AI Research News Feeds for September 25th, 2025

AI RESEARCH PAPERS & ACADEMIC SOURCES

• Challenges and Trends in Egocentric Vision: A Survey
• RG-Attn: Radian Glue Attention for Multi-modality Multi-agent Cooperative Perception
• SoFar: Language-Grounded Orientation Bridges Spatial Reasoning and Object Manipulation
• Deciphering Functions of Neurons in Vision-Language Models
• Generating 360{\deg} Video is What You Need For a 3D Scene
• Imaging Biomarkers for Neurodegenerative Diseases from Detailed Segmentation of Medial Temporal Lobe Subregions on in vivo Brain MRI Using Upsampling Strategy Guided by High-resolution ex vivo MRI
• EDBench: Large-Scale Electron Density Data for Molecular Modeling
• SEM: Enhancing Spatial Understanding for Robust Robot Manipulation
• CellCLIP -- Learning Perturbation Effects in Cell Painting via Text-Guided Contrastive Learning
• HAZEMATCHING: Dehazing Light Microscopy Images with Guided Conditional Flow Matching
• Infrared Image Super-Resolution: Systematic Review, and Future Trends
• To Fold or Not to Fold: Graph Regularized Tensor Train for Visual Data Completion
• VLM See, Robot Do: Human Demo Video to Robot Action Plan via Vision Language Model
• Probabilistic Online Event Downsampling
• A Quad-Step Approach to Uncertainty-Aware Deep Learning for Skin Cancer Classification
• NERO: Explainable Out-of-Distribution Detection with Neuron-level Relevance
• SurgVidLM: Towards Multi-grained Surgical Video Understanding with Large Language Model
• Intervening in Black Box: Concept Bottleneck Model for Enhancing Human Neural Network Mutual Understanding
• Diffusion models for multivariate subsurface generation and efficient probabilistic inversion
• AAPO: Enhancing the Reasoning Capabilities of LLMs with Advantage Momentum
• WikiGap: Promoting Epistemic Equity by Surfacing Knowledge Gaps Between English Wikipedia and other Language Editions
• Localized LoRA: A Structured Low-Rank Approximation for Efficient Fine-Tuning
• OmniSpatial: Towards Comprehensive Spatial Reasoning Benchmark for Vision Language Models
• Urania: Differentially Private Insights into AI Use
• CLOSP: A Unified Semantic Space for SAR, MSI, and Text in Remote Sensing
• Latent Wavelet Diffusion For Ultra-High-Resolution Image Synthesis
• Macroeconomic Forecasting with Large Language Models
• Unifying Symbolic Music Arrangement: Track-Aware Reconstruction and Structured Tokenization
• A GEN AI Framework for Medical Note Generation
• GraphEQA: Using 3D Semantic Scene Graphs for Real-time Embodied Question Answering
• HawkBench: Investigating Resilience of RAG Methods on Stratified Information-Seeking Tasks
• Stepwise Guided Policy Optimization: Coloring your Incorrect Reasoning in GRPO
• LEMUR Neural Network Dataset: Towards Seamless AutoML
• MAME: Multidimensional Adaptive Metamer Exploration with Human Perceptual Feedback
• Enhanced uncertainty quantification variational autoencoders for the solution of Bayesian inverse problems
• Diffusion Classifier-Driven Reward for Offline Preference-based Reinforcement Learning
• A Transformer Model for Predicting Chemical Products from Generic SMARTS Templates with Data Augmentation
• Examining the robustness of Physics-Informed Neural Networks to noise for Inverse Problems
• Deep learning for exoplanet detection and characterization by direct imaging at high contrast
• Pretrained deep models outperform GBDTs in Learning-To-Rank under label scarcity
• Robust Training of Neural Networks at Arbitrary Precision and Sparsity
• Survey of Deep Learning and Physics-Based Approaches in Computational Wave Imaging
• Model-Agnostic AI Framework with Explicit Time Integration for Long-Term Fluid Dynamics Prediction
• LLMs for Cold-Start Cutting Plane Separator Configuration
• Multimodal Representation-disentangled Information Bottleneck for Multimodal Recommendation
• AnchDrive: Bootstrapping Diffusion Policies with Hybrid Trajectory Anchors for End-to-End Driving
• Investigating Security Implications of Automatically Generated Code on the Software Supply Chain
• RAG Security and Privacy: Formalizing the Threat Model and Attack Surface
• Adaptive Event-Triggered Policy Gradient for Multi-Agent Reinforcement Learning
• Tree Search for Language Model Agents
• Reinforcement Learning and Machine ethics:a systematic review
• Multi-Agents are Social Groups: Investigating Social Influence of Multiple Agents in Human-Agent Interactions
• Enhancing Crash Frequency Modeling Based on Augmented Multi-Type Data by Hybrid VAE-Diffusion-Based Generative Neural Networks
• STRIVE: Structured Reasoning for Self-Improvement in Claim Verification
• AutoEval: A Practical Framework for Autonomous Evaluation of Mobile Agents
• Causal Inference under Threshold Manipulation: Bayesian Mixture Modeling and Heterogeneous Treatment Effects
• Eliminating stability hallucinations in llm-based tts models via attention guidance
• CollaPipe: Adaptive Segment-Optimized Pipeline Parallelism for Collaborative LLM Training in Heterogeneous Edge Networks
• Choosing to Be Green: Advancing Green AI via Dynamic Model Selection
• Affective Computing and Emotional Data: Challenges and Implications in Privacy Regulations, The AI Act, and Ethics in Large Language Models
• CyberSOCEval: Benchmarking LLMs Capabilities for Malware Analysis and Threat Intelligence Reasoning
• How People Manage Knowledge in their "Second Brains"- A Case Study with Industry Researchers Using Obsidian
• STAF: Leveraging LLMs for Automated Attack Tree-Based Security Test Generation
• Wrapped Gaussian on the manifold of Symmetric Positive Definite Matrices
• Stein's unbiased risk estimate and Hyv\"arinen's score matching
• Error Propagation in Dynamic Programming: From Stochastic Control to Option Pricing
• Chiseling: Powerful and Valid Subgroup Selection via Interactive Machine Learning
• Hierarchical Bayesian Operator-induced Symbolic Regression Trees for Structural Learning of Scientific Expressions
• Generalized Nonnegative Structured Kruskal Tensor Regression
• Statistical Inference Leveraging Synthetic Data with Distribution-Free Guarantees
• Differentially Private Bootstrap: New Privacy Analysis and Inference Strategies
• Sparse Max-Affine Regression
• A Scalable Nystr\"om-Based Kernel Two-Sample Test with Permutations
• Beyond Grids: Multi-objective Bayesian Optimization With Adaptive Discretization
• The 2020 United States Decennial Census Is More Private Than You (Might) Think
• Unsupervised Cross-Domain 3D Human Pose Estimation via Pseudo-Label-Guided Global Transforms
• ChartQA-X: Generating Explanations for Visual Chart Reasoning
• Revisiting Residual Connections: Orthogonal Updates for Stable and Efficient Deep Networks
• LiDAR MOT-DETR: A LiDAR-based Two-Stage Transformer for 3D Multiple Object Tracking
• Redemption Score: A Multi-Modal Evaluation Framework for Image Captioning via Distributional, Perceptual, and Linguistic Signal Triangulation
• Sparse VideoGen2: Accelerate Video Generation with Sparse Attention via Semantic-Aware Permutation
• EndoBench: A Comprehensive Evaluation of Multi-Modal Large Language Models for Endoscopy Analysis
• To Trust Or Not To Trust Your Vision-Language Model's Prediction
• SpaRC: Sparse Radar-Camera Fusion for 3D Object Detection
• SafeEraser: Enhancing Safety in Multimodal Large Language Models through Multimodal Machine Unlearning
• GaussianSeal: Rooting Adaptive Watermarks for 3D Gaussian Generation Model
• Robust Computer-Vision based Construction Site Detection for Assistive-Technology Applications
• LEDiT: Your Length-Extrapolatable Diffusion Transformer without Positional Encoding
• Adversarial Robustness of Discriminative Self-Supervised Learning in Vision
• Cross-Domain Underwater Image Enhancement Guided by No-Reference Image Quality Assessment: A Transfer Learning Approach
• Multimodal Reference Visual Grounding
• Towards Visual Text Grounding of Multimodal Large Language Model
• Ctrl-Room: Controllable Text-to-3D Room Meshes Generation with Layout Constraints
• Long Video Understanding with Learnable Retrieval in Video-Language Models
• CLIP Can Understand Depth
• MCPDepth: Omnidirectional Depth Estimation via Stereo Matching from Multi-Cylindrical Panoramas
• Positional Prompt Tuning for Efficient 3D Representation Learning
• Lagrangian Motion Fields for Long-term Motion Generation
• Diffusion Curriculum: Synthetic-to-Real Generative Curriculum Learning via Image-Guided Diffusion
• Replay-Free Continual Low-Rank Adaptation with Dynamic Memory
• SMLNet: A SPD Manifold Learning Network for Infrared and Visible Image Fusion
• DreamMix: Decoupling Object Attributes for Enhanced Editability in Customized Image Inpainting
• MeshMosaic: Scaling Artist Mesh Generation via Local-to-Global Assembly
• MultiSoundGen: Video-to-Audio Generation for Multi-Event Scenarios via SlowFast Contrastive Audio-Visual Pretraining and Direct Preference Optimization
• Ensuring Reliable Participation in Subjective Video Quality Tests Across Platforms
• Queryable 3D Scene Representation: A Multi-Modal Framework for Semantic Reasoning and Robotic Task Planning
• KSDiff: Keyframe-Augmented Speech-Aware Dual-Path Diffusion for Facial Animation
• VisualMimic: Visual Humanoid Loco-Manipulation via Motion Tracking and Generation
• An Optimized PatchMatch for Multi-scale and Multi-feature Label Fusion
• Robust superpixels using color and contour features along linear path
• Texture Superpixel Clustering from Patch-based Nearest Neighbor Matching
• Multi-Scale Superpatch Matching using Dual Superpixel Descriptors
• PerFace: Metric Learning in Perceptual Facial Similarity for Enhanced Face Anonymization
• FAST: Foreground-aware Diffusion with Accelerated Sampling Trajectory for Segmentation-oriented Anomaly Synthesis
• A Comprehensive Evaluation of YOLO-based Deer Detection Performance on Edge Devices
• Efficient Encoder-Free Pose Conditioning and Pose Control for Virtual Try-On
• PhysCtrl: Generative Physics for Controllable and Physics-Grounded Video Generation
• EditVerse: Unifying Image and Video Editing and Generation with In-Context Learning
• Frequency-Aware Ensemble Learning for BraTS 2025 Pediatric Brain Tumor Segmentation
• Agentic Scene Policies: Unifying Space, Semantics, and Affordances for Robot Action
• AJAHR: Amputated Joint Aware 3D Human Mesh Recovery
• PU-Gaussian: Point Cloud Upsampling using 3D Gaussian Representation
• ImageNet-trained CNNs are not biased towards texture: Revisiting feature reliance through controlled suppression
• An Anisotropic Cross-View Texture Transfer with Multi-Reference Non-Local Attention for CT Slice Interpolation
• 4D Driving Scene Generation With Stereo Forcing
• A Versatile Foundation Model for AI-enabled Mammogram Interpretation
• A co-evolving agentic AI system for medical imaging analysis
• HiPerformer: A High-Performance Global-Local Segmentation Model with Modular Hierarchical Fusion Strategy
• SHMoAReg: Spark Deformable Image Registration via Spatial Heterogeneous Mixture of Experts and Attention Heads
• Unleashing the Potential of the Semantic Latent Space in Diffusion Models for Image Dehazing
• A Simple Data Augmentation Strategy for Text-in-Image Scientific VQA
• EchoBench: Benchmarking Sycophancy in Medical Large Vision-Language Models
• Smaller is Better: Enhancing Transparency in Vehicle AI Systems via Pruning
• C$^2$MIL: Synchronizing Semantic and Topological Causalities in Multiple Instance Learning for Robust and Interpretable Survival Analysis
• U-Mamba2-SSL for Semi-Supervised Tooth and Pulp Segmentation in CBCT
• Optical Ocean Recipes: Creating Realistic Datasets to Facilitate Underwater Vision Research
• Universal Camouflage Attack on Vision-Language Models for Autonomous Driving
• Interpreting ResNet-based CLIP via Neuron-Attention Decomposition
• When Words Can't Capture It All: Towards Video-Based User Complaint Text Generation with Multimodal Video Complaint Dataset
• SynchroRaMa : Lip-Synchronized and Emotion-Aware Talking Face Generation via Multi-Modal Emotion Embedding
• OmniScene: Attention-Augmented Multimodal 4D Scene Understanding for Autonomous Driving
• CamPVG: Camera-Controlled Panoramic Video Generation with Epipolar-Aware Diffusion
• SDE-DET: A Precision Network for Shatian Pomelo Detection in Complex Orchard Environments
• Improving Generalizability and Undetectability for Targeted Adversarial Attacks on Multimodal Pre-trained Models
• Does the Manipulation Process Matter? RITA: Reasoning Composite Image Manipulations via Reversely-Ordered Incremental-Transition Autoregression
• PS3: A Multimodal Transformer Integrating Pathology Reports with Histology Images and Biological Pathways for Cancer Survival Prediction
• Generative Adversarial Networks Applied for Privacy Preservation in Biometric-Based Authentication and Identification
• PersONAL: Towards a Comprehensive Benchmark for Personalized Embodied Agents
• FreezeVLA: Action-Freezing Attacks against Vision-Language-Action Models
• Adaptive Guidance Semantically Enhanced via Multimodal LLM for Edge-Cloud Object Detection
• Generalized Shortest Path-based Superpixels for 3D Spherical Image Segmentation
• Efficient Cell Painting Image Representation Learning via Cross-Well Aligned Masked Siamese Network
• Aerial-Ground Image Feature Matching via 3D Gaussian Splatting-based Intermediate View Rendering
• CapStARE: Capsule-based Spatiotemporal Architecture for Robust and Efficient Gaze Estimation
• GS-RoadPatching: Inpainting Gaussians via 3D Searching and Placing for Driving Scenes
• nnFilterMatch: A Unified Semi-Supervised Learning Framework with Uncertainty-Aware Pseudo-Label Filtering for Efficient Medical Segmentation
• Talking Head Generation via AU-Guided Landmark Prediction
• Logics-Parsing Technical Report
• Sex-based Bias Inherent in the Dice Similarity Coefficient: A Model Independent Analysis for Multiple Anatomical Structures
• EfficienT-HDR: An Efficient Transformer-Based Framework via Multi-Exposure Fusion for HDR Reconstruction
• BiTAA: A Bi-Task Adversarial Attack for Object Detection and Depth Estimation via 3D Gaussian Splatting
• StrCGAN: A Generative Framework for Stellar Image Restoration
• Adaptive Model Ensemble for Continual Learning
• ThinkFake: Reasoning in Multimodal Large Language Models for AI-Generated Image Detection
• Anatomically Constrained Transformers for Cardiac Amyloidosis Classification
• Learning to Stop: Reinforcement Learning for Efficient Patient-Level Echocardiographic Classification
• Towards Robust In-Context Learning for Medical Image Segmentation via Data Synthesis
• VIMD: Monocular Visual-Inertial Motion and Depth Estimation
• Frequency-domain Multi-modal Fusion for Language-guided Medical Image Segmentation
• PolGS: Polarimetric Gaussian Splatting for Fast Reflective Surface Reconstruction
• CAMILA: Context-Aware Masking for Image Editing with Language Alignment
• Robust RGB-T Tracking via Learnable Visual Fourier Prompt Fine-tuning and Modality Fusion Prompt Generation
• Rectified Decoupled Dataset Distillation: A Closer Look for Fair and Comprehensive Evaluation
• CURE: Centroid-guided Unsupervised Representation Erasure for Facial Recognition Systems
• Synthesizing Artifact Dataset for Pixel-level Detection
• Parameter-Efficient Multi-Task Learning via Progressive Task-Specific Adaptation
• Raw-JPEG Adapter: Efficient Raw Image Compression with JPEG
• The Impact of 2D Segmentation Backbones on Point Cloud Predictions Using 4D Radar
• Bias in the Picture: Benchmarking VLMs with Social-Cue News Images and LLM-as-Judge Assessment
• Enhancing Transformer-Based Vision Models: Addressing Feature Map Anomalies Through Novel Optimization Strategies
• From Prompt to Progression: Taming Video Diffusion Models for Seamless Attribute Transition
• RealitySummary: Exploring On-Demand Mixed Reality Text Summarization and Question Answering using Large Language Models
• Overview of LifeCLEF Plant Identification task 2020
• iFinder: Structured Zero-Shot Vision-Based LLM Grounding for Dash-Cam Video Reasoning
• How Well Can Reasoning Models Identify and Recover from Unhelpful Thoughts?
• Augmenting Multi-Agent Communication with State Delta Trajectory
• The Medium Is Not the Message: Deconfounding Document Embeddings via Linear Concept Erasure
• Detecting Token-Level Hallucinations Using Variance Signals: A Reference-Free Approach
• VisualTrap: A Stealthy Backdoor Attack on GUI Agents via Visual Grounding Manipulation
• Dynamic Parameter Memory: Temporary LoRA-Enhanced LLM for Long-Sequence Emotion Recognition in Conversation
• Date Fragments: A Hidden Bottleneck of Tokenization for Temporal Reasoning
• Safeguarding Privacy of Retrieval Data against Membership Inference Attacks: Is This Query Too Close to Home?
• LASER: Stratified Selective Sampling for Instruction Tuning with Dedicated Scoring Strategy
• Advancing Expert Specialization for Better MoE
• DyePack: Provably Flagging Test Set Contamination in LLMs Using Backdoors
• Threading the Needle: Reweaving Chain-of-Thought Reasoning to Explain Human Label Variation
• LegalSearchLM: Rethinking Legal Case Retrieval as Legal Elements Generation
• RadialRouter: Structured Representation for Efficient and Robust Large Language Models Routing
• DISCO Balances the Scales: Adaptive Domain- and Difficulty-Aware Reinforcement Learning on Imbalanced Data
• Aligned Probing: Relating Toxic Behavior and Model Internals
• Unifying Text Semantics and Graph Structures for Temporal Text-attributed Graphs with Large Language Models
• Inverse Reinforcement Learning with Dynamic Reward Scaling for LLM Alignment
• Playpen: An Environment for Exploring Learning Through Conversational Interaction
• Small or Large? Zero-Shot or Finetuned? Guiding Language Model Choice for Specialized Applications in Healthcare
• Meeseeks: A Feedback-Driven, Iterative Self-Correction Benchmark evaluating LLMs' Instruction Following Capability
• Scent of Knowledge: Optimizing Search-Enhanced Reasoning with Information Foraging
• SAFE: Improving LLM Systems using Sentence-Level In-generation Attribution
• LLMs Reproduce Stereotypes of Sexual and Gender Minorities
• BAP v2: An Enhanced Task Framework for Instruction Following in Minecraft Dialogues
• LLMs as a synthesis between symbolic and distributed approaches to language
• Bridging Information Gaps with Comprehensive Answers: Improving the Diversity and Informativeness of Follow-Up Questions
• HoT: Highlighted Chain of Thought for Referencing Supporting Facts from Inputs
• Large Language Models for Multilingual Previously Fact-Checked Claim Detection
• Language Models Fail to Introspect About Their Knowledge of Language
• Modeling Subjectivity in Cognitive Appraisal with Language Models
• Discrete Diffusion for Reflective Vision-Language-Action Models in Autonomous Driving
• Muse-it: A Tool for Analyzing Music Discourse on Reddit
• TALEC: Teach Your LLM to Evaluate in Specific Domain with In-house Criteria by Criteria Division and Zero-shot Plus Few-shot
• Context-Masked Meta-Prompting for Privacy-Preserving LLM Adaptation in Finance
• Efficient Fine-Tuning of Large Language Models for Automated Medical Documentation
• Evading Toxicity Detection with ASCII-art: A Benchmark of Spatial Attacks on Moderation Systems
• UNComp: Can Matrix Entropy Uncover Sparsity? -- A Compressor Design from an Uncertainty-Aware Perspective
• Blind Men and the Elephant: Diverse Perspectives on Gender Stereotypes in Benchmark Datasets
• Understanding Before Reasoning: Enhancing Chain-of-Thought with Iterative Summarization Pre-Prompting
• SIM-CoT: Supervised Implicit Chain-of-Thought
• Z-Scores: A Metric for Linguistically Assessing Disfluency Removal
• DRES: Benchmarking LLMs for Disfluency Removal
• Morphological Synthesizer for Ge'ez Language: Addressing Morphological Complexity and Resource Limitations
• EmbeddingGemma: Powerful and Lightweight Text Representations
• Language Models that Think, Chat Better
• STARQA: A Question Answering Dataset for Complex Analytical Reasoning over Structured Databases
• Multimodal Language Models with Modality-Specific Experts for Financial Forecasting from Interleaved Sequences of Text and Time Series
• Human-AI Narrative Synthesis to Foster Shared Understanding in Civic Decision-Making
• Probing Gender Bias in Multilingual LLMs: A Case Study of Stereotypes in Persian
• Thinking Augmented Pre-training
• Play by the Type Rules: Inferring Constraints for LLM Functions in Declarative Programs
• Low-Resource English-Tigrinya MT: Leveraging Multilingual Models, Custom Tokenizers, and Clean Evaluation Benchmarks
• Investigating the Representation of Backchannels and Fillers in Fine-tuned Language Models
• Instruction Boundary: Quantifying Biases in LLM Reasoning under Various Coverage
• Feeding Two Birds or Favoring One? Adequacy-Fluency Tradeoffs in Evaluation and Meta-Evaluation of Machine Translation
• Multilingual Hope Speech Detection: A Comparative Study of Logistic Regression, mBERT, and XLM-RoBERTa with Active Learning
• From Input Perception to Predictive Insight: Modeling Model Blind Spots Before They Become Errors
• From Text to Talk: Audio-Language Model Needs Non-Autoregressive Joint Training
• Can Constructions "SCAN" Compositionality ?
• OLaPh: Optimal Language Phonemizer
• Causal Understanding by LLMs: The Role of Uncertainty
• Integrated Framework for LLM Evaluation with Answer Generation
• Less is More: The Effectiveness of Compact Typological Language Representations
• Embedding Domain Knowledge for Large Language Models via Reinforcement Learning from Augmented Generation
• SwissGPC v1.0 -- The Swiss German Podcasts Corpus
• Do Before You Judge: Self-Reference as a Pathway to Better LLM Evaluation
• Future Policy Aware Preference Learning for Mathematical Reasoning
• WEST: LLM based Speech Toolkit for Speech Understanding, Generation, and Interaction
• CorIL: Towards Enriching Indian Language to Indian Language Parallel Corpora and Machine Translation Systems
• The Knowledge-Behaviour Disconnect in LLM-based Chatbots
• DiffNator: Generating Structured Explanations of Time-Series Differences
• Tokenization and Representation Biases in Multilingual Models on Dialectal NLP Tasks
• Responsible AI Technical Report
• Personality Vector: Modulating Personality of Large Language Models by Model Merging
• PART: Progressive Alignment Representation Training for Multilingual Speech-To-Text with LLMs
• CHURRO: Making History Readable with an Open-Weight Large Vision-Language Model for High-Accuracy, Low-Cost Historical Text Recognition
• EnAnchored-X2X: English-Anchored Optimization for Many-to-Many Translation
• bi-GRPO: Bidirectional Optimization for Jailbreak Backdoor Injection on LLMs
• Polarity Detection of Sustainable Detection Goals in News Text
• TianHui: A Domain-Specific Large Language Model for Diverse Traditional Chinese Medicine Scenarios
• Mah\={a}n\={a}ma: A Unique Testbed for Literary Entity Discovery and Linking
• Benchmarking Gaslighting Attacks Against Speech Large Language Models
• SINAI at eRisk@CLEF 2025: Transformer-Based and Conversational Strategies for Depression Detection
• Benchmarking ChatGPT and DeepSeek in April 2025: A Novel Dual Perspective Sentiment Analysis Using Lexicon-Based and Deep Learning Approaches
• Characterizing Knowledge Graph Tasks in LLM Benchmarks Using Cognitive Complexity Frameworks
• A Pipeline to Assess Merging Methods via Behavior and Internals
• Do LLMs Encode Frame Semantics? Evidence from Frame Identification
• Retrieval Augmented Generation based context discovery for ASR
• ExPe: Exact Positional Encodings for Generative Transformer Models with Extrapolating Capabilities
• LLMs4All: A Review on Large Language Models for Research and Applications in Academic Disciplines
• Anatomy of a Feeling: Narrating Embodied Emotions via Large Vision-Language Models
• Evaluating Language Translation Models by Playing Telephone
• AutoSpec: An Agentic Framework for Automatically Drafting Patent Specification
• Projective Kolmogorov Arnold Neural Networks (P-KANs): Entropy-Driven Functional Space Discovery for Interpretable Machine Learning
• A Novel Short-Term Anomaly Prediction for IIoT with Software Defined Twin Network
• First-Extinction Law for Resampling Processes
• Hyperspectral Adapter for Semantic Segmentation with Vision Foundation Models
• How Much of Your Data Can Suck? Thresholds for Domain Performance and Emergent Misalignment in LLMs
• How Model Size, Temperature, and Prompt Style Affect LLM-Human Assessment Score Alignment
• Performance of Large Language Models in Answering Critical Care Medicine Questions
• Diffusion and Flow-based Copulas: Forgetting and Remembering Dependencies
• Convex Regression with a Penalty
• High-Dimensional Statistical Process Control via Manifold Fitting and Learning
• Modeling and Control of Deep Sign-Definite Dynamics with Application to Hybrid Powertrain Control
• Geometric Autoencoder Priors for Bayesian Inversion: Learn First Observe Later
• BioBO: Biology-informed Bayesian Optimization for Perturbation Design
• Anomaly Detection by Clustering DINO Embeddings using a Dirichlet Process Mixture
• Table Detection with Active Learning
• The Syntax and Semantics of einsum
• Predictive Quality Assessment for Mobile Secure Graphics
• Confidence Calibration in Large Language Model-Based Entity Matching
• Stochastic Path Planning in Correlated Obstacle Fields
• Uncertainty in Semantic Language Modeling with PIXELS
• MAGIC: Multi-task Gaussian process for joint imputation and classification in healthcare time series
• Discovery of Sustainable Refrigerants through Physics-Informed RL Fine-Tuning of Sequence Models
• Graph-based Neural Space Weather Forecasting
• EgoBridge: Domain Adaptation for Generalizable Imitation from Egocentric Human Data
• Deep Learning for Clouds and Cloud Shadow Segmentation in Methane Satellite and Airborne Imaging Spectroscopy
• Efficient Online Large-Margin Classification via Dual Certificates
• Formal Safety Verification and Refinement for Generative Motion Planners via Certified Local Stabilization
• A Statistical Mixture-of-Experts Framework for EMG Artifact Removal in EEG: Empirical Insights and a Proof-of-Concept Application
• Hybrid Pipeline SWD Detection in Long-Term EEG Signals
• SpellerSSL: Self-Supervised Learning with P300 Aggregation for Speller BCIs
• Poster: ChatIYP: Enabling Natural Language Access to the Internet Yellow Pages Database
• The Pareto Frontier of Resilient Jet Tagging
• HUNT: High-Speed UAV Navigation and Tracking in Unstructured Environments via Instantaneous Relative Frames
• The Platonic Universe: Do Foundation Models See the Same Sky?
• Anchored Langevin Algorithms
• Quantum Harmonic Analysis and the Structure in Data: Augmentation
• OmniVLA: An Omni-Modal Vision-Language-Action Model for Robot Navigation
• AnySafe: Adapting Latent Safety Filters at Runtime via Safety Constraint Parameterization in the Latent Space
• STL-FFT-STFT-TCN-LSTM: An Effective Wave Height High Accuracy Prediction Model Fusing Time-Frequency Domain Features
• Electric Vehicle Identification from Behind Smart Meter Data
• A Spatio-Temporal Feature Fusion EEG Virtual Channel Signal Generation Network and Its Application in Anxiety Assessment
• A Measurement Report Data-Driven Framework for Localized Statistical Channel Modeling
• ShinkaEvolve: Towards Open-Ended And Sample-Efficient Program Evolution
• LLM-Assisted Topic Reduction for BERTopic on Social Media Data
• Low-Cost Sensor Fusion Framework for Organic Substance Classification and Quality Control Using Classification Methods
• Short-Term Regional Electricity Demand Forecasting in Argentina Using LSTM Networks
• Vision-Based Perception for Autonomous Vehicles in Off-Road Environment Using Deep Learning
• Neural Network Based Framework for Passive Intermodulation Cancellation in MIMO Systems
• When Judgment Becomes Noise: How Design Failures in LLM Judge Benchmarks Silently Undermine Validity
• Alignment-Sensitive Minimax Rates for Spectral Algorithms with Learned Kernels
• Graph Variate Neural Networks
• A Recovery Guarantee for Sparse Neural Networks
• Video models are zero-shot learners and reasoners
• Feature Dynamics as Implicit Data Augmentation: A Depth-Decomposed View on Deep Neural Network Generalization
• Uncovering Graph Reasoning in Decoder-only Transformers with Circuit Tracing
• Spatio-Temporal Directed Graph Learning for Account Takeover Fraud Detection
• Process-Informed Forecasting of Complex Thermal Dynamics in Pharmaceutical Manufacturing
• Graph-Based Spatio-temporal Attention and Multi-Scale Fusion for Clinically Interpretable, High-Fidelity Fetal ECG Extraction
• Time-adaptive H\'enonNets for separable Hamiltonian systems
• Q-Palette: Fractional-Bit Quantizers Toward Optimal Bit Allocation for Efficient LLM Deployment
• Beyond Sharp Minima: Robust LLM Unlearning via Feedback-Guided Multi-Point Optimization
• A HyperGraphMamba-Based Multichannel Adaptive Model for ncRNA Classification
• Energy Use of AI Inference: Efficiency Pathways and Test-Time Compute
• Dynamic Lagging for Time-Series Forecasting in E-Commerce Finance: Mitigating Information Loss with A Hybrid ML Architecture
• Failure Modes of Maximum Entropy RLHF
• Predictive Coding-based Deep Neural Network Fine-tuning for Computationally Efficient Domain Adaptation
• Extended Low-Rank Approximation Accelerates Learning of Elastic Response in Heterogeneous Materials
• PGCLODA: Prompt-Guided Graph Contrastive Learning for Oligopeptide-Infectious Disease Association Prediction
• One Filters All: A Generalist Filter for State Estimation
• You Only Measure Once: On Designing Single-Shot Quantum Machine Learning Models
• Incomplete Data, Complete Dynamics: A Diffusion Approach
• Discovering Association Rules in High-Dimensional Small Tabular Data
• Beyond Slater's Condition in Online CMDPs with Stochastic and Adversarial Constraints
• Probability Signature: Bridging Data Semantics and Embedding Structure in Language Models
• Generative Model Inversion Through the Lens of the Manifold Hypothesis
• An Improved Time Series Anomaly Detection by Applying Structural Similarity
• FairEquityFL -- A Fair and Equitable Client Selection in Federated Learning for Heterogeneous IoV Networks
• Staying on the Manifold: Geometry-Aware Noise Injection
• Practical do-Shapley Explanations with Estimand-Agnostic Causal Inference
• Exploration with Foundation Models: Capabilities, Limitations, and Hybrid Approaches
• MMSE-Calibrated Few-Shot Prompting for Alzheimer's Detection
• TABFAIRGDT: A Fast Fair Tabular Data Generator using Autoregressive Decision Trees
• How deep is your network? Deep vs. shallow learning of transfer operators
• Learnable Sampler Distillation for Discrete Diffusion Models
• From Samples to Scenarios: A New Paradigm for Probabilistic Forecasting
• Faster Than SVD, Smarter Than SGD: The OPLoRA Alternating Update
• RAD: Towards Trustworthy Retrieval-Augmented Multi-modal Clinical Diagnosis
• Pi-Transformer: A Physics-informed Attention Mechanism for Time Series Anomaly Detection
• Learning Robust Penetration-Testing Policies under Partial Observability: A systematic evaluation
• Diffusion-Augmented Contrastive Learning: A Noise-Robust Encoder for Biosignal Representations
• BoreaRL: A Multi-Objective Reinforcement Learning Environment for Climate-Adaptive Boreal Forest Management
• Analyzing Generalization in Pre-Trained Symbolic Regression
• Oversampling and Downsampling with Core-Boundary Awareness: A Data Quality-Driven Approach
• Advancing Universal Deep Learning for Electronic-Structure Hamiltonian Prediction of Materials
• MCGrad:: Multicalibration at Web Scale
• Towards Self-Supervised Foundation Models for Critical Care Time Series
• PromptCoT 2.0: Scaling Prompt Synthesis for Large Language Model Reasoning
• Pure Exploration via Frank-Wolfe Self-Play
• Latent Iterative Refinement Flow: A Geometric-Constrained Approach for Few-Shot Generation
• On the Fragility of Contribution Score Computation in Federated Learning
• Revisiting Performance Claims for Chest X-Ray Models Using Clinical Context
• C${}^2$Prompt: Class-aware Client Knowledge Interaction for Federated Continual Learning
• Frictional Q-Learning
• Sobolev acceleration for neural networks
• PPGFlowECG: Latent Rectified Flow with Cross-Modal Encoding for PPG-Guided ECG Generation and Cardiovascular Disease Detection
• Faster, Smaller, and Smarter: Task-Aware Expert Merging for Online MoE Inference
• RDAR: Reward-Driven Agent Relevance Estimation for Autonomous Driving
• VCRL: Variance-based Curriculum Reinforcement Learning for Large Language Models
• An Efficient Conditional Score-based Filter for High Dimensional Nonlinear Filtering Problems
• On the Rate of Convergence of Kolmogorov-Arnold Network Regression Estimators
• Frame-based Equivariant Diffusion Models for 3D Molecular Generation
• Metriplectic Conditional Flow Matching for Dissipative Dynamics
• Modular Machine Learning with Applications to Genetic Circuit Composition
• Improved Therapeutic Antibody Reformatting through Multimodal Machine Learning
• Adaptive von Mises-Fisher Likelihood Loss for Supervised Deep Time Series Hashing
• TIMED: Adversarial and Autoregressive Refinement of Diffusion-Based Time Series Generation
• Toward Scalable and Structured Global Station Weather Forecasting
• Symbol-Temporal Consistency Self-supervised Learning for Robust Time Series Classification
• Consistent Estimation of Numerical Distributions under Local Differential Privacy by Wavelet Expansion
• Dynamicasome: a molecular dynamics-guided and AI-driven pathogenicity prediction catalogue for all genetic mutations
• FusedANN: Convexified Hybrid ANN via Attribute-Vector Fusion
• Enhancing Credit Default Prediction Using Boruta Feature Selection and DBSCAN Algorithm with Different Resampling Techniques
• Analyzing Uncertainty Quantification in Statistical and Deep Learning Models for Probabilistic Electricity Price Forecasting
• THINNs: Thermodynamically Informed Neural Networks
• Transformer Modeling for Both Scalability and Performance in Multivariate Time Series
• Constraint-Reduced MILP with Local Outlier Factor Modeling for Plausible Counterfactual Explanations in Credit Approval
• Diffusion-Based Impedance Learning for Contact-Rich Manipulation Tasks
• A Unified Noise-Curvature View of Loss of Trainability
• Linear Transformers Implicitly Discover Unified Numerical Algorithms
• Causal Machine Learning for Surgical Interventions
• Intuition to Evidence: Measuring AI's True Impact on Developer Productivity
• SMILES-Inspired Transfer Learning for Quantum Operators in Generative Quantum Eigensolver
• HiCoLoRA: Addressing Context-Prompt Misalignment via Hierarchical Collaborative LoRA for Zero-Shot DST
• Cuffless Blood Pressure Prediction from Speech Sentences using Deep Learning Methods
• ExpFace: Exponential Angular Margin Loss for Deep Face Recognition
• ARCADE: A Real-Time Data System for Hybrid and Continuous Query Processing across Diverse Data Modalities
• Are We Scaling the Right Thing? A System Perspective on Test-Time Scaling
• Where 6G Stands Today: Evolution, Enablers, and Research Gaps
• Large Language Models for Pedestrian Safety: An Application to Predicting Driver Yielding Behavior at Unsignalized Intersections
• RoboSSM: Scalable In-context Imitation Learning via State-Space Models
• MoTiC: Momentum Tightness and Contrast for Few-Shot Class-Incremental Learning
• Selective Classifier-free Guidance for Zero-shot Text-to-speech
• Games Are Not Equal: Classifying Cloud Gaming Contexts for Effective User Experience Measurement
• Thinking While Listening: Simple Test Time Scaling For Audio Classification
• PolicyPad: Collaborative Prototyping of LLM Policies
• DyBBT: Dynamic Balance via Bandit inspired Targeting for Dialog Policy with Cognitive Dual-Systems
• DAWM: Diffusion Action World Models for Offline Reinforcement Learning via Action-Inferred Transitions
• Learning Dynamics of Deep Learning -- Force Analysis of Deep Neural Networks
• A Foundation Chemical Language Model for Comprehensive Fragment-Based Drug Discovery
• Reverse Engineering User Stories from Code using Large Language Models
• Frame-Stacked Local Transformers For Efficient Multi-Codebook Speech Generation
• GuessingGame: Measuring the Informativeness of Open-Ended Questions in Large Language Models
• Knowledge Base-Aware Orchestration: A Dynamic, Privacy-Preserving Method for Multi-Agent Systems
• Advancing Speech Summarization in Multi-modal LLMs with Reinforcement Learning
• Mamba Modulation: On the Length Generalization of Mamba
• ROPA: Synthetic Robot Pose Generation for RGB-D Bimanual Data Augmentation
• Self-evolved Imitation Learning in Simulated World
• A Realistic Evaluation of Cross-Frequency Transfer Learning and Foundation Forecasting Models
• Identifying and Addressing User-level Security Concerns in Smart Homes Using "Smaller" LLMs
• ArtiFree: Detecting and Reducing Generative Artifacts in Diffusion-based Speech Enhancement
• Generative AI as a catalyst for democratic Innovation: Enhancing citizen engagement in participatory budgeting
• AIRwaves at CheckThat! 2025: Retrieving Scientific Sources for Implicit Claims on Social Media with Dual Encoders and Neural Re-Ranking
• The Heterogeneous Multi-Agent Challenge
• A Longitudinal Randomized Control Study of Companion Chatbot Use: Anthropomorphism and Its Mediating Role on Social Impacts
• Semantic-Aware Fuzzing: An Empirical Framework for LLM-Guided, Reasoning-Driven Input Mutation
• TensLoRA: Tensor Alternatives for Low-Rank Adaptation
• OmniFed: A Modular Framework for Configurable Federated Learning from Edge to HPC
• Self-Alignment Learning to Improve Myocardial Infarction Detection from Single-Lead ECG
• FedOC: Multi-Server FL with Overlapping Client Relays in Wireless Edge Networks
• Online Adaptation via Dual-Stage Alignment and Self-Supervision for Fast-Calibration Brain-Computer Interfaces
• Improving Outdoor Multi-cell Fingerprinting-based Positioning via Mobile Data Augmentation
• TimeMosaic: Temporal Heterogeneity Guided Time Series Forecasting via Adaptive Granularity Patch and Segment-wise Decoding
• EngravingGNN: A Hybrid Graph Neural Network for End-to-End Piano Score Engraving
• Probabilistic Runtime Verification, Evaluation and Risk Assessment of Visual Deep Learning Systems
• Learning from Observation: A Survey of Recent Advances
• Data-Driven Reconstruction of Significant Wave Heights from Sparse Observations
• Unsupervised Outlier Detection in Audit Analytics: A Case Study Using USA Spending Data
• Pipeline Parallelism is All You Need for Optimized Early-Exit Based Self-Speculative Decoding
• SLM-Based Agentic AI with P-C-G: Optimized for Korean Tool Use
• Meow: End-to-End Outline Writing for Automatic Academic Survey
• How to inject knowledge efficiently? Knowledge Infusion Scaling Law for Pre-training Large Language Models
• Representation-based Broad Hallucination Detectors Fail to Generalize Out of Distribution
• Uncertainty Quantification of Large Language Models using Approximate Bayesian Computation
• Solving Freshness in RAG: A Simple Recency Prior and the Limits of Heuristic Trend Detection
• The Impact of Structural Changes on Learning Capacity in the Fly Olfactory Neural Circuit
• TriSPrompt: A Hierarchical Soft Prompt Model for Multimodal Rumor Detection with Incomplete Modalities
• RoadMind: Towards a Geospatial AI Expert for Disaster Response
• Benchmarking and Improving LLM Robustness for Personalized Generation
• Anti-Money Laundering Systems Using Deep Learning
• Semantic Representation Attack against Aligned Large Language Models
• DeepACTIF: Efficient Feature Attribution via Activation Traces in Neural Sequence Models
• Analyzing the Impact of Credit Card Fraud on Economic Fluctuations of American Households Using an Adaptive Neuro-Fuzzy Inference System
• The Inadequacy of Offline LLM Evaluations: A Need to Account for Personalization in Model Behavior
• Quantifying Compositionality of Classic and State-of-the-Art Embeddings
• Pluralistic Off-policy Evaluation and Alignment
• CSIYOLO: An Intelligent CSI-based Scatter Sensing Framework for Integrated Sensing and Communication Systems
• Cognitive-Level Adaptive Generation via Capability-Aware Retrieval and Style Adaptation
• Radio Propagation Modelling: To Differentiate or To Deep Learn, That Is The Question
• Multi-population Ensemble Genetic Programming via Cooperative Coevolution and Multi-view Learning for Classification
• Joint Channel Estimation and Computation Offloading in Fluid Antenna-assisted MEC Networks
• Fine-Grained AI Model Caching and Downloading With Coordinated Multipoint Broadcasting in Multi-Cell Edge Networks
• Part-of-speech tagging for Nagamese Language using CRF
• SCORE: A Semantic Evaluation Framework for Generative Document Parsing
• Automated Item Neutralization for Non-Cognitive Scales: A Large Language Model Approach to Reducing Social-Desirability Bias
• Advancing Few-Shot Pediatric Arrhythmia Classification with a Novel Contrastive Loss and Multimodal Learning
• FHIR-AgentBench: Benchmarking LLM Agents for Realistic Interoperable EHR Question Answering
• Readme_AI: Dynamic Context Construction for Large Language Models
• Magnitude Matters: a Superior Class of Similarity Metrics for Holistic Semantic Understanding
• Unveiling the Merits and Defects of LLMs in Automatic Review Generation for Scientific Papers
• A systematic review of trial-matching pipelines using large language models
• Human Activity Recognition Based on Electrocardiogram Data Only
• LibEMER: A novel benchmark and algorithms library for EEG-based Multimodal Emotion Recognition
• Holographic Transformers for Complex-Valued Signal Processing: Integrating Phase Interference into Self-Attention
• Steerable Adversarial Scenario Generation through Test-Time Preference Alignment
• PEPS: Quantum-Inspired Reinforcement Learning for Coherent Reasoning Traces in LLMs
• Formal Verification of Minimax Algorithms
• Federation of Agents: A Semantics-Aware Communication Fabric for Large-Scale Agentic AI
• Design Insights and Comparative Evaluation of a Hardware-Based Cooperative Perception Architecture for Lane Change Prediction
• Scan-do Attitude: Towards Autonomous CT Protocol Management using a Large Language Model Agent
• LLMs as verification oracles for Solidity
• Wavelet Fourier Diffuser: Frequency-Aware Diffusion Model for Reinforcement Learning
• A Federated Fine-Tuning Paradigm of Foundation Models in Heterogenous Wireless Networks
• E2E Learning Massive MIMO for Multimodal Semantic Non-Orthogonal Transmission and Fusion
• Calibrated Reasoning: An Explanatory Verifier for Dynamic and Efficient Problem-Solving
• UserRL: Training Interactive User-Centric Agent via Reinforcement Learning
• The Conductor and the Engine: A Path Towards Co-Designed Reasoning
• Agentic Metacognition: Designing a "Self-Aware" Low-Code Agent for Failure Prediction and Human Handoff
• Analysis of approximate linear programming solution to Markov decision problem with log barrier function
• LatentGuard: Controllable Latent Steering for Robust Refusal of Attacks and Reliable Response Generation
• CON-QA: Privacy-Preserving QA using cloud LLMs in Contract Domain
• Embodied AI: From LLMs to World Models
• MACD: Multi-Agent Clinical Diagnosis with Self-Learned Knowledge for LLM
• From Pheromones to Policies: Reinforcement Learning for Engineered Biological Swarms
• The Indispensable Role of User Simulation in the Pursuit of AGI
• Evaluation-Aware Reinforcement Learning
• Estimating the Self-Consistency of LLMs
• Cognitive Load Limits in Large Language Models: Benchmarking Multi-Hop Reasoning
• Score the Steps, Not Just the Goal: VLM-Based Subgoal Evaluation for Robotic Manipulation
• Nano Bio-Agents (NBA): Small Language Model Agents for Genomics
• What Does Your Benchmark Really Measure? A Framework for Robust Inference of AI Capabilities
• SteinerSQL: Graph-Guided Mathematical Reasoning for Text-to-SQL Generation