AI Research News Feeds for September 26th, 2025

AI RESEARCH PAPERS & ACADEMIC SOURCES

• Frequency-Compensated Network for Daily Arctic Sea Ice Concentration Prediction
• Buffer-free Class-Incremental Learning with Out-of-Distribution Detection
• TempSamp-R1: Effective Temporal Sampling with Reinforcement Fine-Tuning for Video LLMs
• AnyPlace: Learning Generalized Object Placement for Robot Manipulation
• Benchmarking for Practice: Few-Shot Time-Series Crop-Type Classification on the EuroCropsML Dataset
• Estimating Deep Learning energy consumption based on model architecture and training environment
• Reinforcement Learning in Categorical Cybernetics
• FoMo-0D: A Foundation Model for Zero-shot Tabular Outlier Detection
• Expressiveness of Multi-Neuron Convex Relaxations in Neural Network Certification
• TReMu: Towards Neuro-Symbolic Temporal Reasoning for LLM-Agents with Memory in Multi-Session Dialogues
• The Value of Information in Human-AI Decision-making
• RollPacker: Mitigating Long-Tail Rollouts for Fast, Synchronous RL Post-Training
• MPC-based Deep Reinforcement Learning Method for Space Robotic Control with Fuel Sloshing Mitigation
• Physics Informed Neural Networks for design optimisation of diamond particle detectors for charged particle fast-tracking at high luminosity hadron colliders
• IntSR: An Integrated Generative Framework for Search and Recommendation
• Data-driven Neural Networks for Windkessel Parameter Calibration
• The Use of the Simplex Architecture to Enhance Safety in Deep-Learning-Powered Autonomous Systems
• Fine-Tuning LLMs to Analyze Multiple Dimensions of Code Review: A Maximum Entropy Regulated Long Chain-of-Thought Approach
• Adoption, usability and perceived clinical value of a UK AI clinical reference platform (iatroX): a mixed-methods formative evaluation of real-world usage and a 1,223-respondent user survey
• Semantic Edge-Cloud Communication for Real-Time Urban Traffic Surveillance with ViT and LLMs over Mobile Networks
• Data-Centric Elastic Pipeline Parallelism for Efficient Long-Context LLM Training
• A Decision Theoretic Framework for Measuring AI Reliance
• Text-Augmented Multimodal LLMs for Chemical Reaction Condition Recommendation
• Examining the Prevalence and Dynamics of AI-Generated Media in Art Subreddits
• The Asymptotic Behavior of Attention in Transformers
• i-LAVA: Insights on Low Latency Voice-2-Voice Architecture for Agents
• Dual-Path Phishing Detection: Integrating Transformer-Based NLP with Structural URL Analysis
• AnywhereVLA: Language-Conditioned Exploration and Mobile Manipulation
• Automatic Red Teaming LLM-based Agents with Model Context Protocol Tools
• Data-Efficient Time-Dependent PDE Surrogates: Graph Neural Simulators vs. Neural Operators
• A Finite-Time Analysis of TD Learning with Linear Function Approximation without Projections or Strong Convexity
• Improved Scaling Laws in Linear Regression via Data Reuse
• Generalizing while preserving monotonicity in comparison-based preference learning models
• CopulaSMOTE: A Copula-Based Oversampling Approach for Imbalanced Classification in Diabetes Prediction
• An entropy-optimal path to humble AI
• Enhanced Generative Model Evaluation with Clipped Density and Coverage
• Training-Free Stein Diffusion Guidance: Posterior Correction for Sampling Beyond High-Density Regions
• Contextual Combinatorial Bandits with Changing Action Sets via Gaussian Processes
• Rosenthal-type inequalities for linear statistics of Markov chains
• Energy based diffusion generator for efficient sampling of Boltzmann distributions
• Understanding Optimization in Deep Learning with Central Flows
• Bayesian Optimization with Preference Exploration using a Monotonic Neural Network Ensemble
• Regularization can make diffusion models more efficient
• Provably Sample-Efficient Robust Reinforcement Learning with Average Reward
• Why and When Deep is Better than Shallow: An Implementation-Agnostic State-Transition View of Depth Supremacy
• Bridging Arbitrary and Tree Metrics via Differentiable Gromov Hyperbolicity
• Response to Promises and Pitfalls of Deep Kernel Learning
• Incorporating External Controls for Estimating the Average Treatment Effect on the Treated with High-Dimensional Data: Retaining Double Robustness and Ensuring Double Safety
• Towards Complete Causal Explanation with Expert Knowledge
• Hybrid Summary Statistics
• Revenue Maximization Under Sequential Price Competition Via The Estimation Of s-Concave Demand Functions
• Uncertainty-Aware Surrogate-based Amortized Bayesian Inference for Computationally Expensive Models
• Tensor State Space-based Dynamic Multilayer Network Modeling
• Learning to Bid Optimally and Efficiently in Adversarial First-price Auctions
• Empirical PAC-Bayes bounds for Markov chains
• Best-of-$\infty$ -- Asymptotic Performance of Test-Time Compute
• WISER: Segmenting watermarked region - an epidemic change-point perspective
• Breaking the curse of dimensionality for linear rules: optimal predictors over the ellipsoid
• Why Settle for One? Text-to-ImageSet Generation and Evaluation
• 3D-MoRe: Unified Modal-Contextual Reasoning for Embodied Question Answering
• GeMix: Conditional GAN-Based Mixup for Improved Medical Image Augmentation
• CNS-Bench: Benchmarking Image Classifier Robustness Under Continuous Nuisance Shifts
• RIS-LAD: A Benchmark and Model for Referring Low-Altitude Drone Image Segmentation
• VideoPASTA: 7K Preference Pairs That Matter for Video-LLM Alignment
• Learning Flow-Guided Registration for RGB-Event Semantic Segmentation
• Automated Visual Attention Detection using Mobile Eye Tracking in Behavioral Classroom Studies
• Instance-aware Image Colorization with Controllable Textual Descriptions and Segmentation Masks
• CONSIGN: Conformal Segmentation Informed by Spatial Groupings via Decomposition
• MMaDA: Multimodal Large Diffusion Language Models
• MME-VideoOCR: Evaluating OCR-Based Capabilities of Multimodal LLMs in Video Scenarios
• Beyond Quantity: Distribution-Aware Labeling for Visual Grounding
• ProstaTD: Bridging Surgical Triplet from Classification to Fully Supervised Detection
• O-MaMa: Learning Object Mask Matching between Egocentric and Exocentric Views
• LadderMIL: Multiple Instance Learning with Coarse-to-Fine Self-Distillation
• Efficiently Disentangling CLIP for Multi-Object Perception
• HermesFlow: Seamlessly Closing the Gap in Multimodal Understanding and Generation
• Diff-Reg v2: Diffusion-Based Matching Matrix Estimation for Image Matching and 3D Registration
• REArtGS: Reconstructing and Generating Articulated Objects via 3D Gaussian Splatting with Geometric and Motion Constraints
• LLaVA-RadZ: Can Multimodal Large Language Models Effectively Tackle Zero-shot Radiology Recognition?
• Parameter-Efficient Adaptation of Geospatial Foundation Models through Embedding Deflection
• Autoregressive Image Generation with Randomized Parallel Decoding
• Radar-Guided Polynomial Fitting for Metric Depth Estimation
• Improving Brain Disorder Diagnosis with Advanced Brain Function Representation and Kolmogorov-Arnold Networks
• Vim-F: Visual State Space Model Benefiting from Learning in the Frequency Domain
• Model Agnostic Defense against Adversarial Patch Attacks on Object Detection in Unmanned Aerial Vehicles
• SOOD++: Leveraging Unlabeled Data to Boost Oriented Object Detection
• Lightweight Modular Parameter-Efficient Tuning for Open-Vocabulary Object Detection
• Asynchronous Perception Machine For Efficient Test-Time-Training
• Training-Free Layout-to-Image Generation with Marginal Attention Constraints
• Supercharged One-step Text-to-Image Diffusion Models with Negative Prompts
• GVDepth: Zero-Shot Monocular Depth Estimation for Ground Vehicles based on Probabilistic Cue Fusion
• MonSter++: Unified Stereo Matching, Multi-view Stereo, and Real-time Stereo with Monodepth Priors
• Technical report on label-informed logit redistribution for better domain generalization in low-shot classification with foundation models
• MASt3R-Fusion: Integrating Feed-Forward Visual Model with IMU, GNSS for High-Functionality SLAM
• ARMesh: Autoregressive Mesh Generation via Next-Level-of-Detail Prediction
• ArchGPT: Understanding the World's Architectures with Large Multimodal Models
• Autoregressive End-to-End Planning with Time-Invariant Spatial Alignment and Multi-Objective Policy Refinement
• Marching Neurons: Accurate Surface Extraction for Neural Implicit Shapes
• KeyWorld: Key Frame Reasoning Enables Effective and Efficient World Models
• Cross-Modal Instructions for Robot Motion Generation
• CHARM: Control-point-based 3D Anime Hairstyle Auto-Regressive Modeling
• Human-like Navigation in a World Built for Humans
• SuperPatchMatch: an Algorithm for Robust Correspondences using Superpixel Patches
• Retina Vision Transformer (RetinaViT): Introducing Scaled Patches into Vision Transformers
• Quantized Visual Geometry Grounded Transformer
• NewtonGen: Physics-Consistent and Controllable Text-to-Video Generation via Neural Newtonian Dynamics
• SD3.5-Flash: Distribution-Guided Distillation of Generative Flows
• BlockFUL: Enabling Unlearning in Blockchained Federated Learning
• Optimal Transport Based Hyperspectral Unmixing for Highly Mixed Observations
• Equi-RO: A 4D mmWave Radar Odometry via Equivariant Networks
• RAM-NAS: Resource-aware Multiobjective Neural Architecture Search Method for Robot Vision Tasks
• ArtUV: Artist-style UV Unwrapping
• SeamCrafte: Enhancing Mesh Seam Generation for Artist UV Unwrapping via Reinforcement Learning
• SLAM-Free Visual Navigation with Hierarchical Vision-Language Perception and Coarse-to-Fine Semantic Topological Planning
• SlideMamba: Entropy-Based Adaptive Fusion of GNN and Mamba for Enhanced Representation Learning in Digital Pathology
• Hunyuan3D-Omni: A Unified Framework for Controllable Generation of 3D Assets
• Learning to Look: Cognitive Attention Alignment with Vision-Language Models
• Decipher-MR: A Vision-Language Foundation Model for 3D MRI Representations
• Instruction-tuned Self-Questioning Framework for Multimodal Reasoning
• Every Subtlety Counts: Fine-grained Person Independence Micro-Action Recognition via Distributionally Robust Optimization
• Dense Semantic Matching with VGGT Prior
• MedVSR: Medical Video Super-Resolution with Cross State-Space Propagation
• MMR1: Enhancing Multimodal Reasoning with Variance-Aware Sampling and Open Resources
• A Sentinel-3 foundation model for ocean colour
• Does FLUX Already Know How to Perform Physically Plausible Image Composition?
• Vision Transformers: the threat of realistic adversarial patches
• UniTransfer: Video Concept Transfer via Progressive Spatial and Timestep Decomposition
• VideoChat-R1.5: Visual Test-Time Scaling to Reinforce Multimodal Reasoning by Iterative Perception
• Mammo-CLIP Dissect: A Framework for Analysing Mammography Concepts in Vision-Language Models
• MOSS-ChatV: Reinforcement Learning with Process Reasoning Reward for Video Temporal Reasoning
• MotionFlow:Learning Implicit Motion Flow for Complex Camera Trajectory Control in Video Generation
• The Unwinnable Arms Race of AI Image Detection
• WAVECLIP: Wavelet Tokenization for Adaptive-Resolution CLIP
• Can Less Precise Be More Reliable? A Systematic Evaluation of Quantization's Impact on CLIP Beyond Accuracy
• Learning Conformal Explainers for Image Classifiers
• Unlocking Noise-Resistant Vision: Key Architectural Secrets for Robust Models
• Decoding the Surgical Scene: A Scoping Review of Scene Graphs in Surgery
• A Real-Time On-Device Defect Detection Framework for Laser Power-Meter Sensors via Unsupervised Learning
• An Adaptor for Triggering Semi-Supervised Learning to Out-of-Box Serve Deep Image Clustering
• SiNGER: A Clearer Voice Distills Vision Transformers Further
• Fast-SEnSeI: Lightweight Sensor-Independent Cloud Masking for On-board Multispectral Sensors
• A Single Neuron Works: Precise Concept Erasure in Text-to-Image Diffusion Models
• OmniPlantSeg: Species Agnostic 3D Point Cloud Organ Segmentation for High-Resolution Plant Phenotyping Across Modalities
• Background Prompt for Few-Shot Out-of-Distribution Detection
• Stratify or Die: Rethinking Data Splits in Image Segmentation
• EnGraf-Net: Multiple Granularity Branch Network with Fine-Coarse Graft Grained for Classification Task
• Plant identification based on noisy web data: the amazing performance of deep learning (LifeCLEF 2017)
• SD-RetinaNet: Topologically Constrained Semi-Supervised Retinal Lesion and Layer Segmentation in OCT
• Plant identification in an open-world (LifeCLEF 2016)
• The Unanticipated Asymmetry Between Perceptual Optimization and Assessment
• Concepts in Motion: Temporal Bottlenecks for Interpretable Video Classification
• FSMODNet: A Closer Look at Few-Shot Detection in Multispectral Data
• Finding 3D Positions of Distant Objects from Noisy Camera Movement and Semantic Segmentation Sequences
• SwinMamba: A hybrid local-global mamba framework for enhancing semantic segmentation of remotely sensed images
• Revisiting Data Challenges of Computational Pathology: A Pack-based Multiple Instance Learning Framework
• SimDiff: Simulator-constrained Diffusion Model for Physically Plausible Motion Generation
• Recov-Vision: Linking Street View Imagery and Vision-Language Models for Post-Disaster Recovery
• Enhancing Cross-View Geo-Localization Generalization via Global-Local Consistency and Geometric Equivariance
• DENet: Dual-Path Edge Network with Global-Local Attention for Infrared Small Target Detection
• Neptune-X: Active X-to-Maritime Generation for Universal Maritime Object Detection
• FreeInsert: Personalized Object Insertion with Geometric and Style Control
• CompressAI-Vision: Open-source software to evaluate compression methods for computer vision tasks
• Dual-supervised Asymmetric Co-training for Semi-supervised Medical Domain Generalization
• Real-Time Object Detection Meets DINOv3
• Federated Domain Generalization with Domain-specific Soft Prompts Generation
• Poisoning Prompt-Guided Sampling in Video Large Language Models
• Punching Above Precision: Small Quantized Model Distillation with Learnable Regularizer
• Quasi-Synthetic Riemannian Data Generation for Writer-Independent Offline Signature Verification
• Seedream 4.0: Toward Next-generation Multimodal Image Generation
• A Contrastive Learning Framework for Breast Cancer Detection
• Are Foundation Models Ready for Industrial Defect Recognition? A Reality Check on Real-World Data
• Data-Efficient Stream-Based Active Distillation for Scalable Edge Model Deployment
• A Comparative Benchmark of Real-time Detectors for Blueberry Detection towards Precision Orchard Management
• Reflect3r: Single-View 3D Stereo Reconstruction Aided by Mirror Reflections
• SelfBudgeter: Adaptive Token Allocation for Efficient LLM Reasoning
• MMSI-Bench: A Benchmark for Multi-Image Spatial Intelligence
• Fractional Reasoning via Latent Steering Vectors Improves Inference Time Compute
• Blending Supervised and Reinforcement Fine-Tuning with Prefix Sampling
• Reinforcement Fine-Tuning Naturally Mitigates Forgetting in Continual Post-Training
• NoHumansRequired: Autonomous High-Quality Image Editing Triplet Mining
• Can social media provide early warning of retraction? Evidence from critical tweets identified by human annotation and large language models
• TestAgent: Automatic Benchmarking and Exploratory Interaction for Evaluating LLMs in Vertical Domains
• Bias Similarity Measurement: A Black-Box Audit of Fairness Across LLMs
• Reformulation is All You Need: Addressing Malicious Text Features in DNNs
• AdaSVD: Adaptive Singular Value Decomposition for Large Language Models
• Scaling Rich Style-Prompted Text-to-Speech Datasets
• What Makes a Reward Model a Good Teacher? An Optimization Perspective
• On the Perception Bottleneck of VLMs for Chart Understanding
• A Framework for Situating Innovations, Opportunities, and Challenges in Advancing Vertical Systems with Large AI Models
• Process Reward Models That Think
• UDDETTS: Unifying Discrete and Dimensional Emotions for Controllable Emotional Text-to-Speech
• Just-in-time and distributed task representations in language models
• PLaMo 2 Technical Report
• ConsistentChat: Building Skeleton-Guided Consistent Multi-Turn Dialogues for Large Language Models from Scratch
• From Replication to Redesign: Exploring Pairwise Comparisons for LLM-Based Peer Review
• ImpliRet: Benchmarking the Implicit Fact Retrieval Challenge
• When Does Meaning Backfire? Investigating the Role of AMRs in NLI
• THCM-CAL: Temporal-Hierarchical Causal Modelling with Conformal Calibration for Clinical Risk Prediction
• A Simple "Motivation" Can Enhance Reinforcement Finetuning of Large Reasoning Models
• ReasonFlux-PRM: Trajectory-Aware PRMs for Long Chain-of-Thought Reasoning in LLMs
• ARF-RLHF: Adaptive Reward-Following for RLHF through Emotion-Driven Self-Supervision and Trace-Biased Dynamic Optimization
• ixi-GEN: Efficient Industrial sLLMs through Domain Adaptive Continual Pretraining
• Turning Internal Gap into Self-Improvement: Promoting the Generation-Understanding Unification in MLLMs
• Constructions are Revealed in Word Distributions
• Inference-Time Scaling for Generalist Reward Modeling
• Decoding Open-Ended Information Seeking Goals from Eye Movements in Reading
• Ambiguity Resolution in Text-to-Structured Data Mapping
• VerifyBench: Benchmarking Reference-based Reward Systems for Large Language Models
• UNCERTAINTY-LINE: Length-Invariant Estimation of Uncertainty for Large Language Models
• InComeS: Integrating Compression and Selection Mechanisms into LLMs for Efficient Model Editing
• Bayesian Attention Mechanism: A Probabilistic Framework for Positional Encoding and Context Length Extrapolation
• BabyLM's First Constructions: Causal probing provides a signal of learning
• Co-Evolving LLM Coder and Unit Tester via Reinforcement Learning
• Investigating Factuality in Long-Form Text Generation: The Roles of Self-Known and Self-Unknown
• LAMA-UT: Language Agnostic Multilingual ASR through Orthography Unification and Language-Specific Transliteration
• Labeling Free-text Data using Language Model Ensembles
• Improving LLM Unlearning Robustness via Random Perturbations
• Quantifying depressive mental states with large language models
• MathFimer: Enhancing Mathematical Reasoning by Expanding Reasoning Steps through Fill-in-the-Middle Task
• The Validation Gap: A Mechanistic Analysis of How Language Models Compute Arithmetic but Fail to Validate It
• JUREX-4E: Juridical Expert-Annotated Four-Element Knowledge Base for Legal Reasoning
• Collab-Overcooked: Benchmarking and Evaluating Large Language Models as Collaborative Agents
• PMark: Towards Robust and Distortion-free Semantic-level Watermarking with Channel Constraints
• Communication Bias in Large Language Models: A Regulatory Perspective
• Automotive-ENV: Benchmarking Multimodal Agents in Vehicle Interface Systems
• TABLET: A Large-Scale Dataset for Robust Visual Table Understanding
• Sigma: Semantically Informative Pre-training for Skeleton-based Sign Language Understanding
• Evaluating the Evaluators: Metrics for Compositional Text-to-Image Generation
• Hallucination as an Upper Bound: A New Perspective on Text-to-Image Evaluation
• Interactive Recommendation Agent with Active User Commands
• Higher-Order DisCoCat (Peirce-Lambek-Montague semantics)
• ASCIIEval: Benchmarking Models' Visual Perception in Text Strings via ASCII Art
• UniHR: Hierarchical Representation Learning for Unified Knowledge Graph Link Prediction
• LLMTrace: A Corpus for Classification and Fine-Grained Localization of AI-Written Text
• Bounds of Chain-of-Thought Robustness: Reasoning Steps, Embed Norms, and Beyond
• DisCoCLIP: A Distributional Compositional Tensor Network Encoder for Vision-Language Understanding
• The role of synthetic data in Multilingual, Multi-cultural AI systems: Lessons from Indic Languages
• Sycophancy Is Not One Thing: Causal Separation of Sycophantic Behaviors in LLMs
• RLBFF: Binary Flexible Feedback to bridge between Human Feedback & Verifiable Rewards
• SciReasoner: Laying the Scientific Reasoning Ground Across Disciplines
• Intercept Cancer: Cancer Pre-Screening with Large Scale Healthcare Foundation Models
• RadAgents: Multimodal Agentic Reasoning for Chest X-ray Interpretation with Radiologist-like Workflows
• Human Semantic Representations of Social Interactions from Moving Shapes
• Visual Authority and the Rhetoric of Health Misinformation: A Multimodal Analysis of Social Media Videos
• Retrieval over Classification: Integrating Relation Semantics for Multimodal Relation Extraction
• Learning the Wrong Lessons: Syntactic-Domain Spurious Correlations in Language Models
• Who's Laughing Now? An Overview of Computational Humour Generation and Explanation
• GEP: A GCG-Based method for extracting personally identifiable information from chatbots built on small language models
• Eigen-1: Adaptive Multi-Agent Refinement with Monitor-Based RAG for Scientific Reasoning
• CLaw: Benchmarking Chinese Legal Knowledge in Large Language Models - A Fine-grained Corpus and Reasoning Analysis
• SGMem: Sentence Graph Memory for Long-Term Conversational Agents
• Query-Centric Graph Retrieval Augmented Generation
• Un-Doubling Diffusion: LLM-guided Disambiguation of Homonym Duplication
• LLM Output Homogenization is Task Dependent
• Analysis of instruction-based LLMs' capabilities to score and judge text-input problems in an academic setting
• Generative AI for FFRDCs
• Behind RoPE: How Does Causal Mask Encode Positional Information?
• When Instructions Multiply: Measuring and Estimating LLM Capabilities of Multiple Instructions Following
• SoM-1K: A Thousand-Problem Benchmark Dataset for Strength of Materials
• Which Cultural Lens Do Models Adopt? On Cultural Positioning Bias and Agentic Mitigation in LLMs
• PerHalluEval: Persian Hallucination Evaluation Benchmark for Large Language Models
• BESPOKE: Benchmark for Search-Augmented Large Language Model Personalization via Diagnostic Feedback
• VoiceBBQ: Investigating Effect of Content and Acoustics in Social Bias of Spoken Language Model
• Acoustic-based Gender Differentiation in Speech-aware Language Models
• AutoIntent: AutoML for Text Classification
• SFT Doesn't Always Hurt General Capabilities: Revisiting Domain-Specific Fine-Tuning in LLMs
• Few-Shot and Training-Free Review Generation via Conversational Prompting
• Enrich-on-Graph: Query-Graph Alignment for Complex Reasoning with LLM Enriching
• Distilling Many-Shot In-Context Learning into a Cheat Sheet
• Zero-Shot Privacy-Aware Text Rewriting via Iterative Tree Search
• Concise and Sufficient Sub-Sentence Citations for Retrieval-Augmented Generation
• WeFT: Weighted Entropy-driven Fine-Tuning for dLLMs
• Learning to Summarize by Learning to Quiz: Adversarial Agentic Collaboration for Long Document Summarization
• MemLens: Uncovering Memorization in LLMs with Activation Trajectories
• Cross-Linguistic Analysis of Memory Load in Sentence Comprehension: Linear Distance and Structural Density
• Tool Calling for Arabic LLMs: Data Strategies and Instruction Tuning
• Nuclear Diffusion Models for Low-Rank Background Suppression in Videos
• Conditionally Whitened Generative Models for Probabilistic Time Series Forecasting
• ShortCheck: Checkworthiness Detection of Multilingual Short-Form Videos
• SiniticMTError: A Machine Translation Dataset with Error Annotations for Sinitic Languages
• Building Tailored Speech Recognizers for Japanese Speaking Assessment
• Enhancing Molecular Property Prediction with Knowledge from Large Language Models
• RedHerring Attack: Testing the Reliability of Attack Detection
• Overcoming Black-box Attack Inefficiency with Hybrid and Dynamic Select Algorithms
• MI-Fuse: Label Fusion for Unsupervised Domain Adaptation with Closed-Source Large-Audio Language Model
• Probability Distribution Collapse: A Critical Bottleneck to Compact Unsupervised Neural Grammar Induction
• Cryptographic Backdoor for Neural Networks: Boon and Bane
• PALQO: Physics-informed Model for Accelerating Large-scale Quantum Optimization
• Real-Time System for Audio-Visual Target Speech Enhancement
• RAPTOR-GEN: RApid PosTeriOR GENerator for Bayesian Learning in Biomanufacturing
• Identifying Group Anchors in Real-World Group Interactions Under Label Scarcity
• Leveraging Temporally Extended Behavior Sharing for Multi-task Reinforcement Learning
• Extrapolating Phase-Field Simulations in Space and Time with Purely Convolutional Architectures
• Actively Learning Halfspaces without Synthetic Data
• Single Answer is Not Enough: On Generating Ranked Lists with Medical Reasoning Models
• RecIS: Sparse to Dense, A Unified Training Framework for Recommendation Models
• Objective Evaluation of Prosody and Intelligibility in Speech Synthesis via Conditional Prediction of Discrete Tokens
• Fast Estimation of Wasserstein Distances via Regression on Sliced Wasserstein Distances
• Innovative Deep Learning Architecture for Enhanced Altered Fingerprint Recognition
• Large Pre-Trained Models for Bimanual Manipulation in 3D
• Region-of-Interest Augmentation for Mammography Classification under Patient-Level Cross-Validation
• Unsupervised Domain Adaptation with an Unobservable Source Subpopulation
• A Gapped Scale-Sensitive Dimension and Lower Bounds for Offset Rademacher Complexity
• Design, Implementation and Evaluation of a Novel Programming Language Topic Classification Workflow
• A Hierarchical Variational Graph Fused Lasso for Recovering Relative Rates in Spatial Compositional Data
• Implicit Augmentation from Distributional Symmetry in Turbulence Super-Resolution
• No Prior, No Leakage: Revisiting Reconstruction Attacks in Trained Neural Networks
• Copycats: the many lives of a publicly available medical imaging dataset
• An Analytical and AI-discovered Stable, Accurate, and Generalizable Subgrid-scale Closure for Geophysical Turbulence
• Speaker Style-Aware Phoneme Anchoring for Improved Cross-Lingual Speech Emotion Recognition
• Leveraging NTPs for Efficient Hallucination Detection in VLMs
• A Comparative Analysis of Ensemble-Based Machine Learning Approaches with Explainable AI for Multi-Class Intrusion Detection in Drone Networks
• Sample completion, structured correlation, and Netflix problems
• Structuring Collective Action with LLM-Guided Evolution: From Ill-Structured Problems to Executable Heuristics
• SceneWeaver: All-in-One 3D Scene Synthesis with an Extensible and Self-Reflective Agent
• Neural Networks as Surrogate Solvers for Time-Dependent Accretion Disk Dynamics
• Document Summarization with Conformal Importance Guarantees
• Go With The Flow: Churn-Tolerant Decentralized Training of Large Language Models
• AbideGym: Turning Static RL Worlds into Adaptive Challenges
• Tree Search for LLM Agent Reinforcement Learning
• Explaining Fine Tuned LLMs via Counterfactuals A Knowledge Graph Driven Framework
• Federated Flow Matching
• humancompatible.train: Implementing Optimization Algorithms for Stochastically-Constrained Stochastic Optimization Problems
• A Causality-Aware Spatiotemporal Model for Multi-Region and Multi-Pollutant Air Quality Forecasting
• SuperOffload: Unleashing the Power of Large-Scale LLM Training on Superchips
• It's Not You, It's Clipping: A Soft Trust-Region via Probability Smoothing for LLM RL
• Optimal Robust Recourse with $L^p$-Bounded Model Change
• LAVA: Explainability for Unsupervised Latent Embeddings
• CAD-Tokenizer: Towards Text-based CAD Prototyping via Modality-Specific Tokenization
• GRPO is Secretly a Process Reward Model
• DATS: Distance-Aware Temperature Scaling for Calibrated Class-Incremental Learning
• Mixture of Thoughts: Learning to Aggregate What Experts Think, Not Just What They Say
• A Unified Framework for Diffusion Model Unlearning with f-Divergence
• Inverse Reinforcement Learning Using Just Classification and a Few Regressions
• Closed-form $\ell_r$ norm scaling with data for overparameterized linear regression and diagonal linear networks under $\ell_p$ bias
• Towards Foundation Models for Zero-Shot Time Series Anomaly Detection: Leveraging Synthetic Data and Relative Context Discrepancy
• Differential-Integral Neural Operator for Long-Term Turbulence Forecasting
• From Physics to Machine Learning and Back: Part II - Learning and Observational Bias in PHM
• Physics of Learning: A Lagrangian perspective to different learning paradigms
• GeoRef: Referring Expressions in Geometry via Task Formulation, Synthetic Supervision, and Reinforced MLLM-based Solutions
• SPREAD: Sampling-based Pareto front Refinement via Efficient Adaptive Diffusion
• Structure-Attribute Transformations with Markov Chain Boost Graph Domain Adaptation
• ScaleDiff: Scaling Difficult Problems for Advanced Mathematical Reasoning
• TyphoonMLA: A Mixed Naive-Absorb MLA Kernel For Shared Prefix
• GraphUniverse: Enabling Systematic Evaluation of Inductive Generalization
• Teaching RL Agents to Act Better: VLM as Action Advisor for Online Reinforcement Learning
• EvoMail: Self-Evolving Cognitive Agents for Adaptive Spam and Phishing Email Defense
• Sparse Representations Improve Adversarial Robustness of Neural Network Classifiers
• Feature Augmentation of GNNs for ILPs: Local Uniqueness Suffices
• Lossless Compression: A New Benchmark for Time Series Model Evaluation
• MAIFormer: Multi-Agent Inverted Transformer for Flight Trajectory Prediction
• ExMolRL: Phenotype-Target Joint Generation of De Novo Molecules via Multi-Objective Reinforcement Learning
• Mechanism of Task-oriented Information Removal in In-context Learning
• Predicting LLM Reasoning Performance with Small Proxy Model
• DELTA-Code: How Does RL Unlock and Transfer New Programming Algorithms in LLMs?
• Efficient Ensemble Conditional Independence Test Framework for Causal Discovery
• Actor-Critic without Actor
• FORCE: Transferable Visual Jailbreaking Attacks via Feature Over-Reliance CorrEction
• Reinforcement Learning Fine-Tuning Enhances Activation Intensity and Diversity in the Internal Circuitry of LLMs
• GenFacts-Generative Counterfactual Explanations for Multi-Variate Time Series
• Why Attention Fails: The Degeneration of Transformers into MLPs in Time Series Forecasting
• Decoupled-Value Attention for Prior-Data Fitted Networks: GP Inference for Physical Equations
• Alignment Unlocks Complementarity: A Framework for Multiview Circuit Representation Learning
• Knowledgeable Language Models as Black-Box Optimizers for Personalized Medicine
• CLUE: Conflict-guided Localization for LLM Unlearning Framework
• FracAug: Fractional Augmentation boost Graph-level Anomaly Detection under Limited Supervision
• Toward Robust and Efficient ML-Based GPU Caching for Modern Inference
• Learning Ising Models under Hard Constraints using One Sample
• Binary Autoencoder for Mechanistic Interpretability of Large Language Models
• LiLAW: Lightweight Learnable Adaptive Weighting to Meta-Learn Sample Difficulty and Improve Noisy Training
• Aligning Inductive Bias for Data-Efficient Generalization in State Space Models
• FERD: Fairness-Enhanced Data-Free Robustness Distillation
• T2I-Diff: fMRI Signal Generation via Time-Frequency Image Transform and Classifier-Free Denoising Diffusion Models
• Explaining Grokking and Information Bottleneck through Neural Collapse Emergence
• Shaping Initial State Prevents Modality Competition in Multi-modal Fusion: A Two-stage Scheduling Framework via Fast Partial Information Decomposition
• Causal Time Series Generation via Diffusion Models
• Distribution-Controlled Client Selection to Improve Federated Learning Strategies
• Deterministic Discrete Denoising
• Energy saving in off-road vehicles using leakage compensation technique
• Investigating Modality Contribution in Audio LLMs for Music
• Wonder Wins Ways: Curiosity-Driven Exploration through Multi-Agent Contextual Calibration
• Guiding Application Users via Estimation of Computational Resources for Massively Parallel Chemistry Computations
• Theoretical Bounds for Stable In-Context Learning
• Can Federated Learning Safeguard Private Data in LLM Training? Vulnerabilities, Attacks, and Defense Evaluation
• CE-GPPO: Controlling Entropy via Gradient-Preserving Clipping Policy Optimization in Reinforcement Learning
• A Genetic Algorithm for Navigating Synthesizable Molecular Spaces
• Scaling Laws are Redundancy Laws
• The Impact of Audio Watermarking on Audio Anti-Spoofing Countermeasures
• Sig2Model: A Boosting-Driven Model for Updatable Learned Indexes
• A Recovery Theory for Diffusion Priors: Deterministic Analysis of the Implicit Prior Algorithm
• MDBench: Benchmarking Data-Driven Methods for Model Discovery
• Generalizable Diabetes Risk Stratification via Hybrid Machine Learning Models
• The Sensitivity of Variational Bayesian Neural Network Performance to Hyperparameters
• Learning Greens Operators through Hierarchical Neural Networks Inspired by the Fast Multipole Method
• TSKAN: Interpretable Machine Learning for QoE modeling over Time Series Data
• Explicit and Effectively Symmetric Schemes for Neural SDEs
• Function Spaces Without Kernels: Learning Compact Hilbert Space Representations
• Policy Compatible Skill Incremental Learning via Lazy Learning Interface
• Latent Twins
• Training Task Reasoning LLM Agents for Multi-turn Task Planning via Single-turn Reinforcement Learning
• A Theory of Multi-Agent Generative Flow Networks
• FastEagle: Cascaded Drafting for Accelerating Speculative Decoding
• mloz: A Highly Efficient Machine Learning-Based Ozone Parameterization for Climate Sensitivity Simulations
• Bridging Privacy and Utility: Synthesizing anonymized EEG with constraining utility functions
• Efficiently Attacking Memorization Scores
• Offline Goal-conditioned Reinforcement Learning with Quasimetric Representations
• Beyond Visual Similarity: Rule-Guided Multimodal Clustering with explicit domain rules
• Myosotis: structured computation for attention like layer
• Auto-Regressive U-Net for Full-Field Prediction of Shrinkage-Induced Damage in Concrete
• StyleBench: Evaluating thinking styles in Large Language Models
• Model-Based Reinforcement Learning under Random Observation Delays
• SCRA-VQA: Summarized Caption-Rerank for Augmented Large Language Models in Visual Question Answering
• On Theoretical Interpretations of Concept-Based In-Context Learning
• Integrating Object Interaction Self-Attention and GAN-Based Debiasing for Visual Question Answering
• Improving Early Sepsis Onset Prediction Through Federated Learning
• FerretNet: Efficient Synthetic Image Detection via Local Pixel Dependencies
• Deep Learning for Crime Forecasting: The Role of Mobility at Fine-grained Spatiotemporal Scales
• CTI Dataset Construction from Telegram
• Flow Matching in the Low-Noise Regime: Pathologies and a Contrastive Remedy
• Unlocking Financial Insights: An advanced Multimodal Summarization with Multimodal Output Framework for Financial Advisory Videos
• Revolutionizing Precise Low Back Pain Diagnosis via Contrastive Learning
• Even More Kawaii than Real-Person-Driven VTubers? Understanding How Viewers Perceive AI-Driven VTubers
• CaTS-Bench: Can Language Models Describe Numeric Time Series?
• Trustworthy Semantic Communication for Vehicular Networks: Challenges and Solutions
• Security-aware Semantic-driven ISAC via Paired Adversarial Residual Networks
• Verification Limits Code LLM Training
• ImaginationPolicy: Towards Generalizable, Precise and Reliable End-to-End Policy for Robotic Manipulation
• Robust Multi-Omics Integration from Incomplete Modalities Significantly Improves Prediction of Alzheimer's Disease
• FHRFormer: A Self-supervised Transformer Approach for Fetal Heart Rate Inpainting and Forecasting
• TasselNetV4: A vision foundation model for cross-scene, cross-scale, and cross-species plant counting
• Federated Markov Imputation: Privacy-Preserving Temporal Imputation in Multi-Centric ICU Environments
• Imagining Design Workflows in Agentic AI Futures
• AI-Enabled Crater-Based Navigation for Lunar Mapping
• Confidence-guided Refinement Reasoning for Zero-shot Question Answering
• Seeing Through Words, Speaking Through Pixels: Deep Representational Alignment Between Vision and Language Models
• Measuring LLM Sensitivity in Transformer-based Tabular Data Synthesis
• Provenance Analysis of Archaeological Artifacts via Multimodal RAG Systems
• CusEnhancer: A Zero-Shot Scene and Controllability Enhancement Method for Photo Customization via ResInversion
• IConv: Focusing on Local Variation with Channel Independent Convolution for Multivariate Time Series Forecasting
• Towards Atoms of Large Language Models
• DAC-LoRA: Dynamic Adversarial Curriculum for Efficient and Robust Few-Shot Adaptation
• Leveraging What's Overfixed: Post-Correction via LLM Grammatical Error Overcorrection
• Bispectral OT: Dataset Comparison using Symmetry-Aware Optimal Transport
• QAMO: Quality-aware Multi-centroid One-class Learning For Speech Deepfake Detection
• Efficient Construction of Implicit Surface Models From a Single Image for Motion Generation
• Addressing Gradient Misalignment in Data-Augmented Training for Robust Speech Deepfake Detection
• Learning to Align Molecules and Proteins: A Geometry-Aware Approach to Binding Affinity
• Incorporating LLM Embeddings for Variation Across the Human Genome
• Joint Flow Trajectory Optimization For Feasible Robot Motion Generation from Video Demonstrations
• Beyond the Individual: Introducing Group Intention Forecasting with SHOT Dataset
• RobotDancing: Residual-Action Reinforcement Learning Enables Robust Long-Horizon Humanoid Motion Tracking
• Every Character Counts: From Vulnerability to Defense in Phishing Detection
• An LLM-based Agentic Framework for Accessible Network Control
• Experience Deploying Containerized GenAI Services at an HPC Center
• MMG: Mutual Information Estimation via the MMSE Gap in Diffusion
• FS-DFM: Fast and Accurate Long Text Generation with Few-Step Diffusion Language Models
• Personalized Federated Dictionary Learning for Modeling Heterogeneity in Multi-site fMRI Data
• Recidivism and Peer Influence with LLM Text Embeddings in Low Security Correctional Facilities
• Learning Terrain-Specialized Policies for Adaptive Locomotion in Challenging Environments
• A Framework for Rapidly Developing and Deploying Protection Against Large Language Model Attacks
• Look Before you Leap: Estimating LLM Benchmark Scores from Descriptions
• Understanding Mode Switching in Human-AI Collaboration: Behavioral Insights and Predictive Modeling
• CHOIR: A Chatbot-mediated Organizational Memory Leveraging Communication in University Research Labs
• InstructVTON: Optimal Auto-Masking and Natural-Language-Guided Interactive Style Control for Inpainting-Based Virtual Try-On
• Understanding and Improving Adversarial Robustness of Neural Probabilistic Circuits
• GraspFactory: A Large Object-Centric Grasping Dataset
• Perspectra: Choosing Your Experts Enhances Critical Thinking in Multi-Agent Research Ideation
• SwasthLLM: a Unified Cross-Lingual, Multi-Task, and Meta-Learning Zero-Shot Framework for Medical Diagnosis Using Contrastive Representations
• PIRF: Physics-Informed Reward Fine-Tuning for Diffusion Models
• MechStyle: Augmenting Generative AI with Mechanical Simulation to Create Stylized and Structurally Viable 3D Models
• Dynamic Reasoning Chains through Depth-Specialized Mixture-of-Experts in Transformer Architectures
• Hierarchical Resolution Transformers: A Wavelet-Inspired Architecture for Multi-Scale Language Understanding
• Shared Neural Space: Unified Precomputed Feature Encoding for Multi-Task and Cross Domain Vision
• CoSupFormer : A Contrastive Supervised learning approach for EEG signal Classification
• AI-Specific Code Smells: From Specification to Detection
• Boosting Zero-Shot VLN via Abstract Obstacle Map-Based Waypoint Prediction with TopoGraph-and-VisitInfo-Aware Prompting
• MARS: toward more efficient multi-agent collaboration for LLM reasoning
• Complexity-Driven Policy Optimization
• Variational Low-Rank Adaptation for Personalized Impaired Speech Recognition
• Defending against Stegomalware in Deep Neural Networks with Permutation Symmetry
• Adversarial Defense in Cybersecurity: A Systematic Review of GANs for Threat Detection and Mitigation
• A Taxonomy of Data Risks in AI and Quantum Computing (QAI) - A Systematic Review
• Wartime Media Dynamics in Emerging Democracies: Case Study of Pakistani Media in May 2025 Indo-Pak Conflict
• ACCeLLiuM: Supervised Fine-Tuning for Automated OpenACC Pragma Generation
• USB-Rec: An Effective Framework for Improving Conversational Recommendation Capability of Large Language Model
• Lightweight MobileNetV1+GRU for ECG Biometric Authentication: Federated and Adversarial Evaluation
• MARS: A Malignity-Aware Backdoor Defense in Federated Learning
• R1-Fuzz: Specializing Language Models for Textual Fuzzing via Reinforcement Learning
• Dynamic ReAct: Scalable Tool Selection for Large-Scale MCP Environments
• Can You Trust Your Copilot? A Privacy Scorecard for AI Coding Assistants
• The Secret Agenda: LLMs Strategically Lie and Our Current Safety Tools Are Blind
• Blueprints of Trust: AI System Cards for End to End Transparency and Governance
• Centralized vs. Decentralized Security for Space AI Systems? A New Look
• Data-Efficient ASR Personalization for Non-Normative Speech Using an Uncertainty-Based Phoneme Difficulty Score for Guided Sampling
• Grounding AI Explanations in Experience: A Reflective Cognitive Architecture for Clinical Decision Support
• VC-Agent: An Interactive Agent for Customized Video Dataset Collection
• SAGE: A Realistic Benchmark for Semantic Understanding
• Interpreting Public Sentiment in Diplomacy Events: A Counterfactual Analysis Framework Using Large Language Models
• AI-driven formative assessment and adaptive learning in data-science education: Evaluating an LLM-powered virtual teaching assistant
• CFD-LLMBench: A Benchmark Suite for Evaluating Large Language Models in Computational Fluid Dynamics
• Assessing Classical Machine Learning and Transformer-based Approaches for Detecting AI-Generated Research Text
• ConceptViz: A Visual Analytics Approach for Exploring Concepts in Large Language Models
• SKILL-RAG: Self-Knowledge Induced Learning and Filtering for Retrieval-Augmented Generation
• Beyond Global Emotion: Fine-Grained Emotional Speech Synthesis with Dynamic Word-Level Modulation
• Combinatorial Creativity: A New Frontier in Generalization Abilities
• Disagreements in Reasoning: How a Model's Thinking Process Dictates Persuasion in Multi-Agent Systems
• Recon-Act: A Self-Evolving Multi-Agent Browser-Use System via Web Reconnaissance, Tool Generation, and Task Execution
• 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
• RL Squeezes, SFT Expands: A Comparative Study of Reasoning LLMs
• ToMPO: Training LLM Strategic Decision Making from a Multi-Agent Perspective
• Embodied Representation Alignment with Mirror Neurons
• Distributed Specialization: Rare-Token Neurons in Large Language Models
• A Fano-Style Accuracy Upper Bound for LLM Single-Pass Reasoning in Multi-Hop QA
• What Do LLM Agents Do When Left Alone? Evidence of Spontaneous Meta-Cognitive Patterns
• Fairy: Interactive Mobile Assistant to Real-world Tasks via LMM-based Multi-agent
• Parallel Thinking, Sequential Answering: Bridging NAR and AR for Efficient Reasoning
• Meta-Memory: Retrieving and Integrating Semantic-Spatial Memories for Robot Spatial Reasoning
• LogReasoner: Empowering LLMs with Expert-like Coarse-to-Fine Reasoning for Log Analysis Tasks
• DeFacto: Counterfactual Thinking with Images for Enforcing Evidence-Grounded and Faithful Reasoning
• GALAX: Graph-Augmented Language Model for Explainable Reinforcement-Guided Subgraph Reasoning in Precision Medicine
• Beyond Stars: Bridging the Gap Between Ratings and Review Sentiment with LLM
• AOT*: Efficient Synthesis Planning via LLM-Empowered AND-OR Tree Search
• CORE: Full-Path Evaluation of LLM Agents Beyond Final State
• Who Gets Cited Most? Benchmarking Long-Context Language Models on Scientific Articles
• CLAUSE: Agentic Neuro-Symbolic Knowledge Graph Reasoning via Dynamic Learnable Context Engineering
• An Approach to Checking Correctness for Agentic Systems
• LATTS: Locally Adaptive Test-Time Scaling
• Philosophy-informed Machine Learning
• InsightGUIDE: An Opinionated AI Assistant for Guided Critical Reading of Scientific Literature
• Reconstruction-Based Adaptive Scheduling Using AI Inferences in Safety-Critical Systems
• Adaptive Approach to Enhance Machine Learning Scheduling Algorithms During Runtime Using Reinforcement Learning in Metascheduling Applications
• A Compound Classification System Based on Fuzzy Relations Applied to the Noise-Tolerant Control of a Bionic Hand via EMG Signal Recognition
• SAMULE: Self-Learning Agents Enhanced by Multi-level Reflection
• Adaptive Cybersecurity Architecture for Digital Product Ecosystems Using Agentic AI
• Accelerate Creation of Product Claims Using Generative AI
• An Automated Retrieval-Augmented Generation LLaMA-4 109B-based System for Evaluating Radiotherapy Treatment Plans