this issue. In this framework, we first detect vehicles using YOLOv5, followed by depth estimation and 3D transformation to simulate vehicle proximity and critical driving scenarios better. This allows for targeted modification of vehicle dynamics data to reflect potentially hazardous situations. Compared to the simulated or artificially generated data, our augmentation methods can generate safety-critical driving data with minimal compromise on image authenticity. Experiments using KITTI datasets demonstrate that a downstream self-driving algorithm trained on this augmented dataset performs superiorly compared to the baselines, which include SMOGN and importance sampling.
2501.02144
Establishing baselines for generative discovery of inorganic crystals
cond-mat.mtrl-sci cs.AI physics.chem-ph
Generative artificial intelligence offers a promising avenue for materials discovery, yet its advantages over traditional methods remain unclear. In this work, we introduce and benchmark two baseline approaches - random enumeration of charge-balanced prototypes and data-driven ion exchange of known compounds - against three generative models: a variational autoencoder, a large language model, and a diffusion model. Our results show that established methods such as ion exchange perform comparably well in generating stable materials, although many of these materials tend to closely resemble known compounds. In contrast, generative models excel at proposing novel structural frameworks and, when sufficient training data is available, can more effectively target properties such as electronic band gap and bulk modulus while maintaining a high stability rate. To enhance the performance of both the baseline and generative approaches, we implement a post-generation screening step in which all proposed structures are passed through stability and property filters from pre-trained machine learning models including universal interatomic potentials. This low-cost filtering step leads to substantial improvement in the success rates of all methods, remains computationally efficient, and ultimately provides a practical pathway toward more effective generative strategies for materials discovery.
2501.02146
Plasma-CycleGAN: Plasma Biomarker-Guided MRI to PET Cross-modality Translation Using Conditional CycleGAN
cs.CV cs.AI q-bio.NC
Cross-modality translation between MRI and PET imaging is challenging due to the distinct mechanisms underlying these modalities. Blood-based biomarkers (BBBMs) are revolutionizing Alzheimer's disease (AD) detection by identifying patients and quantifying brain amyloid levels. However, the potential of BBBMs to enhance PET image synthesis remains unexplored. In this paper, we performed a thorough study on the effect of incorporating BBBM into deep generative models. By evaluating three widely used cross-modality translation models, we found that BBBMs integration consistently enhances the generative quality across all models. By visual inspection of the generated results, we observed that PET images generated by CycleGAN exhibit the best visual fidelity. Based on these findings, we propose Plasma-CycleGAN, a novel generative model based on CycleGAN, to synthesize PET images from MRI using BBBMs as conditions. This is the first approach to integrate BBBMs in conditional cross-modality translation between MRI and PET.
2501.02147
Exploring Secure Machine Learning Through Payload Injection and FGSM Attacks on ResNet-50
cs.CR cs.LG
This paper investigates the resilience of a ResNet-50 image classification model under two prominent security threats: Fast Gradient Sign Method (FGSM) adversarial attacks and malicious payload injection. Initially, the model attains a 53.33% accuracy on clean images. When subjected to FGSM perturbations, its overall accuracy remains unchanged; however, the model's confidence in incorrect predictions notably increases. Concurrently, a payload injection scheme is successfully executed in 93.33% of the tested samples, revealing how stealthy attacks can manipulate model predictions without degrading visual quality. These findings underscore the vulnerability of even high-performing neural networks and highlight the urgency of developing more robust defense mechanisms for security-critical applications.
2501.02149
Attribute-Based Robotic Grasping with Data-Efficient Adaptation
cs.RO cs.AI
Robotic grasping is one of the most fundamental robotic manipulation tasks and has been the subject of extensive research. However, swiftly teaching a robot to grasp a novel target object in clutter remains challenging. This paper attempts to address the challenge by leveraging object attributes that facilitate recognition, grasping, and rapid adaptation to new domains. In this work, we present an end-to-end encoder-decoder network to learn attribute-based robotic grasping with data-efficient adaptation capability. We first pre-train the end-to-end model with a variety of basic objects to learn generic attribute representation for recognition and grasping. Our approach fuses the embeddings of a workspace image and a query text using a gated-attention mechanism and learns to predict instance grasping affordances. To train the joint embedding space of visual and textual attributes, the robot utilizes object persistence before and after grasping. Our model is self-supervised in a simulation that only uses basic objects of various colors and shapes but generalizes to novel objects in new environments. To further facilitate generalization, we propose two adaptation methods, adversarial adaption and one-grasp adaptation. Adversarial adaptation regulates the image encoder using augmented data of unlabeled images, whereas one-grasp adaptation updates the overall end-to-end model using augmented data from one grasp trial. Both adaptation methods are data-efficient and considerably improve instance grasping performance. Experimental results in both simulation and the real world demonstrate that our approach achieves over 81% instance grasping success rate on unknown objects, which outperforms several baselines by large margins.
2501.02151
From Images to Detection: Machine Learning for Blood Pattern Classification
cs.CV stat.AP
Bloodstain Pattern Analysis (BPA) helps us understand how bloodstains form, with a focus on their size, shape, and distribution. This aids in crime scene reconstruction and provides insight into victim positions and crime investigation. One challenge in BPA is distinguishing between different types of bloodstains, such as those from firearms, impacts, or other mechanisms. Our study focuses on differentiating impact spatter bloodstain patterns from gunshot bloodstain patterns. We distinguish patterns by extracting well-designed individual stain features, applying effective data consolidation methods, and selecting boosting classifiers. As a result, we have developed a model that excels in both accuracy and efficiency. In addition, we use outside data sources from previous studies to discuss the challenges and future directions for BPA.
2501.02152
Table as Thought: Exploring Structured Thoughts in LLM Reasoning
cs.AI cs.CL
Large language models' reasoning abilities benefit from methods that organize their thought processes, such as chain-of-thought prompting, which employs a sequential structure to guide the reasoning process step-by-step. However, existing approaches focus primarily on organizing the sequence of thoughts, leaving structure in individual thought steps underexplored. To address this gap, we propose Table as Thought, a framework inspired by cognitive neuroscience theories on human thought. Table as Thought organizes reasoning within a tabular schema, where rows represent sequential thought steps and columns capture critical constraints and contextual information to enhance reasoning. The reasoning process iteratively populates the table until self-verification ensures completeness and correctness. Our experiments show that Table as Thought excels in planning tasks and demonstrates a strong potential for enhancing LLM performance in mathematical reasoning compared to unstructured thought baselines. This work provides a novel exploration of refining thought representation within LLMs, paving the way for advancements in reasoning and AI cognition.
2501.02153
Resolving the Exploration-Exploitation Dilemma in Evolutionary Algorithms: A Novel Human-Centered Framework
cs.NE math.OC
Evolutionary Algorithms (EAs) are powerful tools for tackling complex computational problems, yet effectively managing the exploitation and exploration dynamics -- crucial for robust search navigation -- remains a persistent challenge for EA designers, and leads to the so-called exploration-exploitation dilemma. In this paper, a new human-centered framework is proposed to resolve this dilemma. Unlike the traditional approach, the search process will not be compromised of a single-phase nor the decision-maker tuning efforts will be distributed among the algorithm's traditional parameters such as defining new evolutionary operators internal to the algorithm to influence its search navigation. Instead, a human-centered two-phase search process, compromised of a global search phase followed by a local phase will be utilized. In this framework, the designer plays the central role in directing the algorithm's search navigation through the focused tuning efforts of a new Search Space Size Control parameter external to the algorithm which proves itself to be the dominant parameter in-effect to the algorithm's effective search navigation. The framework is applicable to any search algorithm. We demonstrate its effectiveness on 14 well-known benchmark problems in unconstrained optimization.
2501.02156
The Race to Efficiency: A New Perspective on AI Scaling Laws
cs.LG cs.AI cs.PF
As large-scale AI models expand, training becomes costlier and sustaining progress grows harder. Classical scaling laws (e.g., Kaplan et al. (2020), Hoffmann et al. (2022)) predict training loss from a static compute budget yet neglect time and efficiency, prompting the question: how can we balance ballooning GPU fleets with rapidly improving hardware and algorithms? We introduce the relative-loss equation, a time- and efficiency-aware framework that extends classical AI scaling laws. Our model shows that, without ongoing efficiency gains, advanced performance could demand millennia of training or unrealistically large GPU fleets. However, near-exponential progress remains achievable if the "efficiency-doubling rate" parallels Moore's Law. By formalizing this race to efficiency, we offer a quantitative roadmap for balancing front-loaded GPU investments with incremental improvements across the AI stack. Empirical trends suggest that sustained efficiency gains can push AI scaling well into the coming decade, providing a new perspective on the diminishing returns inherent in classical scaling.
2501.02157
Personalized Graph-Based Retrieval for Large Language Models
cs.CL
As large language models (LLMs) evolve, their ability to deliver personalized and context-aware responses offers transformative potential for improving user experiences. Existing personalization approaches, however, often rely solely on user history to augment the prompt, limiting their effectiveness in generating tailored outputs, especially in cold-start scenarios with sparse data. To address these limitations, we propose Personalized Graph-based Retrieval-Augmented Generation (PGraphRAG), a framework that leverages user-centric knowledge graphs to enrich personalization. By directly integrating structured user knowledge into the retrieval process and augmenting prompts with user-relevant context, PGraphRAG enhances contextual understanding and output quality. We also introduce the Personalized Graph-based Benchmark for Text Generation, designed to evaluate personalized text generation tasks in real-world settings where user history is sparse or unavailable. Experimental results show that PGraphRAG significantly outperforms state-of-the-art personalization methods across diverse tasks, demonstrating the unique advantages of graph-based retrieval for personalization.
2501.02158
Joint Optimization for 4D Human-Scene Reconstruction in the Wild
cs.CV
Reconstructing human motion and its surrounding environment is crucial for understanding human-scene interaction and predicting human movements in the scene. While much progress has been made in capturing human-scene interaction in constrained environments, those prior methods can hardly reconstruct the natural and diverse human motion and scene context from web videos. In this work, we propose JOSH, a novel optimization-based method for 4D human-scene reconstruction in the wild from monocular videos. JOSH uses techniques in both dense scene reconstruction and human mesh recovery as initialization, and then it leverages the human-scene contact constraints to jointly optimize the scene, the camera poses, and the human motion. Experiment results show JOSH achieves better results on both global human motion estimation and dense scene reconstruction by joint optimization of scene geometry and human motion. We further design a more efficient model, JOSH3R, and directly train it with pseudo-labels from web videos. JOSH3R outperforms other optimization-free methods by only training with labels predicted from JOSH, further demonstrating its accuracy and generalization ability.
2501.02166
ROLO-SLAM: Rotation-Optimized LiDAR-Only SLAM in Uneven Terrain with Ground Vehicle
cs.RO cs.CV
LiDAR-based SLAM is recognized as one effective method to offer localization guidance in rough environments. However, off-the-shelf LiDAR-based SLAM methods suffer from significant pose estimation drifts, particularly components relevant to the vertical direction, when passing to uneven terrains. This deficiency typically leads to a conspicuously distorted global map. In this article, a LiDAR-based SLAM method is presented to improve the accuracy of pose estimations for ground vehicles in rough terrains, which is termed Rotation-Optimized LiDAR-Only (ROLO) SLAM. The method exploits a forward location prediction to coarsely eliminate the location difference of consecutive scans, thereby enabling separate and accurate determination of the location and orientation at the front-end. Furthermore, we adopt a parallel-capable spatial voxelization for correspondence-matching. We develop a spherical alignment-guided rotation registration within each voxel to estimate the rotation of vehicle. By incorporating geometric alignment, we introduce the motion constraint into the optimization formulation to enhance the rapid and effective estimation of LiDAR's translation. Subsequently, we extract several keyframes to construct the submap and exploit an alignment from the current scan to the submap for precise pose estimation. Meanwhile, a global-scale factor graph is established to aid in the reduction of cumulative errors. In various scenes, diverse experiments have been conducted to evaluate our method. The results demonstrate that ROLO-SLAM excels in pose estimation of ground vehicles and outperforms existing state-of-the-art LiDAR SLAM frameworks.
2501.02167
Generating Multimodal Images with GAN: Integrating Text, Image, and Style
cs.CV
In the field of computer vision, multimodal image generation has become a research hotspot, especially the task of integrating text, image, and style. In this study, we propose a multimodal image generation method based on Generative Adversarial Networks (GAN), capable of effectively combining text descriptions, reference images, and style information to generate images that meet multimodal requirements. This method involves the design of a text encoder, an image feature extractor, and a style integration module, ensuring that the generated images maintain high quality in terms of visual content and style consistency. We also introduce multiple loss functions, including adversarial loss, text-image consistency loss, and style matching loss, to optimize the generation process. Experimental results show that our method produces images with high clarity and consistency across multiple public datasets, demonstrating significant performance improvements compared to existing methods. The outcomes of this study provide new insights into multimodal image generation and present broad application prospects.
2501.02169
The Integration of Blockchain and Artificial Intelligence for Secure Healthcare Systems
cs.CY cs.AI
Verisign reported a 125 percent increase in data breaches within the healthcare sector in the United States during 2022, with 18.2 million patient records being impacted. Growing healthcare data volumes and diversification mean that medical information is becoming more valuable. Many Health Centers use various technologies to ease the classification, storage, and exchange of big data. This use can also make the health data of the users at risk and vulnerable. AI and blockchain are among the leading technologies at hand. With AI, data-driven operations and big data efficiency have been improved with respect to traditional techniques. Due to its potential to bring about improvements in health services and lower medical costs, this AI technology is regularly used in healthcare. Blockchain helps protect transactions on sharing information and private privacy as long as the exchange of knowledge is that of the standard. The objective of this analysis is to investigate the research and unique contributions since 2008 regarding blockchain-integrated AI and healthcare systems. The work sheds light on applied AI-based healthcare schemes with machine, ballistic, and acrylic learning and disparate blockchain structures. The use of technology in order to ensure patient data security and manage medical information effectively in healthcare settings offers a highly successful position for both healthcare providers and patients. From 2018 to 2021, the best year was 2021 to grow, enhancing everything to examine the download of the device and the counting of Google Academies, for which the joining perspective was borrowed; local research experts were asked, identified articles in recent years, and read reviews of large research grants.
2501.02172
Multifractal Terrain Generation for Evaluating Autonomous Off-Road Ground Vehicles
cs.RO
We present a multifractal artificial terrain generation method that uses the 3D Weierstrass-Mandelbrot function to control roughness. By varying the fractal dimension used in terrain generation across three different values, we generate 60 unique off-road terrains. We use gradient maps to categorize the roughness of each terrain, consisting of low-, semi-, and high-roughness areas. To test how the fractal dimension affects the difficulty of vehicle traversals, we measure the success rates, vertical accelerations, pitch and roll rates, and traversal times of an autonomous ground vehicle traversing 20 randomized straight-line paths in each terrain. As we increase the fractal dimension from 2.3 to 2.45 and from 2.45 to 2.6, we find that the median area of low-roughness terrain decreases 13.8% and 7.16%, the median area of semi-rough terrain increases 11.7% and 5.63%, and the median area of high-roughness terrain increases 1.54% and 3.33%, all respectively. We find that the median success rate of the vehicle decreases 22.5% and 25% as the fractal dimension increases from 2.3 to 2.45 and from 2.45 to 2.6, respectively. Successful traversal results show that the median root-mean-squared vertical accelerations, median root-mean-squared pitch and roll rates, and median traversal times all increase with the fractal dimension.
2501.02173
The Efficiency vs. Accuracy Trade-off: Optimizing RAG-Enhanced LLM Recommender Systems Using Multi-Head Early Exit
cs.IR cs.LG
The deployment of Large Language Models (LLMs) in recommender systems for predicting Click-Through Rates (CTR) necessitates a delicate balance between computational efficiency and predictive accuracy. This paper presents an optimization framework that combines Retrieval-Augmented Generation (RAG) with an innovative multi-head early exit architecture to concurrently enhance both aspects. By integrating Graph Convolutional Networks (GCNs) as efficient retrieval mechanisms, we are able to significantly reduce data retrieval times while maintaining high model performance. The early exit strategy employed allows for dynamic termination of model inference, utilizing real-time predictive confidence assessments across multiple heads. This not only quickens the responsiveness of LLMs but also upholds or improves their accuracy, making it ideal for real-time application scenarios. Our experiments demonstrate how this architecture effectively decreases computation time without sacrificing the accuracy needed for reliable recommendation delivery, establishing a new standard for efficient, real-time LLM deployment in commercial systems.
2501.02174
TACTIC: Task-Agnostic Contrastive pre-Training for Inter-Agent Communication
cs.MA
The "sight range dilemma" in cooperative Multi-Agent Reinforcement Learning (MARL) presents a significant challenge: limited observability hinders team coordination, while extensive sight ranges lead to distracted attention and reduced performance. While communication can potentially address this issue, existing methods often struggle to generalize across different sight ranges, limiting their effectiveness. We propose TACTIC, Task-Agnostic Contrastive pre-Training strategy Inter-Agent Communication. TACTIC is an adaptive communication mechanism that enhances agent coordination even when the sight range during execution is vastly different from that during training. The communication mechanism encodes messages and integrates them with local observations, generating representations grounded in the global state using contrastive learning. By learning to generate and interpret messages that capture important information about the whole environment, TACTIC enables agents to effectively "see" more through communication, regardless of their sight ranges. We comprehensively evaluate TACTIC on the SMACv2 benchmark across various scenarios with broad sight ranges. The results demonstrate that TACTIC consistently outperforms traditional state-of-the-art MARL techniques with and without communication, in terms of generalizing to sight ranges different from those seen in training, particularly in cases of extremely limited or extensive observability.
2501.02176
Molecule-dynamic-based Aging Clock and Aging Roadmap Forecast with Sundial
q-bio.QM cs.LG
Addressing the unavoidable bias inherent in supervised aging clocks, we introduce Sundial, a novel framework that models molecular dynamics through a diffusion field, capturing both the population-level aging process and the individual-level relative aging order. Sundial enables unbiasedestimation of biological age and the forecast of aging roadmap. Fasteraging individuals from Sundial exhibit a higher disease risk compared to those identified from supervised aging clocks. This framework opens new avenues for exploring key topics, including age- and sex-specific aging dynamics and faster yet healthy aging paths.
2501.02178
The Application of Large Language Models in Recommendation Systems
cs.IR
The integration of Large Language Models into recommendation frameworks presents key advantages for personalization and adaptability of experiences to the users. Classic methods of recommendations, such as collaborative filtering and content-based filtering, are seriously limited in the solution of cold-start problems, sparsity of data, and lack of diversity in information considered. LLMs, of which GPT-4 is a good example, have emerged as powerful tools that enable recommendation frameworks to tap into unstructured data sources such as user reviews, social interactions, and text-based content. By analyzing these data sources, LLMs improve the accuracy and relevance of recommendations, thereby overcoming some of the limitations of traditional approaches. This work discusses applications of LLMs in recommendation systems, especially in electronic commerce, social media platforms, streaming services, and educational technologies. This showcases how LLMs enrich recommendation diversity, user engagement, and the system's adaptability; yet it also looks into the challenges connected to their technical implementation. This can also be presented as a study that shows the potential of LLMs for changing user experiences and making innovation possible in industries.
2501.02180
Phase Retrieval by Quaternionic Reweighted Amplitude Flow on Image Reconstruction
cs.CV math.CV
Quaternionic signal processing provides powerful tools for efficiently managing color signals by preserving the intrinsic correlations among signal dimensions through quaternion algebra. In this paper, we address the quaternionic phase retrieval problem by systematically developing novel algorithms based on an amplitude-based model. Specifically, we propose the Quaternionic Reweighted Amplitude Flow (QRAF) algorithm, which is further enhanced by three of its variants: incremental, accelerated, and adapted QRAF algorithms. In addition, we introduce the Quaternionic Perturbed Amplitude Flow (QPAF) algorithm, which has linear convergence. Extensive numerical experiments on both synthetic data and real images, demonstrate that our proposed methods significantly improve recovery performance and computational efficiency compared to state-of-the-art approaches.
2501.02181
SMDP-Based Dynamic Batching for Improving Responsiveness and Energy Efficiency of Batch Services
cs.DC cs.LG cs.SY eess.SY
For servers incorporating parallel computing resources, batching is a pivotal technique for providing efficient and economical services at scale. Parallel computing resources exhibit heightened computational and energy efficiency when operating with larger batch sizes. However, in the realm of online services, the adoption of a larger batch size may lead to longer response times. This paper aims to provide a dynamic batching scheme that delicately balances latency and efficiency. The system is modeled as a batch service queue with size-dependent service times. Then, the design of dynamic batching is formulated as a semi-Markov decision process (SMDP) problem, with the objective of minimizing the weighted sum of average response time and average power consumption. A method is proposed to derive an approximate optimal SMDP solution, representing the chosen dynamic batching policy. By introducing an abstract cost to reflect the impact of "tail" states, the space complexity and the time complexity of the procedure can decrease by 63.5% and 98%, respectively. Numerical results showcase the superiority of SMDP-based batching policies across various parameter setups. Additionally, the proposed scheme exhibits noteworthy flexibility in balancing power consumption and latency.
2501.02182
AdaMixup: A Dynamic Defense Framework for Membership Inference Attack Mitigation
cs.LG cs.AI
Membership inference attacks have emerged as a significant privacy concern in the training of deep learning models, where attackers can infer whether a data point was part of the training set based on the model's outputs. To address this challenge, we propose a novel defense mechanism, AdaMixup. AdaMixup employs adaptive mixup techniques to enhance the model's robustness against membership inference attacks by dynamically adjusting the mixup strategy during training. This method not only improves the model's privacy protection but also maintains high performance. Experimental results across multiple datasets demonstrate that AdaMixup significantly reduces the risk of membership inference attacks while achieving a favorable trade-off between defensive efficiency and model accuracy. This research provides an effective solution for data privacy protection and lays the groundwork for future advancements in mixup training methods.
2501.02184
Model-Free and Real-Time Bioinspired Unicycle-Based Source Seeking: Differential Wheeled Robotic Experiments
cs.RO math.OC
Bioinspred robots aimed at source-seeking are often studied, and their controls designed, using unicycle modeling and formulation. This is true not only for model-based controllers, but also for model-free, real-time control methods such as extremum seeking control (ESC). In this paper, we propose a unicycle-based ESC design applicable to differential wheeled robots that: (1) is very simple design, based on one simple control-affine law, and without state integrators; (2) attenuates oscillations known to persist in ESC designs (i.e., fully stop at the source); and (3) operates in a model-free, real-time setting, tolerating environmental/sensor noise. We provide simulation and real-world robotic experimental results for fixed and moving light source seeking by a differential wheeled robot using our proposed design. Results indicate clear advantages of our proposed design when compared to the literature, including attenuation of undesired oscillations, improved convergence speed, and better handling of noise.
2501.02187
An Efficient Quadratic Penalty Method for a Class of Graph Clustering Problems
math.OC cs.SI
Community-based graph clustering is one of the most popular topics in the analysis of complex social networks. This type of clustering involves grouping vertices that are considered to share more connections, whereas vertices in different groups share fewer connections. A successful clustering result forms densely connected induced subgraphs. This paper studies a specific form of graph clustering problems that can be formulated as semi-assignment problems, where the objective function exhibits block properties. We reformulate these problems as sparse-constrained optimization problems and relax them to continuous optimization models. We apply a quadratic penalty method to the relaxation problem and solve the nonlinear quadratic penalty subproblem with simple box constraints using a projected gradient method based on the active set. Extensive numerical results indicate that our method provides more accurate clustering results for solving graph clustering problems at a faster speed, both for synthetic graphs and real-world network datasets, particularly in large-scale cases.
2501.02189
Benchmark Evaluations, Applications, and Challenges of Large Vision Language Models: A Survey
cs.CV cs.AI cs.CL cs.LG cs.RO
Multimodal Vision Language Models (VLMs) have emerged as a transformative technology at the intersection of computer vision and natural language processing, enabling machines to perceive and reason about the world through both visual and textual modalities. For example, models such as CLIP, Claude, and GPT-4V demonstrate strong reasoning and understanding abilities on visual and textual data and beat classical single modality vision models on zero-shot classification. Despite their rapid advancements in research and growing popularity in applications, a comprehensive survey of existing studies on VLMs is notably lacking, particularly for researchers aiming to leverage VLMs in their specific domains. To this end, we provide a systematic overview of VLMs in the following aspects: model information of the major VLMs developed over the past five years (2019-2024); the main architectures and training methods of these VLMs; summary and categorization of the popular benchmarks and evaluation metrics of VLMs; the applications of VLMs including embodied agents, robotics, and video generation; the challenges and issues faced by current VLMs such as hallucination, fairness, and safety. Detailed collections including papers and model repository links are listed in https://github.com/zli12321/Awesome-VLM-Papers-And-Models.git.
2501.02191
On LLM-Enhanced Mixed-Type Data Imputation with High-Order Message Passing
cs.LG cs.SI
Missing data imputation, which aims to impute the missing values in the raw datasets to achieve the completeness of datasets, is crucial for modern data-driven models like large language models (LLMs) and has attracted increasing interest over the past decades. Despite its importance, existing solutions for missing data imputation either 1) only support numerical and categorical data or 2) show an unsatisfactory performance due to their design prioritizing text data and the lack of key properties for tabular data imputation. In this paper, we propose UnIMP, a Unified IMPutation framework that leverages LLM and high-order message passing to enhance the imputation of mixed-type data including numerical, categorical, and text data. Specifically, we first introduce a cell-oriented hypergraph to model the table. We then propose BiHMP, an efficient Bidirectional High-order Message-Passing network to aggregate global-local information and high-order relationships on the constructed hypergraph while capturing the inter-column heterogeneity and intra-column homogeneity. To effectively and efficiently align the capacity of the LLM with the information aggregated by BiHMP, we introduce Xfusion, which, together with BiHMP, acts as adapters for the LLM. We follow a pre-training and fine-tuning pipeline to train UnIMP, integrating two optimizations: chunking technique, which divides tables into smaller chunks to enhance efficiency; and progressive masking technique, which gradually adapts the model to learn more complex data patterns. Both theoretical proofs and empirical experiments on 10 real world datasets highlight the superiority of UnIMP over existing techniques.
2501.02192
EvoPath: Evolutionary Meta-path Discovery with Large Language Models for Complex Heterogeneous Information Networks
cs.SI
Heterogeneous Information Networks (HINs) encapsulate diverse entity and relation types, with meta-paths providing essential meta-level semantics for knowledge reasoning, although their utility is constrained by discovery challenges. While Large Language Models (LLMs) offer new prospects for meta-path discovery due to their extensive knowledge encoding and efficiency, their adaptation faces challenges such as corpora bias, lexical discrepancies, and hallucination. This paper pioneers the mitigation of these challenges by presenting EvoPath, an innovative framework that leverages LLMs to efficiently identify high-quality meta-paths. EvoPath is carefully designed, with each component aimed at addressing issues that could lead to potential knowledge conflicts. With a minimal subset of HIN facts, EvoPath iteratively generates and evolves meta-paths by dynamically replaying meta-paths in the buffer with prioritization based on their scores. Comprehensive experiments on three large, complex HINs with hundreds of relations demonstrate that our framework, EvoPath, enables LLMs to generate high-quality meta-paths through effective prompting, confirming its superior performance in HIN reasoning tasks. Further ablation studies validate the effectiveness of each module within the framework.
2501.02194
Ensemble-based Deep Multilayer Community Search
cs.SI
Multilayer graphs, consisting of multiple interconnected layers, are widely used to model diverse relationships in the real world. A community is a cohesive subgraph that offers valuable insights for analyzing (multilayer) graphs. Recently, there has been an emerging trend focused on searching query-driven communities within the multilayer graphs. However, existing methods for multilayer community search are either 1) rule-based, which suffer from structure inflexibility; or 2) learning-based, which rely on labeled data or fail to capture layer-specific characteristics. To address these, we propose EnMCS, an Ensemble-based unsupervised (i.e., label-free) Multilayer Community Search framework. EnMCS contains two key components, i.e., HoloSearch which identifies potential communities in each layer while integrating both layer-shared and layer-specific information, and EMerge which is an Expectation-Maximization (EM)-based method that synthesizes the potential communities from each layer into a consensus community. Specifically, HoloSearch first employs a graph-diffusion-based model that integrates three label-free loss functions to learn layer-specific and layer-shared representations for each node. Communities in each layer are then identified based on nodes that exhibit high similarity in layer-shared representations while demonstrating low similarity in layer-specific representations w.r.t. the query nodes. To account for the varying layer-specific characteristics of each layer when merging communities, EMerge models the error rates of layers and true community as latent variables. It then employs the EM algorithm to simultaneously minimize the error rates of layers and predict the final consensus community through iterative maximum likelihood estimation. Experiments over 10 real-world datasets highlight the superiority of EnMCS in terms of both efficiency and effectiveness.
2501.02196
CPTuning: Contrastive Prompt Tuning for Generative Relation Extraction
cs.CL cs.AI
Generative relation extraction (RE) commonly involves first reformulating RE as a linguistic modeling problem easily tackled with pre-trained language models (PLM) and then fine-tuning a PLM with supervised cross-entropy loss. Although having achieved promising performance, existing approaches assume only one deterministic relation between each pair of entities without considering real scenarios where multiple relations may be valid, i.e., entity pair overlap, causing their limited applications. To address this problem, we introduce a novel contrastive prompt tuning method for RE, CPTuning, which learns to associate a candidate relation between two in-context entities with a probability mass above or below a threshold, corresponding to whether the relation exists. Beyond learning schema, CPTuning also organizes RE as a verbalized relation generation task and uses Trie-constrained decoding to ensure a model generates valid relations. It adaptively picks out the generated candidate relations with a high estimated likelihood in inference, thereby achieving multi-relation extraction. We conduct extensive experiments on four widely used datasets to validate our method. Results show that T5-large fine-tuned with CPTuning significantly outperforms previous methods, regardless of single or multiple relations extraction.
2501.02197
Majorization-Minimization Dual Stagewise Algorithm for Generalized Lasso
stat.ML cs.LG stat.CO
The generalized lasso is a natural generalization of the celebrated lasso approach to handle structural regularization problems. Many important methods and applications fall into this framework, including fused lasso, clustered lasso, and constrained lasso. To elevate its effectiveness in large-scale problems, extensive research has been conducted on the computational strategies of generalized lasso. However, to our knowledge, most studies are under the linear setup, with limited advances in non-Gaussian and non-linear models. We propose a majorization-minimization dual stagewise (MM-DUST) algorithm to efficiently trace out the full solution paths of the generalized lasso problem. The majorization technique is incorporated to handle different convex loss functions through their quadratic majorizers. Utilizing the connection between primal and dual problems and the idea of ``slow-brewing'' from stagewise learning, the minimization step is carried out in the dual space through a sequence of simple coordinate-wise updates on the dual coefficients with a small step size. Consequently, selecting an appropriate step size enables a trade-off between statistical accuracy and computational efficiency. We analyze the computational complexity of MM-DUST and establish the uniform convergence of the approximated solution paths. Extensive simulation studies and applications with regularized logistic regression and Cox model demonstrate the effectiveness of the proposed approach.
2501.02198
Fresh-CL: Feature Realignment through Experts on Hypersphere in Continual Learning
cs.LG cs.CV
Continual Learning enables models to learn and adapt to new tasks while retaining prior knowledge. Introducing new tasks, however, can naturally lead to feature entanglement across tasks, limiting the model's capability to distinguish between new domain data. In this work, we propose a method called Feature Realignment through Experts on hyperSpHere in Continual Learning (Fresh-CL). By leveraging predefined and fixed simplex equiangular tight frame (ETF) classifiers on a hypersphere, our model improves feature separation both intra and inter tasks. However, the projection to a simplex ETF shifts with new tasks, disrupting structured feature representation of previous tasks and degrading performance. Therefore, we propose a dynamic extension of ETF through mixture of experts, enabling adaptive projections onto diverse subspaces to enhance feature representation. Experiments on 11 datasets demonstrate a 2% improvement in accuracy compared to the strongest baseline, particularly in fine-grained datasets, confirming the efficacy of combining ETF and MoE to improve feature distinction in continual learning scenarios.
2501.02199
Can ChatGPT implement finite element models for geotechnical engineering applications?
math.NA cs.AI cs.NA
This study assesses the capability of ChatGPT to generate finite element code for geotechnical engineering applications from a set of prompts. We tested three different initial boundary value problems using a hydro-mechanically coupled formulation for unsaturated soils, including the dissipation of excess pore water pressure through fluid mass diffusion in one-dimensional space, time-dependent differential settlement of a strip footing, and gravity-driven seepage. For each case, initial prompting involved providing ChatGPT with necessary information for finite element implementation, such as balance and constitutive equations, problem geometry, initial and boundary conditions, material properties, and spatiotemporal discretization and solution strategies. Any errors and unexpected results were further addressed through prompt augmentation processes until the ChatGPT-generated finite element code passed the verification/validation test. Our results demonstrate that ChatGPT required minimal code revisions when using the FEniCS finite element library, owing to its high-level interfaces that enable efficient programming. In contrast, the MATLAB code generated by ChatGPT necessitated extensive prompt augmentations and/or direct human intervention, as it involves a significant amount of low-level programming required for finite element analysis, such as constructing shape functions or assembling global matrices. Given that prompt engineering for this task requires an understanding of the mathematical formulation and numerical techniques, this study suggests that while a large language model may not yet replace human programmers, it can greatly assist in the implementation of numerical models.
2501.02200
Learning Evolution via Optimization Knowledge Adaptation
cs.NE cs.AI cs.CV cs.LG
Evolutionary algorithms (EAs) maintain populations through evolutionary operators to discover diverse solutions for complex tasks while gathering valuable knowledge, such as historical population data and fitness evaluations. However, traditional EAs face challenges in dynamically adapting to expanding knowledge bases, hindering the efficient exploitation of accumulated information and limiting adaptability to new situations. To address these issues, we introduce an Optimization Knowledge Adaptation Evolutionary Model (OKAEM), which features dynamic parameter adjustment using accumulated knowledge to enhance its optimization capabilities. OKAEM employs attention mechanisms to model the interactions among individuals, fitness landscapes, and genetic components separately, thereby parameterizing the evolutionary operators of selection, crossover, and mutation. These powerful learnable operators enable OKAEM to benefit from pre-learned extensive prior knowledge and self-tune with real-time evolutionary insights. Experimental results demonstrate that OKAEM: 1) exploits prior knowledge for significant performance gains across various knowledge transfer settings; 2) achieves competitive performance through self-tuning alone, even without prior knowledge; 3) outperforms state-of-the-art black-box baselines in a vision-language model tuning case; 4) can improve its optimization capabilities with growing knowledge; 5) is capable of emulating principles of natural selection and genetic recombination.
2501.02201
Accounting for Focus Ambiguity in Visual Questions
cs.CV
No existing work on visual question answering explicitly accounts for ambiguity regarding where the content described in the question is located in the image. To fill this gap, we introduce VQ-FocusAmbiguity, the first VQA dataset that visually grounds each region described in the question that is necessary to arrive at the answer. We then provide an analysis showing how our dataset for visually grounding `questions' is distinct from visually grounding `answers', and characterize the properties of the questions and segmentations provided in our dataset. Finally, we benchmark modern models for two novel tasks: recognizing whether a visual question has focus ambiguity and localizing all plausible focus regions within the image. Results show that the dataset is challenging for modern models. To facilitate future progress on these tasks, we publicly share the dataset with an evaluation server at https://focusambiguity.github.io/.
2501.02205
Digital Twin Calibration with Model-Based Reinforcement Learning
cs.LG
This paper presents a novel methodological framework, called the Actor-Simulator, that incorporates the calibration of digital twins into model-based reinforcement learning for more effective control of stochastic systems with complex nonlinear dynamics. Traditional model-based control often relies on restrictive structural assumptions (such as linear state transitions) and fails to account for parameter uncertainty in the model. These issues become particularly critical in industries such as biopharmaceutical manufacturing, where process dynamics are complex and not fully known, and only a limited amount of data is available. Our approach jointly calibrates the digital twin and searches for an optimal control policy, thus accounting for and reducing model error. We balance exploration and exploitation by using policy performance as a guide for data collection. This dual-component approach provably converges to the optimal policy, and outperforms existing methods in extensive numerical experiments based on the biopharmaceutical manufacturing domain.
2501.02207
Self-Supervised Learning for Detecting AI-Generated Faces as Anomalies
cs.CV
The detection of AI-generated faces is commonly approached as a binary classification task. Nevertheless, the resulting detectors frequently struggle to adapt to novel AI face generators, which evolve rapidly. In this paper, we describe an anomaly detection method for AI-generated faces by leveraging self-supervised learning of camera-intrinsic and face-specific features purely from photographic face images. The success of our method lies in designing a pretext task that trains a feature extractor to rank four ordinal exchangeable image file format (EXIF) tags and classify artificially manipulated face images. Subsequently, we model the learned feature distribution of photographic face images using a Gaussian mixture model. Faces with low likelihoods are flagged as AI-generated. Both quantitative and qualitative experiments validate the effectiveness of our method. Our code is available at \url{https://github.com/MZMMSEC/AIGFD_EXIF.git}.
2501.02208
Robust Multi-Dimensional Scaling via Accelerated Alternating Projections
stat.ML cs.LG math.OC
We consider the robust multi-dimensional scaling (RMDS) problem in this paper. The goal is to localize point locations from pairwise distances that may be corrupted by outliers. Inspired by classic MDS theories, and nonconvex works for the robust principal component analysis (RPCA) problem, we propose an alternating projection based algorithm that is further accelerated by the tangent space projection technique. For the proposed algorithm, if the outliers are sparse enough, we can establish linear convergence of the reconstructed points to the original points after centering and rotation alignment. Numerical experiments verify the state-of-the-art performances of the proposed algorithm.
2501.02211
Examining the Robustness of Homogeneity Bias to Hyperparameter Adjustments in GPT-4
cs.CV cs.CL cs.LG
Vision-Language Models trained on massive collections of human-generated data often reproduce and amplify societal stereotypes. One critical form of stereotyping reproduced by these models is homogeneity bias-the tendency to represent certain groups as more homogeneous than others. We investigate how this bias responds to hyperparameter adjustments in GPT-4, specifically examining sampling temperature and top p which control the randomness of model outputs. By generating stories about individuals from different racial and gender groups and comparing their similarities using vector representations, we assess both bias robustness and its relationship with hyperparameter values. We find that (1) homogeneity bias persists across most hyperparameter configurations, with Black Americans and women being represented more homogeneously than White Americans and men, (2) the relationship between hyperparameters and group representations shows unexpected non-linear patterns, particularly at extreme values, and (3) hyperparameter adjustments affect racial and gender homogeneity bias differently-while increasing temperature or decreasing top p can reduce racial homogeneity bias, these changes show different effects on gender homogeneity bias. Our findings suggest that while hyperparameter tuning may mitigate certain biases to some extent, it cannot serve as a universal solution for addressing homogeneity bias across different social group dimensions.