Papers

New on arXiv

Tue 16 Sep 2025

391 papers

Optimizing Agricultural Research: A RAG-Based Approach to Mycorrhizal Fungi InformationRetrieval-augmented generation → Time series and agricultural RAG
Membership Inference Attack against Large Language Model-based Recommendation Systems: A New Distillation-based ParadigmLLM safety and security → Membership inference attacks
Linear probes rely on textual evidence: Results from leakage mitigation studies in language modelsMechanistic interpretability → Probing deception, bias and beliefs
LATTS: Locally Adaptive Test-Time ScalingLLM reasoning → Test-time compute and efficient thinking
FastMTP: Accelerating LLM Inference with Enhanced Multi-Token PredictionEfficient LLM inference and serving → Speculative decoding
MiniCPM-V 4.5: Cooking Efficient MLLMs via Architecture, Data, and Training RecipeMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
Improving MLLM Training Efficiency via Stage-Aware SparsityMultimodal LLMs / vision-language models → Visual token pruning and efficiency
A Multi-Agent LLM Defense Pipeline Against Prompt Injection AttacksLLM agents and tool use → Agent threats, guardrails and governance +1
The Sum Leaks More Than Its Parts: Compositional Privacy Risks and Mitigations in Multi-Agent CollaborationLLM agents and tool use → Agent threats, guardrails and governance
Towards Robust Agentic CUDA Kernel Benchmarking, Verification, and OptimizationLLM reasoning → Reasoning for agents, planning and code +1
Agentic JWT: A Secure Delegation Protocol for Autonomous AI AgentsLLM agents and tool use → Agent threats, guardrails and governance
CoBRA: Programming Cognitive Bias in Social Agents Using Classic Social Science ExperimentsLLM agents and tool use → Social simulation and multi-agent debate
Zero-sum turn games using Q-learning: finite computation with security guaranteesDeep reinforcement learning methods → RL for games and combinatorial optimization
TreeIRL: Safe Urban Driving with Tree Search and Inverse Reinforcement LearningEnd-to-end autonomous driving → Imitation-learning end-to-end planners +2
Adaptive Multi-Mode Out-of-Distribution Detection for Trajectory Prediction in Autonomous VehiclesEnd-to-end autonomous driving → End-to-end system architecture
Cross-Distribution Diffusion Priors-Driven Iterative Reconstruction for Sparse-View CTDiffusion models → Medical imaging and segmentation
Dense-Jump Flow Matching with Non-Uniform Time Scheduling for Robotic Policies: Mitigating Multi-Step Inference DegradationDiffusion models → Diffusion and flow policies for robotics
Using Visual Language Models to Control Bionic Hands: Assessment of Object Perception and Grasp InferenceMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
AI Agents with Human-Like Collaborative Tools: Adaptive Strategies for Enhanced Problem-SolvingLLM agents and tool use → General LLM agent frameworks and benchmarks
ColonCrafter: A Depth Estimation Model for Colonoscopy Videos Using Diffusion PriorsDiffusion models → Medical imaging and segmentation
AERIS: Argonne Earth Systems Model for Reliable and Skillful PredictionsDiffusion models → Flow matching, time series and science
DEFT-VTON: Efficient Virtual Try-On with Consistent Generalised H-TransformDiffusion models → 3D, point cloud and scene generation
BiasMap: Leveraging Cross-Attentions to Discover and Mitigate Hidden Social Biases in Text-to-Image GenerationDiffusion models → Image generation, text-to-image and editing
An LLM Agentic Approach for Legal-Critical Software: A Case Study for Tax Prep SoftwareLLM agents and tool use → Software engineering coding agents
SteeringSafety: Benchmarking Representation Steering in LLMs Across Safety PerspectivesMechanistic interpretability → Steering and probing for LLM safety
DuetUI: A Bidirectional Context Loop for Human-Agent Co-Generation of Task-Oriented InterfacesLLM agents and tool use → General LLM agent frameworks and benchmarks
EdiVal-Agent: An Object-Centric Framework for Automated, Fine-Grained Evaluation of Multi-Turn EditingLLM agents and tool use → Multimodal and embodied agents
TICL: Text-Embedding KNN For Speech In-Context Learning Unlocks Speech Recognition Abilities of Large Multimodal ModelsMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
From Next Token Prediction to (STRIPS) World ModelsWorld models → World models for language agents
VEGA: Electric Vehicle Navigation Agent via Physics-Informed Neural Operator and Proximal Policy OptimizationDeep reinforcement learning methods → RL for energy, networks and operations
ASTREA: Introducing Agentic Intelligence for Orbital Thermal AutonomyLLM agents and tool use → General LLM agent frameworks and benchmarks
The Art of Saying "Maybe": A Conformal Lens for Uncertainty Benchmarking in VLMsMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
An Empirical Analysis of VLM-based OOD Detection: Mechanisms, Advantages, and SensitivityMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
Agent2: An Agent-Generates-Agent Framework for Reinforcement Learning AutomationLLM agents and tool use → Reinforcement learning for LLM agents +1
3D Aware Region Prompted Vision Language ModelMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Do Activation Verbalization Methods Convey Privileged Information?Mechanistic interpretability → Knowledge and representations inside LLMs
ReSum: Unlocking Long-Horizon Search Intelligence via Context SummarizationLLM agents and tool use → Agent memory systems
Towards General Agentic Intelligence via Environment ScalingLLM agents and tool use → General LLM agent frameworks and benchmarks
Scaling Agents via Continual Pre-trainingLLM agents and tool use → General LLM agent frameworks and benchmarks
WebResearcher: Unleashing unbounded reasoning capability in Long-Horizon AgentsLLM reasoning → Knowledge- and tool-augmented reasoning
WebSailor-V2: Bridging the Chasm to Proprietary Agents via Synthetic Data and Scalable Reinforcement LearningLLM agents and tool use → Web, search, and research agents +2
Post-Hoc Split-Point Self-Consistency Verification for Efficient, Unified Quantification of Aleatoric and Epistemic Uncertainty in Deep LearningLLM reasoning → Test-time compute and efficient thinking
Evaluating LLM Alignment on Personality Inference from Real-World Interview DataLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
Metacognitive Reuse: Turning Recurring LLM Reasoning Into Concise BehaviorsLLM reasoning → Test-time compute and efficient thinking
Simulating Clinical AI Assistance using Multimodal LLMs: A Case Study in Diabetic RetinopathyMultimodal LLMs / vision-language models → Medical vision-language models
Single-stream Policy OptimizationLLM reasoning → Reinforcement learning for reasoning
End4: End-to-end Denoising Diffusion for Diffusion-Based Inpainting DetectionDiffusion models → 3D, point cloud and scene generation
More performant and scalable: Rethinking contrastive vision-language pre-training of radiology in the LLM eraLLM reasoning → Clinical and biomedical applications +1
TeraSim-World: Worldwide Safety-Critical Data Synthesis for End-to-End Autonomous DrivingEnd-to-end autonomous driving → Generative simulation for driving
Enhancing Video Large Language Models with Structured Multi-Video Collaborative ReasoningLLM reasoning → Multimodal reasoning
FinSearchComp: Towards a Realistic, Expert-Level Evaluation of Financial Search and ReasoningLLM reasoning → Reasoning for agents, planning and code +1
MSDNet: Efficient 4D Radar Super-Resolution via Multi-Stage DistillationKnowledge distillation → Detection and dense prediction distillation
Agentic AI for Financial Crime ComplianceLLM agents and tool use → Financial trading and forecasting agents
Discovering Mathematical Equations with Diffusion Language ModelDiffusion models → Diffusion foundations and sampling
Reasoning with Preference Constraints: A Benchmark for Language Models in Many-to-One Matching MarketsLLM reasoning → Reasoning benchmarks and evaluation
Empowering Multi-Robot Cooperation via Sequential World ModelsWorld models → Latent dynamics models and model-based RL +1
GLAD: Global-Local Aware Dynamic Mixture-of-Experts for Multi-Talker ASRMixture-of-experts → Multimodal MoE models
Shaping Explanations: Semantic Reward Modeling with Encoder-Only Transformers for GRPOLLM reasoning → Reinforcement learning for reasoning +1
When Inverse Data Outperforms: Exploring the Pitfalls of Mixed Data in Multi-Stage Fine-TuningLLM alignment / RLHF / preference optimisation → Process reward models for reasoning
HERO: Rethinking Visual Token Early Dropping in High-Resolution Large Vision-Language ModelsLLM reasoning → Multimodal reasoning +1
MIA-EPT: Membership Inference Attack via Error Prediction for Tabular DataDiffusion models → Robustness, privacy and adversarial aspects +1
Perception Before Reasoning: Two-Stage Reinforcement Learning for Visual Reasoning in Vision-Language ModelsLLM reasoning → Multimodal reasoning +1
xOffense: An Autonomous Multi-Agent Framework for Penetration Testing with Domain-Adapted Large Language ModelsLLM agents and tool use → Agent threats, guardrails and governance
A Visualized Framework for Event Cooperation with Generative AgentsLLM agents and tool use → Social simulation and multi-agent debate
ReTrack: Data Unlearning in Diffusion Models through Redirecting the Denoising TrajectoryDiffusion models → Robustness, privacy and adversarial aspects +1
Dual-Stage Reweighted MoE for Long-Tailed Egocentric Mistake DetectionMixture-of-experts → Vision mixture-of-experts
Jailbreaking Large Language Models Through Content ConcretizationLLM safety and security → Jailbreaking
Rethinking the Evaluation of Alignment Methods: Insights into Diversity, Generalisation, and SafetyLLM reasoning → Reinforcement learning for reasoning +1
The Anatomy of Alignment: Decomposing Preference Optimization by Steering Sparse FeaturesLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
HLSMAC: A New StarCraft Multi-Agent Challenge for High-Level Strategic Decision-MakingLLM agents and tool use → Reinforcement learning for LLM agents
All Roads Lead to Rome: Graph-Based Confidence Estimation for Large Language Model ReasoningLLM reasoning → Test-time compute and efficient thinking
Cross-Layer Vision Smoothing: Enhancing Visual Understanding via Sustained Focus on Key Objects in Large Vision-Language ModelsMultimodal LLMs / vision-language models → Visual token pruning and efficiency
Runge-Kutta Approximation and Decoupled Attention for Rectified Flow Inversion and Semantic EditingDiffusion models → Image generation, text-to-image and editing
LTA-thinker: Latent Thought-Augmented Training Framework for Large Language Models on Complex ReasoningLLM reasoning → Test-time compute and efficient thinking
Tool-R1: Sample-Efficient Reinforcement Learning for Agentic Tool UseLLM agents and tool use → Reinforcement learning for LLM agents
GRATE: a Graph transformer-based deep Reinforcement learning Approach for Time-efficient autonomous robot ExplorationDeep reinforcement learning methods → RL for robot navigation and UAV control
Contrastive Representation Learning for Robust Sim-to-Real Transfer of Adaptive Humanoid LocomotionDeep reinforcement learning methods → RL for robot locomotion and manipulation
Bayesian Signal Separation via Plug-and-Play Diffusion-Within-Gibbs SamplingDiffusion models → Inverse problems and image restoration
Safe Reinforcement Learning using Action Projection: Safeguard the Policy or the Environment?Deep reinforcement learning methods → Safe and constrained RL
DiffHash: Text-Guided Targeted Attack via Diffusion Models against Deep Hashing Image RetrievalDiffusion models → Robustness, privacy and adversarial aspects
H2R: Hierarchical Hindsight Reflection for Multi-Task LLM AgentsLLM agents and tool use → Agent memory systems
CECT-Mamba: a Hierarchical Contrast-enhanced-aware Model for Pancreatic Tumor Subtyping from Multi-phase CECTState-space sequence models (S4, Mamba) → Mamba for medical image analysis
Integrating Trajectory Optimization and Reinforcement Learning for Quadrupedal Jumping with Terrain-Adaptive LandingDeep reinforcement learning methods → RL for robot locomotion and manipulation
MEGAN: Mixture of Experts for Robust Uncertainty Estimation in Endoscopy VideosMixture-of-experts → Medical and biomedical MoE
InfoGain-RAG: Boosting Retrieval-Augmented Generation via Document Information Gain-based Reranking and FilteringRetrieval-augmented generation → Generation and context integration
Myopic Optimality: why reinforcement learning portfolio management strategies lose moneyDeep reinforcement learning methods → Adversarial robustness and RL in finance
What Makes a Good Generated Image? Investigating Human and Multimodal LLM Image Preference AlignmentMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning +1
Zero-shot Graph Reasoning via Retrieval Augmented Framework with LLMsLLM reasoning → Knowledge- and tool-augmented reasoning +1
Generalizable Holographic Reconstruction via Amplitude-Only Diffusion PriorsDiffusion models → Inverse problems and image restoration
Defense-to-Attack: Bypassing Weak Defenses Enables Stronger Jailbreaks in Vision-Language ModelsLLM safety and security → Jailbreaking
SPGen: Spherical Projection as Consistent and Flexible Representation for Single Image 3D Shape GenerationDiffusion models → 3D, point cloud and scene generation
HistoryBankQA: Multilingual Temporal Question Answering on Historical EventsLLM reasoning → Knowledge- and tool-augmented reasoning
EvoEmpirBench: Dynamic Spatial Reasoning with Agent-ExpVerLLM reasoning → Spatial and embodied reasoning
AsyMoE: Leveraging Modal Asymmetry for Enhanced Expert Specialization in Large Vision-Language ModelsMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning +1
Mitigating Strategy Preference Bias in Emotional Support Conversation via Uncertainty EstimationsLLM alignment / RLHF / preference optimisation → Medical LLM alignment
Large Language Models Imitate Logical Reasoning, but at what Cost?LLM reasoning → Reasoning benchmarks and evaluation
Maps for Autonomous Driving: Full-process Survey and FrontiersEnd-to-end autonomous driving → End-to-end system architecture
DoubleAgents: Human-Agent Alignment in a Socially Embedded WorkflowLLM agents and tool use → General LLM agent frameworks and benchmarks
GBV-SQL: Guided Generation and SQL2Text Back-Translation Validation for Multi-Agent Text2SQLLLM agents and tool use → Web, search, and research agents
Analogy-Driven Financial Chain-of-Thought (AD-FCoT): A Prompting Approach for Financial Sentiment AnalysisLLM reasoning → Reasoning for agents, planning and code

All 391