Papers

New on arXiv

Fri 13 Feb 2026

340 papers

Association Is Not Similarity: Learning Corpus-Specific Associations for Multi-Hop RetrievalLLM reasoning → Knowledge- and tool-augmented reasoning
Evolving Demonstration Optimization for Chain-of-Thought Feature TransformationLLM reasoning → Reinforcement learning for reasoning
Information-Theoretic Privacy Control for Sequential Multi-Agent LLM SystemsLLM agents and tool use → Agent threats, guardrails and governance
Beyond Binary Preferences: A Principled Framework for Reward Modeling with Ordinal FeedbackLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF
Generalized Discrete Diffusion with Self-CorrectionDiffusion models → Discrete and text diffusion
Safety Training May Persist Through Helpfulness Optimization in LLM AgentsLLM agents and tool use → Agent threats, guardrails and governance +1
Learning to Remember: End-to-End Training of Memory Agents for Long-Context ReasoningLLM reasoning → Reasoning for agents, planning and code +2
Doc-to-LoRA: Learning to Instantly Internalize ContextsLLM reasoning → Reasoning benchmarks and evaluation
Safe-SDL:Establishing Safety Boundaries and Control Mechanisms for AI-Driven Self-Driving LaboratoriesEnd-to-end autonomous driving → Large language models for driving
Frequency-Enhanced Hilbert Scanning Mamba for Short-Term Arctic Sea Ice Concentration PredictionState-space sequence models (S4, Mamba) → Time series forecasting with SSMs
Arming Data Agents with Tribal KnowledgeLLM agents and tool use → Web, search, and research agents
Think Deep, Not Just Long: Measuring LLM Reasoning Effort via Deep-Thinking TokensLLM reasoning → Test-time compute and efficient thinking
SPILLage: Agentic Oversharing on the WebLLM agents and tool use → Web, search, and research agents
SpargeAttention2: Trainable Sparse Attention via Hybrid Top-k+Top-p Masking and Distillation Fine-TuningDiffusion models → Video generation
OMNI-LEAK: Orchestrator Multi-Agent Network Induced Data LeakageLLM agents and tool use → Agent threats, guardrails and governance +1
NeuroWeaver: An Autonomous Evolutionary Agent for Exploring the Programmatic Space of EEG Analysis PipelinesLLM agents and tool use → Medical and scientific domain agents
How Multimodal Large Language Models Support Access to Visual Information: A Diary Study With Blind and Low Vision PeopleMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Backdooring Bias in Large Language ModelsLLM safety and security → Backdoor and data poisoning attacks
On-Policy Supervised Fine-Tuning for Efficient ReasoningLLM reasoning → Reinforcement learning for reasoning
Unsafer in Many Turns: Benchmarking and Defending Multi-Turn Safety Risks in Tool-Using AgentsLLM agents and tool use → Agent threats, guardrails and governance
An Online Reference-Free Evaluation Framework for Flowchart Image-to-Code GenerationMultimodal LLMs / vision-language models → Chart understanding and anomaly detection
G2CP: A Graph-Grounded Communication Protocol for Verifiable and Efficient Multi-Agent ReasoningLLM reasoning → Reasoning for agents, planning and code +1
Nanbeige4.1-3B: A Small General Model that Reasons, Aligns, and ActsLLM reasoning → Reinforcement learning for reasoning +2
Assessing Spear-Phishing Website Generation in Large Language Model Coding AgentsLLM agents and tool use → Agent threats, guardrails and governance
The Diffusion Duet: Harmonizing Dual Channels with Wavelet Suppression for Image SeparationDiffusion models → Inverse problems and image restoration
AdaCorrection: Adaptive Offset Cache Correction for Accurate Diffusion TransformersDiffusion models → Diffusion foundations and sampling
Steerable Vision-Language-Action Policies for Embodied Reasoning and Hierarchical ControlLLM reasoning → Spatial and embodied reasoning +2
Fix Before Search: Benchmarking Agentic Query Visual Pre-processing in Multimodal Retrieval-augmented GenerationRetrieval-augmented generation → Multimodal RAG +1
Realistic Face Reconstruction from Facial Embeddings via Diffusion ModelsDiffusion models → Robustness, privacy and adversarial aspects
Asynchronous Verified Semantic Caching for Tiered LLM ArchitecturesLLM agents and tool use → Software engineering coding agents
In-Context Autonomous Network Incident Response: An End-to-End Large Language Model Agent ApproachLLM agents and tool use → Agent threats, guardrails and governance
Quantization-Robust LLM Unlearning via Low-Rank AdaptationLLM safety and security → Machine unlearning
Order Matters in Retrosynthesis: Structure-aware Generation via Reaction-Center-Guided Discrete Flow MatchingDiffusion models → Flow matching, time series and science
Awakening Dormant Users: Generative Recommendation with Counterfactual Functional Role ReasoningLLM reasoning → LLM reasoning for recommendation
Automating UI Optimization through Multi-Agentic ReasoningLLM agents and tool use → GUI and mobile agents
R-Diverse: Mitigating Diversity Illusion in Self-Play LLM TrainingLLM reasoning → Reinforcement learning for reasoning +1
Consistency of Large Reasoning Models Under Multi-Turn AttacksLLM reasoning → Reasoning for agents, planning and code
UniManip: General-Purpose Zero-Shot Robotic Manipulation with Agentic Operational GraphLLM agents and tool use → Multimodal and embodied agents
Agentic AI for Robot Control: Flexible but still FragileLLM agents and tool use → Multimodal and embodied agents
Native Extrapolation Awareness in Flow-Based Conditional GenerationDiffusion models → Flow matching, time series and science
Curriculum-DPO++: Direct Preference Optimization via Data and Model Curricula for Text-to-Image GenerationLLM alignment / RLHF / preference optimisation → Diffusion and multimodal model alignment
How Swarms Differ: Challenges in Collective Behaviour ComparisonOffline RL and imitation learning from data → Imitation learning methods
SciAgentGym: Benchmarking Multi-Step Scientific Tool-use in LLM AgentsLLM reasoning → Reasoning for agents, planning and code +1
Learning Native Continuation for Action Chunking Flow PoliciesVision-language-action models → Efficient VLA inference and adaptation
RGAlign-Rec: Ranking-Guided Alignment for Latent Query Reasoning in Recommendation SystemsLLM reasoning → LLM reasoning for recommendation
HSD: Training-Free Acceleration for Document Parsing Vision-Language Models with Hierarchical Speculative DecodingMultimodal LLMs / vision-language models → Document understanding and multimodal RAG +1
Human Tool: An MCP-Style Framework for Human-Agent CollaborationLLM agents and tool use → General LLM agent frameworks and benchmarks
Neighborhood Blending: A Lightweight Inference-Time Defense Against Membership Inference AttacksLLM safety and security → Membership inference attacks
TFTF: Training-Free Targeted Flow for Conditional SamplingDiffusion models → Flow matching, time series and science
Never say never: Exploring the effects of available knowledge on agent persuasiveness in controlled physiotherapy motivation dialoguesLLM agents and tool use → Social simulation and multi-agent debate
Reliable Thinking with ImagesLLM reasoning → Multimodal reasoning +1
RADAR: Revealing Asymmetric Development of Abilities in MLLM Pre-trainingLLM reasoning → Multimodal reasoning +1
BaziQA-Benchmark: Evaluating Symbolic and Temporally Compositional Reasoning in Large Language ModelsLLM reasoning → Reasoning benchmarks and evaluation
BrowseComp-V3: A Visual, Vertical, and Verifiable Benchmark for Multimodal Browsing AgentsMultimodal LLMs / vision-language models → Document understanding and multimodal RAG +1
WebClipper: Efficient Evolution of Web Agents with Graph-based Trajectory PruningLLM agents and tool use → Web, search, and research agents
Amortized Reasoning Tree Search: Decoupling Proposal and Decision in Large Language ModelsLLM reasoning → Reinforcement learning for reasoning
Vital Trace: Protocol-Constrained Patient-State Reasoning for Longitudinal Clinical TrajectoriesLLM reasoning → Clinical and biomedical applications
FLAC: Maximum Entropy RL via Kinetic Energy Regularized Bridge MatchingDiffusion models → Diffusion and flow policies for robotics
PixelRush: Ultra-Fast, Training-Free High-Resolution Image Generation via One-step DiffusionDiffusion models → Diffusion foundations and sampling
Closing the Loop: A Control-Theoretic Framework for Provably Stable Time Series Forecasting with LLMsLLM reasoning → Time series reasoning
Towards reconstructing experimental sparse-view X-ray CT data with diffusion modelsDiffusion models → Medical imaging and segmentation
VimRAG: Navigating Massive Visual Context in Retrieval-Augmented Generation via Multimodal Memory GraphRetrieval-augmented generation → Multimodal RAG +1
Scaling Single Human Demonstrations for Imitation Learning using Generative Foundational ModelsOffline RL and imitation learning from data → Robot manipulation imitation learning
Training Dense Retrievers with Multiple Positive PassagesRetrieval-augmented generation → Generation and context integration
ADEPT: RL-Aligned Agentic Decoding of Emotion via Evidence Probing Tools -- From Consensus Learning to Ambiguity-Driven Emotion ReasoningLLM reasoning → Audio-language model reasoning +1
ReFilter: Improving Robustness of Retrieval-Augmented Generation via Gated FilterRetrieval-augmented generation → Generation and context integration
Mixture of Predefined Experts: Maximizing Data Usage on Vertical Federated LearningMixture-of-experts → Federated mixture-of-experts
MedXIAOHE: A Comprehensive Recipe for Building Medical MLLMsMultimodal LLMs / vision-language models → Medical vision-language models
ALOE: Action-Level Off-Policy Evaluation for Vision-Language-Action Model Post-TrainingVision-language-action models → Reinforcement learning and post-training
Xiaomi-Robotics-0: An Open-Sourced Vision-Language-Action Model with Real-Time ExecutionVision-language-action models → Generalist VLA models and benchmarks
Flow Matching from Viewpoint of Proximal OperatorsDiffusion models → Flow matching, time series and science
Motion Prior Distillation in Time Reversal Sampling for Generative InbetweeningDiffusion models → Video generation
SLA2: Sparse-Linear Attention with Learnable Routing and QATDiffusion models → Video generation
SkillsBench: Benchmarking How Well Agent Skills Work Across Diverse TasksLLM agents and tool use → Agent skill learning and evolution
Evaluating Robustness of Reasoning Models on Parameterized Logical ProblemsLLM reasoning → Reasoning benchmarks and evaluation
Think Fast and Slow: Step-Level Cognitive Depth Adaptation for LLM AgentsLLM reasoning → Reasoning for agents, planning and code +1
Learning Ordinal Probabilistic Reward from PreferencesLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF
IndicFairFace: Balanced Indian Face Dataset for Auditing and Mitigating Geographical Bias in Vision-Language ModelsMultimodal LLMs / vision-language models → Social bias and fairness in VLMs
Beyond Normalization: Rethinking the Partition Function as a Difficulty Scheduler for RLVRLLM reasoning → Reinforcement learning for reasoning
ImageRAGTurbo: Towards One-step Text-to-Image Generation with Retrieval-Augmented Diffusion ModelsRetrieval-augmented generation → Multimodal RAG +1
AI Agents for Inventory Control: Human-LLM-OR ComplementarityLLM agents and tool use → Social simulation and multi-agent debate
Beyond Imitation: Reinforcement Learning-Based Sim-Real Co-Training for VLA ModelsVision-language-action models → Reinforcement learning and post-training
Formalizing the Sampling Design Space of Diffusion-Based Generative Models via Adaptive Solvers and Wasserstein-Bounded TimestepsDiffusion models → Diffusion foundations and sampling
Vision Token Reduction via Attention-Driven Self-Compression for Efficient Multimodal Large Language ModelsMultimodal LLMs / vision-language models → Visual token pruning and efficiency +1
When Environments Shift: Safe Planning with Generative Priors and Robust Conformal PredictionDiffusion models → Diffusion and flow policies for robotics
Self-EvolveRec: Self-Evolving Recommender Systems with LLM-based Directional FeedbackSynthetic data and self-improvement for LLMs → Self-evolving LLM agents
Multi-Head Attention as a Source of Catastrophic Forgetting in MoE TransformersMixture-of-experts → Mixture of LoRA experts
Can I Have Your Order? Monte-Carlo Tree Search for Slot Filling Ordering in Diffusion Language ModelsDiffusion models → Discrete and text diffusion
VI-CuRL: Stabilizing Verifier-Independent RL Reasoning via Confidence-Guided Variance ReductionLLM reasoning → Reinforcement learning for reasoning
Monte Carlo Tree Search with Reasoning Path Refinement for Small Language Models in Conversational Text-to-NoSQLLLM reasoning → Knowledge- and tool-augmented reasoning
To Mix or To Merge: Toward Multi-Domain Reinforcement Learning for Large Language ModelsLLM reasoning → Reinforcement learning for reasoning
The Constant Eye: Benchmarking and Bridging Appearance Robustness in Autonomous DrivingEnd-to-end autonomous driving → End-to-end system architecture
SD-MoE: Spectral Decomposition for Effective Expert SpecializationMixture-of-experts → MoE large language models
Decoder-only Conformer with Modality-aware Sparse Mixtures of Experts for ASRMixture-of-experts → Multimodal MoE models
Scaling Web Agent Training through Automatic Data Generation and Fine-grained EvaluationLLM agents and tool use → Web, search, and research agents
Self-Supervised JEPA-based World Models for LiDAR Occupancy Completion and ForecastingWorld models → World models for autonomous driving
AMPS: Adaptive Modality Preference Steering via Functional EntropyMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
CRAFT: Adapting VLA Models to Contact-rich Manipulation via Force-aware Curriculum Fine-tuningVision-language-action models → Tactile, force-aware and humanoid VLA
Reasoning to Rank: An End-to-End Solution for Exploiting Large Language Models for RecommendationLLM reasoning → LLM reasoning for recommendation
Flow-Factory: A Unified Framework for Reinforcement Learning in Flow-Matching ModelsDiffusion models → Diffusion and flow policies for robotics

All 340