AI Research News Feeds for October 23rd, 2025

AI RESEARCH PAPERS & ACADEMIC SOURCES

• Optimized 3D Gaussian Splatting using Coarse-to-Fine Image Frequency Modulation
• Advancing Image Super-resolution Techniques in Remote Sensing: A Comprehensive Survey
• Lost in the Maze: Overcoming Context Limitations in Long-Horizon Agentic Search
• Dynamic Evaluation for Oversensitivity in LLMs
• Are they lovers or friends? Evaluating LLMs' Social Reasoning in English and Korean Dialogues
• Re:Member: Emotional Question Generation from Personal Memories
• When Can We Trust LLMs in Mental Health? Large-Scale Benchmarks for Reliable LLM Evaluation
• From Memorization to Generalization: Fine-Tuning Large Language Models for Biomedical Term-to-Identifier Normalization
• Tibetan Language and AI: A Comprehensive Survey of Resources, Methods and Challenges
• "You Are Rejected!": An Empirical Study of Large Language Models Taking Hiring Evaluations
• Think Straight, Stop Smart: Structured Reasoning for Efficient Multi-Hop RAG
• Multi-Faceted Evaluation of Tool-Augmented Dialogue Systems
• DiSRouter: Distributed Self-Routing for LLM Selections
• Modality Matching Matters: Calibrating Language Distances for Cross-Lingual Transfer in URIEL+
• SheetBrain: A Neuro-Symbolic Agent for Accurate Reasoning over Complex and Large Spreadsheets
• Difficulty-Controllable Multiple-Choice Question Generation Using Large Language Models and Direct Preference Optimization
• TheMCPCompany: Creating General-purpose Agents with Task-specific Tools
• JointCQ: Improving Factual Hallucination Detection with Joint Claim and Query Generation
• KORE: Enhancing Knowledge Injection for Large Multimodal Models via Knowledge-Oriented Augmentations and Constraints
• HAD: HAllucination Detection Language Models Based on a Comprehensive Hallucination Taxonomy
• Slot Filling as a Reasoning Task for SpeechLLMs
• Algorithmic Fairness in NLP: Persona-Infused LLMs for Human-Centric Hate Speech Detection
• Local Obfuscation by GLINER for Impartial Context Aware Lineage: Development and evaluation of PII Removal system
• Modeling Turn-Taking with Semantically Informed Gestures
• LoongRL:Reinforcement Learning for Advanced Reasoning over Long Contexts
• Sign Language Translation with Sentence Embedding Supervision
• SONAR-SLT: Multilingual Sign Language Translation via Language-Agnostic Sentence Embedding Supervision
• Spatio-temporal Sign Language Representation and Translation
• BLiSS 1.0: Evaluating Bilingual Learner Competence in Second Language Small Language Models
• MINED: Probing and Updating with Multimodal Time-Sensitive Knowledge for Large Multimodal Models
• Machine Text Detectors are Membership Inference Attacks
• What is the Best Sequence Length for BABYLM?
• Lookahead Routing for Large Language Models
• Which Evaluation for Which Model? A Taxonomy for Speech Model Assessment
• Conditions for Catastrophic Forgetting in Multilingual Translation
• PBBQ: A Persian Bias Benchmark Dataset Curated with Human-AI Collaboration for Large Language Models
• CrossNews-UA: A Cross-lingual News Semantic Similarity Benchmark for Ukrainian, Polish, Russian, and English
• LLavaCode: Compressed Code Representations for Retrieval-Augmented Code Generation
• DiffAdapt: Difficulty-Adaptive Reasoning for Token-Efficient LLM Inference
• CoSense-LLM: Semantics at the Edge with Cost- and Uncertainty-Aware Cloud-Edge Cooperation
• From Answers to Guidance: A Proactive Dialogue System for Legal Documents
• Adapting Multilingual Models to Code-Mixed Tasks via Model Merging
• ToolDreamer: Instilling LLM Reasoning Into Tool Retrievers
• The Art of Asking: Multilingual Prompt Optimization for Synthetic Data
• Towards Better Health Conversations: The Benefits of Context-seeking
• OpenGuardrails: An Open-Source Context-Aware AI Guardrails Platform
• Aligning Multilingual News for Stock Return Prediction
• [De|Re]constructing VLMs' Reasoning in Counting
• olmOCR 2: Unit Test Rewards for Document OCR
• LV-Eval: A Balanced Long-Context Benchmark with 5 Length Levels Up to 256K
• CausalRAG: Integrating Causal Graphs into Retrieval-Augmented Generation
• Can Large Language Models be Effective Online Opinion Miners?
• metaTextGrad: Automatically optimizing language model optimizers
• ETT: Expanding the Long Context Understanding Capability of LLMs at Test-Time
• PixelWorld: How Far Are We from Perceiving Everything as Pixels?
• ScholaWrite: A Dataset of End-to-End Scholarly Writing Process
• WikiVideo: Article Generation from Multiple Videos
• Efficient Interleaved Speech Modeling through Knowledge Distillation
• Dimensionality Reduction for Remote Sensing Data Analysis: A Systematic Review of Methods and Applications
• Ninja Codes: Neurally Generated Fiducial Markers for Stealthy 6-DoF Tracking
• MoAlign: Motion-Centric Representation Alignment for Video Diffusion Models
• UniHPR: Unified Human Pose Representation via Singular Value Contrastive Learning
• Advancing Brain Tumor Segmentation via Attention-based 3D U-Net Architecture and Digital Image Processing
• FootFormer: Estimating Stability from Visual Input
• Malaria Detection from Blood Cell Images Using XceptionNet
• Video Consistency Distance: Enhancing Temporal Consistency for Image-to-Video Generation via Reward-Based Fine-Tuning
• MoE-GS: Mixture of Experts for Dynamic Gaussian Splatting
• SFGFusion: Surface Fitting Guided 3D Object Detection with 4D Radar and Camera Fusion
• Space Object Detection using Multi-frame Temporal Trajectory Completion Method
• Background Fades, Foreground Leads: Curriculum-Guided Background Pruning for Efficient Foreground-Centric Collaborative Perception
• Advances in 4D Representation: Geometry, Motion, and Interaction
• SCEESR: Semantic-Control Edge Enhancement for Diffusion-Based Super-Resolution
• MobiAct: Efficient MAV Action Recognition Using MobileNetV4 with Contrastive Learning and Knowledge Distillation
• D2D: Detector-to-Differentiable Critic for Improved Numeracy in Text-to-Image Generation
• Vision-Based Mistake Analysis in Procedural Activities: A Review of Advances and Challenges
• Unified Reinforcement and Imitation Learning for Vision-Language Models
• Exploring Scale Shift in Crowd Localization under the Context of Domain Generalization
• BrainMCLIP: Brain Image Decoding with Multi-Layer feature Fusion of CLIP
• A Training-Free Framework for Open-Vocabulary Image Segmentation and Recognition with EfficientNet and CLIP
• DaMo: Data Mixing Optimizer in Fine-tuning Multimodal LLMs for Mobile Phone Agents
• DARE: A Deformable Adaptive Regularization Estimator for Learning-Based Medical Image Registration
• AegisRF: Adversarial Perturbations Guided with Sensitivity for Protecting Intellectual Property of Neural Radiance Fields
• Seeing Across Views: Benchmarking Spatial Reasoning of Vision-Language Models in Robotic Scenes
• Multi-Camera Worker Tracking in Logistics Warehouse Considering Wide-Angle Distortion
• Reasoning Like Experts: Leveraging Multimodal Large Language Models for Drawing-based Psychoanalysis
• Predicting before Reconstruction: A generative prior framework for MRI acceleration
• PRGCN: A Graph Memory Network for Cross-Sequence Pattern Reuse in 3D Human Pose Estimation
• Mitigating representation bias caused by missing pixels in methane plume detection
• Towards Single-Source Domain Generalized Object Detection via Causal Visual Prompts
• PoseCrafter: Extreme Pose Estimation with Hybrid Video Synthesis
• The Intricate Dance of Prompt Complexity, Quality, Diversity, and Consistency in T2I Models
• HAD: Hierarchical Asymmetric Distillation to Bridge Spatio-Temporal Gaps in Event-Based Object Tracking
• Can You Trust What You See? Alpha Channel No-Box Attacks on Video Object Detection
• VGD: Visual Geometry Gaussian Splatting for Feed-Forward Surround-view Driving Reconstruction
• Addressing the Depth-of-Field Constraint: A New Paradigm for High Resolution Multi-Focus Image Fusion
• Digitizing Paper ECGs at Scale: An Open-Source Algorithm for Clinical Research
• Decomposed Attention Fusion in MLLMs for Training-Free Video Reasoning Segmentation
• CBDiff:Conditional Bernoulli Diffusion Models for Image Forgery Localization
• Beyond sparse denoising in frames: minimax estimation with a scattering transform
• Pragmatic Heterogeneous Collaborative Perception via Generative Communication Mechanism
• Augmenting Moment Retrieval: Zero-Dependency Two-Stage Learning
• MedReason-R1: Learning to Reason for CT Diagnosis with Reinforcement Learning and Local Zoom
• Re-Activating Frozen Primitives for 3D Gaussian Splatting
• Curvilinear Structure-preserving Unpaired Cross-domain Medical Image Translation
• Explainable Face Presentation Attack Detection via Ensemble-CAM
• LyTimeT: Towards Robust and Interpretable State-Variable Discovery
• Adaptive Distribution-aware Quantization for Mixed-Precision Neural Networks
• OmniMotion-X: Versatile Multimodal Whole-Body Motion Generation
• Class-Aware Prototype Learning with Negative Contrast for Test-Time Adaptation of Vision-Language Models
• How to Evaluate Monocular Depth Estimation?
• Is This Tracker On? A Benchmark Protocol for Dynamic Tracking
• GRASPLAT: Enabling dexterous grasping through novel view synthesis
• GigaBrain-0: A World Model-Powered Vision-Language-Action Model
• Automated Morphological Analysis of Neurons in Fluorescence Microscopy Using YOLOv8
• LBL: Logarithmic Barrier Loss Function for One-class Classification
• LookUp3D: Data-Driven 3D Scanning
• Brain3D: Generating 3D Objects from fMRI
• ComDrive: Comfort-Oriented End-to-End Autonomous Driving
• Adversarial Attacks on LiDAR-Based Tracking Across Road Users: Robustness Evaluation and Target-Aware Black-Box Method
• Learning Differential Pyramid Representation for Tone Mapping
• VLsI: Verbalized Layers-to-Interactions from Large to Small Vision Language Models
• ASAP: Advancing Semantic Alignment Promotes Multi-Modal Manipulation Detecting and Grounding
• FLARE: Feed-forward Geometry, Appearance and Camera Estimation from Uncalibrated Sparse Views
• Video-R1: Reinforcing Video Reasoning in MLLMs
• MMLA: Multi-Environment, Multi-Species, Low-Altitude Drone Dataset
• 3D Visual Illusion Depth Estimation
• Spiking Neural Networks Need High Frequency Information
• See through the Dark: Learning Illumination-affined Representations for Nighttime Occupancy Prediction
• Uni-MuMER: Unified Multi-Task Fine-Tuning of Vision-Language Model for Handwritten Mathematical Expression Recognition
• Doctor Approved: Generating Medically Accurate Skin Disease Images through AI-Expert Feedback
• Towards foundational LiDAR world models with efficient latent flow matching
• Latent Diffusion Models with Masked AutoEncoders
• Towards Depth Foundation Model: Recent Trends in Vision-Based Depth Estimation
• Breaking the Discretization Barrier of Continuous Physics Simulation Learning
• Discretized Gaussian Representation for Tomographic Reconstruction
• Vectorization of Persistence Diagrams for Topological Data Analysis in R and Python Using TDAvec Package
• When Do Transformers Learn Heuristics for Graph Connectivity?
• CONFEX: Uncertainty-Aware Counterfactual Explanations with Conformal Guarantees
• The Tail Tells All: Estimating Model-Level Membership Inference Vulnerability Without Reference Models
• GaLLoP: Gradient-based Sparse Learning on Low-Magnitude Parameters
• Environment Inference for Learning Generalizable Dynamical System
• Blackbox Model Provenance via Palimpsestic Membership Inference
• Transformers are almost optimal metalearners for linear classification
• The Feasibility of Training Sovereign Language Models in the Global South: A Study of Brazil and Mexico
• CityAQVis: Integrated ML-Visualization Sandbox Tool for Pollutant Estimation in Urban Regions Using Multi-Source Data (Software Article)
• When Models Can't Follow: Testing Instruction Adherence Across 256 LLMs
• Foundation Models for Discovery and Exploration in Chemical Space
• Evaluating LLM Story Generation through Large-scale Network Analysis of Social Structures
• Impartial Selection with Predictions
• Calibrated Principal Component Regression
• Learning noisy tissue dynamics across time scales
• Signature Kernel Scoring Rule as Spatio-Temporal Diagnostic for Probabilistic Forecasting
• A Graph Signal Processing Framework for Hallucination Detection in Large Language Models
• Training-Free Spectral Fingerprints of Voice Processing in Transformers
• HAMLOCK: HArdware-Model LOgically Combined attacK
• Extreme Event Aware ($\eta$-) Learning
• Transfer Learning Beyond the Standard Model
• RLBoost: Harvesting Preemptible Resources for Cost-Efficient Reinforcement Learning on LLMs
• Synthesizability Prediction of Crystalline Structures with a Hierarchical Transformer and Uncertainty Quantification
• Hierarchical DLO Routing with Reinforcement Learning and In-Context Vision-language Models
• Magnetic field estimation using Gaussian process regression for interactive wireless power system design
• Topology of Currencies: Persistent Homology for FX Co-movements: A Comparative Clustering Study
• Transformers are Inherently Succinct
• Nonmonotone subgradient methods based on a local descent lemma
• Autobidding Arena: unified evaluation of the classical and RL-based autobidding algorithms
• MoE-Prism: Disentangling Monolithic Experts for Elastic MoE Services via Model-System Co-Designs
• AMAuT: A Flexible and Efficient Multiview Audio Transformer Framework Trained from Scratch
• On the hardness of RL with Lookahead
• Using Temperature Sampling to Effectively Train Robot Learning Policies on Imbalanced Datasets
• Square root Cox's survival analysis by the fittest linear and neural networks model
• A Derandomization Framework for Structure Discovery: Applications in Neural Networks and Beyond
• From See to Shield: ML-Assisted Fine-Grained Access Control for Visual Data
• Exploring "Many in Few" and "Few in Many" Properties in Long-Tailed, Highly-Imbalanced IC Defect Classification
• PCP-GAN: Property-Constrained Pore-scale image reconstruction via conditional Generative Adversarial Networks
• Re-evaluating Minimum Bayes Risk Decoding for Automatic Speech Recognition
• Online Two-Stage Submodular Maximization
• Learning Upper Lower Value Envelopes to Shape Online RL: A Principled Approach
• Uncertainty evaluation of segmentation models for Earth observation
• Remarks on a recent preprint of Chernikov and Towsner
• Bridging Earth and Space: A Survey on HAPS for Non-Terrestrial Networks
• Zhyper: Factorized Hypernetworks for Conditioned LLM Fine-Tuning
• Exploring the Effect of DNN Depth on Adversarial Attacks in Network Intrusion Detection Systems
• Pico-Banana-400K: A Large-Scale Dataset for Text-Guided Image Editing
• Hubble: a Model Suite to Advance the Study of LLM Memorization
• DR-VIDAL -- Doubly Robust Variational Information-theoretic Deep Adversarial Learning for Counterfactual Prediction and Treatment Effect Estimation on Real World Data
• Source-Free Domain Adaptation for SSVEP-based Brain-Computer Interfaces
• AtomSurf : Surface Representation for Learning on Protein Structures
• Phase-driven Domain Generalizable Learning for Nonstationary Time Series
• Survey of Graph Neural Network for Internet of Things and NextG Networks
• Bootstrap Sampling Rate Greater than 1.0 May Improve Random Forest Performance
• Learning to Learn with Contrastive Meta-Objective
• AdaptGrad: Adaptive Sampling to Reduce Noise
• Deep Linear Probe Generators for Weight Space Learning
• LoRA vs Full Fine-tuning: An Illusion of Equivalence
• DNN Modularization via Activation-Driven Training
• by Using Less: Distributed Learning with Energy-Constrained Devices
• Joint Hierarchical Representation Learning of Samples and Features via Informed Tree-Wasserstein Distance
• Finite Sample Identification of Partially Observed Bilinear Dynamical Systems
• A recursive Bayesian neural network for constitutive modeling of sands under monotonic and cyclic loading
• Learning Reward Machines from Partially Observed Policies
• Training-Free Constrained Generation With Stable Diffusion Models
• Optimizing Asynchronous Federated Learning: A Delicate Trade-Off Between Model-Parameter Staleness and Update Frequency
• Learning to Add, Multiply, and Execute Algorithmic Instructions Exactly with Neural Networks
• Long-term Causal Inference via Modeling Sequential Latent Confounding
• Using (Not-so) Large Language Models to Generate Simulation Models in a Formal DSL: A Study on Reaction Networks
• Learning Spatially Adaptive $\ell_1$-Norms Weights for Convolutional Synthesis Regularization
• Kolmogorov-Arnold Attention: Is Learnable Attention Better For Vision Transformers?
• Customizing Spider Silk: Generative Models with Mechanical Property Conditioning for Protein Engineering
• TunnElQNN: A Hybrid Quantum-classical Neural Network for Efficient Learning
• CATransformers: Carbon Aware Transformers Through Joint Model-Hardware Optimization
• InfiFPO: Implicit Model Fusion via Preference Optimization in Large Language Models
• Discrete Neural Flow Samplers with Locally Equivariant Transformer
• Beyond Masked and Unmasked: Discrete Diffusion Models via Partial Masking
• AmorLIP: Efficient Language-Image Pretraining via Amortization
• Diffusion-Based Hierarchical Graph Neural Networks for Simulating Nonlinear Solid Mechanics
• SEC-bench: Automated Benchmarking of LLM Agents on Real-World Software Security Tasks
• Heavy-Ball Momentum Method in Continuous Time and Discretization Error Analysis
• Generating Directed Graphs with Dual Attention and Asymmetric Encoding
• Unfolding Generative Flows with Koopman Operators: Fast and Interpretable Sampling
• Online Conformal Prediction with Efficiency Guarantees
• Semi-off-Policy Reinforcement Learning for Vision-Language Slow-Thinking Reasoning
• What Expressivity Theory Misses: Message Passing Complexity for GNNs
• Rebalancing with Calibrated Sub-classes (RCS): A Statistical Fusion-based Framework for Robust Imbalanced Classification across Modalities
• LASeR: Learning to Adaptively Select Reward Models with Multi-Armed Bandits
• VisDiff: SDF-Guided Polygon Generation for Visibility Reconstruction and Recognition
• Stable Matching with Ties: Approximation Ratios and Learning
• The Streaming Batch Model for Efficient and Fault-Tolerant Heterogeneous Execution
• Advancing Carbon Capture using AI: Design of permeable membrane and estimation of parameters for Carbon Capture using linear regression and membrane-based equations
• MsEdF: A Multi-stream Encoder-decoder Framework for Remote Sensing Image Captioning
• Evolution of Information in Interactive Decision Making: A Case Study for Multi-Armed Bandits
• Fast, Modular, and Differentiable Framework for Machine Learning-Enhanced Molecular Simulations
• A Comprehensive Benchmark for RNA 3D Structure-Function Modeling
• Smoothed Distance Kernels for MMDs and Applications in Wasserstein Gradient Flows
• Sub-optimality of the Separation Principle for Quadratic Control from Bilinear Observations
• Rank-One Modified Value Iteration
• Non-Stationary Lipschitz Bandits
• Nearly-Linear Time Private Hypothesis Selection with the Optimal Approximation Factor
• QiMeng-MuPa: Mutual-Supervised Learning for Sequential-to-Parallel Code Translation
• QPPG: Quantum-Preconditioned Policy Gradient for Link Adaptation in Rayleigh Fading Channels
• Stable Minima of ReLU Neural Networks Suffer from the Curse of Dimensionality: The Neural Shattering Phenomenon
• Trajectory learning for ensemble forecasts via the continuous ranked probability score: a Lorenz '96 case study
• Are Modern Speech Enhancement Systems Vulnerable to Adversarial Attacks?
• Transformer-Based Low-Resource Language Translation: A Study on Standard Bengali to Sylheti
• Integrating Transparent Models, LLMs, and Practitioner-in-the-Loop: A Case of Nonprofit Program Evaluation
• Scaf-GRPO: Scaffolded Group Relative Policy Optimization for Enhancing LLM Reasoning
• Semantic World Models
• Hummer: Towards Limited Competitive Preference Dataset
• Reasoning Models Better Express Their Confidence
• Follow the STARs: Dynamic $\omega$-Regular Shielding of Learned Policies
• ROTATE: Regret-driven Open-ended Training for Ad Hoc Teamwork
• Measuring Data Science Automation: A Survey of Evaluation Tools for AI Assistants and Agents
• The Endless Tuning. An Artificial Intelligence Design To Avoid Human Replacement and Trace Back Responsibilities
• IM-Chat: A Multi-agent LLM Framework Integrating Tool-Calling and Diffusion Modeling for Knowledge Transfer in Injection Molding Industry
• Data Efficient Adaptation in Large Language Models via Continuous Low-Rank Fine-Tuning
• FP-IRL: Fokker-Planck Inverse Reinforcement Learning -- A Physics-Constrained Approach to Markov Decision Processes
• Embedding in Recommender Systems: A Survey
• Towards Context-Aware Domain Generalization: Understanding the Benefits and Limits of Marginal Transfer Learning
• LICO: Large Language Models for In-Context Molecular Optimization
• Estimating Long-term Heterogeneous Dose-response Curve: Generalization Bound Leveraging Optimal Transport Weights
• PARCO: Parallel AutoRegressive Models for Multi-Agent Combinatorial Optimization
• Unveiling Transformer Perception by Exploring Input Manifolds
• Learning Linear Attention in Polynomial Time
• Model-based Large Language Model Customization as Service
• Benchmarking Large Language Models with Integer Sequence Generation Tasks
• Fast MRI for All: Bridging Access Gaps by Training without Raw Data
• Open-World Drone Active Tracking with Goal-Centered Rewards
• Explainable fault and severity classification for rolling element bearings using Kolmogorov-Arnold networks
• Graph Representation Learning with Diffusion Generative Models
• ModServe: Modality- and Stage-Aware Resource Disaggregation for Scalable Multimodal Model Serving
• An Efficient Local Search Approach for Polarized Community Discovery in Signed Networks
• Probing Perceptual Constancy in Large Vision-Language Models
• Towards Enhanced Image Generation Via Multi-modal Chain of Thought in Unified Generative Models
• FairGen: Controlling Sensitive Attributes for Fair Generations in Diffusion Models via Adaptive Latent Guidance
• Merge then Realign: Simple and Effective Modality-Incremental Continual Learning for Multimodal LLMs
• EgoBlind: Towards Egocentric Visual Assistance for the Blind
• PTFA: An LLM-based Agent that Facilitates Online Consensus Building through Parallel Thinking
• GeoBenchX: Benchmarking LLMs in Agent Solving Multistep Geospatial Tasks
• NAACL2025 Tutorial: Adaptation of Large Language Models
• Merging Embedded Topics with Optimal Transport for Online Topic Modeling on Data Streams
• Quantum Natural Language Processing: A Comprehensive Review of Models, Methods, and Applications
• LongCodeBench: Evaluating Coding LLMs at 1M Context Windows
• Memorization-Compression Cycles Improve Generalization
• A Multi-Task Foundation Model for Wireless Channel Representation Using Contrastive and Masked Autoencoder Learning
• CtrlDiff: Boosting Large Diffusion Language Models with Dynamic Block Prediction and Controllable Generation
• Improving planning and MBRL with temporally-extended actions
• An Analysis of Concept Bottleneck Models: Measuring, Understanding, and Mitigating the Impact of Noisy Annotations
• Evaluating NLP Embedding Models for Handling Science-Specific Symbolic Expressions in Student Texts
• High-order Equivariant Flow Matching for Density Functional Theory Hamiltonian Prediction
• MLR-Bench: Evaluating AI Agents on Open-Ended Machine Learning Research
• Learning from Videos for 3D World: Enhancing MLLMs with 3D Vision Geometry Priors
• QoQ-Med: Building Multimodal Clinical Foundation Models with Domain-Aware GRPO Training
• Horizon Reduction Makes RL Scalable
• TimeWak: Temporal Chained-Hashing Watermark for Time Series Data
• GUARD: Guided Unlearning and Retention via Data Attribution for Large Language Models
• Flexible-length Text Infilling for Discrete Diffusion Models
• One-Step Diffusion for Detail-Rich and Temporally Consistent Video Super-Resolution
• Improved Exploration in GFlownets via Enhanced Epistemic Neural Networks
• With Limited Data for Multimodal Alignment, Let the STRUCTURE Guide You
• Chiron-o1: Igniting Multimodal Large Language Models towards Generalizable Medical Reasoning via Mentor-Intern Collaborative Search
• Understanding Reasoning in Thinking Language Models via Steering Vectors
• Pay Attention to Small Weights
• PRING: Rethinking Protein-Protein Interaction Prediction from Pairs to Graphs
• Where are we with calibration under dataset shift in image classification?
• Position: Many generalization measures for deep learning are fragile
• Towards Universal Solvers: Using PGD Attack in Active Learning to Increase Generalizability of Neural Operators as Knowledge Distillation from Numerical PDE Solvers
• An Encode-then-Decompose Approach to Unsupervised Time Series Anomaly Detection on Contaminated Training Data--Extended Version
• Category learning in deep neural networks: Information content and geometry of internal representations
• Empowering Decision Trees via Shape Function Branching
• POLAR: Policy-based Layerwise Reinforcement Learning Method for Stealthy Backdoor Attacks in Federated Learning
• Weight Decay may matter more than muP for Learning Rate Transfer in Practice
• MetaCluster: Enabling Deep Compression of Kolmogorov-Arnold Network
• Learning Peer Influence Probabilities with Linear Contextual Bandits
• Subliminal Corruption: Mechanisms, Thresholds, and Interpretability
• Feature Space Adaptation for Robust Model Fine-Tuning
• Instance-Dependent Regret Bounds for Nonstochastic Linear Partial Monitoring
• Preliminary Use of Vision Language Model Driven Extraction of Mouse Behavior Towards Understanding Fear Expression
• Natural Gradient VI: Guarantees for Non-Conjugate Models
• A Communication-Efficient Decentralized Actor-Critic Algorithm
• Enhancing Graph Neural Networks: A Mutual Learning Approach
• Controllable Machine Unlearning via Gradient Pivoting
• Brain-Inspired Perspective on Configurations: Unsupervised Similarity and Early Cognition
• Understanding the Implicit Biases of Design Choices for Time Series Foundation Models
• Interpret Policies in Deep Reinforcement Learning using SILVER with RL-Guided Labeling: A Model-level Approach to High-dimensional and Multi-action Environments
• Mixing Configurations for Downstream Prediction
• Data Efficient Any Transformer-to-Mamba Distillation via Attention Bridge
• Knowledge Distillation of Uncertainty using Deep Latent Factor Model
• QiMeng-SALV: Signal-Aware Learning for Verilog Code Generation
• Loopholing Discrete Diffusion: Deterministic Bypass of the Sampling Wall
• FrogDeepSDM: Improving Frog Counting and Occurrence Prediction Using Multimodal Data and Pseudo-Absence Imputation
• Calibration and Discrimination Optimization Using Clusters of Learned Representation
• A Markov Decision Process for Variable Selection in Branch & Bound
• Scalable LinUCB: Low-Rank Design Matrix Updates for Recommenders with Large Action Spaces
• ConvXformer: Differentially Private Hybrid ConvNeXt-Transformer for Inertial Navigation
• Optimization Benchmark for Diffusion Models on Dynamical Systems
• LMFD: Latent Monotonic Feature Discovery
• Learning Noise-Resilient and Transferable Graph-Text Alignment via Dynamic Quality Assessment
• CPSVD: Enhancing Large Language Model Compression via Column-Preserving Singular Value Decomposition
• ARA: Adaptive Rank Allocation for Efficient Large Language Model SVD Compression
• Iterative Training of Physics-Informed Neural Networks with Fourier-enhanced Features
• LLM Unlearning with LLM Beliefs
• Revisiting the Relation Between Robustness and Universality
• g-DPO: Scalable Preference Optimization for Protein Language Models
• ELUTQ: Efficient LUT-Aware Quantization for Deploying Large Language Models on Edge Devices
• Energy-Efficient and Dequantization-Free Q-LLMs: A Spiking Neural Network Approach to Salient Value Mitigation
• Teaming LLMs to Detect and Mitigate Hallucinations
• Bi-Level Decision-Focused Causal Learning for Large-Scale Marketing Optimization: Bridging Observational and Experimental Data
• The Confusing Instance Principle for Online Linear Quadratic Control
• A Climate-Aware Deep Learning Framework for Generalizable Epidemic Forecasting
• Learning and Simulating Building Evacuation Patterns for Enhanced Safety Design Using Generative Models
• Matrix-Free Least Squares Solvers: Values, Gradients, and What to Do With Them
• Latent Space Factorization in LoRA
• Overlap-weighted orthogonal meta-learner for treatment effect estimation over time
• Policy Learning with Abstention
• Fast Inference via Hierarchical Speculative Decoding
• SEMPO: Lightweight Foundation Models for Time Series Forecasting
• Statistical Inference for Linear Functionals of Online Least-squares SGD when $t \gtrsim d^{1+\delta}$
• BATIS: Bayesian Approaches for Targeted Improvement of Species Distribution Models
• A Graph Engine for Guitar Chord-Tone Soloing Education
• Explainable e-sports win prediction through Machine Learning classification in streaming
• RLIE: Rule Generation with Logistic Regression, Iterative Refinement, and Evaluation for Large Language Models
• Memo: Training Memory-Efficient Embodied Agents with Reinforcement Learning
• Misalignment Bounty: Crowdsourcing AI Agent Misbehavior
• Beyond Reactivity: Measuring Proactive Problem Solving in LLM Agents
• Benchmarking World-Model Learning
• A Unified Formal Theory on the Logical Limits of Symbol Grounding
• What is Implementation Science; and Why It Matters for Bridging the Artificial Intelligence Innovation-to-Application Gap in Medical Imaging
• LLM Bazaar: A Service Design for Supporting Collaborative Learning with an LLM-Powered Multi-Party Collaboration Infrastructure
• Contextual Augmentation for Entity Linking using Large Language Models
• Small Language Models Offer Significant Potential for Science Community
• CodeCRDT: Observation-Driven Coordination for Multi-Agent LLM Code Generation
• CosmoCore Affective Dream-Replay Reinforcement Learning for Code Generation
• AI for Distributed Systems Design: Scalable Cloud Optimization Through Repeated LLMs Sampling And Simulators
• Evaluating LLMs for Career Guidance: Comparative Analysis of Computing Competency Recommendations Across Ten African Countries
• DuoLens: A Framework for Robust Detection of Machine-Generated Multilingual Text and Code
• 3D Optimization for AI Inference Scaling: Balancing Accuracy, Cost, and Latency
• Improving Topic Modeling of Social Media Short Texts with Rephrasing: A Case Study of COVID-19 Related Tweets
• Learning from the Best, Differently: A Diversity-Driven Rethinking on Data Selection
• Large Connectome Model: An fMRI Foundation Model of Brain Connectomes Empowered by Brain-Environment Interaction in Multitask Learning Landscape
• Prospects for Using Artificial Intelligence to Understand Intrinsic Kinetics of Heterogeneous Catalytic Reactions
• ADPO: Anchored Direct Preference Optimization
• Context-aware Fairness Evaluation and Mitigation in LLMs
• MMAO-Bench: MultiModal All in One Benchmark Reveals Compositional Law between Uni-modal and Omni-modal in OmniModels
• Misinformation Detection using Large Language Models with Explainability
• Benchmarking On-Device Machine Learning on Apple Silicon with MLX
• Noise-corrected GRPO: From Noisy Rewards to Unbiased Gradients
• Application of Reduced-Order Models for Temporal Multiscale Representations in the Prediction of Dynamical Systems
• BAPO: Stabilizing Off-Policy Reinforcement Learning for LLMs via Balanced Policy Optimization with Adaptive Clipping
• A Justice Lens on Fairness and Ethics Courses in Computing Education: LLM-Assisted Multi-Perspective and Thematic Evaluation
• StutterZero and StutterFormer: End-to-End Speech Conversion for Stuttering Transcription and Correction
• NeuroAda: Activating Each Neuron's Potential for Parameter-Efficient Fine-Tuning
• ProfBench: Multi-Domain Rubrics requiring Professional Knowledge to Answer and Judge
• $\nabla$-SDF: Learning Euclidean Signed Distance Functions Online with Gradient-Augmented Octree Interpolation and Neural Residual
• Robust Driving QA through Metadata-Grounded Context and Task-Specific Prompts
• $\Delta$t-Mamba3D: A Time-Aware Spatio-Temporal State-Space Model for Breast Cancer Risk Prediction
• Plural Voices, Single Agent: Towards Inclusive AI in Multi-User Domestic Spaces
• Prior-informed optimization of treatment recommendation via bandit algorithms trained on large language model-processed historical records
• FlexiDataGen: An Adaptive LLM Framework for Dynamic Semantic Dataset Generation in Sensitive Domains
• CLiVR: Conversational Learning System in Virtual Reality with AI-Powered Patients
• "Over-the-Hood" AI Inclusivity Bugs and How 3 AI Product Teams Found and Fixed Them
• REPAIR Approach for Social-based City Reconstruction Planning in case of natural disasters
• PoSh: Using Scene Graphs To Guide LLMs-as-a-Judge For Detailed Image Descriptions
• Local Guidance for Configuration-Based Multi-Agent Pathfinding
• What Makes a Good Curriculum? Disentangling the Effects of Data Ordering on LLM Mathematical Reasoning
• That's Deprecated! Understanding, Detecting, and Steering Knowledge Conflicts in Language Models for Code Generation
• A Novel Approach to Breast Cancer Segmentation using U-Net Model with Attention Mechanisms and FedProx
• Steering Autoregressive Music Generation with Recursive Feature Machines
• A Cross-Environment and Cross-Embodiment Path Planning Framework via a Conditional Diffusion Model
• InvarGC: Invariant Granger Causality for Heterogeneous Interventional Time Series under Latent Confounding
• X-Ego: Acquiring Team-Level Tactical Situational Awareness via Cross-Egocentric Contrastive Video Representation Learning
• When Facts Change: Probing LLMs on Evolving Knowledge with evolveQA
• News-Aware Direct Reinforcement Trading for Financial Markets
• Imbalanced Gradients in RL Post-Training of Multi-Task LLMs
• Interpretable Question Answering with Knowledge Graphs
• PruneHal: Reducing Hallucinations in Multi-modal Large Language Models through Adaptive KV Cache Pruning
• Rethinking Driving World Model as Synthetic Data Generator for Perception Tasks
• An Active Diffusion Neural Network for Graphs
• No Intelligence Without Statistics: The Invisible Backbone of Artificial Intelligence
• SPOT: Scalable Policy Optimization with Trees for Markov Decision Processes
• See, Think, Act: Online Shopper Behavior Simulation with VLM Agents
• FnRGNN: Distribution-aware Fairness in Graph Neural Network
• LAPRAD: LLM-Assisted PRotocol Attack Discovery
• Social World Model-Augmented Mechanism Design Policy Learning
• Enhancing Early Alzheimer Disease Detection through Big Data and Ensemble Few-Shot Learning
• Knowledge and Common Knowledge of Strategies
• Collaborative penetration testing suite for emerging generative AI algorithms
• Online Handwritten Signature Verification Based on Temporal-Spatial Graph Attention Transformer
• Enabling Reconfiguration-Communication Overlap for Collective Communication in Optical Networks
• Balancing Rewards in Text Summarization: Multi-Objective Reinforcement Learning via HyperVolume Optimization
• SORA-ATMAS: Adaptive Trust Management and Multi-LLM Aligned Governance for Future Smart Cities
• Seabed-Net: A multi-task network for joint bathymetry estimation and seabed classification from remote sensing imagery in shallow waters
• Metadata Extraction Leveraging Large Language Models
• Every Attention Matters: An Efficient Hybrid Architecture for Long-Context Reasoning
• To Use or to Refuse? Re-Centering Student Agency with Generative AI in Engineering Design Education
• Foundation Model Forecasts: Form and Function
• A New Type of Adversarial Examples
• Learning To Defer To A Population With Limited Demonstrations
• M3-SLU: Evaluating Speaker-Attributed Reasoning in Multimodal Large Language Models
• AgenticMath: Enhancing LLM Reasoning via Agentic-based Math Data Generation
• The Massive Legal Embedding Benchmark (MLEB)
• ColorAgent: Building A Robust, Personalized, and Interactive OS Agent
• ToMMeR -- Efficient Entity Mention Detection from Large Language Models
• EchoFake: A Replay-Aware Dataset for Practical Speech Deepfake Detection
• Monitoring LLM-based Multi-Agent Systems Against Corruptions via Node Evaluation
• FairNet: Dynamic Fairness Correction without Performance Loss via Contrastive Conditional LoRA
• Neural Variational Dropout Processes
• Universal Quantitative Abstraction: Categorical Duality and Logical Completeness for Probabilistic Systems
• HybridEP: Scaling Expert Parallelism to Cross-Datacenter Scenario via Hybrid Expert/Data Transmission
• A Concrete Roadmap towards Safety Cases based on Chain-of-Thought Monitoring
• Graph Unlearning Meets Influence-aware Negative Preference Optimization
• KnowMol: Advancing Molecular Large Language Models with Multi-Level Chemical Knowledge
• VideoAgentTrek: Computer Use Pretraining from Unlabeled Videos
• Using Non-Expert Data to Robustify Imitation Learning via Offline Reinforcement Learning
• CARES: Context-Aware Resolution Selector for VLMs
• Modeling realistic human behavior using generative agents in a multimodal transport system: Software architecture and Application to Toulouse
• From Prototypes to Sparse ECG Explanations: SHAP-Driven Counterfactuals for Multivariate Time-Series Multi-class Classification
• Optimizing the Unknown: Black Box Bayesian Optimization with Energy-Based Model and Reinforcement Learning
• Insights into the Unknown: Federated Data Diversity Analysis on Molecular Data
• Demonstrating Real Advantage of Machine-Learning-Enhanced Monte Carlo for Combinatorial Optimization
• A Matter of Time: Revealing the Structure of Time in Vision-Language Models
• Multi-modal Co-learning for Earth Observation: Enhancing single-modality models via modality collaboration
• Detecting Latin in Historical Books with Large Language Models: A Multimodal Benchmark
• A Goal-Driven Survey on Root Cause Analysis
• XBench: A Comprehensive Benchmark for Visual-Language Explanations in Chest Radiography
• Human-Agent Collaborative Paper-to-Page Crafting for Under $0.1
• Style Attack Disguise: When Fonts Become a Camouflage for Adversarial Intent
• From Forecasting to Planning: Policy World Model for Collaborative State-Action Prediction
• Unraveling Emotions with Pre-Trained Models
• Study of Training Dynamics for Memory-Constrained Fine-Tuning
• I Spy With My Model's Eye: Visual Search as a Behavioural Test for MLLMs
• Directive, Metacognitive or a Blend of Both? A Comparison of AI-Generated Feedback Types on Student Engagement, Confidence, and Outcomes
• Are Large Language Models Sensitive to the Motives Behind Communication?
• Serverless GPU Architecture for Enterprise HR Analytics: A Production-Scale BDaaS Implementation
• Toward Agentic Software Engineering Beyond Code: Framing Vision, Values, and Vocabulary
• Do Prompts Reshape Representations? An Empirical Study of Prompting Effects on Embeddings
• Enabling Granular Subgroup Level Model Evaluations by Generating Synthetic Medical Time Series
• Learning Affordances at Inference-Time for Vision-Language-Action Models
• A Survey on Cache Methods in Diffusion Models: Toward Efficient Multi-Modal Generation
• SmartSwitch: Advancing LLM Reasoning by Overcoming Underthinking via Promoting Deeper Thought Exploration
• AdaSPEC: Selective Knowledge Distillation for Efficient Speculative Decoders
• On Controlled Change: Generative AI's Impact on Professional Authority in Journalism
• Test-time Verification via Optimal Transport: Coverage, ROC, & Sub-optimality
• Timely Clinical Diagnosis through Active Test Selection
• Rectifying Shortcut Behaviors in Preference-based Reward Learning
• The MUSE Benchmark: Probing Music Perception and Auditory Relational Reasoning in Audio LLMS
• A Multi-faceted Analysis of Cognitive Abilities: Evaluating Prompt Methods with Large Language Models on the CONSORT Checklist
• The Zero-Step Thinking: An Empirical Study of Mode Selection as Harder Early Exit in Reasoning Models
• WebGraphEval: Multi-Turn Trajectory Evaluation for Web Agents using Graph Representation
• ChatGPT Unveils Its Limits: Principles of Law Deliver Checkmate
• An Argumentative Explanation Framework for Generalized Reason Model with Inconsistent Precedents
• Learning to Make Friends: Coaching LLM Agents toward Emergent Social Ties
• Continual Knowledge Adaptation for Reinforcement Learning
• MSC-Bench: A Rigorous Benchmark for Multi-Server Tool Orchestration
• NeSyPr: Neurosymbolic Proceduralization For Efficient Embodied Reasoning
• DAIL: Beyond Task Ambiguity for Language-Conditioned Reinforcement Learning
• HSCodeComp: A Realistic and Expert-level Benchmark for Deep Search Agents in Hierarchical Rule Application
• AgentSense: LLMs Empower Generalizable and Explainable Web-Based Participatory Urban Sensing