Papers

New on arXiv

Fri 17 Jul 2026

319 papers

One Axis, No Brake: Self-Knowledge Limits the Filtering of Harmful Peer Conformity in LLMsLLM agents and tool use → General LLM agent frameworks and benchmarks
Compiled Agency: Frontier General-Purpose Coding Agents Build Winning Game Players from Bare Interaction - from Flappy Bird to StarCraft II and CivilizationLLM agents and tool use → Software engineering coding agents
Lost in Perception: Isolating Perceptual and Reasoning Failures in Multimodal Physics and Geometry ReasoningLLM reasoning → Multimodal reasoning +1
Beyond Right and Wrong: Evaluating Second-order Social Reasoning in Large Language ModelsLLM reasoning → Reasoning benchmarks and evaluation
Retrieval-Augmented Generation in LLMs for Mental Health: Quantifying the Incremental Contribution of Retrieval Within a Layered Safety ArchitectureRetrieval-augmented generation → Medical and biomedical RAG
MegaSlide-DiT: Memory-Centric Adaptation and Deformable Local Attention for Efficient Video DiffusionDiffusion models → Video generation
LazyMem: Retrieve Broadly, Construct Selectively for Efficient Long-Term Agent MemoryLLM agents and tool use → Agent memory systems
Beyond Sequential Interaction: Benchmarking Parallel Execution and Coordination for GUI AgentsLLM agents and tool use → GUI and mobile agents
Co-Harness: Co-Evolving Harnesses and Model Weights for LLM AgentsLLM agents and tool use → Software engineering coding agents
Chronofy: A Temporal-Logical Decay Architecture for Information Validity in Time-Aware Retrieval-Augmented GenerationRetrieval-augmented generation → Generation and context integration
KeySI: An Interaction Framework for Tuning Text Embeddings Based on Human FeedbackLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
Binding Drift in Multi-Step Tool-Augmented AgentsLLM agents and tool use → Tool learning and function calling
Agentic Calibration of Grey-Box Simulation Models: An LLM-Driven AlternativeLLM agents and tool use → General LLM agent frameworks and benchmarks
TD-DPO: Difference-Aware Preference Optimization for Mitigating Sycophancy in Clinical Autism Intervention DialogueLLM alignment / RLHF / preference optimisation → Medical LLM alignment
From Optimal Policies to Individual Differences: Rethinking Reinforcement Learning for BiologyDeep reinforcement learning methods → RL for games and combinatorial optimization
Building2Building: A Large Scale Benchmark for Generalizable Real-World Reinforcement LearningDeep reinforcement learning methods → RL for energy, networks and operations
Geometry-Enhanced Portion Estimation for Multimodal LLMsMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
Differentiable Reinforcement Learning for Path Tracking by an Agile Fish-Like RobotDeep reinforcement learning methods → RL for robot navigation and UAV control
Foresight Residual RL for Long-Horizon Robot Manipulation with Vision-Language-Action ModelsVision-language-action models → Reinforcement learning and post-training
Certifiable Safe Model-Based Reinforcement Learning with Control-Affine Dynamics ApproximationDeep reinforcement learning methods → Safe and constrained RL
WeedExpert-R1: Incentivizing Botanical Reasoning in MLLMs with Reinforcement Learning for Precision Weed GroundingLLM reasoning → Clinical and biomedical applications
Auto-Scaling Heterogeneous Neural Processing Units for Energy and Cost-Efficient LLM ServingEfficient LLM inference and serving → Serving systems and request scheduling
Enabling Spatially Fine-Grained DVFS in Neural Processing Units for Energy-Efficient LLM ServingEfficient LLM inference and serving → Serving systems and request scheduling
Retrieval is Enough: Training-Free Interpretability with a Tool-Using AgentLLM agents and tool use → Web, search, and research agents +1
One Modality to Forget Them All: Enhancing Cross-Modal Unlearning in Vision-Language ModelsMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness +1
RIMS: Preference Optimization via Smoothed Multi-pair Aggregation for Small-Scale LLM Retrieval-Augmented GenerationRetrieval-augmented generation → RAG robustness and adaptive retrieval +1
Multi-level context Modeling for consistent expert selection in Mixture-of-ExpertsMixture-of-experts → Expert routing strategies
Think, Plan, Paint: Layout-Aware Reasoning for Controllable Image Generation in Unified ModelsLLM reasoning → Multimodal reasoning +1
Apple-π: Benchmarking Thinking with Video Towards Law-Grounded Physical IntelligenceWorld models → Latent predictive world models +1
AoA: Theorem Proving Agent over Abstract Syntax Tree of Redesigned LanguageLLM agents and tool use → Educational agents and math problem solving
Spatial Transport of Integration Error in Generative ODEsDiffusion models → Diffusion foundations and sampling
PhysAgent: Reflective Agentic Physics Control for Physically Plausible Video GenerationLLM agents and tool use → Multimodal and embodied agents
Knowing the Self, Understanding the World: A Dual-Cognition Benchmark for UAV Spatio-temporal Reasoning with MLLMsLLM reasoning → Spatial and embodied reasoning +1
FVAttn: Adaptive Sparse Attention with Runtime Load Balancing for Video GenerationDiffusion models → Video generation
Searching Videos as Trees: Self-Correcting Agents for Grounded Long Video QALLM agents and tool use → Multimodal and embodied agents
PagedWeight: Efficient MoE LLM Serving with Dynamic Quality-Aware Weight QuantizationMixture-of-experts → Efficient MoE training and inference systems +1
Physics-enhanced reinforcement learning for real-time optimal control of dynamical systemsDeep reinforcement learning methods → RL theory and sample complexity
When Does Muon Help Agentic Reinforcement Learning?LLM agents and tool use → Reinforcement learning for LLM agents
When Do Multi-Agent Systems Help? An Information Bottleneck PerspectiveLLM agents and tool use → General LLM agent frameworks and benchmarks
ToolSciVer: Multimodal Scientific Claim Verification with Visual Tool Augmented Reinforcement LearningDeep reinforcement learning methods → RL training of language and agent models
The Honest Quorum Problem: Epistemic Byzantine Fault Tolerance for Agentic InfrastructureLLM agents and tool use → Agent memory systems
Audio-Visual Flamingo: Open Audio-Visual Intelligence for Long and Complex VideosLLM reasoning → Audio-language model reasoning
Attention-Guided Saliency Maps for Interpreting Visualization Literacy in VLMsMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Understanding Reasoning from Pretraining to Post-TrainingLLM reasoning → Reinforcement learning for reasoning
How Do VLMs Fail? Vision-Operation Misalignment in Compositional VQALLM reasoning → Multimodal reasoning +1
DADiff: Diffusion-Driven Cross-Domain Policy Adaptation for Reinforcement LearningDeep reinforcement learning methods → Deep RL algorithms and representations
JoyNexus: Service-Oriented Multi-Tenant Post-Training for VLA ModelsVision-language-action models → Efficient VLA inference and adaptation
Frontier Language Models Struggle to Copy: Text Can Be Better Viewed in 2DLLM reasoning → Reasoning benchmarks and evaluation
LLM-Powered Agentic AI for 5G/6G Networks: A Tutorial and Survey on Architectures, Protocols, and StandardizationLLM agents and tool use → Agentic AI for wireless networks
When Model Merging Rivals Joint Multi-Task Reinforcement Learning: A Task-Vector Geometry AnalysisDeep reinforcement learning methods → Deep RL algorithms and representations
Frontier AI performance across the business disciplines: a case-grounded benchmark of knowledge work and analytical reasoningLLM reasoning → Reasoning benchmarks and evaluation +1
Loop the Loopies!Mixture-of-experts → Expert routing strategies
CLaC@FinMMEval 2026 Task 3: Sentiment-Augmented Deep Reinforcement Learning for Active Trading -- An Alpha-Reward ApproachDeep reinforcement learning methods → Adversarial robustness and RL in finance
AI Watermark Evidence Fails Forensic Readiness: An Empirical EvaluationLLM safety and security → Watermarking generated content and models
Robustness of Reinforcement Learning-Based Congestion Management in Low-Voltage GridsDeep reinforcement learning methods → RL for energy, networks and operations
Refusal is Not Safety! Benchmarking Latent Safety Risks of LLM-Driven Content HumorizationLLM safety and security → Safety alignment, refusal and guardrails
A Human-Centric Evaluation of a Retrieval-Augmented Generation System for Explaining Quebec Insurance ContractsRetrieval-augmented generation → Legal and financial RAG
TARS: A Theory-of-Mind Agent for Personalized In-IDE Code ComprehensionLLM agents and tool use → Software engineering coding agents
More with Less: a Large Scale Remote Sensing VLM with a Simple RecipeMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Learning Reach-Avoid Task with Reinforcement Learning: Vectorized Simulation and BenchmarkDeep reinforcement learning methods → RL for robot locomotion and manipulation
On the Failure of Boundary-Seeking Distillation in Bottlenecked Generative ArchitecturesKnowledge distillation → General knowledge distillation methods
DSWorld: A Data Science World Model for Efficient Autonomous AgentsWorld models → World models for language agents +1
Orbis 2: A Hierarchical World Model for DrivingWorld models → World models for autonomous driving +2
Induction in Both Directions: A Mechanistic Analysis of In-Context Learning in Masked Diffusion Language ModelsDiffusion models → Discrete and text diffusion
Dynamics-Aware Meta-Imitation for Generalization to Unseen Robotic ManipulationOffline RL and imitation learning from data → Robot manipulation imitation learning
PriEco-DRL: Joint Optimization of Electric-Bus Eco-Driving and Transit-Priority Adaptive Signals via Deep Reinforcement LearningDeep reinforcement learning methods → RL for energy, networks and operations
Agentic Synthesis against Counterexample-Supplemented SketchesLLM agents and tool use → Software engineering coding agents
Von Mises-Fisher Mixture Model with Dynamic Shrinkage for Realistic Test-Time TransductionMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
Test-Time Noise Guided Adaptation for Realistic Autoregressive Video GenerationDiffusion models → Video generation
Knowledge-Centric Agents for Workflow Generation in ComfyUILLM agents and tool use → Software engineering coding agents
QUADS: Stabilizing NVFP4 Reinforcement Learning for MoE via QUantization-error Alignment across Dual SidesMixture-of-experts → MoE in reinforcement learning and robotics
AgentFAIR: A Multi-Agent Collaborative Framework for FAIRness Evaluation of Geospatial DatasetsLLM agents and tool use → Multimodal and embodied agents
NeurOWL: An LLM-Based Neural-symbolic Framework for Incomplete OWL Ontology ReasoningLLM reasoning → Reasoning benchmarks and evaluation
Making Agent-Mediated Contributions Governable: A Project-Level Governance Manifest for Open-Source AI CollaborationLLM agents and tool use → Software engineering coding agents
GeoChrono: Benchmarking and Rethinking Long-Term Temporal Understanding in Remote SensingLLM reasoning → Spatial and embodied reasoning
SkillNav: Score-Level Skill Intervention for Zero-Shot Object Goal NavigationMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Debiasing Text-to-Image Evaluation via Implicit Cultural Alignment Reward ModelingLLM alignment / RLHF / preference optimisation → Diffusion and multimodal model alignment
Better Starts, Better Ends: Bootstrapped Iterative Self-Reasoning Distillation for Compressed ReasoningLLM reasoning → Prompting and chain-of-thought reasoning +1
IoU-PD: IoU-Aware Privileged Distillation for Visual Grounding with Multimodal Large Language ModelsMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Natural Backdoor Attacks on Speech Recognition ModelsLLM safety and security → Audio jailbreaks and speech backdoors
Behavioral Controllability of Agentic Models for Information Extraction: From Fixed Workflows to Reflective AgentsLLM agents and tool use → Web, search, and research agents
AC-VLA: Robust Out-of-Distribution Action Execution via Compositional LearningVision-language-action models → Efficient VLA inference and adaptation
Efficient Difficulty-Aware Dynamic Routing for Diffusion-Based Real-World Image Super-ResolutionDiffusion models → Inverse problems and image restoration
GoStop: Reinforcement Learning for Adaptive Temporal Aggregation in Event-Based Feature TrackingDeep reinforcement learning methods → RL training of language and agent models
SpeechGuard: Online Defense against Backdoor Attacks on Speech Recognition ModelsLLM safety and security → Audio jailbreaks and speech backdoors
Toward a mechanistic understanding of inference in visual cortex and diffusion modelsDiffusion models → Diffusion foundations and sampling
Efficient Frame Selection for Long Videos at Test Time with Attention-Based MLLM SelectorsMultimodal LLMs / vision-language models → Video understanding with MLLMs
S1-Omni: A Unified Multimodal Reasoning Model for Scientific Understanding, Prediction, and GenerationLLM reasoning → Multimodal reasoning
Understanding Agent-Reactive Bugs at the Model-Harness Boundary: An Empirical Study of LLM Agent Issue ReportsLLM agents and tool use → Software engineering coding agents
Beyond Detection: Agentic Attack Synthesis and Simulation for Smart ContractsLLM agents and tool use → Agent threats, guardrails and governance
PE-Field 4D: Video Generation Models as CanvasDiffusion models → Video generation
Model Merging for Medical LVLMs: A Benchmark and a Winner-Take-All ApproachMultimodal LLMs / vision-language models → Medical vision-language models
ToolVerse: Unlocking Massive Environments and Long-Horizon Tasks for Agentic Reinforcement LearningLLM reasoning → Reasoning for agents, planning and code +1
Do Agents Dream of False Memories? Black-box Visual Attacks on Long-term Memory in Multimodal AI AgentsLLM agents and tool use → Agent threats, guardrails and governance
Adaptive Multi-Step Lookahead Decoding for Diffusion Language ModelsDiffusion models → Discrete and text diffusion
DiTango: Cost-Effective Parallel Diffusion Generation with Selective Attention State ReuseDiffusion models → Diffusion foundations and sampling
Difference-Based Relational Learning for Zero-Shot Object-Goal Visual Navigation With Direct Sim-to-Real TransferDeep reinforcement learning methods → RL for robot navigation and UAV control
IMBench: A Benchmark for Intuitive Robotic ManipulationVision-language-action models → Failure recovery and robustness
Think at 5 Hz, Act at 20 Hz: Asynchronous Fast-Slow Vision-Language-Action Inference for Closed-Loop DrivingLLM reasoning → Chain-of-thought reasoning for driving +2
Region-Grounded Vision-Language Learning for Detection-Guided Mammographic Lesion ClassificationMultimodal LLMs / vision-language models → Medical vision-language models

All 319