Papers

New on arXiv

Wed 15 Apr 2026

402 papers

TCAR-Gen: Temporal Graph Retrieval with Evidence Fusion for Knowledge-Grounded GenerationLLM reasoning → Knowledge- and tool-augmented reasoning +1
A Multi-Domain Red Teaming Framework for Safety, Robustness, and Fairness Evaluation of Medical Large Language ModelsLLM safety and security → Safety alignment, refusal and guardrails
TrustLDM: Benchmarking Trustworthiness in Language Diffusion ModelsDiffusion models → Discrete and text diffusion
NaP-Control: Navigating Diffusion Prior for Versatile and Fast Character ControlDiffusion models → Diffusion and flow policies for robotics
Agentic Graph Retrieval-Augmented Generation for Auditable Commercial Registry AnalysisRetrieval-augmented generation → Knowledge-graph and structured-knowledge RAG +1
BackFlush: Knowledge-Free Backdoor Detection and Elimination with Watermark Preservation in Large Language ModelsLLM safety and security → Backdoor and data poisoning attacks
HadAgent: Harness-Aware Decentralized Agentic AI Serving with Proof-of-Inference Blockchain ConsensusLLM agents and tool use → Agent threats, guardrails and governance
Positive-Only Drifting Policy OptimizationDeep reinforcement learning methods → RL with diffusion and flow models
SmoGVLM: A Small, Graph-enhanced Vision-Language ModelMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Penny Wise, Pixel Foolish: Bypassing Price Constraints in Multimodal Agents via Visual Adversarial PerturbationsMultimodal LLMs / vision-language models → Chart understanding and anomaly detection
BARD: Bridging AutoRegressive and Diffusion Vision-Language Models Via Highly Efficient Progressive Block Merging and Stage-Wise DistillationMultimodal LLMs / vision-language models → Image editing and quality assessment
Seeing the imagined: latent functional alignment in visual imagery decoding from fMRI dataDiffusion models → Medical imaging and segmentation
An Agentic Workflow for Detecting Personally Identifiable Information in Crash NarrativesLLM agents and tool use → Agent threats, guardrails and governance
LogJack: Indirect Prompt Injection Through Cloud Logs Against LLM Debugging AgentsLLM safety and security → Prompt injection and agent security
SoK: Security of Autonomous LLM Agents in Agentic CommerceLLM agents and tool use → Agent threats, guardrails and governance
Seeing Through Circuits: Faithful Mechanistic Interpretability for Vision TransformersMechanistic interpretability → Circuits and mechanistic analysis
Evo-MedAgent: Beyond One-Shot Diagnosis with Agents That Remember, Reflect, and ImproveLLM agents and tool use → Medical and scientific domain agents
Scouting By Reward: VLM-TO-IRL-Driven Player Selection For EsportsOffline RL and imitation learning from data → Imitation learning methods
AIBuildAI: An AI Agent for Automatically Building AI ModelsLLM agents and tool use → Web, search, and research agents
CooperDrive: Enhancing Driving Decisions Through Cooperative PerceptionEnd-to-end autonomous driving → Cooperative and V2X driving
Robustness Analysis of Machine Learning Models for IoT Intrusion Detection Under Data Poisoning AttacksLLM safety and security → Backdoor and data poisoning attacks
Geometric Routing Enables Causal Expert Control in Mixture of ExpertsMixture-of-experts → Expert routing strategies
Demonstration of Pneuma-Seeker: Agentic System for Reifying and Fulfilling Information Needs on Tabular DataLLM agents and tool use → Web, search, and research agents
Equifinality in Mixture of Experts: Routing Topology Does Not Determine Language Modeling QualityMixture-of-experts → Expert routing strategies
Empirical Evaluation of Policy-Based Reinforcement Learning for Dynamic Service Control in an M/M/1 QueueDeep reinforcement learning methods → RL for energy, networks and operations
A Unified Model and Document Representation for On-Device Retrieval-Augmented GenerationRetrieval-augmented generation → RAG pipeline design and evaluation
Credo: Declarative Control of LLM Pipelines via Beliefs and PoliciesLLM agents and tool use → General LLM agent frameworks and benchmarks
SpaceMind: A Modular and Self-Evolving Embodied Vision-Language Agent Framework for Autonomous On-orbit ServicingLLM agents and tool use → Multimodal and embodied agents +1
Timescale Separation Enables Deep Reinforcement Learning Control of Rotating Detonation Engine Mode TransitionsDeep reinforcement learning methods → RL theory and sample complexity
Coalition Formation in LLM Agent Networks: Stability Analysis and Convergence GuaranteesLLM agents and tool use → Social simulation and multi-agent debate
Step-level Denoising-time Diffusion Alignment with Multiple ObjectivesDiffusion models → Diffusion and flow policies for robotics +1
The Cost of Language: Centroid Erasure Exposes and Exploits Modal Competition in Multimodal Language ModelsMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
APEX-MEM: Agentic Semi-Structured Memory with Temporal Reasoning for Long-Term Conversational AILLM reasoning → Reasoning for agents, planning and code +1
Shuffle the Context: RoPE-Perturbed Self-Distillation for Long-Context AdaptationRetrieval-augmented generation → Generation and context integration +1
Mamba-SSM with LLM Reasoning for Feature Selection: Faithfulness-Aware Biomarker DiscoveryLLM reasoning → Clinical and biomedical applications +1
Thermodynamic Diffusion Inference with Minimal Digital ConditioningDiffusion models → Diffusion foundations and sampling
Challenges and Future Directions in Agentic Reverse Engineering SystemsLLM agents and tool use → Agent threats, guardrails and governance
Seeing Through Experts Eyes A Foundational Vision Language Model Trained on Radiologists Gaze and ReasoningLLM reasoning → Clinical and biomedical applications +1
DharmaOCR: Specialized Small Language Models for Structured OCR that outperform Open-Source and Commercial BaselinesLLM alignment / RLHF / preference optimisation → Diffusion and multimodal model alignment
Aerial Multi-Functional RIS in Fluid Antennas-Aided Full-Duplex Networks: A Self-Optimized Hybrid Deep Reinforcement Learning ApproachDeep reinforcement learning methods → RL for robot navigation and UAV control
HY-World 2.0: A Multi-Modal World Model for Reconstructing, Generating, and Simulating 3D WorldsWorld models → Video-generation world models and simulators
Enhancing LLM-based Search Agents via Contribution Weighted Group Relative Policy OptimizationLLM reasoning → Reinforcement learning for reasoning +1
Reinforcement Learning via Value Gradient FlowOffline RL and imitation learning from data → Diffusion and flow policies for offline RL +1
GUI-Perturbed: Domain Randomization Reveals Systematic Brittleness in GUI Grounding ModelsLLM reasoning → Multimodal reasoning
ReviewGrounder: Improving Review Substantiveness with Rubric-Guided, Tool-Integrated AgentsLLM agents and tool use → Web, search, and research agents
Evaluation of Agents under Simulated AI Marketplace DynamicsLLM agents and tool use → General LLM agent frameworks and benchmarks
Calibrate-Then-Delegate: Safety Monitoring with Risk and Budget Guarantees via Model CascadesLLM safety and security → Safety alignment, refusal and guardrails
Awakening Dormant Experts:Counterfactual Routing to Mitigate MoE HallucinationsMixture-of-experts → Expert routing strategies
One Token per Highly Selective Frame: Towards Extreme Compression for Long Video UnderstandingMultimodal LLMs / vision-language models → Video understanding with MLLMs
ROSE: Retrieval-Oriented Segmentation EnhancementMultimodal LLMs / vision-language models → Document understanding and multimodal RAG
SpatialEvo: Self-Evolving Spatial Intelligence via Deterministic Geometric EnvironmentsLLM reasoning → Spatial and embodied reasoning
From P(y|x) to P(y): Investigating Reinforcement Learning in Pre-train SpaceLLM reasoning → Reinforcement learning for reasoning +1
LongCoT: Benchmarking Long-Horizon Chain-of-Thought ReasoningLLM reasoning → Prompting and chain-of-thought reasoning
Don't Let the Video Speak: Audio-Contrastive Preference Optimization for Audio-Visual Language ModelsMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs +1
HiVLA: A Visual-Grounded-Centric Hierarchical Embodied Manipulation SystemLLM reasoning → Spatial and embodied reasoning +3
Correct Prediction, Wrong Steps? Consensus Reasoning Knowledge Graph for Robust Chain-of-Thought SynthesisLLM reasoning → Prompting and chain-of-thought reasoning
TREX: Automating LLM Fine-tuning via Agent-Driven Tree-based ExplorationLLM agents and tool use → General LLM agent frameworks and benchmarks
TIP: Token Importance in On-Policy DistillationKnowledge distillation → On-policy and self-distillation for LLMs
Training-Free Semantic Multi-Object Tracking with Vision-Language ModelsMultimodal LLMs / vision-language models → Video understanding with MLLMs
OneHOI: Unifying Human-Object Interaction Generation and EditingDiffusion models → Video generation
A Comparative Study of Dynamic Programming and Reinforcement Learning in Finite Horizon Dynamic PricingDeep reinforcement learning methods → Adversarial robustness and RL in finance
π-Play: Multi-Agent Self-Play via Privileged Self-Distillation without External DataKnowledge distillation → On-policy and self-distillation for LLMs
Enhancing Local Life Service Recommendation with Agentic Reasoning in Large Language ModelLLM reasoning → LLM reasoning for recommendation +1
Decoding the Delta: Unifying Remote Sensing Change Detection and Understanding with Multimodal Large Language ModelsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Seek-and-Solve: Benchmarking MLLMs for Visual Clue-Driven Reasoning in Daily ScenariosLLM reasoning → Multimodal reasoning +1
Hierarchical Reinforcement Learning with Runtime Safety Shielding for Power Grid OperationDeep reinforcement learning methods → RL for energy, networks and operations
POINTS-Seeker: An Open Recipe for Multimodal Search Agents with Visual Memory ManagementMultimodal LLMs / vision-language models → Document understanding and multimodal RAG
MAny: Merge Anything for Multimodal Continual Instruction TuningMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
Memory Transfer Learning: How Memories are Transferred Across Domains in Coding AgentsLLM agents and tool use → Agent memory systems
Diffusion Language Models for Speech RecognitionDiffusion models → Discrete and text diffusion
Remote Sensing Image Super-Resolution for Imbalanced Textures: A Texture-Aware Diffusion FrameworkDiffusion models → Inverse problems and image restoration
Reward Design for Physical Reasoning in Vision-Language ModelsLLM reasoning → Multimodal reasoning +1
Provably Efficient Offline-to-Online Value Adaptation with General Function ApproximationOffline RL and imitation learning from data → Offline RL algorithms +1
HINTBench: Horizon-agent Intrinsic Non-attack Trajectory BenchmarkLLM agents and tool use → Agent threats, guardrails and governance
Goal2Skill: Long-Horizon Manipulation with Adaptive Planning and ReflectionVision-language-action models → Reasoning and chain-of-thought in VLA
ASTRA: Enhancing Multi-Subject Generation with Retrieval-Augmented Pose Guidance and Disentangled Position EmbeddingRetrieval-augmented generation → Multimodal RAG
Blind Bitstream-corrupted Video Recovery via Metadata-guided Diffusion ModelDiffusion models → Inverse problems and image restoration
DiPO: Disentangled Perplexity Policy Optimization for Fine-grained Exploration-Exploitation Trade-OffLLM reasoning → Reinforcement learning for reasoning
Beyond Conservative Automated Driving in Multi-Agent Scenarios via Coupled Model Predictive Control and Deep Reinforcement LearningEnd-to-end autonomous driving → Reinforcement learning for driving
GeoAgentBench: A Dynamic Execution Benchmark for Tool-Augmented Agents in Spatial AnalysisLLM agents and tool use → Multimodal and embodied agents
MCPThreatHive: Automated Threat Intelligence for Model Context Protocol EcosystemsLLM agents and tool use → Agent threats, guardrails and governance
DiffMagicFace: Identity Consistent Facial Editing of Real VideosDiffusion models → Image generation, text-to-image and editing
Robust Reward Modeling for Large Language Models via Causal DecompositionLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF
MUSE: Multi-Domain Chinese User Simulation via Self-Evolving Profiles and Rubric-Guided AlignmentLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
Beyond State Consistency: Behavior Consistency in Text-Based World ModelsWorld models → World models for language agents
UI-Copilot: Advancing Long-Horizon GUI Automation via Tool-Integrated Policy OptimizationLLM agents and tool use → GUI and mobile agents
Gaslight, Gatekeep, V1-V3: Early Visual Cortex Alignment Shields Vision-Language Models from Sycophantic ManipulationMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Failure Identification in Imitation Learning Via Statistical and Semantic FilteringOffline RL and imitation learning from data → Language-conditioned policies and VLAs
ToolOmni: Enabling Open-World Tool Use via Agentic learning with Proactive Retrieval and Grounded ExecutionLLM agents and tool use → Tool learning and function calling
QuantileMark: A Message-Symmetric Multi-bit Watermark for LLMsLLM safety and security → Watermarking generated content and models
Soft Q(λ): A multi-step off-policy method for entropy regularised reinforcement learning using eligibility tracesDeep reinforcement learning methods → RL theory and sample complexity
From Anchors to Supervision: Memory-Graph Guided Corpus-Free Unlearning for Large Language ModelsLLM safety and security → Machine unlearning
Who Gets Flagged? The Pluralistic Evaluation Gap in AI Content WatermarkingLLM safety and security → Watermarking generated content and models
Design and Behavior of Sparse Mixture-of-Experts Layers in CNN-based Semantic SegmentationMixture-of-experts → Vision mixture-of-experts
The cognitive companion: a lightweight parallel monitoring architecture for detecting and recovering from reasoning degradation in LLM agentsLLM reasoning → Test-time compute and efficient thinking +1
Rethinking AI Hardware: A Three-Layer Cognitive Architecture for Autonomous AgentsLLM reasoning → Reasoning for agents, planning and code +1
MedRCube: A Multidimensional Framework for Fine-Grained and In-Depth Evaluation of MLLMs in Medical ImagingMultimodal LLMs / vision-language models → Medical vision-language models
Vision-Language-Action Jump-Starting for Reinforcement Learning Robotic AgentsVision-language-action models → Reinforcement learning and post-training +1
Doc-V*:Coarse-to-Fine Interactive Visual Reasoning for Multi-Page Document VQALLM reasoning → Multimodal reasoning
On Cost-Effective LLM-as-a-Judge Improvement TechniquesLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment

All 402