Papers

New on arXiv

Thu 20 Aug 2026

357 papers

ChronosAttack: Adversarial Tool Scheduling Attacks on LLM AgentsLLM agents and tool use → Agent threats, guardrails and governance
Agentic AI Cybersecurity FrameworkLLM agents and tool use → Agent threats, guardrails and governance
Reachable Global Optimization in AI Systems: How Global Is Global?LLM reasoning → Reasoning for agents, planning and code
Agentic Governance and Adversarial Verification for Policy-Constrained LLM Healthcare Appeal GenerationLLM agents and tool use → Medical and scientific domain agents
PonyEval: Evaluating LLM-Based Program Repair for Capability-Safe and Actor-Oriented Pony SoftwareLLM agents and tool use → Software engineering coding agents
Towards Adaptive Interaction Strategies for Human Companion Robot via Deep Reinforcement LearningDeep reinforcement learning methods → RL for robot navigation and UAV control
Balancing Reasoning and Hardware Constraints in RAG Pipelines for Ukrainian Multi-Domain Document UnderstandingLLM reasoning → Knowledge- and tool-augmented reasoning +1
Success Leaves Detours: Learning Executable Walkthroughs for Long-Horizon AgentsLLM agents and tool use → Software engineering coding agents
SCAFFOLD: A Large-Scale Structured Dataset of Computer Science Research Figures with Diagram QA and Chain-of-Thought Reasoning TracesLLM reasoning → Knowledge- and tool-augmented reasoning
FigmaTrace: Capturing Creative Nuances in Human Figma Design WorkflowsMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
When Failures Propagate: Causal Failure Attribution in Agentic Retrieval-Augmented GenerationRetrieval-augmented generation → Agentic RAG, agents and memory +1
Dual-Cache Latent Space Communication between Heterogeneous Language ModelsLLM agents and tool use → Software engineering coding agents
Evaluating Skills, Not Just Agents: Agentic Continuous Evaluation of SkillsLLM agents and tool use → Agent skill learning and evolution
Difficulty-Aware Semantic-ID Optimization for Generative RecommendationLLM reasoning → LLM reasoning for recommendation
Testing and Evaluation of Agentic AI Systems In Military Command and ControlLLM agents and tool use → Agent threats, guardrails and governance
AutoMOOSE: Use Case and Logical Views of Agentic Phase-Field Simulation SoftwareLLM agents and tool use → Medical and scientific domain agents
AgentDecarbonizer: Carbon-Aware Execution for AI AgentsLLM agents and tool use → Software engineering coding agents
Consilience: Conformally Calibrated Communication Control for Hidden-Profile Multi-Agent ReasoningLLM reasoning → Reasoning for agents, planning and code +1
Beyond End-to-End Success: Diagnosing Failures in Long-Horizon Security LLM AgentsLLM agents and tool use → Software engineering coding agents
Consistency Models for Fast MRI Reconstruction Using Regularization by DenoisingDiffusion models → Medical imaging and segmentation
Logic-VLA: A Temporal Logic Conditioned Vision-Language-Action ModelVision-language-action models → Aerial and drone VLA
aiXamine: Unified Black-Box Evaluation of Cross-Dimensional Trade-offs in LLM Safety, Security, and PrivacyLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
Volumetric Radiology AI in the Era of Multimodal Large Language ModelsMultimodal LLMs / vision-language models → Medical vision-language models
LiLiCorr: Lightweight Likelihood Correlation of Parallel Drafts for Speculative DecodingEfficient LLM inference and serving → Speculative decoding
FL-MAESTRO: Multi-Agent LLM Orchestration for Resource-Constrained Federated LearningLLM agents and tool use → Agentic AI for wireless networks
DiffVC-ONE: Diffusion-based Generative Video Compression with One-Step Video Diffusion TransformerDiffusion models → Diffusion for compression and wireless
Towards Traffic Modelling of Multi-Agent Systems: The Role of Coordination TopologyLLM agents and tool use → General LLM agent frameworks and benchmarks
Annotations as Rollouts: Efficient and Scalable Reinforcement Learning for Video MLLMsLLM reasoning → Multimodal reasoning +1
Metag: A dataset to build agentic meta-reviewing capabilitiesLLM agents and tool use → Web, search, and research agents
Terminal Agents: A Survey of AI Agents in Command-Line EnvironmentsLLM agents and tool use → Software engineering coding agents
AEGIS: Preventing Cross-Domain Resource Abuse in MCPLLM agents and tool use → Agent threats, guardrails and governance
EndoLIFT: Language-Disambiguated Latent-Conditioned Rectified Flow for Bidirectional Endoscopic ControlDiffusion models → Medical imaging and segmentation
Aggregating Visual Information with Optimal Transport for VideoLM Token CompressionEfficient LLM inference and serving → Visual token pruning in multimodal models
Peer-Voted LLM-Agent Stress Tests Find Feed-Induced Lexical Convergence but No Reliable Matched-Exposure Advantage for Distributed SourcesLLM agents and tool use → Social simulation and multi-agent debate
An LLM agent for end-to-end computational materials discoveryLLM agents and tool use → Medical and scientific domain agents
RISE: Adaptive Imagination for World Action ModelsWorld models → Embodied and robot world models
ConceptGuard: Benchmarking Context-Sensitive Unlearning in Large Language ModelsLLM safety and security → Machine unlearning
An Agentic Approach for Active Data Collection, Travel Behavior Modeling, and Weather-Sensitive Demand PredictionLLM agents and tool use → Multimodal and embodied agents
AI4AI-Bench: Benchmarking LLM Agents in Algorithmic Design for Recursive Self-ImprovementLLM agents and tool use → General LLM agent frameworks and benchmarks
Projecting BrowseComp-Plus onto ClimbMix: Toward More Realistic Corpora for Agentic SearchLLM agents and tool use → Web, search, and research agents
MidTool: Mid-training Data Synthesis for Agentic Tool UseLLM agents and tool use → Reinforcement learning for LLM agents +1
ACE-Ego-Hand: Repurposing Video Diffusion Models for Occlusion-Robust Egocentric 3D Hand Motion RecoveryDiffusion models → Diffusion and flow policies for robotics
Phantom Gains: Auditing Self-Improvement Against a Measured NullSynthetic data and self-improvement for LLMs → Self-rewarding alignment and self-training
Towards Surgical World-Action Modeling: A Preliminary Joint Visual-Trajectory Forecasting for Surgical Motion PlanningWorld models → Medical and surgical world models
Break It Down, Pass It On: Cross-Task Skill Transfer in LLM AgentsLLM agents and tool use → Agent skill learning and evolution
Learning When to Think: Adaptive Reasoning for Test-Time Compute AllocationLLM reasoning → Test-time compute and efficient thinking
What Makes a Good Fiqh Retriever? Answer Retrieval for Arabic Islamic JurisprudenceRetrieval-augmented generation → RAG robustness and adaptive retrieval
Rule-Compliant Visual Spatial Planning for Multimodal Large Language ModelsLLM reasoning → Spatial and embodied reasoning +1
RoMAN-Flow: Taming Autoregressive Normalizing Flows for Offline Reinforcement Learning in Robotic ManipulationOffline RL and imitation learning from data → Real-world robot offline RL
From Agent Behaviour to Agent-Friendly Documentation: An Empirical Study of How Coding Agents Discover, Read, and Write Technical DocumentationLLM agents and tool use → Software engineering coding agents
DARS: Dual-Level Credit Assignment RL with Structured Reasoning for Instruction-Based Image EditingLLM reasoning → Multimodal reasoning +2
Feature Evolution and Migration during Vision Transformer TrainingMechanistic interpretability → Interpreting vision and multimodal models
Multi-Agent Orchestration with the Common-Sense Reasoning Capabilities of LLMs for Autonomous DrivingLLM reasoning → Chain-of-thought reasoning for driving +2
Discrete Diffusion Inference-Time Control with Nested Sequential Monte CarloDiffusion models → Discrete and text diffusion
ArmorOCR: Grounded Adversarial Visual Perception via Observation-Transferred Self-DistillationMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness +1
SAE-Xplainers: Rule-Based Feature Interpretation for Extreme Earth EventsMechanistic interpretability → Interpreting time-series and weather models
DECOWAM: Decoupled Whole-Body World-Action Model for Legged Mobile ManipulationWorld models → Embodied and robot world models
Flow Matching-Based PET Image ReconstructionDiffusion models → Medical imaging and segmentation
Planning-Oriented End-to-End Autonomous Driving: Architectures, Evaluation, and Emerging ParadigmsWorld models → World models for autonomous driving +1
Reward-Guided Autoregressive Graph Generation for Efficient Multi-Agent Communication Topology DesignLLM agents and tool use → General LLM agent frameworks and benchmarks +1
TrustRAG: Blockchain-Enhanced RAG via Committee-Based Credibility ScoringRetrieval-augmented generation → Privacy and security of RAG
Backstepping-Guided Reinforcement Learning for Wide-Range Saint-Venant Canal RegulationDeep reinforcement learning methods → RL theory and sample complexity
Evidence-Gated Task and Motion Planning with Vision-Language ModelsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Orthogonal JEPA: Factorized Predictive States for Latent World ModelsWorld models → Latent predictive world models
Let's Scale Step by Step: Compute-Efficient Hyperparameter Transfer for Large-Scale Mixture-of-ExpertsMixture-of-experts → MoE theory and gating functions
EchoCoT: Extracting Hidden Chain-of-Thought from Large Reasoning ModelsLLM reasoning → Prompting and chain-of-thought reasoning
Auditing Cross-Lingual Fairness in Language Model WatermarkingLLM safety and security → Watermarking generated content and models
A three-dimensional typology of agency for advanced AI systemsLLM agents and tool use → General LLM agent frameworks and benchmarks
Emergence of cooperation: A reputation-modulated reinforcement learningDeep reinforcement learning methods → RL for games and combinatorial optimization
Manifold Drift in Flow Preference Optimization: A Root Cause of Reward HackingDiffusion models → Flow matching, time series and science +1
Optimal Skill Selection for LLM Agents with Provable Bicriteria GuaranteesLLM agents and tool use → Agent skill learning and evolution
COPA: Continual Preference Optimization for Adaptive Prompt Injection DefenseLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
ReguSim: Evaluating LLM Agent Rule Grounding in Financial ComplianceLLM agents and tool use → Financial trading and forecasting agents
Flow Matching Meets 3D Curvilinear Structure Segmentation in Medical ImagingDiffusion models → Medical imaging and segmentation
Bringing analytic rigor to agentic AI for science: The Brain Researcher platform for neuroimaging data analysisLLM agents and tool use → Medical and scientific domain agents
MaliciousSkillBench: A Comprehensive Benchmark for Malicious Agent Skill DetectionLLM agents and tool use → Agent threats, guardrails and governance
EXIMO: VLM Guided Exploration of VLA PoliciesVision-language-action models → Reinforcement learning and post-training
Write Once, Run Everywhere: The Axon DSL for Shape-Safe and Framework-Agnostic LLM ArchitecturesEfficient LLM inference and serving → Serving systems and request scheduling
EnvHarness: Awakening Static Worlds for Agent LearningLLM agents and tool use → Reinforcement learning for LLM agents
A knowledge-guided agentic framework for mitigating patient-context ambiguity in health queriesLLM agents and tool use → Medical and scientific domain agents
DIFFCZSL: Compositional Zero-Shot Learning Regularized by Diffusion RepresentationsMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
Repo0: Design-Driven Zero-to-All Code GenerationLLM agents and tool use → Software engineering coding agents
SAPO: Single-Rollout Autoregressive Policy Optimization for Agentic Reinforcement LearningLLM agents and tool use → Reinforcement learning for LLM agents
Adaptive Probabilistic Shielding by Learning MDPs for Safe Reinforcement LearningDeep reinforcement learning methods → Safe and constrained RL
Causal Reasoning with Bipartite Graphical Causal ModelsLLM reasoning → Causal reasoning in LLMs
Energy-Mamba: A Physics-Constrained State-Space Model for Medical Image ClassificationState-space sequence models (S4, Mamba) → Mamba for medical image analysis
Answer-Level Trust Selection for Physical Vision-Language ReasoningLLM reasoning → Multimodal reasoning +1
ADAPT: Physics-Aware Diffusion-based World Models for Adaptive Predictive Transferable HVAC ControlWorld models → Latent predictive world models
MileGPO: Milestone Inference with Local Evidence for Graph-Based Policy Optimization of Long-Horizon LLM AgentsLLM agents and tool use → Reinforcement learning for LLM agents
Stopping and Routing LLM Judge PanelsLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
SWE-bench Science: Can Coding Agents Resolve Engineering Tasks in Science?LLM agents and tool use → Software engineering coding agents
Towards general embodied intelligence: integrating large language models, knowledge bases, and reasoning capabilities to build the next generation of AI agentsLLM reasoning → Spatial and embodied reasoning +1
PRAXIS: Graph-Grounded Tacit Knowledge for Domain Code GenerationLLM agents and tool use → Software engineering coding agents
An Irreducible Quantum Advantage in Aligning World Models with RealityWorld models → World models for language agents
Credit Without Ground Truth: Auditing Step-Level Credit Assignment in LLM Agents Against Executed ReplayLLM agents and tool use → Agent threats, guardrails and governance
FlashPrefill V2: Block-Sparse Prefill Attention for Long-Context LLM ServingEfficient LLM inference and serving → KV cache optimization
Truncate Bad, Upweight Good: BoN-Style Distillation via Rank-Based ClassificationLLM alignment / RLHF / preference optimisation → Process reward models for reasoning
One Success Isn't Reliability: Thinkingbox, a Sandbox and Benchmark for Agents in Stateful Business WorkflowsLLM agents and tool use → Software engineering coding agents
Question-Guided Evidence Acquisition for Multimodal Visual Question AnsweringMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
TempJail: Temporal Jailbreak Attack against Large Vision-Language Models via Subtitle SchedulingMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness +1

All 357