AI Research News Feeds for September 29th, 2025

AI RESEARCH PAPERS & ACADEMIC SOURCES

• APTx Neuron: A Unified Trainable Neuron Architecture Integrating Activation and Computation
• Multimodal Recurrent Ensembles for Predicting Brain Responses to Naturalistic Movies (Algonauts 2025)
• Recent Advancements in Microscopy Image Enhancement using Deep Learning: A Survey
• SeamCrafter: Enhancing Mesh Seam Generation for Artist UV Unwrapping via Reinforcement Learning
• FERD: Fairness-Enhanced Data-Free Robustness Distillation
• Differential-Integral Neural Operator for Long-Term Turbulence Forecasting
• TAMMs: Temporal-Aware Multimodal Model for Satellite Image Change Understanding and Forecasting
• Shape-for-Motion: Precise and Consistent Video Editing with 3D Proxy
• Reasoning to Edit: Hypothetical Instruction-Based Image Editing with Visual Reasoning
• SIU3R: Simultaneous Scene Understanding and 3D Reconstruction Beyond Feature Alignment
• Investigating Redundancy in Multimodal Large Language Models with Multiple Vision Encoders
• pFedMMA: Personalized Federated Fine-Tuning with Multi-Modal Adapter for Vision-Language Models
• NarrLV: Towards a Comprehensive Narrative-Centric Evaluation for Long Video Generation
• STQE: Spatial-Temporal Attribute Quality Enhancement for G-PCC Compressed Dynamic Point Clouds
• DriveAgent-R1: Advancing VLM-based Autonomous Driving with Active Perception and Hybrid Thinking
• $A^2R^2$: Advancing Img2LaTeX Conversion via Visual Reasoning with Attention-Guided Refinement
• Content-Aware Mamba for Learned Image Compression
• Small Dents, Big Impact: A Dataset and Deep Learning Approach for Vehicle Dent Detection
• Re-Densification Meets Cross-Scale Propagation: Real-Time Neural Compression of LiDAR Point Clouds
• Draw-In-Mind: Rebalancing Designer-Painter Roles in Unified Multimodal Models Benefits Image Editing
• GLEAM: Learning to Match and Explain in Cross-View Geo-Localization
• Deep Learning for Clouds and Cloud Shadow Segmentation in Methane Satellite and Airborne Imaging Spectroscopy
• Diffence: Fencing Membership Privacy With Diffusion Models
• Metric-Guided Conformal Bounds for Probabilistic Image Reconstruction
• STHN: Deep Homography Estimation for UAV Thermal Geo-localization with Satellite Imagery
• Diverse Subset Selection via Norm-Based Sampling and Orthogonality
• Excavating in the Wild: The GOOSE-Ex Dataset for Semantic Segmentation
• DOTA: Distributional Test-Time Adaptation of Vision-Language Models
• Unstable Unlearning: The Hidden Risk of Concept Resurgence in Diffusion Models
• Rare-to-Frequent: Unlocking Compositional Generation Power of Diffusion Models on Rare Concepts with LLM Guidance
• Surgical Vision World Model
• Learning Personalized Driving Styles via Reinforcement Learning from Human Feedback
• Texture or Semantics? Vision-Language Models Get Lost in Font Recognition
• Can Diffusion Models Disentangle? A Theoretical Perspective
• Geometry aware inference of steady state PDEs using Equivariant Neural Fields representations
• Plan-R1: Safe and Feasible Trajectory Planning as Language Modeling
• Mobi-$\pi$: Mobilizing Your Robot Learning Policy
• iTACO: Interactable Digital Twins of Articulated Objects from Casually Captured RGBD Videos
• NeuVAS: Neural Implicit Surfaces for Variational Shape Modeling
• Rethinking the Embodied Gap in Vision-and-Language Navigation: A Holistic Study of Physical and Visual Disparities
• Multi-View Hypercomplex Learning for Breast Cancer Screening
• Frequency-Domain Refinement with Multiscale Diffusion for Super Resolution
• Leveraging Model Guidance to Extract Training Data from Personalized Diffusion Models
• Pose Prior Learner: Unsupervised Categorical Prior Learning for Pose Estimation
• Diffusion Curriculum: Synthetic-to-Real Data Curriculum via Image-Guided Diffusion
• Large Pre-Training Datasets Don't Always Guarantee Robustness after Fine-Tuning
• TAPTRv3: Spatial and Temporal Context Foster Robust Tracking of Any Point in Long Video
• DynamicControl: Adaptive Condition Selection for Improved Text-to-Image Generation
• Expanding Performance Boundaries of Open-Source Multimodal Models with Model, Data, and Test-Time Scaling
• Self-Guidance: Boosting Flow and Diffusion Generation on Their Own
• LOGen: Toward Lidar Object Generation by Point Diffusion
• UIP2P: Unsupervised Instruction-based Image Editing via Edit Reversibility Constraint
• Unforgettable Lessons from Forgettable Images: Intra-Class Memorability Matters in Computer Vision
• LiT: Delving into a Simple Linear Diffusion Transformer for Image Generation
• Single-weight Model Editing for Post-hoc Spurious Correlation Neutralization
• PDV: Prompt Directional Vectors for Zero-shot Composed Image Retrieval
• VidCRAFT3: Camera, Object, and Lighting Control for Image-to-Video Generation
• GeoDANO: Geometric VLM with Domain Agnostic Vision Encoder
• SPEED: Scalable, Precise, and Efficient Concept Erasure for Diffusion Models
• Semantic Consistent Language Gaussian Splatting for Point-Level Open-vocabulary Querying
• DanceText: A Training-Free Layered Framework for Controllable Multilingual Text Transformation in Images
• CARL: Camera-Agnostic Representation Learning for Spectral Image Analysis
• Image Recognition with Online Lightweight Vision Transformer: A Survey
• Unified Multimodal Chain-of-Thought Reward Model through Reinforcement Fine-Tuning
• OS-W2S: An Automatic Labeling Engine for Language-Guided Open-Set Aerial Object Detection
• Intentional Gesture: Deliver Your Intentions with Gestures for Speech
• Octic Vision Transformers: Quicker ViTs Through Equivariance
• PhyMAGIC: Physical Motion-Aware Generative Inference with Confidence-guided LLM
• Deeper Diffusion Models Amplify Bias
• DVD-Quant: Data-free Video Diffusion Transformers Quantization
• ChartGalaxy: A Dataset for Infographic Chart Understanding and Generation
• CARE: Confidence-aware Ratio Estimation for Medical Biomarkers
• Mamba-Driven Topology Fusion for Monocular 3D Human Pose Estimation
• Towards Scalable Language-Image Pre-training for 3D Medical Imaging
• Pose-free 3D Gaussian splatting via shape-ray estimation
• Physics-Guided Motion Loss for Video Generation Model
• ReSpace: Text-Driven 3D Indoor Scene Synthesis and Editing with Preference Alignment
• Dual Branch VideoMamba with Gated Class Token Fusion for Violence Detection
• Astraea: A Token-wise Acceleration Framework for Video Diffusion Transformers
• DriveAction: A Benchmark for Exploring Human-like Driving Decisions in VLA Models
• Structure before the Machine: Input Space is the Prerequisite for Concepts
• HiSin: A Sinogram-Aware Framework for Efficient High-Resolution Inpainting
• VidBridge-R1: Bridging QA and Captioning for RL-based Video Understanding Models with Intermediate Proxy Tasks
• LEO-VL: Efficient Scene Representation for Scalable 3D Vision-Language Learning
• Think With Videos For Agentic Long-Video Understanding
• video-SALMONN 2: Caption-Enhanced Audio-Visual Large Language Models
• HyCoVAD: A Hybrid SSL-LLM Model for Complex Video Anomaly Detection
• JanusVLN: Decoupling Semantics and Spatiality with Dual Implicit Memory for Vision-Language Navigation
• SpikeMatch: Semi-Supervised Learning with Temporal Dynamics of Spiking Neural Networks
• Vision-Language Alignment from Compressed Image Representations using 2D Gaussian Splatting
• LongLive: Real-time Interactive Long Video Generation
• SPARK: Synergistic Policy And Reward Co-Evolving Framework
• CCNeXt: An Effective Self-Supervised Stereo Depth Estimation Approach
• UML-CoT: Structured Reasoning and Planning with Unified Modeling Language for Robotic Room Cleaning
• LABELING COPILOT: A Deep Research Agent for Automated Data Curation in Computer Vision
• Training-Free Synthetic Data Generation with Dual IP-Adapter Guidance
• Scale-Wise VAR is Secretly Discrete Diffusion
• Hierarchical Representation Matching for CLIP-based Class-Incremental Learning
• Learning Human-Perceived Fakeness in AI-Generated Videos via Multimodal LLMs
• CapRL: Stimulating Dense Image Caption Capabilities via Reinforcement Learning
• RefAM: Attention Magnets for Zero-Shot Referral Segmentation
• SGAligner++: Cross-Modal Language-Aided 3D Scene Graph Alignment
• Cross-Modal Retrieval with Cauchy-Schwarz Divergence
• Language-in-the-Loop Culvert Inspection on the Erie Canal
• Are Hallucinations Bad Estimations?
• VISION: Prompting Ocean Vertical Velocity Reconstruction from Incomplete Observations
• SlimDiff: Training-Free, Activation-Guided Hands-free Slimming of Diffusion Models
• DistillKac: Few-Step Image Generation via Damped Wave Equations
• TRiCo: Triadic Game-Theoretic Co-Training for Robust Semi-Supervised Learning
• Patch-Based Diffusion for Data-Efficient, Radiologist-Preferred MRI Reconstruction
• ControlHair: Physically-based Video Diffusion for Controllable Dynamic Hair Rendering
• Visual Multi-Agent System: Mitigating Hallucination Snowballing via Visual Flow
• Perception-Consistency Multimodal Large Language Models Reasoning via Caption-Regularized Policy Optimization
• Closing the Oracle Gap: Increment Vector Transformation for Class Incremental Learning
• Comparative Analysis of GAN and Diffusion for MRI-to-CT translation
• Enriching Knowledge Distillation with Intra-Class Contrastive Learning
• Guidance Watermarking for Diffusion Models
• Rigidity-Aware 3D Gaussian Deformation from a Single Image
• Aerial Path Planning for Urban Geometry and Texture Co-Capture
• COMPASS: Robust Feature Conformal Prediction for Medical Segmentation Metrics
• Clinical Uncertainty Impacts Machine Learning Evaluations
• RoboView-Bias: Benchmarking Visual Bias in Embodied Agents for Robotic Manipulation
• Deep Learning-Based Cross-Anatomy CT Synthesis Using Adapted nnResU-Net with Anatomical Feature Prioritized Loss
• Adaptive Dual-Mode Distillation with Incentive Schemes for Scalable, Heterogeneous Federated Learning on Non-IID Data
• JointDiff: Bridging Continuous and Discrete in Multi-Agent Trajectory Generation
• Activation Function Design Sustains Plasticity in Continual Learning
• MINT-RVAE: Multi-Cues Intention Prediction of Human-Robot Interaction using Human Pose and Emotion Information from RGB-only Camera Data
• Learn the Ropes, Then Trust the Wins: Self-imitation with Progressive Exploration for Agentic Reinforcement Learning
• WoW: Towards a World omniscient World model Through Embodied Interaction
• VoiceAssistant-Eval: Benchmarking AI Assistants across Listening, Speaking, and Viewing
• Pixel Motion Diffusion is What We Need for Robot Control
• See, Point, Fly: A Learning-Free VLM Framework for Universal Unmanned Aerial Navigation
• Self-Supervised Point Cloud Completion based on Multi-View Augmentations of Single Partial Point Cloud
• REFINE-CONTROL: A Semi-supervised Distillation Method For Conditional Image Generation
• Joint graph entropy knowledge distillation for point cloud classification and robustness against corruptions
• MultiMat: Multimodal Program Synthesis for Procedural Materials using Large Multimodal Models
• DragGANSpace: Latent Space Exploration and Control for GANs
• MinerU2.5: A Decoupled Vision-Language Model for Efficient High-Resolution Document Parsing
• Towards Faithful Reasoning in Remote Sensing: A Perceptually-Grounded GeoSpatial Chain-of-Thought for Vision-Language Models
• Polysemous Language Gaussian Splatting via Matching-based Mask Lifting
• UrbanFeel: A Comprehensive Benchmark for Temporal and Perceptual Understanding of City Scenes through Human Perspective
• A Tale of Two Experts: Cooperative Learning for Source-Free Unsupervised Domain Adaptation
• FlashEdit: Decoupling Speed, Structure, and Semantics for Precise Image Editing
• Beyond Classification Accuracy: Neural-MedBench and the Need for Deeper Reasoning Benchmarks
• UniMapGen: A Generative Framework for Large-Scale Map Construction from Multi-modal Data
• GS-2M: Gaussian Splatting for Joint Mesh Reconstruction and Material Decomposition
• MesaTask: Towards Task-Driven Tabletop Scene Generation via 3D Spatial Reasoning
• Rule-Based Reinforcement Learning for Document Image Classification with Vision Language Models
• Jailbreaking on Text-to-Video Models via Scene Splitting Strategy
• HiGS: History-Guided Sampling for Plug-and-Play Enhancement of Diffusion Models
• Johnson-Lindenstrauss Lemma Guided Network for Efficient 3D Medical Segmentation
• NIFTY: a Non-Local Image Flow Matching for Texture Synthesis
• RAPID^3: Tri-Level Reinforced Acceleration Policies for Diffusion Transformer
• Pedestrian Attribute Recognition via Hierarchical Cross-Modality HyperGraph Learning
• CircuitSense: A Hierarchical Circuit System Benchmark Bridging Visual Comprehension and Symbolic Reasoning in Engineering Design Process
• HierLight-YOLO: A Hierarchical and Lightweight Object Detection Network for UAV Photography
• Effectiveness of Large Multimodal Models in Detecting Disinformation: Experimental Results
• GPT-4 for Occlusion Order Recovery
• Gradient-based multi-focus image fusion with focus-aware saliency enhancement
• Text Adversarial Attacks with Dynamic Outputs
• Integrating Background Knowledge in Medical Semantic Segmentation with Logic Tensor Networks
• Closing the Safety Gap: Surgical Concept Erasure in Visual Autoregressive Models
• RAU: Reference-based Anatomical Understanding with Vision Language Models
• FreqDebias: Towards Generalizable Deepfake Detection via Consistency-Driven Frequency Debiasing
• LucidFlux: Caption-Free Universal Image Restoration via a Large-Scale Diffusion Transformer
• Explaining multimodal LLMs via intra-modal token interactions
• U-MAN: U-Net with Multi-scale Adaptive KAN Network for Medical Image Segmentation
• $\gamma$-Quant: Towards Learnable Quantization for Low-bit Pattern Recognition
• SSVIF: Self-Supervised Segmentation-Oriented Visible and Infrared Image Fusion
• B\'ezier Meets Diffusion: Robust Generation Across Domains for Medical Image Segmentation
• PSTTS: A Plug-and-Play Token Selector for Efficient Event-based Spatio-temporal Representation Learning
• Group Critical-token Policy Optimization for Autoregressive Image Generation
• Where MLLMs Attend and What They Rely On: Explaining Autoregressive Token Generation
• Color Names in Vision-Language Models
• EfficientDepth: A Fast and Detail-Preserving Monocular Depth Estimation Model
• Category Discovery: An Open-World Perspective
• MIRG-RL: Multi-Image Reasoning and Grounding with Reinforcement Learning
• LongScape: Advancing Long-Horizon Embodied World Models with Context-Aware MoE
• MoWM: Mixture-of-World-Models for Embodied Planning via Latent-to-Pixel Feature Modulation
• DiTraj: training-free trajectory control for video diffusion transformer
• A Comprehensive Evaluation of Transformer-Based Question Answering Models and RAG-Enhanced Design
• Dynamic Novel View Synthesis in High Dynamic Range
• SRHand: Super-Resolving Hand Images and 3D Shapes via View/Pose-aware Neural Image Representations and Explicit 3D Meshes
• Deepfakes: we need to re-think the concept of "real" images
• Unlocking the Essence of Beauty: Advanced Aesthetic Reasoning with Relative-Absolute Policy Optimization
• StableDub: Taming Diffusion Prior for Generalized and Efficient Visual Dubbing
• Drag4D: Align Your Motion with Text-Driven 3D Scene Generation
• Syncphony: Synchronized Audio-to-Video Generation with Diffusion Transformers
• LG-CD: Enhancing Language-Guided Change Detection through SAM2 Adaptation
• TDEdit: A Unified Diffusion Framework for Text-Drag Guided Image Manipulation
• Enhancing Vehicle Detection under Adverse Weather Conditions with Contrastive Learning
• Taming Flow-based I2V Models for Creative Video Editing
• Multi-View Crowd Counting With Self-Supervised Learning
• Spatial Reasoning in Foundation Models: Benchmarking Object-Centric Spatial Understanding
• PANICL: Mitigating Over-Reliance on Single Prompt in Visual In-Context Learning
• SingRef6D: Monocular Novel Object Pose Estimation with a Single RGB Reference
• DynaNav: Dynamic Feature and Layer Selection for Efficient Visual Navigation
• SemanticControl: A Training-Free Approach for Handling Loosely Aligned Visual Conditions in ControlNet
• Customizing Visual Emotion Evaluation for MLLMs: An Open-vocabulary, Multifaceted, and Scalable Approach
• MultiCrafter: High-Fidelity Multi-Subject Generation via Spatially Disentangled Attention and Identity-Aware Reinforcement Learning
• PartSAM: A Scalable Promptable Part Segmentation Model Trained on Native 3D Data
• No-Reference Image Contrast Assessment with Customized EfficientNet-B0
• Geo-R1: Improving Few-Shot Geospatial Referring Expression Understanding with Reinforcement Fine-Tuning
• Benchmarking and Mitigate Psychological Sycophancy in Medical Vision-Language Models
• Resolving Ambiguity in Gaze-Facilitated Visual Assistant Interaction Paradigm
• From Bias to Balance: Exploring and Mitigating Spatial Bias in LVLMs
• Mind-the-Glitch: Visual Correspondence for Detecting Inconsistencies in Subject-Driven Generation
• WAVE: Learning Unified & Versatile Audio-Visual Embeddings with Multimodal LLM
• ERGO: Efficient High-Resolution Visual Understanding for Vision-Language Models
• DualFocus: Depth from Focus with Spatio-Focal Dual Variational Constraints
• Rate-Distortion Optimized Communication for Collaborative Perception
• FailureAtlas:Mapping the Failure Landscape of T2I Models via Active Exploration
• Exposing Hallucinations To Suppress Them: VLMs Representation Editing With Generative Anchors
• CoFFT: Chain of Foresight-Focus Thought for Visual Language Models
• Lightweight Structured Multimodal Reasoning for Clinical Scene Understanding in Robotics
• EgoInstruct: An Egocentric Video Dataset of Face-to-face Instructional Interactions with Multi-modal LLM Benchmarking
• High-Quality Sound Separation Across Diverse Categories via Visually-Guided Generative Modeling
• SpecXNet: A Dual-Domain Convolutional Network for Robust Deepfake Detection
• Large Material Gaussian Model for Relightable 3D Generation
• On the Status of Foundation Models for SAR Imagery
• UISim: An Interactive Image-Based UI Simulator for Dynamic Mobile Environments
• LFA-Net: A Lightweight Network with LiteFusion Attention for Retinal Vessel Segmentation
• Incorporating Scene Context and Semantic Labels for Enhanced Group-level Emotion Recognition
• KG-SAM: Injecting Anatomical Knowledge into Segment Anything Models via Conditional Random Fields
• UniVid: Unifying Vision Tasks with Pre-trained Video Generation Models
• CubistMerge: Spatial-Preserving Token Merging For Diverse ViT Backbones
• Training-Free Multimodal Deepfake Detection via Graph Reasoning
• Prompt-guided Representation Disentanglement for Action Recognition
• DeHate: A Stable Diffusion-based Multimodal Approach to Mitigate Hate Speech in Images
• R-Stitch: Dynamic Trajectory Stitching for Efficient Reasoning
• Resource Consumption Red-Teaming for Large Vision-Language Models
• Sparse but Wrong: Incorrect L0 Leads to Incorrect Features in Sparse Autoencoders
• EigenBench: A Comparative Behavioral Measure of Value Alignment
• Cognitive Load Limits in Large Language Models: Benchmarking Multi-Hop Reasoning
• TrustJudge: Inconsistencies of LLM-as-a-Judge and How to Alleviate Them
• Expanding Reasoning Potential in Foundation Model by Learning Diverse Chains of Thought Patterns
• Random Direct Preference Optimization for Radiography Report Generation
• Improving Autism Detection with Multimodal Behavioral Analysis
• KV-Efficient VLA: A Method of Speed up Vision Language Model with RNN-Gated Chunked KV Cache
• Phrase-grounded Fact-checking for Automatically Generated Chest X-ray Reports
• MDF-MLLM: Deep Fusion Through Cross-Modal Feature Alignment for Contextually Aware Fundoscopic Image Classification
• Multimodal Prompt Decoupling Attack on the Safety Filters in Text-to-Image Models
• A Mutual Learning Method for Salient Object Detection with intertwined Multi-Supervision--Revised
• MAJORScore: A Novel Metric for Evaluating Multimodal Relevance via Joint Representation
• Safety Assessment of Scaffolding on Construction Site using AI
• Automated Prompt Generation for Creative and Counterfactual Text-to-image Synthesis
• In silico Deep Learning Protocols for Label-Free Super-Resolution Microscopy: A Comparative Study of Network Architectures and SNR Dependence
• Dynamic Multi-Target Fusion for Efficient Audio-Visual Navigation
• SAEmnesia: Erasing Concepts in Diffusion Models with Sparse Autoencoders
• Coreset selection based on Intra-class diversity
• The LongiMam model for improved breast cancer risk prediction using longitudinal mammograms
• Assessing the Alignment of Popular CNNs to the Brain for Valence Appraisal
• Debugging Concept Bottleneck Models through Removal and Retraining
• ShipwreckFinder: A QGIS Tool for Shipwreck Detection in Multibeam Sonar Data
• Do Sparse Subnetworks Exhibit Cognitively Aligned Attention? Effects of Pruning on Saliency Map Fidelity, Sparsity, and Concept Coherence
• TUN3D: Towards Real-World Scene Understanding from Unposed Images
• Large AI Model-Enabled Generative Semantic Communications for Image Transmission
• mmHSense: Multi-Modal and Distributed mmWave ISAC Datasets for Human Sensing
• Skeleton Sparsification and Densification Scale-Spaces
• Downscaling climate projections to 1 km with single-image super resolution
• JaiLIP: Jailbreaking Vision-Language Models via Loss Guided Image Perturbation
• Overview of ExpertLifeCLEF 2018: how far automated identification systems are from the best experts?
• QuadGPT: Native Quadrilateral Mesh Generation with Autoregressive Models
• DyME: Dynamic Multi-Concept Erasure in Diffusion Models with Bi-Level Orthogonal LoRA Adaptation
• VideoJudge: Bootstrapping Enables Scalable Supervision of MLLM-as-a-Judge for Video Understanding
• Residual Vector Quantization For Communication-Efficient Multi-Agent Perception
• Gender Stereotypes in Professional Roles Among Saudis: An Analytical Study of AI-Generated Images Using Language Models
• Reasoning-Enhanced Domain-Adaptive Pretraining of Multimodal Large Language Models for Short Video Content Moderation
• Learning GUI Grounding with Spatial Reasoning from Visual Feedback
• X-CoT: Explainable Text-to-Video Retrieval via LLM-based Chain-of-Thought Reasoning
• Unsupervised Defect Detection for Surgical Instruments
• No Alignment Needed for Generation: Learning Linearly Separable Representations in Diffusion Models
• Enhancing Contrastive Learning for Geolocalization by Discovering Hard Negatives on Semivariograms
• X-Streamer: Unified Human World Modeling with Audiovisual Interaction
• What Happens Next? Anticipating Future Motion by Generating Point Trajectories
• Temporal vs. Spatial: Comparing DINOv3 and V-JEPA2 Feature Representations for Video Action Analysis
• VLCE: A Knowledge-Enhanced Framework for Image Description in Disaster Assessment
• A Data-driven Typology of Vision Models from Integrated Representational Metrics
• FantasyWorld: Geometry-Consistent World Modeling via Unified Video and 3D Prediction
• MORPH: Shape-agnostic PDE Foundation Models
• MS-YOLO: Infrared Object Detection for Edge Deployment via MobileNetV4 and SlideLoss
• Motion-Aware Transformer for Multi-Object Tracking
• DeLiVR: Differential Spatiotemporal Lie Bias for Efficient Video Deraining
• InfiMed: Low-Resource Medical MLLMs with Advancing Understanding and Reasoning
• RARE: Retrieval-Aware Robustness Evaluation for Retrieval-Augmented Generation Systems
• Probing Neural Topology of Large Language Models
• Resisting Contextual Interference in RAG via Parametric-Knowledge Reinforcement
• Personalized LLM Decoding via Contrasting Personal Preference
• MUCAR: Benchmarking Multilingual Cross-Modal Ambiguity Resolution for Multimodal Large Language Models
• KaLM-Embedding-V2: Superior Training Techniques and Data Inspire A Versatile Embedding Model
• VAT-KG: Knowledge-Intensive Multimodal Knowledge Graph Dataset for Retrieval-Augmented Generation
• WildSpeech-Bench: Benchmarking End-to-End SpeechLLMs in the Wild
• Why Reinforcement Fine-Tuning Enables MLLMs Preserve Prior Knowledge Better: A Data Perspective
• Unveiling the Potential of Diffusion Large Language Model in Controllable Generation
• Evaluating Memory in LLM Agents via Incremental Multi-Turn Interactions
• Learn Globally, Speak Locally: Bridging the Gaps in Multilingual Reasoning
• What Factors Affect LLMs and RLLMs in Financial Question Answering?
• KV Cache Steering for Controlling Frozen LLMs
• The Imitation Game: Turing Machine Imitator is Length Generalizable Reasoner
• LoopServe: An Adaptive Dual-phase LLM Inference Acceleration System for Multi-Turn Dialogues
• Persona-Augmented Benchmarking: Evaluating LLMs Across Diverse Writing Styles
• DAMR: Efficient and Adaptive Context-Aware Knowledge Graph Question Answering with LLM-Guided MCTS
• MLP Memory: A Retriever-Pretrained Memory for Large Language Models
• Conflict-Aware Soft Prompting for Retrieval-Augmented Generation
• Influence-driven Curriculum Learning for Pre-training on Limited Data
• Dream to Chat: Model-based Reinforcement Learning on Dialogues with User Belief Modeling
• JudgeAgent: Knowledge-wise and Dynamic LLM Evaluation with Agent-as-Interviewer
• CMRAG: Co-modality-based visual document retrieval and question answering
• Chain or tree? Re-evaluating complex reasoning from the perspective of a matrix of thought
• Towards an AI Musician: Synthesizing Sheet Music Problems for Musical Reasoning
• Positional Encoding via Token-Aware Phase Attention
• DivLogicEval: A Framework for Benchmarking Logical Reasoning Evaluation in Large Language Models
• Distribution-Aligned Decoding for Efficient LLM Task Adaptation
• RPG: A Repository Planning Graph for Unified and Scalable Codebase Generation
• HiCoLoRA: Addressing Context-Prompt Misalignment via Hierarchical Collaborative LoRA for Zero-Shot DST
• A Critical Look At Tokenwise Reward-Guided Text Generation
• Large Language Models versus Classical Machine Learning: Performance in COVID-19 Mortality Prediction Using High-Dimensional Tabular Data
• On the Within-class Variation Issue in Alzheimer's Disease Detection
• Development and Validation of a Large Language Model for Generating Fully-Structured Radiology Reports
• Rare-to-Frequent: Unlocking Compositional Generation Power of Diffusion Models on Rare Concepts with LLM Guidance
• Training-Free Bayesianization for Low-Rank Adapters of Large Language Models
• Detecting and Interpreting NSFW Prompts in Text-to-Image Models through Uncovering Harmful Semantics
• Process Reinforcement through Implicit Rewards
• GeoDANO: Geometric VLM with Domain Agnostic Vision Encoder
• Taxonomy, Opportunities, and Challenges of Representation Engineering for Large Language Models
• Recursive Training Loops in LLMs: How training data properties modulate distribution shift in generated data?
• TokUR: Token-Level Uncertainty Estimation for Large Language Model Reasoning
• Feature Hedging: Correlated Features Break Narrow Sparse Autoencoders
• The Polar Express: Optimal Matrix Sign Methods and Their Application to the Muon Algorithm
• ChartGalaxy: A Dataset for Infographic Chart Understanding and Generation
• Domain-Aware Tensor Network Structure Search
• Think With Videos For Agentic Long-Video Understanding
• Security Degradation in Iterative AI Code Generation -- A Systematic Analysis of the Paradox
• video-SALMONN 2: Caption-Enhanced Audio-Visual Large Language Models
• Latent Concept Disentanglement in Transformer-based Language Models
• MultiVox: A Benchmark for Evaluating Voice Assistants for Multimodal Interactions
• Rethinking the Embodied Gap in Vision-and-Language Navigation: A Holistic Study of Physical and Visual Disparities
• From Roots to Rewards: Dynamic Tree Reasoning with Reinforcement Learning
• The Invisible Leash: Why RLVR May or May Not Escape Its Origin
• Library Hallucinations in LLMs: Risk Analysis Grounded in Developer Queries
• InfiMed-Foundation: Pioneering Advanced Multimodal Medical Models with Compute-Efficient Pre-Training and Multi-Stage Fine-Tuning
• PRIME: Planning and Retrieval-Integrated Memory for Enhanced Reasoning
• Can Synthetic Query Rewrites Capture User Intent Better than Humans in Retrieval-Augmented Generation?
• Bridging Kolmogorov Complexity and Deep Learning: Asymptotically Optimal Description Length Objectives for Transformers
• MDAR: A Multi-scene Dynamic Audio Reasoning Benchmark
• IIET: Efficient Numerical Transformer via Implicit Iterative Euler Method
• Mental Health Impacts of AI Companions: Triangulating Social Media Quasi-Experiments, User Perspectives, and Relational Theory
• Does AI Coaching Prepare us for Workplace Negotiations?
• Dynamic Experts Search: Enhancing Reasoning in Mixture-of-Experts LLMs at Test Time
• EPO: Entropy-regularized Policy Optimization for LLM Agents Reinforcement Learning
• Learn the Ropes, Then Trust the Wins: Self-imitation with Progressive Exploration for Agentic Reinforcement Learning
• Vision-Language Alignment from Compressed Image Representations using 2D Gaussian Splatting
• IA2: Alignment with ICL Activations Improves Supervised Fine-Tuning
• LABELING COPILOT: A Deep Research Agent for Automated Data Curation in Computer Vision
• Towards Efficient Online Exploration for Reinforcement Learning with Human Feedback
• Learning Human-Perceived Fakeness in AI-Generated Videos via Multimodal LLMs
• CapRL: Stimulating Dense Image Caption Capabilities via Reinforcement Learning
• See, Point, Fly: A Learning-Free VLM Framework for Universal Unmanned Aerial Navigation
• Constituency Parsing using LLMs
• TEXT2AFFORD: Probing Object Affordance Prediction abilities of Language Models solely from Text
• Sharing Matters: Analysing Neurons Across Languages and Tasks in LLMs
• LLMAEL: Large Language Models are Good Context Augmenters for Entity Linking
• Position IDs Matter: An Enhanced Position Layout for Efficient Context Compression in Large Language Models
• Stuffed Mamba: Oversized States Lead to the Inability to Forget
• Improving the Language Understanding Capabilities of Large Language Models Using Reinforcement Learning
• Vulnerability of LLMs to Vertically Aligned Text Manipulations
• Semantic Component Analysis: Introducing Multi-Topic Distributions to Clustering-Based Topic Modeling
• $100K or 100 Days: Trade-offs when Pre-Training with Academic Resources
• AtomR: Atomic Operator-Empowered Large Language Models for Heterogeneous Knowledge Reasoning
• Can LLMs be Good Graph Judge for Knowledge Graph Construction?
• Demystifying Domain-adaptive Post-training for Financial LLMs
• Demystifying Multilingual Chain-of-Thought in Process Reward Modeling
• Elucidating Mechanisms of Demographic Bias in LLMs for Healthcare
• LoRA-MGPO: Mitigating Double Descent in Low-Rank Adaptation via Momentum-Guided Perturbation Optimization
• RuCCoD: Towards Automated ICD Coding in Russian
• How LLMs Fail to Support Fact-Checking
• Adaptively profiling models with task elicitation
• Generator-Assistant Stepwise Rollback Framework for Large Language Model Agent
• Improving LLM-as-a-Judge Inference with the Judgment Distribution
• InftyThink: Breaking the Length Limits of Long-Context Reasoning in Large Language Models
• Cost-Optimal Grouped-Query Attention for Long-Context Modeling
• Retrieval-Augmented Generation with Hierarchical Knowledge
• CLASH: Evaluating Language Models on Judging High-Stakes Dilemmas from Multiple Perspectives
• SOLAR: Towards Characterizing Subjectivity of Individuals through Modeling Value Conflicts and Trade-offs
• MrGuard: A Multilingual Reasoning Guardrail for Universal LLM Safety
• Comparing Uncertainty Measurement and Mitigation Methods for Large Language Models: A Systematic Review
• LLM-OptiRA: LLM-Driven Optimization of Resource Allocation for Non-Convex Problems in Wireless Communications
• Follow the Path: Reasoning over Knowledge Graph Paths to Improve LLM Factuality
• SuperCoder: Assembly Program Superoptimization with Large Language Models
• HiddenBench: Assessing Collective Reasoning in Multi-Agent LLMs via Hidden Profile Tasks
• ZeroTuning: Unlocking the Initial Token's Power to Enhance Large Language Models Without Training
• HBO: Hierarchical Balancing Optimization for Fine-Tuning Large Language Models
• ExpertSteer: Intervening in LLMs through Expert Knowledge
• Shadow-FT: Tuning Instruct Model via Training on Paired Base Model
• Language-Specific Latent Process Hinders Cross-Lingual Performance
• UltraEdit: Training-, Subject-, and Memory-Free Lifelong Editing in Language Models
• UniErase: Towards Balanced and Precise Unlearning in Language Models
• Beyond Early-Token Bias: Model-Specific and Language-Specific Position Effects in Multilingual LLMs
• Attributing Response to Context: A Jensen-Shannon Divergence Driven Mechanistic Study of Context Attribution in Retrieval-Augmented Generation
• Beyond Static Testbeds: An Interaction-Centric Agent Simulation Platform for Dynamic Recommender Systems
• Unlearning Isn't Deletion: Investigating Reversibility of Machine Unlearning in LLMs
• BP-Seg: A graphical model approach to unsupervised and non-contiguous text segmentation using belief propagation
• From Tokens to Thoughts: How LLMs and Humans Trade Compression for Meaning
• Prompting is not Enough: Exploring Knowledge Integration and Controllable Generation on Large Language Models
• BiomedSQL: Text-to-SQL for Scientific Reasoning on Biomedical Knowledge Bases
• EmoBench-UA: A Benchmark Dataset for Emotion Detection in Ukrainian
• Table-R1: Inference-Time Scaling for Table Reasoning
• FeatBench: Evaluating Coding Agents on Feature Implementation for Vibe Coding
• FLEXI: Benchmarking Full-duplex Human-LLM Speech Interaction
• Safety Compliance: Rethinking LLM Safety Reasoning through the Lens of Compliance
• Beyond Textual Context: Structural Graph Encoding with Adaptive Space Alignment to alleviate the hallucination of LLMs
• Bridging Fairness and Explainability: Can Input-Based Explanations Promote Fairness in Hate Speech Detection?
• Advancing Natural Language Formalization to First Order Logic with Fine-tuned LLMs
• Transformers Can Learn Connectivity in Some Graphs but Not Others
• The InviTE Corpus: Annotating Invectives in Tudor English Texts for Computational Modeling
• Conversational Implicatures: Modelling Relevance Theory Probabilistically
• CHRONOBERG: Capturing Language Evolution and Temporal Awareness in Foundation Models
• Exploratory Semantic Reliability Analysis of Wind Turbine Maintenance Logs using Large Language Models
• What Is The Political Content in LLMs' Pre- and Post-Training Data?
• Chimera: Diagnosing Shortcut Learning in Visual-Language Understanding
• Detecting (Un)answerability in Large Language Models with Linear Directions
• Evaluating the Limits of Large Language Models in Multilingual Legal Reasoning
• NeLLCom-Lex: A Neural-agent Framework to Study the Interplay between Lexical Systems and Language Use
• Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
• JGU Mainz's Submission to the WMT25 Shared Task on LLMs with Limited Resources for Slavic Languages: MT and QA
• Representing LLMs in Prompt Semantic Task Space
• We Think, Therefore We Align LLMs to Helpful, Harmless and Honest Before They Go Wrong
• InfiR2: A Comprehensive FP8 Training Recipe for Reasoning-Enhanced Language Models
• Think Socially via Cognitive Reasoning
• Retrieval-Augmented Guardrails for AI-Drafted Patient-Portal Messages: Error Taxonomy Construction and Large-Scale Evaluation
• Fine-Grained Detection of Context-Grounded Hallucinations Using LLMs
• ArabJobs: A Multinational Corpus of Arabic Job Ads
• From Formal Language Theory to Statistical Learning: Finite Observability of Subregular Languages
• Capturing Opinion Shifts in Deliberative Discourse through Frequency-based Quantum deep learning methods
• From tests to effect sizes: Quantifying uncertainty and statistical variability in multilingual and multitask NLP evaluation benchmarks
• StateX: Enhancing RNN Recall via Post-training State Expansion
• Variational Reasoning for Language Models
• Language Models Can Learn from Verbal Feedback Without Scalar Rewards
• Death of the Novel(ty): Beyond n-Gram Novelty as a Metric for Textual Creativity
• WebGen-Agent: Enhancing Interactive Website Generation with Multi-Level Feedback and Step-Level Reinforcement Learning
• VoiceAssistant-Eval: Benchmarking AI Assistants across Listening, Speaking, and Viewing
• Accelerate Creation of Product Claims Using Generative AI
• HetaRAG: Hybrid Deep Retrieval-Augmented Generation across Heterogeneous Data Stores
• Towards mitigating information leakage when evaluating safety monitors
• Random Direct Preference Optimization for Radiography Report Generation
• ReGeS: Reciprocal Retrieval-Generation Synergy for Conversational Recommender Systems
• LLMs for Bayesian Optimization in Scientific Domains: Are We There Yet?
• ARTI-6: Towards Six-dimensional Articulatory Speech Encoding
• VideoJudge: Bootstrapping Enables Scalable Supervision of MLLM-as-a-Judge for Video Understanding
• Gender Stereotypes in Professional Roles Among Saudis: An Analytical Study of AI-Generated Images Using Language Models
• Are Hallucinations Bad Estimations?
• LLM Agent Meets Agentic AI: Can LLM Agents Simulate Customers to Evaluate Agentic-AI-based Shopping Assistants?
• Uncertainty-Aware Knowledge Tracing Models
• C-QUERI: Congressional Questions, Exchanges, and Responses in Institutions Dataset
• Learning GUI Grounding with Spatial Reasoning from Visual Feedback
• Leveraging Big Data Frameworks for Spam Detection in Amazon Reviews
• AUDDT: Audio Unified Deepfake Detection Benchmark Toolkit
• InvBench: Can LLMs Accelerate Program Verification with Invariant Synthesis?
• UISim: An Interactive Image-Based UI Simulator for Dynamic Mobile Environments
• UltraHorizon: Benchmarking Agent Capabilities in Ultra Long-Horizon Scenarios
• DeHate: A Stable Diffusion-based Multimodal Approach to Mitigate Hate Speech in Images
• Compiling by Proving: Language-Agnostic Automatic Optimization from Formal Semantics
• SBFA: Single Sneaky Bit Flip Attack to Break Large Language Models
• What Makes LLM Agent Simulations Useful for Policy? Insights From an Iterative Design Engagement in Emergency Preparedness
• You Can't Steal Nothing: Mitigating Prompt Leakages in LLMs via System Vectors
• AgentPack: A Dataset of Code Changes, Co-Authored by Agents and Humans
• Evaluating Open-Source Large Language Models for Technical Telecom Question Answering
• RISK: A Framework for GUI Agents in E-commerce Risk Management
• From Bias to Balance: Exploring and Mitigating Spatial Bias in LVLMs
• ERGO: Efficient High-Resolution Visual Understanding for Vision-Language Models
• The Thinking Spectrum: An Emperical Study of Tunable Reasoning in LLMs through Model Merging
• A2R: An Asymmetric Two-Stage Reasoning Framework for Parallel Reasoning
• Speak Your Mind: The Speech Continuation Task as a Probe of Voice-Based Model Bias
• SecureAgentBench: Benchmarking Secure Code Generation under Realistic Vulnerability Scenarios
• MinerU2.5: A Decoupled Vision-Language Model for Efficient High-Resolution Document Parsing
• Thinking in Many Modes: How Composite Reasoning Elevates Large Language Model Performance with Limited Data
• In Their Own Words: Reasoning Traces Tailored for Small Models Make Them Better Reasoners
• Influence Guided Context Selection for Effective Retrieval-Augmented Generation
• Context Is What You Need: The Maximum Effective Context Window for Real World Limits of LLMs
• How Large Language Models Need Symbolism
• One Model, Many Morals: Uncovering Cross-Linguistic Misalignments in Computational Moral Reasoning
• LLM-Based Support for Diabetes Diagnosis: Opportunities, Scenarios, and Challenges with GPT-5
• Diagnosing the Performance Trade-off in Moral Alignment: A Case Study on Gender Stereotypes
• A State-of-the-Art SQL Reasoning Model using RLVR
• Learning to Reason with Mixture of Tokens
• Dual-Head Reasoning Distillation: Improving Classifier Accuracy with Train-Time-Only Reasoning
• On Code-Induced Reasoning in LLMs
• Agribot: agriculture-specific question answer system
• Domain-Aware Speaker Diarization On African-Accented English
• Generation-Time vs. Post-hoc Citation: A Holistic Evaluation of LLM Attribution
• Comparative Personalization for Multi-document Summarization
• Vision Language Models Cannot Plan, but Can They Formalize?
• "Be My Cheese?": Assessing Cultural Nuance in Multilingual LLM Translations
• Multi-Objective Reinforcement Learning for Large Language Model Optimization: Visionary Perspective
• OjaKV: Context-Aware Online Low-Rank KV Cache Compression with Oja's Rule
• Towards Transparent AI: A Survey on Explainable Language Models
• ReviewScore: Misinformed Peer Review Detection with Large Language Models
• GRAB: A Risk Taxonomy--Grounded Benchmark for Unsupervised Topic Discovery in Financial Disclosures
• Think-on-Graph 3.0: Efficient and Adaptive LLM Reasoning on Heterogeneous Graphs via Multi-Agent Dual-Evolving Context Retrieval
• ProPerSim: Developing Proactive and Personalized AI Assistants through User-Assistant Simulation
• How Accurate Are LLMs at Multi-Question Answering on Conversational Transcripts?
• Self-Speculative Biased Decoding for Faster Live Translation
• Thinking with Sound: Audio Chain-of-Thought Enables Multimodal Reasoning in Large Audio-Language Models
• SynerGen: Contextualized Generative Recommender for Unified Search and Recommendation
• Navigating the Impact of Structured Output Format on Large Language Models through the Compass of Causal Inference
• Evaluating and Improving Cultural Awareness of Reward Models for LLM Alignment
• Redefining Machine Simultaneous Interpretation: From Incremental Translation to Human-Like Strategies
• Towards Minimal Causal Representations for Human Multimodal Language Understanding
• Can LLMs Solve and Generate Linguistic Olympiad Puzzles?
• ResT: Reshaping Token-Level Policy Gradients for Tool-Use Large Language Models
• Semantic Agreement Enables Efficient Open-Ended LLM Cascades
• Following the TRACE: A Structured Path to Empathetic Response Generation with Multi-Agent Models
• KnowMT-Bench: Benchmarking Knowledge-Intensive Long-Form Question Answering in Multi-Turn Dialogues
• Enhancing Low-Rank Adaptation with Structured Nonlinear Transformations
• LUMINA: Detecting Hallucinations in RAG System with Context-Knowledge Signals
• No Prompt Left Behind: Exploiting Zero-Variance Prompts in LLM Reinforcement Learning via Entropy-Guided Advantage Shaping
• QoNext: Towards Next-generation QoE for Foundation Models
• Elastic MoE: Unlocking the Inference-Time Scalability of Mixture-of-Experts
• A Large-Scale Dataset and Citation Intent Classification in Turkish with LLMs
• AutoSCORE: Enhancing Automated Scoring with Multi-Agent Large Language Models via Structured Component Recognition
• SimulSense: Sense-Driven Interpreting for Efficient Simultaneous Speech Translation
• Why Chain of Thought Fails in Clinical Text Understanding
• Debiasing Large Language Models in Thai Political Stance Detection via Counterfactual Calibration
• MotivGraph-SoIQ: Integrating Motivational Knowledge Graphs and Socratic Dialogue for Enhanced LLM Ideation
• Black-Box Hallucination Detection via Consistency Under the Uncertain Expression
• GraphSearch: An Agentic Deep Searching Workflow for Graph Retrieval-Augmented Generation
• From Outliers to Topics in Language Models: Anticipating Trends in News Corpora
• Taxonomy of Comprehensive Safety for Clinical Agents
• Fuzzy Reasoning Chain (FRC): An Innovative Reasoning Framework from Fuzziness to Clarity
• RedNote-Vibe: A Dataset for Capturing Temporal Dynamics of AI-Generated Text in Social Media
• The QCET Taxonomy of Standard Quality Criterion Names and Definitions for the Evaluation of NLP Systems
• Fine-tuning Done Right in Model Editing
• COSPADI: Compressing LLMs via Calibration-Guided Sparse Dictionary Learning
• Multilingual Dialogue Generation and Localization with Dialogue Act Scripting
• S2J: Bridging the Gap Between Solving and Judging Ability in Generative Reward Models
• Think Right, Not More: Test-Time Scaling for Numerical Claim Verification
• Universal Legal Article Prediction via Tight Collaboration between Supervised Classification Model and LLM
• Multilingual Vision-Language Models, A Survey
• FoodSEM: Large Language Model Specialized in Food Named-Entity Linking
• R-Capsule: Compressing High-Level Plans for Efficient Large Language Model Reasoning
• Bridging Draft Policy Misalignment: Group Tree Optimization for Speculative Decoding
• NFDI4DS Shared Tasks for Scholarly Document Processing
• From Long to Lean: Performance-aware and Adaptive Chain-of-Thought Compression via Multi-round Refinement
• Mixture of Detectors: A Compact View of Machine-Generated Text Detection
• Context Parametrization with Compositional Adapters
• When Does Reasoning Matter? A Controlled Study of Reasoning's Contribution to Model Performance
• The Outputs of Large Language Models are Meaningless
• Question-Driven Analysis and Synthesis: Building Interpretable Thematic Trees with LLMs for Text Clustering and Controllable Generation
• StableToken: A Noise-Robust Semantic Speech Tokenizer for Resilient SpeechLLMs
• A Novel Differential Feature Learning for Effective Hallucination Detection and Classification
• Intuition emerges in Maximum Caliber models at criticality
• Mitigating Exponential Mixed Frequency Growth through Frequency Selection
• Partially Functional Dynamic Backdoor Diffusion-based Causal Model
• EigenBench: A Comparative Behavioral Measure of Value Alignment
• Decentralized Stochastic Nonconvex Optimization under the Relaxed Smoothness
• Scaling to Multimodal and Multichannel Heart Sound Classification: Fine-Tuning Wav2Vec 2.0 with Synthetic and Augmented Biosignals
• Recent Advancements in Microscopy Image Enhancement using Deep Learning: A Survey
• DivLogicEval: A Framework for Benchmarking Logical Reasoning Evaluation in Large Language Models
• Audio Super-Resolution with Latent Bridge Models
• Residual Off-Policy RL for Finetuning Behavior Cloning Policies
• Cognitive Load Limits in Large Language Models: Benchmarking Multi-Hop Reasoning
• Benchmarking LLMs in Web API Integration Tasks
• Thinking Augmented Pre-training
• IntSR: An Integrated Generative Framework for Search and Recommendation
• Data-driven Neural Networks for Windkessel Parameter Calibration
• Pre-Training Representations of Binary Code Using Contrastive Learning
• Data-driven Piecewise Affine Decision Rules for Stochastic Programming with Covariate Information
• QECO: A QoE-Oriented Computation Offloading Algorithm based on Deep Reinforcement Learning for Mobile Edge Computing
• Diffence: Fencing Membership Privacy With Diffusion Models
• DoDo-Code: an Efficient Levenshtein Distance Embedding-based Code for 4-ary IDS Channel
• Online Resource Allocation with Average Budget Constraints
• Discretization Error of Fourier Neural Operators
• On the Within-class Variation Issue in Alzheimer's Disease Detection
• Leveraging Model Guidance to Extract Training Data from Personalized Diffusion Models
• A Survey on LLM-based Code Generation for Low-Resource and Domain-Specific Programming Languages
• Stuffed Mamba: Oversized States Lead to the Inability to Forget
• GraphSCENE: On-Demand Critical Scenario Generation for Autonomous Vehicles in Simulation
• $100K or 100 Days: Trade-offs when Pre-Training with Academic Resources
• Conditional Latent Space Molecular Scaffold Optimization for Accelerated Molecular Design
• Effectively Leveraging Momentum Terms in Stochastic Line Search Frameworks for Fast Optimization of Finite-Sum Problems
• Training-Free Bayesianization for Low-Rank Adapters of Large Language Models
• Demystifying Domain-adaptive Post-training for Financial LLMs
• IP$^{2}$-RSNN: Bi-level Intrinsic Plasticity Enables Learning-to-learn in Recurrent Spiking Neural Networks
• Forecasting the future development in quality and value of professional football players
• VidCRAFT3: Camera, Object, and Lighting Control for Image-to-Video Generation
• Adaptively profiling models with task elicitation
• Surgical Vision World Model
• Cost-Optimal Grouped-Query Attention for Long-Context Modeling
• Learning Personalized Driving Styles via Reinforcement Learning from Human Feedback
• Ethical AI for Young Digital Citizens: A Call to Action on Privacy Governance
• Detecting Scarce and Sparse Anomalous: Solving Dual Imbalance in Multi-Instance Learning
• Do Data Valuations Make Good Data Prices?
• Multi-Agent Reinforcement Learning for Greenhouse Gas Offset Credit Markets
• Can Code Language Models Learn Clarification-Seeking Behaviors?
• CARL: Camera-Agnostic Representation Learning for Spectral Image Analysis
• LLM-OptiRA: LLM-Driven Optimization of Resource Allocation for Non-Convex Problems in Wireless Communications
• From Grunts to Lexicons: Emergent Language from Cooperative Foraging
• UltraEdit: Training-, Subject-, and Memory-Free Lifelong Editing in Language Models
• Octic Vision Transformers: Quicker ViTs Through Equivariance
• Beyond Early-Token Bias: Model-Specific and Language-Specific Position Effects in Multilingual LLMs
• Attributing Response to Context: A Jensen-Shannon Divergence Driven Mechanistic Study of Context Attribution in Retrieval-Augmented Generation
• Unlearning Isn't Deletion: Investigating Reversibility of Machine Unlearning in LLMs
• BP-Seg: A graphical model approach to unsupervised and non-contiguous text segmentation using belief propagation
• Distillation-Enabled Knowledge Alignment Protocol for Semantic Communication in AI Agent Networks
• BiomedSQL: Text-to-SQL for Scientific Reasoning on Biomedical Knowledge Bases
• Transfer learning for multifidelity simulation-based inference in cosmology
• Mobi-$\pi$: Mobilizing Your Robot Learning Policy
• Two failure modes of deep transformers and how to avoid them: a unified theory of signal propagation at initialisation
• Scalable In-Context Q-Learning
• Dual Branch VideoMamba with Gated Class Token Fusion for Violence Detection
• SNR and Resource Adaptive Deep JSCC for Distributed IoT Image Classification
• Security Degradation in Iterative AI Code Generation -- A Systematic Analysis of the Paradox
• HEIST: A Graph Foundation Model for Spatial Transcriptomics and Proteomics Data
• MUCAR: Benchmarking Multilingual Cross-Modal Ambiguity Resolution for Multimodal Large Language Models
• pFedMMA: Personalized Federated Fine-Tuning with Multi-Modal Adapter for Vision-Language Models
• Learn Globally, Speak Locally: Bridging the Gaps in Multilingual Reasoning
• A Unified Empirical Risk Minimization Framework for Flexible N-Tuples Weak Supervision
• APTx Neuron: A Unified Trainable Neuron Architecture Integrating Activation and Computation
• Tricks and Plug-ins for Gradient Boosting in Image Classification
• Hallucination to Truth: A Review of Fact-Checking and Factuality Evaluation in Large Language Models
• Beyond the Proxy: Trajectory-Distilled Guidance for Offline GFlowNet Training
• Practical estimation of the optimal classification error with soft labels and calibration
• Spectral-inspired Operator Learning with Limited Data and Unknown Physics
• SDPO: Importance-Sampled Direct Preference Optimization for Stable Diffusion Training
• Model-Preserving Adaptive Rounding
• Domain-Aware Tensor Network Structure Search
• Mamba Integrated with Physics Principles Masters Long-term Chaotic System Forecasting
• Intercept Cancer: Cancer Pre-Screening with Large Scale Healthcare Foundation Models
• RsGCN: Subgraph-Based Rescaling Enhances Generalization of GCNs for Solving Traveling Salesman Problems
• WeightLoRA: Keep Only Necessary Adapters
• Sign-SGD is the Golden Gate between Multi-Node to Single-Node Learning: Significant Boost via Parameter-Free Optimization
• OrthoGrad Improves Neural Calibration
• Spectral Graph Neural Networks are Incomplete on Graphs with a Simple Spectrum
• AMPED: Adaptive Multi-objective Projection for balancing Exploration and skill Diversification
• Caterpillar GNN: Replacing Message Passing with Efficient Aggregation
• Aircraft Trajectory Dataset Augmentation in Latent Space
• Exploiting Block Coordinate Descent for Cost-Effective LLM Model Training
• RL-Obfuscation: Can Language Models Learn to Evade Latent-Space Monitors?
• SecP-Tuning: Efficient Privacy-Preserving Prompt Tuning for Large Language Models via MPC
• Latent Concept Disentanglement in Transformer-based Language Models
• Online Multi-Agent Control with Adversarial Disturbances
• On the Necessity of Output Distribution Reweighting for Effective Class Unlearning
• Beyond Simple Graphs: Neural Multi-Objective Routing on Multigraphs
• Whom to Trust? Adaptive Collaboration in Personalized Federated Learning
• Neural-Network solver of ideal MHD equilibria
• Lightweight MSA Design Advances Protein Folding From Evolutionary Embeddings
• Relative Entropy Pathwise Policy Optimization
• The Invisible Leash: Why RLVR May or May Not Escape Its Origin
• SpectrumWorld: Artificial Intelligence Foundation for Spectroscopy
• Tricks and Plug-ins for Gradient Boosting with Transformers
• Graph is a Natural Regularization: Revisiting Vector Quantization for Graph Representation Learning
• ERIS: An Energy-Guided Feature Disentanglement Framework for Out-of-Distribution Time Series Classification
• Multi-Channel Differential Transformer for Cross-Domain Sleep Stage Classification with Heterogeneous EEG and EOG
• Sparse but Wrong: Incorrect L0 Leads to Incorrect Features in Sparse Autoencoders
• In-Context Algorithm Emulation in Fixed-Weight Transformers
• Scalable Option Learning in High-Throughput Environments
• Challenges in Non-Polymeric Crystal Structure Prediction: Why a Geometric, Permutation-Invariant Loss is Needed
• AI for Scientific Discovery is a Social Problem
• Towards a Physics Foundation Model
• A Variational Framework for Residual-Based Adaptivity in Neural PDE Solvers and Operator Learning
• GPU Temperature Simulation-Based Testing for In-Vehicle Deep Learning Frameworks
• TimeMosaic: Temporal Heterogeneity Guided Time Series Forecasting via Adaptive Granularity Patch and Segment-wise Decoding
• The Limit Points of (Optimistic) Gradient Descent in Min-Max Optimization
• Multi-View Hypercomplex Learning for Breast Cancer Screening
• Efficient Epistemic Uncertainty Estimation in Regression Ensemble Models Using Pairwise-Distance Estimators
• VDFD: Multi-Agent Value Decomposition Framework with Disentangled World Model
• Fast Partition-Based Cross-Validation With Centering and Scaling for $\mathbf{X}^\mathbf{T}\mathbf{X}$ and $\mathbf{X}^\mathbf{T}\mathbf{Y}$
• Metric-Guided Conformal Bounds for Probabilistic Image Reconstruction
• A Notion of Uniqueness for the Adversarial Bayes Classifier
• Machine Learning-Assisted Sustainable Remanufacturing, Reusing and Recycling for Lithium-ion Batteries
• Diverse Subset Selection via Norm-Based Sampling and Orthogonality
• A Critical Look At Tokenwise Reward-Guided Text Generation
• VeriFlow: Modeling Distributions for Neural Network Verification
• Large Language Models versus Classical Machine Learning: Performance in COVID-19 Mortality Prediction Using High-Dimensional Tabular Data
• Closed-Form Interpretation of Neural Network Latent Spaces with Symbolic Gradients
• DOTA: Distributional Test-Time Adaptation of Vision-Language Models
• Degree-Conscious Spiking Graph for Cross-Domain Adaptation
• Unstable Unlearning: The Hidden Risk of Concept Resurgence in Diffusion Models
• Measurability in the Fundamental Theorem of Statistical Learning
• Capacity-Aware Planning and Scheduling in Budget-Constrained Multi-Agent MDPs: A Meta-RL Approach
• Rare-to-Frequent: Unlocking Compositional Generation Power of Diffusion Models on Rare Concepts with LLM Guidance
• Machine Unlearning for Speaker-Agnostic Detection of Gender-Based Violence Condition in Speech
• How Strategic Agents Respond: Comparing Analytical Models with LLM-Generated Responses in Strategic Classification
• Avoiding $\mathbf{exp(R_{max})}$ scaling in RLHF through Preference-based Exploration
• Efficient Prior Selection in Gaussian Process Bandits with Thompson Sampling
• Process Reinforcement through Implicit Rewards
• GNN-DT: Graph Neural Network Enhanced Decision Transformer for Efficient Optimization in Dynamic Environments
• ReciNet: Reciprocal Space-Aware Long-Range Modeling for Crystalline Property Prediction
• Mechanisms of Projective Composition of Diffusion Models
• LDC-MTL: Balancing Multi-Task Learning through Scalable Loss Discrepancy Control
• Beyond Shallow Behavior: Task-Efficient Value-Based Multi-Task Offline MARL via Skill Discovery
• Fused Partial Gromov-Wasserstein for Structured Objects
• Tokenizing Single-Channel EEG with Time-Frequency Motif Learning
• Taxonomy, Opportunities, and Challenges of Representation Engineering for Large Language Models
• Multi-View Causal Discovery without Non-Gaussianity: Identifiability and Algorithms
• BPINN-EM-Post: Bayesian Physics-Informed Neural Network based Stochastic Electromigration Damage Analysis in the Post-void Phase
• MNT-TNN: Spatiotemporal Traffic Data Imputation via Compact Multimode Nonlinear Transform-based Tensor Nuclear Norm
• Can Diffusion Models Disentangle? A Theoretical Perspective
• CSF: Fixed-outline Floorplanning Based on the Conjugate Subgradient Algorithm Assisted by Q-Learning
• Recursive Training Loops in LLMs: How training data properties modulate distribution shift in generated data?
• Sparsity Forcing: Reinforcing Token Sparsity of MLLMs
• Geometry aware inference of steady state PDEs using Equivariant Neural Fields representations
• Trial and Trust: Addressing Byzantine Attacks with Comprehensive Defense Strategy
• Exploiting the Asymmetric Uncertainty Structure of Pre-trained VLMs on the Unit Hypersphere
• Efficient Orthogonal Fine-Tuning with Principal Subspace Adaptation
• Quantization Meets Reasoning: Exploring and Mitigating Degradation of Low-Bit LLMs in Mathematical Reasoning
• TokUR: Token-Level Uncertainty Estimation for Large Language Model Reasoning
• Feature Hedging: Correlated Features Break Narrow Sparse Autoencoders
• Latent Veracity Inference for Identifying Errors in Stepwise Reasoning
• Structured Relational Representations
• Implicit bias produces neural scaling laws in learning curves, from perceptrons to deep networks
• Forward-only Diffusion Probabilistic Models
• Learning Flexible Forward Trajectories for Masked Molecular Diffusion
• The Polar Express: Optimal Matrix Sign Methods and Their Application to the Muon Algorithm
• SPAR: Self-supervised Placement-Aware Representation Learning for Distributed Sensing
• Bottlenecked Transformers: Periodic KV Cache Consolidation for Generalised Reasoning
• Navigate the Unknown: Enhancing LLM Reasoning with Intrinsic Motivation Guided Exploration
• FFT-based Dynamic Subspace Selection for Low-Rank Adaptive Optimization of Large Language Models
• Can LLMs Alleviate Catastrophic Forgetting in Graph Continual Learning? A Systematic Study
• HD-PiSSA: High-Rank Distributed Orthogonal Adaptation
• ExPLAIND: Unifying Model, Data, and Training Attribution to Study Model Behavior
• Scale-Wise VAR is Secretly Discrete Diffusion
• Variational Reasoning for Language Models
• Language Models Can Learn from Verbal Feedback Without Scalar Rewards
• See, Point, Fly: A Learning-Free VLM Framework for Universal Unmanned Aerial Navigation
• Adaptive Policy Learning to Additional Tasks
• A Random Matrix Perspective of Echo State Networks: From Precise Bias--Variance Characterization to Optimal Regularization
• Exploring the Early Universe with Deep Learning
• Comparative Analysis of GAN and Diffusion for MRI-to-CT translation
• Direct Bias-Correction Term Estimation for Propensity Scores and Average Treatment Effect Estimation
• Incorporating priors in learning: a random matrix study under a teacher-student framework
• Multi-Agent Path Finding via Offline RL and LLM Collaboration
• DragGANSpace: Latent Space Exploration and Control for GANs
• COMPASS: Robust Feature Conformal Prediction for Medical Segmentation Metrics
• Clinical Uncertainty Impacts Machine Learning Evaluations
• Structured Sparse Transition Matrices to Enable State Tracking in State-Space Models
• HiGS: History-Guided Sampling for Plug-and-Play Enhancement of Diffusion Models
• NIFTY: a Non-Local Image Flow Matching for Texture Synthesis
• Preventing Model Collapse Under Overparametrization: Optimal Mixing Ratios for Interpolation Learning and Ridge Regression
• Transformers Can Learn Connectivity in Some Graphs but Not Others
• Multi-channel convolutional neural quantum embedding
• Multidimensional Uncertainty Quantification via Optimal Transport
• Integrating Background Knowledge in Medical Semantic Segmentation with Logic Tensor Networks
• NeuroScalar: A Deep Learning Framework for Fast, Accurate, and In-the-Wild Cycle-Level Performance Prediction
• Learning to Ball: Composing Policies for Long-Horizon Basketball Moves
• Universal Inverse Distillation for Matching Models with Real-Data Supervision (No GANs)
• CausalKANs: interpretable treatment effect estimation with Kolmogorov-Arnold networks
• Evaluating the Limits of Large Language Models in Multilingual Legal Reasoning
• Estimating the Empowerment of Language Model Agents
• Representing LLMs in Prompt Semantic Task Space
• TrueGradeAI: Retrieval-Augmented and Bias-Resistant AI for Transparent and Explainable Digital Assessments
• REMA: A Unified Reasoning Manifold Framework for Interpreting Large Language Model
• Smoothing-Based Conformal Prediction for Balancing Efficiency and Interpretability
• Debiased Front-Door Learners for Heterogeneous Effects
• Metrics for Parametric Families of Networks
• ConQuER: Modular Architectures for Control and Bias Mitigation in IQP Quantum Generative Models
• Linear Causal Representation Learning by Topological Ordering, Pruning, and Disentanglement
• Nearly Tight Regret Bounds for Profit Maximization in Bilateral Trade
• Dynamic Experts Search: Enhancing Reasoning in Mixture-of-Experts LLMs at Test Time
• Effective Policy Learning for Multi-Agent Online Coordination Beyond Submodular Objectives
• From Formal Language Theory to Statistical Learning: Finite Observability of Subregular Languages
• Benefits and Pitfalls of Reinforcement Learning for Language Model Planning: A Theoretical Perspective
• SPARK: Synergistic Policy And Reward Co-Evolving Framework
• StateX: Enhancing RNN Recall via Post-training State Expansion
• Towards Efficient Online Exploration for Reinforcement Learning with Human Feedback
• Training-Free Synthetic Data Generation with Dual IP-Adapter Guidance
• New Algorithmic Directions in Optimal Transport and Applications for Product Spaces
• Agribot: agriculture-specific question answer system
• AutoClimDS: Climate Data Science Agentic AI -- A Knowledge Graph is All You Need
• Domain-Aware Speaker Diarization On African-Accented English
• No Alignment Needed for Generation: Learning Linearly Separable Representations in Diffusion Models
• EEG-Based Consumer Behaviour Prediction: An Exploration from Classical Machine Learning to Graph Neural Networks
• General Pruning Criteria for Fast SBL
• IndiSeek learns information-guided disentangled representations
• What Happens Next? Anticipating Future Motion by Generating Point Trajectories
• Automated and Interpretable Survival Analysis from Multimodal Data
• VLCE: A Knowledge-Enhanced Framework for Image Description in Disaster Assessment
• Multi-Objective Reinforcement Learning for Large Language Model Optimization: Visionary Perspective
• Effective continuous equations for adaptive SGD: a stochastic analysis view
• OjaKV: Context-Aware Online Low-Rank KV Cache Compression with Oja's Rule
• Guiding Audio Editing with Audio Language Model
• InvBench: Can LLMs Accelerate Program Verification with Invariant Synthesis?
• MobiLLM: An Agentic AI Framework for Closed-Loop Threat Mitigation in 6G Open RANs
• Automated Machine Learning Pipeline for Training and Analysis Using Large Language Models
• A regret minimization approach to fixed-point iterations
• Automating Sensor Characterization with Bayesian Optimization
• Generating Stable Placements via Physics-guided Diffusion Models
• MORPH: Shape-agnostic PDE Foundation Models
• HuLA: Prosody-Aware Anti-Spoofing with Multi-Task Learning for Expressive and Emotional Synthetic Speech
• SADA: Safe and Adaptive Inference with Multiple Black-Box Predictions
• Multi-modal Bayesian Neural Network Surrogates with Conjugate Last-Layer Estimation
• Align2Speak: Improving TTS for Low Resource Languages via ASR-Guided Online Preference Optimization
• UISim: An Interactive Image-Based UI Simulator for Dynamic Mobile Environments
• Noise-to-Notes: Diffusion-based Generation and Refinement for Automatic Drum Transcription
• Self-Speculative Biased Decoding for Faster Live Translation
• Retrieval-of-Thought: Efficient Reasoning via Reusing Thoughts
• Reinforcement Learning Based Traffic Signal Design to Minimize Queue Lengths
• CubistMerge: Spatial-Preserving Token Merging For Diverse ViT Backbones
• Lifelong Learning with Behavior Consolidation for Vehicle Routing
• Navigating the Impact of Structured Output Format on Large Language Models through the Compass of Causal Inference
• SBFA: Single Sneaky Bit Flip Attack to Break Large Language Models
• Causal-EPIG: A Prediction-Oriented Active Learning Framework for CATE Estimation
• No Prompt Left Behind: Exploiting Zero-Variance Prompts in LLM Reinforcement Learning via Entropy-Guided Advantage Shaping
• Elastic MoE: Unlocking the Inference-Time Scalability of Mixture-of-Experts
• Error Analysis of Discrete Flow with Generator Matching
• Sequential 1-bit Mean Estimation with Near-Optimal Sample Complexity
• Outlier Detection in Plantar Pressure: Human-Centered Comparison of Statistical Parametric Mapping and Explainable Machine Learning
• Learnable Conformal Prediction with Context-Aware Nonconformity Functions for Robotic Planning and Perception
• FlowDrive: moderated flow matching with data balancing for trajectory planning
• ERGO: Efficient High-Resolution Visual Understanding for Vision-Language Models
• Bilinear relational structure fixes reversal curse and enables consistent model editing
• A Nonparametric Discrete Hawkes Model with a Collapsed Gaussian-Process Prior
• GSM-Agent: Understanding Agentic Reasoning Using Controllable Environments
• From Parameters to Behavior: Unsupervised Compression of the Policy Space
• Machine learning approaches to seismic event classification in the Ostrava region
• EPO: Entropy-regularized Policy Optimization for LLM Agents Reinforcement Learning
• The Lie of the Average: How Class Incremental Learning Evaluation Deceives You?
• Transport Based Mean Flows for Generative Modeling
• Learn the Ropes, Then Trust the Wins: Self-imitation with Progressive Exploration for Agentic Reinforcement Learning
• Quantile Advantage Estimation for Entropy-Safe Reasoning
• IA2: Alignment with ICL Activations Improves Supervised Fine-Tuning
• A Theoretical Analysis of Discrete Flow Matching Generative Models
• Learning Admissible Heuristics for A*: Theory and Practice
• Assessment of deep learning models integrated with weather and environmental variables for wildfire spread prediction and a case study of the 2023 Maui fires
• Interpretable Spectral Features Predict Conductivity in Self-Driving Doped Conjugated Polymer Labs
• Seismic Velocity Inversion from Multi-Source Shot Gathers Using Deep Segmentation Networks: Benchmarking U-Net Variants and SeismoLabV3+
• Cycle is All You Need: More Is Different
• From Embeddings to Equations: Genetic-Programming Surrogates for Interpretable Transformer Classification
• SGNNBench: A Holistic Evaluation of Spiking Graph Neural Network on Large-scale Graph
• Towards mitigating information leakage when evaluating safety monitors
• Spiking Neural Networks for Mental Workload Classification with a Multimodal Approach
• Accurate typhoon intensity forecasts using a non-iterative spatiotemporal transformer model
• Improving Autism Detection with Multimodal Behavioral Analysis
• Data-driven approach to the design of complexing agents for trivalent transuranium elements
• ReGeS: Reciprocal Retrieval-Generation Synergy for Conversational Recommender Systems
• Coreset selection based on Intra-class diversity
• The LongiMam model for improved breast cancer risk prediction using longitudinal mammograms
• Debugging Concept Bottleneck Models through Removal and Retraining
• Do Sparse Subnetworks Exhibit Cognitively Aligned Attention? Effects of Pruning on Saliency Map Fidelity, Sparsity, and Concept Coherence
• mmHSense: Multi-Modal and Distributed mmWave ISAC Datasets for Human Sensing
• Downscaling climate projections to 1 km with single-image super resolution
• Near-Optimal Experiment Design in Linear non-Gaussian Cyclic Models
• DyME: Dynamic Multi-Concept Erasure in Diffusion Models with Bi-Level Orthogonal LoRA Adaptation
• Foundation models for high-energy physics
• A State-of-the-Art SQL Reasoning Model using RLVR
• Enhanced Generative Machine Listener
• Learning to Reason with Mixture of Tokens
• Context-Aware Hybrid Routing in Bluetooth Mesh Networks Using Multi-Model Machine Learning and AODV Fallback
• Functional Encryption in Secure Neural Network Training: Data Leakage and Practical Mitigations
• ASSESS: A Semantic and Structural Evaluation Framework for Statement Similarity
• Wavelet-Induced Rotary Encodings: RoPE Meets Graphs
• Erase or Hide? Suppressing Spurious Unlearning Neurons for Robust Unlearning
• Towards a more realistic evaluation of machine learning models for bearing fault diagnosis
• Fine-Grained Uncertainty Decomposition in Large Language Models: A Spectral Approach
• Unlocking the Power of Mixture-of-Experts for Task-Aware Time Series Analytics
• Conditional Denoising Diffusion Autoencoders for Wireless Semantic Communications
• A Multi-Level Framework for Multi-Objective Hypergraph Partitioning: Combining Minimum Spanning Tree and Proximal Gradient
• Aurora: Towards Universal Generative Multimodal Time Series Forecasting
• HEAPr: Hessian-based Efficient Atomic Expert Pruning in Output Space
• SoDaDE: Solvent Data-Driven Embeddings with Small Transformer Models
• Adaptive Policy Backbone via Shared Network
• Progressive Weight Loading: Accelerating Initial Inference and Gradually Boosting Performance on Resource-Constrained Environments
• Distributed Associative Memory via Online Convex Optimization
• Spectral Collapse Drives Loss of Plasticity in Deep Continual Learning
• SurvDiff: A Diffusion Model for Generating Synthetic Data in Survival Analysis
• Context and Diversity Matter: The Emergence of In-Context Learning in World Models
• Stochastic activations
• Neural Feature Geometry Evolves as Discrete Ricci Flow
• Investigating Faithfulness in Large Audio Language Models
• Role-Aware Multi-modal federated learning system for detecting phishing webpages
• Enhancing Credit Risk Prediction: A Meta-Learning Framework Integrating Baseline Models, LASSO, and ECOC for Superior Accuracy
• (Sometimes) Less is More: Mitigating the Complexity of Rule-based Representation for Interpretable Classification
• SpinGPT: A Large-Language-Model Approach to Playing Poker Correctly
• Improving accuracy in short mortality rate series: Exploring Multi-step Forecasting Approaches in Hybrid Systems
• ReLAM: Learning Anticipation Model for Rewarding Visual Robotic Manipulation
• MoveFM-R: Advancing Mobility Foundation Models via Language-driven Semantic Reasoning
• Fast-Forward Lattice Boltzmann: Learning Kinetic Behaviour with Physics-Informed Neural Operators
• One Prompt Fits All: Universal Graph Adaptation for Pretrained Models
• Partial Parameter Updates for Efficient Distributed Training
• Learning from Delayed Feedback in Games via Extra Prediction
• The Flood Complex: Large-Scale Persistent Homology on Millions of Points
• Global Convergence in Neural ODEs: Impact of Activation Functions
• Bridging Kolmogorov Complexity and Deep Learning: Asymptotically Optimal Description Length Objectives for Transformers
• Overclocking Electrostatic Generative Models
• Physics-informed GNN for medium-high voltage AC power flow with edge-aware attention and line search correction operator
• Nonlinear Optimization with GPU-Accelerated Neural Network Constraints
• IIET: Efficient Numerical Transformer via Implicit Iterative Euler Method
• Learning the Neighborhood: Contrast-Free Multimodal Self-Supervised Molecular Graph Pretraining
• Bayesian Transfer Operators in Reproducing Kernel Hilbert Spaces
• OFMU: Optimization-Driven Framework for Machine Unlearning
• A Machine Learning Pipeline for Multiple Sclerosis Biomarker Discovery: Comparing explainable AI and Traditional Statistical Approaches
• Dual Optimistic Ascent (PI Control) is the Augmented Lagrangian Method in Disguise
• Adaptive Dual-Mode Distillation with Incentive Schemes for Scalable, Heterogeneous Federated Learning on Non-IID Data
• JointDiff: Bridging Continuous and Discrete in Multi-Agent Trajectory Generation
• ECHO: Toward Contextual Seq2Seq Paradigms in Large EEG Models
• Learning to Price Bundles: A GCN Approach for Mixed Bundling
• Activation Function Design Sustains Plasticity in Continual Learning
• Preference-Guided Learning for Sparse-Reward Multi-Agent Reinforcement Learning
• On the Complexity Theory of Masked Discrete Diffusion: From $\mathrm{poly}(1/\epsilon)$ to Nearly $\epsilon$-Free
• Beyond Johnson-Lindenstrauss: Uniform Bounds for Sketched Bilinear Forms
• Graph of Agents: Principled Long Context Modeling by Emergent Multi-Agent Collaboration
• MolSpectLLM: A Molecular Foundation Model Bridging Spectroscopy, Molecule Elucidation, and 3D Structure Generation
• Beyond RAG vs. Long-Context: Learning Distraction-Aware Retrieval for Efficient Knowledge Grounding
• Abductive Logical Rule Induction by Bridging Inductive Logic Programming and Multimodal Large Language Models
• Zubov-Net: Adaptive Stability for Neural ODEs Reconciling Accuracy with Robustness
• Position: The Hidden Costs and Measurement Gaps of Reinforcement Learning with Verifiable Rewards
• Why High-rank Neural Networks Generalize?: An Algebraic Framework with RKHSs
• Closing the Oracle Gap: Increment Vector Transformation for Class Incremental Learning
• Discrete Guidance Matching: Exact Guidance for Discrete Flow Matching
• Multiplicative-Additive Constrained Models:Toward Joint Visualization of Interactive and Independent Effects
• Generation Properties of Stochastic Interpolation under Finite Training Set
• Extracting Actionable Insights from Building Energy Data using Vision LLMs on Wavelet and 3D Recurrence Representations
• Statistical Advantage of Softmax Attention: Insights from Single-Location Regression
• Structural Information-based Hierarchical Diffusion for Offline Reinforcement Learning
• Active Attacks: Red-teaming LLMs via Adaptive Environments
• Think Smart, Not Hard: Difficulty Adaptive Reasoning for Large Audio Language Models
• GRAM-TDI: adaptive multimodal representation learning for drug target interaction prediction
• Stage-wise Dynamics of Classifier-Free Guidance in Diffusion Models
• Goal-Guided Efficient Exploration via Large Language Model in Reinforcement Learning
• Concept-SAE: Active Causal Probing of Visual Model Behavior
• AEGIS: Authentic Edge Growth In Sparsity for Link Prediction in Edge-Sparse Bipartite Knowledge Graphs
• Task-Adaptive Parameter-Efficient Fine-Tuning for Weather Foundation Models
• Teaching Transformers to Solve Combinatorial Problems through Efficient Trial & Error
• MCGM: Multi-stage Clustered Global Modeling for Long-range Interactions in Molecules
• OrtSAE: Orthogonal Sparse Autoencoders Uncover Atomic Features
• Latent Diffusion : Multi-Dimension Stable Diffusion Latent Space Explorer
• Convexity-Driven Projection for Point Cloud Dimensionality Reduction
• MO-GRPO: Mitigating Reward Hacking of Group Relative Policy Optimization on Multi-Objective Problems
• BrainPro: Towards Large-scale Brain State-aware EEG Representation Learning
• Enriching Knowledge Distillation with Intra-Class Contrastive Learning
• Towards Understanding Feature Learning in Parameter Transfer
• The Rogue Scalpel: Activation Steering Compromises LLM Safety
• Non-Linear Trajectory Modeling for Multi-Step Gradient Inversion Attacks in Federated Learning
• SHAKE-GNN: Scalable Hierarchical Kirchhoff-Forest Graph Neural Network
• Reinforcement Learning for Durable Algorithmic Recourse
• Modeling Psychological Profiles in Volleyball via Mixed-Type Bayesian Networks
• Countering adversarial evasion in regression analysis
• Learning More with Less: A Dynamic Dual-Level Down-Sampling Framework for Efficient Policy Optimization
• Mind the Missing: Variable-Aware Representation Learning for Irregular EHR Time Series using Large Language Models
• Slicing Wasserstein Over Wasserstein Via Functional Optimal Transport
• Pushing Toward the Simplex Vertices: A Simple Remedy for Code Collapse in Smoothed Vector Quantization
• Lightweight error mitigation strategies for post-training N:M activation sparsity in LLMs
• Efficiency Boost in Decentralized Optimization: Reimagining Neighborhood Aggregation with Minimal Overhead
• Learning Equivariant Functions via Quadratic Forms
• Mechanistic Independence: A Principle for Identifiable Disentangled Representations
• Kernel Regression of Multi-Way Data via Tensor Trains with Hadamard Overparametrization: The Dynamic Graph Flow Case
• Reversible GNS for Dissipative Fluids with Consistent Bidirectional Dynamics
• A Law of Data Reconstruction for Random Features (and Beyond)
• Automatic Discovery of One Parameter Subgroups of $SO(n)$
• Fairness-Aware Reinforcement Learning (FAReL): A Framework for Transparent and Balanced Sequential Decision-Making
• Limitations on Safe, Trusted, Artificial General Intelligence
• DriftLite: Lightweight Drift Control for Inference-Time Scaling of Diffusion Models
• Differentiable Structure Learning for General Binary Data
• RED-DiffEq: Regularization by denoising diffusion models for solving inverse PDE problems with application to full waveform inversion
• A Systematic Review of Conformal Inference Procedures for Treatment Effect Estimation: Methods and Challenges
• MMPlanner: Zero-Shot Multimodal Procedural Planning with Chain-of-Thought Object State Reasoning
• Logic of Hypotheses: from Zero to Full Knowledge in Neurosymbolic Integration
• DIM: Enforcing Domain-Informed Monotonicity in Deep Neural Networks
• Neuroprobe: Evaluating Intracranial Brain Responses to Naturalistic Stimuli
• SlotFM: A Motion Foundation Model with Slot Attention for Diverse Downstream Tasks
• Scalable Second-order Riemannian Optimization for $K$-means Clustering
• Prophecy: Inferring Formal Properties from Neuron Activations
• SpecMER: Fast Protein Generation with K-mer Guided Speculative Decoding
• Wav2Arrest 2.0: Long-Horizon Cardiac Arrest Prediction with Time-to-Event Modeling, Identity-Invariance, and Pseudo-Lab Alignment
• Exact Subgraph Isomorphism Network for Predictive Graph Mining
• Downscaling human mobility data based on demographic socioeconomic and commuting characteristics using interpretable machine learning methods
• PQFed: A Privacy-Preserving Quality-Controlled Federated Learning Framework
• A Unifying Framework for Parallelizing Sequential Models with Linear Dynamical Systems
• Information-Theoretic Bayesian Optimization for Bilevel Optimization Problems
• Uncovering Alzheimer's Disease Progression via SDE-based Spatio-Temporal Graph Deep Learning on Longitudinal Brain Networks
• POLO: Preference-Guided Multi-Turn Reinforcement Learning for Lead Optimization
• Brain PathoGraph Learning
• HyperCore: Coreset Selection under Noise via Hypersphere Models
• SubZeroCore: A Submodular Approach with Zero Training for Coreset Selection
• Reparameterizing 4DVAR with neural fields
• Machine Learning and AI Applied to fNIRS Data Reveals Novel Brain Activity Biomarkers in Stable Subclinical Multiple Sclerosis
• Beyond Formula Complexity: Effective Information Criterion Improves Performance and Interpretability for Symbolic Regression
• FastGRPO: Accelerating Policy Optimization via Concurrency-aware Speculative Decoding and Online Draft Learning
• Exploring the Relationships Between Physiological Signals During Automated Fatigue Detection
• ChaosNexus: A Foundation Model for Universal Chaotic System Forecasting with Multi-scale Representations
• Scaling Laws for Neural Material Models
• Sharpness-Aware Minimization Can Hallucinate Minimizers
• High-Probability Analysis of Online and Federated Zero-Order Optimisation
• Neural Operators for Mathematical Modeling of Transient Fluid Flow in Subsurface Reservoir Systems
• GraphPFN: A Prior-Data Fitted Graph Foundation Model
• SlimDiff: Training-Free, Activation-Guided Hands-free Slimming of Diffusion Models
• Chasing the Tail: Effective Rubric-based Reward Modeling for Large Language Model Post-Training
• Contrastive Mutual Information Learning: Toward Robust Representations without Positive-Pair Augmentations
• DistillKac: Few-Step Image Generation via Damped Wave Equations
• Uncertainty-Aware Knowledge Tracing Models
• $\mathbf{Li_2}$: A Framework on Dynamics of Feature Emergence and Delayed Generalization
• TRiCo: Triadic Game-Theoretic Co-Training for Robust Semi-Supervised Learning
• Preemptive Detection and Steering of LLM Misalignment via Latent Reachability
• Expert-guided Clinical Text Augmentation via Query-Based Model Collaboration
• A circuit for predicting hierarchical structure in-context in Large Language Models
• Evidence for Limited Metacognition in LLMs
• Machine Learning. The Science of Selection under Uncertainty
• Interpretable time series analysis with Gumbel dynamics
• Leveraging Big Data Frameworks for Spam Detection in Amazon Reviews
• GenUQ: Predictive Uncertainty Estimates via Generative Hyper-Networks
• Task-Agnostic Federated Continual Learning via Replay-Free Gradient Projection
• Causal Abstraction Inference under Lossy Representations
• LANCE: Low Rank Activation Compression for Efficient On-Device Continual Learning
• PreLoRA: Hybrid Pre-training of Vision Transformers with Full Training and Low-Rank Adapters
• Shoot from the HIP: Hessian Interatomic Potentials without derivatives
• Blockwise Hadamard high-Rank Adaptation for Parameter-Efficient LLM Fine-Tuning
• Understanding and Enhancing Mask-Based Pretraining towards Universal Representations
• Hallucination to Truth: A Review of Fact-Checking and Factuality Evaluation in Large Language Models
• GTPO and GRPO-S: Token and Sequence-Level Reward Shaping with Policy Entropy
• StreetReaderAI: Making Street View Accessible Using Context-Aware Multimodal AI
• Scalable Option Learning in High-Throughput Environments
• Draw-In-Mind: Rebalancing Designer-Painter Roles in Unified Multimodal Models Benefits Image Editing
• JudgeAgent: Knowledge-wise and Dynamic LLM Evaluation with Agent-as-Interviewer
• Chain or tree? Re-evaluating complex reasoning from the perspective of a matrix of thought
• Positional Encoding via Token-Aware Phase Attention
• Justice in Judgment: Unveiling (Hidden) Bias in LLM-assisted Peer Reviews
• Towards a Physics Foundation Model
• Constructive Conflict-Driven Multi-Agent Reinforcement Learning for Strategic Diversity
• Recent Advancements in Microscopy Image Enhancement using Deep Learning: A Survey
• Comparing RAG and GraphRAG for Page-Level Retrieval Question Answering on Math Textbook
• Dendritic Resonate-and-Fire Neuron for Effective and Efficient Long Sequence Modeling
• Automated Facility Enumeration for Building Compliance Checking using Door Detection and Large Language Models
• Diffusion-Augmented Contrastive Learning: A Noise-Robust Encoder for Biosignal Representations
• AnchDrive: Bootstrapping Diffusion Policies with Hybrid Trajectory Anchors for End-to-End Driving
• Discovering and Analyzing Stochastic Processes to Reduce Waste in Food Retail
• Impact of Loss Weight and Model Complexity on Physics-Informed Neural Networks for Computational Fluid Dynamics
• LLMs for Bayesian Optimization in Scientific Domains: Are We There Yet?
• Object Identification Under Known Dynamics: A PIRNN Approach for UAV Classification
• Null-Space Filtering for Data-Free Continual Model Merging: Preserving Transparency, Promoting Fidelity
• Forecasting Seismic Waveforms: A Deep Learning Approach for Einstein Telescope
• Talking Trees: Reasoning-Assisted Induction of Decision Trees for Tabular Data
• Score-based Idempotent Distillation of Diffusion Models
• Are Hallucinations Bad Estimations?
• d2: Improved Techniques for Training Reasoning Diffusion Language Models
• VISION: Prompting Ocean Vertical Velocity Reconstruction from Incomplete Observations
• Filtering with Confidence: When Data Augmentation Meets Conformal Prediction
• SDPO: Importance-Sampled Direct Preference Optimization for Stable Diffusion Training
• DORAEMON: Decentralized Ontology-aware Reliable Agent with Enhanced Memory Oriented Navigation
• Model-Preserving Adaptive Rounding
• Mamba Integrated with Physics Principles Masters Long-term Chaotic System Forecasting
• InfiMed: Low-Resource Medical MLLMs with Advancing Understanding and Reasoning
• Probing Neural Topology of Large Language Models
• Physics-Guided Motion Loss for Video Generation Model
• CapSpeech: Enabling Downstream Applications in Style-Captioned Text-to-Speech
• Dual Branch VideoMamba with Gated Class Token Fusion for Violence Detection
• Resisting Contextual Interference in RAG via Parametric-Knowledge Reinforcement
• DriveAction: A Benchmark for Exploring Human-like Driving Decisions in VLA Models
• AMPED: Adaptive Multi-objective Projection for balancing Exploration and skill Diversification
• Position: Simulating Society Requires Simulating Thought
• VidBridge-R1: Bridging QA and Captioning for RL-based Video Understanding Models with Intermediate Proxy Tasks
• Think With Videos For Agentic Long-Video Understanding
• Security Degradation in Iterative AI Code Generation -- A Systematic Analysis of the Paradox
• Exploiting Block Coordinate Descent for Cost-Effective LLM Model Training
• Personalized LLM Decoding via Contrasting Personal Preference
• Latent Concept Disentanglement in Transformer-based Language Models
• TAMMs: Temporal-Aware Multimodal Model for Satellite Image Change Understanding and Forecasting
• On the Necessity of Output Distribution Reweighting for Effective Class Unlearning
• Beyond Simple Graphs: Neural Multi-Objective Routing on Multigraphs
• Why Reinforcement Fine-Tuning Enables MLLMs Preserve Prior Knowledge Better: A Data Perspective
• Investigating Redundancy in Multimodal Large Language Models with Multiple Vision Encoders
• Evaluating Memory in LLM Agents via Incremental Multi-Turn Interactions
• Learn Globally, Speak Locally: Bridging the Gaps in Multilingual Reasoning
• Lightweight MSA Design Advances Protein Folding From Evolutionary Embeddings
• KV Cache Steering for Controlling Frozen LLMs
• Rethinking the Embodied Gap in Vision-and-Language Navigation: A Holistic Study of Physical and Visual Disparities
• LoopServe: An Adaptive Dual-phase LLM Inference Acceleration System for Multi-Turn Dialogues
• The Invisible Leash: Why RLVR May or May Not Escape Its Origin
• R-Stitch: Dynamic Trajectory Stitching for Efficient Reasoning
• DAMR: Efficient and Adaptive Context-Aware Knowledge Graph Question Answering with LLM-Guided MCTS
• Comparing Uncertainty Measurement and Mitigation Methods for Large Language Models: A Systematic Review
• Geometry aware inference of steady state PDEs using Equivariant Neural Fields representations
• CARL: Camera-Agnostic Representation Learning for Spectral Image Analysis
• Neural Orchestration for Multi-Agent Systems: A Deep Learning Framework for Optimal Agent Selection in Multi-Domain Task Environments
• Formalising Human-in-the-Loop: Computational Reductions, Failure Modes, and Legal-Moral Responsibility
• Follow the Path: Reasoning over Knowledge Graph Paths to Improve LLM Factuality
• SuperCoder: Assembly Program Superoptimization with Large Language Models
• HiddenBench: Assessing Collective Reasoning in Multi-Agent LLMs via Hidden Profile Tasks
• Quantization Meets Reasoning: Exploring and Mitigating Degradation of Low-Bit LLMs in Mathematical Reasoning
• TokUR: Token-Level Uncertainty Estimation for Large Language Model Reasoning
• ZeroTuning: Unlocking the Initial Token's Power to Enhance Large Language Models Without Training
• Feature Hedging: Correlated Features Break Narrow Sparse Autoencoders
• Latent Veracity Inference for Identifying Errors in Stepwise Reasoning
• Structured Relational Representations
• Shadow-FT: Tuning Instruct Model via Training on Paired Base Model
• Learning Hierarchical Domain Models Through Environment-Grounded Interaction
• VocalAgent: Large Language Models for Vocal Health Diagnostics with Safety-Aware Evaluation
• UltraEdit: Training-, Subject-, and Memory-Free Lifelong Editing in Language Models
• Intentional Gesture: Deliver Your Intentions with Gestures for Speech
• Octic Vision Transformers: Quicker ViTs Through Equivariance
• UniErase: Towards Balanced and Precise Unlearning in Language Models
• Attributing Response to Context: A Jensen-Shannon Divergence Driven Mechanistic Study of Context Attribution in Retrieval-Augmented Generation
• Beyond Static Testbeds: An Interaction-Centric Agent Simulation Platform for Dynamic Recommender Systems
• Learning Flexible Forward Trajectories for Masked Molecular Diffusion
• Unlearning Isn't Deletion: Investigating Reversibility of Machine Unlearning in LLMs
• The Polar Express: Optimal Matrix Sign Methods and Their Application to the Muon Algorithm
• Bottlenecked Transformers: Periodic KV Cache Consolidation for Generalised Reasoning
• BP-Seg: A graphical model approach to unsupervised and non-contiguous text segmentation using belief propagation
• From Tokens to Thoughts: How LLMs and Humans Trade Compression for Meaning
• FFT-based Dynamic Subspace Selection for Low-Rank Adaptive Optimization of Large Language Models
• Can LLMs Alleviate Catastrophic Forgetting in Graph Continual Learning? A Systematic Study
• HD-PiSSA: High-Rank Distributed Orthogonal Adaptation
• Prompting is not Enough: Exploring Knowledge Integration and Controllable Generation on Large Language Models
• Beyond the Proxy: Trajectory-Distilled Guidance for Offline GFlowNet Training
• BiomedSQL: Text-to-SQL for Scientific Reasoning on Biomedical Knowledge Bases
• Spectral-inspired Operator Learning with Limited Data and Unknown Physics
• Large Language Models versus Classical Machine Learning: Performance in COVID-19 Mortality Prediction Using High-Dimensional Tabular Data
• Closed-Form Interpretation of Neural Network Latent Spaces with Symbolic Gradients
• On the Within-class Variation Issue in Alzheimer's Disease Detection
• DOTA: Distributional Test-Time Adaptation of Vision-Language Models
• Leveraging Model Guidance to Extract Training Data from Personalized Diffusion Models
• Degree-Conscious Spiking Graph for Cross-Domain Adaptation
• Stuffed Mamba: Oversized States Lead to the Inability to Forget
• Improving the Language Understanding Capabilities of Large Language Models Using Reinforcement Learning
• Diffusion Curriculum: Synthetic-to-Real Data Curriculum via Image-Guided Diffusion
• Capacity-Aware Planning and Scheduling in Budget-Constrained Multi-Agent MDPs: A Meta-RL Approach
• Large Pre-Training Datasets Don't Always Guarantee Robustness after Fine-Tuning
• Rare-to-Frequent: Unlocking Compositional Generation Power of Diffusion Models on Rare Concepts with LLM Guidance
• Conditional Latent Space Molecular Scaffold Optimization for Accelerated Molecular Design
• Can LLMs be Good Graph Judge for Knowledge Graph Construction?
• Training-Free Bayesianization for Low-Rank Adapters of Large Language Models
• Demystifying Domain-adaptive Post-training for Financial LLMs
• How Strategic Agents Respond: Comparing Analytical Models with LLM-Generated Responses in Strategic Classification
• Avoiding $\mathbf{exp(R_{max})}$ scaling in RLHF through Preference-based Exploration
• Process Reinforcement through Implicit Rewards
• VidCRAFT3: Camera, Object, and Lighting Control for Image-to-Video Generation
• Beyond Shallow Behavior: Task-Efficient Value-Based Multi-Task Offline MARL via Skill Discovery
• Tokenizing Single-Channel EEG with Time-Frequency Motif Learning
• RuCCoD: Towards Automated ICD Coding in Russian
• How LLMs Fail to Support Fact-Checking
• Adaptively profiling models with task elicitation
• InftyThink: Breaking the Length Limits of Long-Context Reasoning in Large Language Models
• Cost-Optimal Grouped-Query Attention for Long-Context Modeling
• Retrieval-Augmented Generation with Hierarchical Knowledge
• Ethical AI for Young Digital Citizens: A Call to Action on Privacy Governance
• Detecting Scarce and Sparse Anomalous: Solving Dual Imbalance in Multi-Instance Learning
• Can Diffusion Models Disentangle? A Theoretical Perspective
• Recursive Training Loops in LLMs: How training data properties modulate distribution shift in generated data?
• Prima.cpp: Fast 30-70B LLM Inference on Heterogeneous and Low-Resource Home Clusters
• CLASH: Evaluating Language Models on Judging High-Stakes Dilemmas from Multiple Perspectives
• CapRL: Stimulating Dense Image Caption Capabilities via Reinforcement Learning
• Toward a Physics of Deep Learning and Brains
• VoiceAssistant-Eval: Benchmarking AI Assistants across Listening, Speaking, and Viewing
• See, Point, Fly: A Learning-Free VLM Framework for Universal Unmanned Aerial Navigation
• A critical review of methods and challenges in large language models
• Attributing Responsibility in AI-Induced Incidents: A Computational Reflective Equilibrium Framework for Accountability
• Development and Validation of a Large Language Model for Generating Fully-Structured Radiology Reports
• Grounding Multimodal LLMs to Embodied Agents that Ask for Help with Reinforcement Learning
• A Domain-Agnostic Scalable AI Safety Ensuring Framework
• Reasoning BO: Enhancing Bayesian Optimization with Long-Context Reasoning Power of LLMs
• From Grunts to Lexicons: Emergent Language from Cooperative Foraging
• The First Impression Problem: Internal Bias Triggers Overthinking in Reasoning Models
• XBOUND: Exploring Capability Boundaries of Device-Control Agents at the State Level
• Darwin Godel Machine: Open-Ended Evolution of Self-Improving Agents
• Scalable In-Context Q-Learning
• Mitigating Hallucination Through Theory-Consistent Symmetric Multimodal Preference Optimization
• AgentOrchestra: Orchestrating Hierarchical Multi-Agent Intelligence with the Tool-Environment-Agent(TEA) Protocol
• LocationReasoner: Evaluating LLMs on Real-World Site Selection Reasoning
• From Roots to Rewards: Dynamic Tree Reasoning with Reinforcement Learning
• MMSearch-Plus: Benchmarking Provenance-Aware Search for Multimodal Browsing Agents
• Towards Agentic OS: An LLM Agent Framework for Linux Schedulers
• EigenBench: A Comparative Behavioral Measure of Value Alignment
• The Anatomy of Alignment: Decomposing Preference Optimization by Steering Sparse Features
• The STAR-XAI Protocol: A Framework for Inducing and Verifying Agency, Reasoning, and Reliability in AI Agents
• Multi-View Hypercomplex Learning for Breast Cancer Screening
• Efficient Epistemic Uncertainty Estimation in Regression Ensemble Models Using Pairwise-Distance Estimators
• VDFD: Multi-Agent Value Decomposition Framework with Disentangled World Model
• Biospheric AI
• Machine Learning-Assisted Sustainable Remanufacturing, Reusing and Recycling for Lithium-ion Batteries
• Diverse Subset Selection via Norm-Based Sampling and Orthogonality
• VeriFlow: Modeling Distributions for Neural Network Verification
• CHRONOBERG: Capturing Language Evolution and Temporal Awareness in Foundation Models
• What Is The Political Content in LLMs' Pre- and Post-Training Data?
• Zero-Effort Image-to-Music Generation: An Interpretable RAG-based VLM Approach
• SpinGPT: A Large-Language-Model Approach to Playing Poker Correctly
• Deep Learning-Based Cross-Anatomy CT Synthesis Using Adapted nnResU-Net with Anatomical Feature Prioritized Loss
• RAU: Reference-based Anatomical Understanding with Vision Language Models
• Explaining multimodal LLMs via intra-modal token interactions
• Partial Parameter Updates for Efficient Distributed Training
• An Ontology for Unified Modeling of Tasks, Actions, Environments, and Capabilities in Personal Service Robotics
• Global Convergence in Neural ODEs: Impact of Activation Functions
• Chimera: Diagnosing Shortcut Learning in Visual-Language Understanding
• Learning to Ball: Composing Policies for Long-Horizon Basketball Moves
• Bridging Kolmogorov Complexity and Deep Learning: Asymptotically Optimal Description Length Objectives for Transformers
• Physics-informed GNN for medium-high voltage AC power flow with edge-aware attention and line search correction operator
• MDAR: A Multi-scene Dynamic Audio Reasoning Benchmark
• Learning the Neighborhood: Contrast-Free Multimodal Self-Supervised Molecular Graph Pretraining
• Evaluating the Limits of Large Language Models in Multilingual Legal Reasoning
• Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
• OFMU: Optimization-Driven Framework for Machine Unlearning
• A Machine Learning Pipeline for Multiple Sclerosis Biomarker Discovery: Comparing explainable AI and Traditional Statistical Approaches
• Ontological foundations for contrastive explanatory narration of robot plans
• Mental Health Impacts of AI Companions: Triangulating Social Media Quasi-Experiments, User Perspectives, and Relational Theory
• InfiR2: A Comprehensive FP8 Training Recipe for Reasoning-Enhanced Language Models
• Does AI Coaching Prepare us for Workplace Negotiations?
• ConQuER: Modular Architectures for Control and Bias Mitigation in IQP Quantum Generative Models
• Activation Function Design Sustains Plasticity in Continual Learning
• Retrieval-Augmented Guardrails for AI-Drafted Patient-Portal Messages: Error Taxonomy Construction and Large-Scale Evaluation
• From Parameters to Behavior: Unsupervised Compression of the Policy Space
• Learn the Ropes, Then Trust the Wins: Self-imitation with Progressive Exploration for Agentic Reinforcement Learning
• Quantile Advantage Estimation for Entropy-Safe Reasoning
• Vision-Language Alignment from Compressed Image Representations using 2D Gaussian Splatting
• IA2: Alignment with ICL Activations Improves Supervised Fine-Tuning
• A Theoretical Analysis of Discrete Flow Matching Generative Models
• Learning Admissible Heuristics for A*: Theory and Practice
• StateX: Enhancing RNN Recall via Post-training State Expansion
• Towards Efficient Online Exploration for Reinforcement Learning with Human Feedback
• Variational Reasoning for Language Models
• Language Models Can Learn from Verbal Feedback Without Scalar Rewards
• Death of the Novel(ty): Beyond n-Gram Novelty as a Metric for Textual Creativity
• WebGen-Agent: Enhancing Interactive Website Generation with Multi-Level Feedback and Step-Level Reinforcement Learning
• Hierarchical Representation Matching for CLIP-based Class-Incremental Learning
• Learning Human-Perceived Fakeness in AI-Generated Videos via Multimodal LLMs
• Thinking in Many Modes: How Composite Reasoning Elevates Large Language Model Performance with Limited Data
• Polysemous Language Gaussian Splatting via Matching-based Mask Lifting
• Fairness-Aware Reinforcement Learning (FAReL): A Framework for Transparent and Balanced Sequential Decision-Making
• FeatBench: Evaluating Coding Agents on Feature Implementation for Vibe Coding
• ASSESS: A Semantic and Structural Evaluation Framework for Statement Similarity
• Safety Compliance: Rethinking LLM Safety Reasoning through the Lens of Compliance
• Beyond Textual Context: Structural Graph Encoding with Adaptive Space Alignment to alleviate the hallucination of LLMs
• Secure and Efficient Access Control for Computer-Use Agents via Context Space
• Beyond Classification Accuracy: Neural-MedBench and the Need for Deeper Reasoning Benchmarks
• Wavelet-Induced Rotary Encodings: RoPE Meets Graphs
• A Global Analysis of Cyber Threats to the Energy Sector: "Currents of Conflict" from a Geopolitical Perspective
• Leveraging Large Language Models for Robot-Assisted Learning of Morphological Structures in Preschool Children with Language Vulnerabilities
• Bridging Fairness and Explainability: Can Input-Based Explanations Promote Fairness in Hate Speech Detection?
• Jailbreaking on Text-to-Video Models via Scene Splitting Strategy
• HEAPr: Hessian-based Efficient Atomic Expert Pruning in Output Space
• HiGS: History-Guided Sampling for Plug-and-Play Enhancement of Diffusion Models
• Adaptive Policy Backbone via Shared Network
• Progressive Weight Loading: Accelerating Initial Inference and Gradually Boosting Performance on Resource-Constrained Environments
• Pedestrian Attribute Recognition via Hierarchical Cross-Modality HyperGraph Learning
• Spectral Collapse Drives Loss of Plasticity in Deep Continual Learning
• Advancing Natural Language Formalization to First Order Logic with Fine-tuned LLMs
• Transformers Can Learn Connectivity in Some Graphs but Not Others
• SurvDiff: A Diffusion Model for Generating Synthetic Data in Survival Analysis
• Context and Diversity Matter: The Emergence of In-Context Learning in World Models
• Stochastic activations
• Forecasting the Future with Yesterday's Climate: Temperature Bias in AI Weather and Climate Models
• REFINE-CONTROL: A Semi-supervised Distillation Method For Conditional Image Generation
• From Long to Lean: Performance-aware and Adaptive Chain-of-Thought Compression via Multi-round Refinement
• Pushing Toward the Simplex Vertices: A Simple Remedy for Code Collapse in Smoothed Vector Quantization
• Lightweight error mitigation strategies for post-training N:M activation sparsity in LLMs
• Teaching AI to Feel: A Collaborative, Full-Body Exploration of Emotive Communication
• Efficiency Boost in Decentralized Optimization: Reimagining Neighborhood Aggregation with Minimal Overhead
• Learning Equivariant Functions via Quadratic Forms
• MimicDreamer: Aligning Human and Robot Demonstrations for Scalable VLA Training
• The Outputs of Large Language Models are Meaningless
• Reversible GNS for Dissipative Fluids with Consistent Bidirectional Dynamics
• Question-Driven Analysis and Synthesis: Building Interpretable Thematic Trees with LLMs for Text Clustering and Controllable Generation
• Impact of Collective Behaviors of Autonomous Vehicles on Urban Traffic Dynamics: A Multi-Agent Reinforcement Learning Approach
• VizGen: Data Exploration and Visualization from Natural Language via a Multi-Agent AI Architecture
• Automatic Discovery of One Parameter Subgroups of $SO(n)$
• Rigidity-Aware 3D Gaussian Deformation from a Single Image
• SAGE: Scene Graph-Aware Guidance and Execution for Long-Horizon Manipulation Tasks
• Why Chain of Thought Fails in Clinical Text Understanding
• SemanticControl: A Training-Free Approach for Handling Loosely Aligned Visual Conditions in ControlNet
• Unveiling Many Faces of Surrogate Models for Configuration Tuning: A Fitness Landscape Analysis Perspective
• Debiasing Large Language Models in Thai Political Stance Detection via Counterfactual Calibration
• Active Attacks: Red-teaming LLMs via Adaptive Environments
• FlowDrive: moderated flow matching with data balancing for trajectory planning
• No-Reference Image Contrast Assessment with Customized EfficientNet-B0
• From Superficial Outputs to Superficial Learning: Risks of Large Language Models in Education
• Geo-R1: Improving Few-Shot Geospatial Referring Expression Understanding with Reinforcement Fine-Tuning
• Benchmarking and Mitigate Psychological Sycophancy in Medical Vision-Language Models
• Hybrid Diffusion for Simultaneous Symbolic and Continuous Planning
• Developing Vision-Language-Action Model from Egocentric Videos
• ERGO: Efficient High-Resolution Visual Understanding for Vision-Language Models
• Black-Box Hallucination Detection via Consistency Under the Uncertain Expression
• Lightweight Structured Multimodal Reasoning for Clinical Scene Understanding in Robotics
• Latent Diffusion : Multi-Dimension Stable Diffusion Latent Space Explorer
• Fuzzy Reasoning Chain (FRC): An Innovative Reasoning Framework from Fuzziness to Clarity
• An Adaptive ICP LiDAR Odometry Based on Reliable Initial Pose
• Decoding Deception: Understanding Automatic Speech Recognition Vulnerabilities in Evasion and Poisoning Attacks
• The QCET Taxonomy of Standard Quality Criterion Names and Definitions for the Evaluation of NLP Systems
• The Rogue Scalpel: Activation Steering Compromises LLM Safety
• Action-aware Dynamic Pruning for Efficient Vision-Language-Action Manipulation
• SecureAgentBench: Benchmarking Secure Code Generation under Realistic Vulnerability Scenarios
• Reinforcement Learning for Durable Algorithmic Recourse
• Learning More with Less: A Dynamic Dual-Level Down-Sampling Framework for Efficient Policy Optimization
• The AI_INFN Platform: Artificial Intelligence Development in the Cloud
• Universal Legal Article Prediction via Tight Collaboration between Supervised Classification Model and LLM
• Multi-Agent Path Finding via Offline RL and LLM Collaboration
• R-Capsule: Compressing High-Level Plans for Efficient Large Language Model Reasoning
• Bridging Draft Policy Misalignment: Group Tree Optimization for Speculative Decoding
• DIM: Enforcing Domain-Informed Monotonicity in Deep Neural Networks
• MORPH: Shape-agnostic PDE Foundation Models
• SlotFM: A Motion Foundation Model with Slot Attention for Diverse Downstream Tasks
• QueryGym: Step-by-Step Interaction with Relational Databases
• Optimizing the non-Clifford-count in unitary synthesis using Reinforcement Learning
• Not My Agent, Not My Boundary? Elicitation of Personal Privacy Boundaries in AI-Delegated Information Sharing
• Developing Strategies to Increase Capacity in AI Education
• UISim: An Interactive Image-Based UI Simulator for Dynamic Mobile Environments
• Uncovering Alzheimer's Disease Progression via SDE-based Spatio-Temporal Graph Deep Learning on Longitudinal Brain Networks
• POLO: Preference-Guided Multi-Turn Reinforcement Learning for Lead Optimization
• LFA-Net: A Lightweight Network with LiteFusion Attention for Retinal Vessel Segmentation
• Self-Speculative Biased Decoding for Faster Live Translation
• Brain PathoGraph Learning
• HyperCore: Coreset Selection under Noise via Hypersphere Models
• SubZeroCore: A Submodular Approach with Zero Training for Coreset Selection
• Backdoor Attribution: Elucidating and Controlling Backdoor in Language Models
• Beyond Structure: Invariant Crystal Property Prediction with Pseudo-Particle Ray Diffraction
• Unbiased Binning: Fairness-aware Attribute Representation
• FastGRPO: Accelerating Policy Optimization via Concurrency-aware Speculative Decoding and Online Draft Learning
• Evaluating and Improving Cultural Awareness of Reward Models for LLM Alignment
• ChaosNexus: A Foundation Model for Universal Chaotic System Forecasting with Multi-scale Representations
• DiTraj: training-free trajectory control for video diffusion transformer
• Can Large Language Models Autoformalize Kinematics?
• Beyond Johnson-Lindenstrauss: Uniform Bounds for Sketched Bilinear Forms
• Graph of Agents: Principled Long Context Modeling by Emergent Multi-Agent Collaboration
• Enhancing Low-Rank Adaptation with Structured Nonlinear Transformations
• Unlocking the Essence of Beauty: Advanced Aesthetic Reasoning with Relative-Absolute Policy Optimization
• No Prompt Left Behind: Exploiting Zero-Variance Prompts in LLM Reinforcement Learning via Entropy-Guided Advantage Shaping
• Position: The Hidden Costs and Measurement Gaps of Reinforcement Learning with Verifiable Rewards
• You Can't Steal Nothing: Mitigating Prompt Leakages in LLMs via System Vectors
• Elastic MoE: Unlocking the Inference-Time Scalability of Mixture-of-Experts
• A Large-Scale Dataset and Citation Intent Classification in Turkish with LLMs
• AutoSCORE: Enhancing Automated Scoring with Multi-Agent Large Language Models via Structured Component Recognition
• EqDiff-CT: Equivariant Conditional Diffusion model for CT Image Synthesis from CBCT
• Generation Properties of Stochastic Interpolation under Finite Training Set
• One Model, Many Morals: Uncovering Cross-Linguistic Misalignments in Computational Moral Reasoning
• ARTI-6: Towards Six-dimensional Articulatory Speech Encoding
• A State-of-the-Art SQL Reasoning Model using RLVR
• Enhanced Generative Machine Listener
• Gender Stereotypes in Professional Roles Among Saudis: An Analytical Study of AI-Generated Images Using Language Models
• Score-based Idempotent Distillation of Diffusion Models
• Are Hallucinations Bad Estimations?
• Learning to Reason with Mixture of Tokens
• Neural Operators for Mathematical Modeling of Transient Fluid Flow in Subsurface Reservoir Systems
• Dual-Head Reasoning Distillation: Improving Classifier Accuracy with Train-Time-Only Reasoning
• Chasing the Tail: Effective Rubric-based Reward Modeling for Large Language Model Post-Training
• New Algorithmic Directions in Optimal Transport and Applications for Product Spaces
• DistillKac: Few-Step Image Generation via Damped Wave Equations
• $\mathbf{Li_2}$: A Framework on Dynamics of Feature Emergence and Delayed Generalization
• Shortcut Flow Matching for Speech Enhancement: Step-Invariant flows via single stage training
• Preemptive Detection and Steering of LLM Misalignment via Latent Reachability
• Agribot: agriculture-specific question answer system
• Psychological and behavioural responses in human-agent vs. human-human interactions: a systematic review and meta-analysis
• Domain-Aware Speaker Diarization On African-Accented English
• No Alignment Needed for Generation: Learning Linearly Separable Representations in Diffusion Models
• Enhancing Contrastive Learning for Geolocalization by Discovering Hard Negatives on Semivariograms
• What Happens Next? Anticipating Future Motion by Generating Point Trajectories
• Temporal vs. Spatial: Comparing DINOv3 and V-JEPA2 Feature Representations for Video Action Analysis
• Multi-Objective Reinforcement Learning for Large Language Model Optimization: Visionary Perspective
• LANCE: Low Rank Activation Compression for Efficient On-Device Continual Learning
• OjaKV: Context-Aware Online Low-Rank KV Cache Compression with Oja's Rule
• Guiding Audio Editing with Audio Language Model
• A Data-driven Typology of Vision Models from Integrated Representational Metrics
• InvBench: Can LLMs Accelerate Program Verification with Invariant Synthesis?
• MobiLLM: An Agentic AI Framework for Closed-Loop Threat Mitigation in 6G Open RANs
• Limitations on Safe, Trusted, Artificial General Intelligence
• Logic of Hypotheses: from Zero to Full Knowledge in Neurosymbolic Integration
• StepORLM: A Self-Evolving Framework With Generative Process Supervision For Operations Research Language Models
• UniMIC: Token-Based Multimodal Interactive Coding for Human-AI Collaboration
• Dynamic Experts Search: Enhancing Reasoning in Mixture-of-Experts LLMs at Test Time
• Benefits and Pitfalls of Reinforcement Learning for Language Model Planning: A Theoretical Perspective
• From Search to Reasoning: A Five-Level RAG Capability Framework for Enterprise Data
• PIR-RAG: A System for Private Information Retrieval in Retrieval-Augmented Generation
• Assessment of deep learning models integrated with weather and environmental variables for wildfire spread prediction and a case study of the 2023 Maui fires
• Seismic Velocity Inversion from Multi-Source Shot Gathers Using Deep Segmentation Networks: Benchmarking U-Net Variants and SeismoLabV3+
• Cross-Modal Retrieval with Cauchy-Schwarz Divergence
• Cycle is All You Need: More Is Different
• From Embeddings to Equations: Genetic-Programming Surrogates for Interpretable Transformer Classification
• SGNNBench: A Holistic Evaluation of Spiking Graph Neural Network on Large-scale Graph
• Random Direct Preference Optimization for Radiography Report Generation
• KV-Efficient VLA: A Method of Speed up Vision Language Model with RNN-Gated Chunked KV Cache
• Domain-Informed Genetic Superposition Programming: A Case Study on SFRC Beams
• Phrase-grounded Fact-checking for Automatically Generated Chest X-ray Reports
• A Novel Differential Feature Learning for Effective Hallucination Detection and Classification
• MDF-MLLM: Deep Fusion Through Cross-Modal Feature Alignment for Contextually Aware Fundoscopic Image Classification
• Influence Guided Context Selection for Effective Retrieval-Augmented Generation
• Multimodal Prompt Decoupling Attack on the Safety Filters in Text-to-Image Models
• Context Is What You Need: The Maximum Effective Context Window for Real World Limits of LLMs
• A Mutual Learning Method for Salient Object Detection with intertwined Multi-Supervision--Revised
• MAJORScore: A Novel Metric for Evaluating Multimodal Relevance via Joint Representation
• Design and Implementation of a Secure RAG-Enhanced AI Chatbot for Smart Tourism Customer Service: Defending Against Prompt Injection Attacks -- A Case Study of Hsinchu, Taiwan
• Safety Assessment of Scaffolding on Construction Site using AI
• ReGeS: Reciprocal Retrieval-Generation Synergy for Conversational Recommender Systems
• Automated Prompt Generation for Creative and Counterfactual Text-to-image Synthesis
• In silico Deep Learning Protocols for Label-Free Super-Resolution Microscopy: A Comparative Study of Network Architectures and SNR Dependence
• Dynamic Multi-Target Fusion for Efficient Audio-Visual Navigation
• SAEmnesia: Erasing Concepts in Diffusion Models with Sparse Autoencoders
• Toward a Realistic Encoding Model of Auditory Affective Understanding in the Brain
• Do Sparse Subnetworks Exhibit Cognitively Aligned Attention? Effects of Pruning on Saliency Map Fidelity, Sparsity, and Concept Coherence
• Towards Adapting Federated & Quantum Machine Learning for Network Intrusion Detection: A Survey
• MIXRAG : Mixture-of-Experts Retrieval-Augmented Generation for Textual Graph Understanding and Question Answering
• Large AI Model-Enabled Generative Semantic Communications for Image Transmission
• How Large Language Models Need Symbolism
• Near-Optimal Experiment Design in Linear non-Gaussian Cyclic Models
• PhenoMoler: Phenotype-Guided Molecular Optimization via Chemistry Large Language Model
• DyME: Dynamic Multi-Concept Erasure in Diffusion Models with Bi-Level Orthogonal LoRA Adaptation
• Foundation models for high-energy physics
• Can AI Perceive Physical Danger and Intervene?
• Align2Speak: Improving TTS for Low Resource Languages via ASR-Guided Online Preference Optimization
• Retrieval-of-Thought: Efficient Reasoning via Reusing Thoughts
• Lifelong Learning with Behavior Consolidation for Vehicle Routing
• UltraHorizon: Benchmarking Agent Capabilities in Ultra Long-Horizon Scenarios
• Benchmarking MLLM-based Web Understanding: Reasoning, Robustness and Safety
• D-Artemis: A Deliberative Cognitive Framework for Mobile GUI Multi-Agents
• ProRe: A Proactive Reward System for GUI Agents via Reasoner-Actor Collaboration
• DS-STAR: Data Science Agent via Iterative Planning and Verification
• Axiomatic Choice and the Decision-Evaluation Paradox
• DeepTravel: An End-to-End Agentic Reinforcement Learning Framework for Autonomous Travel Planning Agents
• Reimagining Agent-based Modeling with Large Language Model Agents via Shachi
• TRACE: Learning to Compute on Graphs
• GenesisGeo: Technical Report
• DyRo-MCTS: A Robust Monte Carlo Tree Search Approach to Dynamic Job Shop Scheduling
• Outlier Detection in Plantar Pressure: Human-Centered Comparison of Statistical Parametric Mapping and Explainable Machine Learning
• CoBel-World: Harnessing LLM Reasoning to Build a Collaborative Belief World for Optimizing Embodied Multi-Agent Collaboration
• RISK: A Framework for GUI Agents in E-commerce Risk Management
• Bilinear relational structure fixes reversal curse and enables consistent model editing
• GSM-Agent: Understanding Agentic Reasoning Using Controllable Environments
• The Thinking Spectrum: An Emperical Study of Tunable Reasoning in LLMs through Model Merging
• A2R: An Asymmetric Two-Stage Reasoning Framework for Parallel Reasoning
• Generalizing Multi-Objective Search via Objective-Aggregation Functions
• Ground-Truthing AI Energy Consumption: Validating CodeCarbon Against External Measurements
• Log2Plan: An Adaptive GUI Automation Framework Integrated with Task Mining Approach
• Clinical Uncertainty Impacts Machine Learning Evaluations
• Evaluating LLMs for Combinatorial Optimization: One-Phase and Two-Phase Heuristics for 2D Bin-Packing
• InfiMed-Foundation: Pioneering Advanced Multimodal Medical Models with Compute-Efficient Pre-Training and Multi-Stage Fine-Tuning
• Structured Sparse Transition Matrices to Enable State Tracking in State-Space Models
• Large Language Models as Nondeterministic Causal Models
• PRIME: Planning and Retrieval-Integrated Memory for Enhanced Reasoning
• Do LLM Agents Know How to Ground, Recover, and Assess? A Benchmark for Epistemic Competence in Information-Seeking Agents
• EMMA: Generalizing Real-World Robot Manipulation via Generative Visual Transfer
• Guiding Evolution of Artificial Life Using Vision-Language Models
• GeoSketch: A Neural-Symbolic Approach to Geometric Multimodal Reasoning with Auxiliary Line Construction and Affine Transformation
• InfiAgent: Self-Evolving Pyramid Agent Framework for Infinite Scenarios
• Estimating the Empowerment of Language Model Agents
• TrueGradeAI: Retrieval-Augmented and Bias-Resistant AI for Transparent and Explainable Digital Assessments
• REMA: A Unified Reasoning Manifold Framework for Interpreting Large Language Model
• The Emergence of Altruism in Large-Language-Model Agents Society
• Towards mitigating information leakage when evaluating safety monitors
• Correct Reasoning Paths Visit Shared Decision Pivots
• AutoClimDS: Climate Data Science Agentic AI -- A Knowledge Graph is All You Need
• EEG-Based Consumer Behaviour Prediction: An Exploration from Classical Machine Learning to Graph Neural Networks
• GeoEvolve: Automating Geospatial Model Discovery via Multi-Agent Large Language Models
• Automated and Interpretable Survival Analysis from Multimodal Data
• Semantic F1 Scores: Fair Evaluation Under Fuzzy Class Boundaries