Papers

New on arXiv

Sat 6 Jun 2026

237 papers

AI Writes Code, Humans Pay the Debt. An Empirical Study on the Sustainability and Evolution of Agent-Generated CodeLLM agents and tool use → Software engineering coding agents
Position: Multi-Agent Systems Should Prioritize Concurrency ControlLLM agents and tool use → Software engineering coding agents
cMoLLM at Scale: Horizontal Scaling Laws for Mixture-of-LLMsMixture-of-experts → MoE large language models
EmbodimentSemantic: A Spatial Scene-Graph Dataset and Benchmark for Vision-Language Models on Embodied Manipulation TrajectoriesMultimodal LLMs / vision-language models → Embodied and 3D vision-language models +1
IntentKV: Cross-Turn Intent-Aware KV Cache Pruning for Agent InferenceLLM agents and tool use → Software engineering coding agents +1
Mix, Don't Pick: Why Synthetic Corpus Composition Matters for Time Series Foundation Model PretrainingSynthetic data and self-improvement for LLMs → Model collapse from synthetic data
Bypassing Copyright Protection in Diffusion-based Customization via Two-Stage Latent Feature OptimizationDiffusion models → Robustness, privacy and adversarial aspects
LongMoE: Longitudinal Multimodal Learning via Trajectory-Aware Mixture-of-ExpertsMixture-of-experts → Medical and biomedical MoE
Emergence World: A Platform for Evaluating Long-Horizon Multi-Agent AutonomyLLM agents and tool use → Social simulation and multi-agent debate
Pre-Intervention Prediction of Sparse Autoencoder Steering Side EffectsMechanistic interpretability → SAE features and their applications
Forward-Free Diffusion Language Models with BPTT-Free Looped RefinementDiffusion models → Discrete and text diffusion
Bayesian-Agent: Posterior-Guided Skill Evolution Across LLM Agent HarnessesLLM agents and tool use → Agent skill learning and evolution
CATPO: Critique-Augmented Tree Policy OptimizationLLM reasoning → Reinforcement learning for reasoning
Benchmarking Open-Ended Multi-Agent Coordination in Language AgentsLLM agents and tool use → General LLM agent frameworks and benchmarks
"So There's a Catch-22 Here": How Early Adopters Who Build Multi-Agent LLM Systems Conceptualize TransparencyLLM agents and tool use → General LLM agent frameworks and benchmarks
Neuro-Symbolic Injection of LTLf Constraints in Autoregressive Reinforcement Learning PoliciesOffline RL and imitation learning from data → Offline RL algorithms +1
To Nuke or Not to Nuke: LLMs' (Missing) Ethical Reasoning and Actions in a High-Stakes Decision-Making SimulationLLM reasoning → Reasoning for agents, planning and code
Where the Score Lives: A Wavelet View of DiffusionDiffusion models → Diffusion foundations and sampling
QueryGraph: Reliable Multi-Tool Query Execution Planning via LLM-Based Graph GenerationLLM agents and tool use → Web, search, and research agents
TLRD: Teaching LLMs to Reason over Tabular Data with Tri-Level Rationale DistillationLLM reasoning → Prompting and chain-of-thought reasoning
MotionVLA: Injecting Geometric Motion into Vision-Language-Action ModelVision-language-action models → Reasoning and chain-of-thought in VLA
Beyond Agent Architecture: Execution Assumptions and Reproducibility in LLM-Based Trading SystemsLLM agents and tool use → Financial trading and forecasting agents
QnRL: Quantum-Native Reinforcement LearningDeep reinforcement learning methods → Quantum reinforcement learning
Causal Agent Replay: Counterfactual Attribution for LLM-Agent FailuresLLM agents and tool use → Software engineering coding agents
Toward Human-Centered Multi-Agent Systems: Integrating Cognition, Culture, Values, and Cooperation in AI AgentsLLM agents and tool use → General LLM agent frameworks and benchmarks
TIDE: Task-Isolated Diffusion for Unified Video Editing and GenerationDiffusion models → Video generation
Differentially Private Synthetic Data via APIs 4: Tabular DataSynthetic data and self-improvement for LLMs → Privacy-preserving tabular data synthesis
Building Comparative Motivation Profiles with Instrumental InterventionsMechanistic interpretability → Activation steering methods
Light-WAM: Efficient World Action Models with State-Fusion Action DecodingWorld models → Embodied and robot world models
When No Answer Is Correct: Diagnosing Absent Answer Detection for MLLMs in Video UnderstandingMultimodal LLMs / vision-language models → Video understanding with MLLMs
SciTrace: Trajectory-Aware Safety Reasoning for Scientific Discovery AgentsLLM reasoning → Reasoning for agents, planning and code +1
Test-Time Scaling in Multimodal Foundation Models: A Comprehensive Survey of Generation and ReasoningLLM reasoning → Test-time compute and efficient thinking
CG-MambaNet: A spatiotemporal framework for cross-patient epileptic seizure prediction using CNN-GCN-Mamba-BiLSTM with event-level clinical evaluationState-space sequence models (S4, Mamba) → Physiological signal modeling
Agentic Neuro-Symbolic Planning and Commissioning for Human-in-the-Loop Industrial Robotics with Digital TwinsLLM agents and tool use → Multimodal and embodied agents
Online Agent-as-a-Judge: Situation-Generating Evaluation for Interactive AgentsLLM agents and tool use → General LLM agent frameworks and benchmarks
Learning from Human Driving: A Human-in-the-Loop Online Behavior Cloning Framework for Autonomous DrivingEnd-to-end autonomous driving → Imitation-learning end-to-end planners
Silent Failure in LLM Agent Systems: The Entropy Principle and the Inevitable Disorder of Autonomous AgentsLLM agents and tool use → General LLM agent frameworks and benchmarks
RAPID: Layer-Wise Redundancy-Aware Pruning and Importance-Driven Token Merging for Efficient ViTEfficient LLM inference and serving → Token pruning and merging for ViTs
SynthICL: Scalable In-context Imitation Learning with Synthetic DataOffline RL and imitation learning from data → Robot manipulation imitation learning
Decision-Aware Memory Cards: Counterfactual-Inspired Context Selection and Compression for Tool-Using LLM AgentsLLM agents and tool use → Agent memory systems
Property-Informed Diffusion-Based Text-to-Microstructure GenerationDiffusion models → Flow matching, time series and science
Biological Reasoning-Informed Regression for Interpretable Regulatory DNA Activity PredictionLLM reasoning → Clinical and biomedical applications
SAGE: An LLM-driven Self Reflective Agentic Framework for Fraud DetectionLLM agents and tool use → Agent threats, guardrails and governance
Learning Predictive Control with Deep Koopman Operators for Autonomous Vehicle Motion PlanningEnd-to-end autonomous driving → Reinforcement learning for driving +1
LCAM: A Framework for Diagnosing Interactional Alignment Failures in Con-versational AILLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
One Stone, Three Birds: Self-adaptive Optimal Transport for Multi-VLM Selection, Adaptation, and EnsemblingMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Think Before You Act: Intention-Guided Reasoning for LLM-Based Location PredictionLLM reasoning → Spatial and embodied reasoning
Reinforcement learning in linear embedding space unlocks generalizable control across soft robot configurationsDeep reinforcement learning methods → RL for robot locomotion and manipulation
vla.cpp: A Unified Inference Runtime for Vision-Language-Action ModelsVision-language-action models → Efficient VLA inference and adaptation
A Multimodal Agentic Pathology Co-pilot via Evidence Grounded ReasoningLLM agents and tool use → Medical and scientific domain agents
When Languages Disagree: Self-Evolving Multilingual LLM JudgesSynthetic data and self-improvement for LLMs → Self-rewarding alignment and self-training
VideoWeaver: Evaluating and Evolving Skills for Agentic Long Video GenerationLLM agents and tool use → Multimodal and embodied agents
ConSteer-RL: Steering Reasoning Capabilities in Large Language Models via Confidence-Aware Reinforcement LearningLLM reasoning → Reinforcement learning for reasoning
Aligned but Not Partner-Specific: Distinguishing How Multimodal LLM Agents Succeed in Reference Games Without Human-Like ConventionsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models +1
Support Vector Rubrics: Closing the Gap Between Self-Generated and Human RubricsSynthetic data and self-improvement for LLMs → Self-rewarding alignment and self-training
DICE: Entropy-Regularized Equilibrium Selection for Stable Multi-Agent LLM CoordinationLLM agents and tool use → Social simulation and multi-agent debate
Robust-U1: Can MLLMs Self-Recover Corrupted Visual Content for Robust Understanding?Multimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Diffusion Language Model Parallel Decoding via Product-of-Experts BridgeDiffusion models → Discrete and text diffusion +1
When Behavioral Safety Evaluation Fails: A Representation-Level PerspectiveLLM safety and security → Safety alignment, refusal and guardrails
MuJoCo-Drones-Gym: A GPU-Accelerated Multi-Drone Simulator for Control and Reinforcement LearningDeep reinforcement learning methods → RL for robot navigation and UAV control
DyCo-RL: Dynamic Cross-Modal Coordination for Visual ReasoningLLM reasoning → Multimodal reasoning +1
Variational Proximal Policy OptimizationLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF +2
Vision-Language Asymmetry in Bistable Image CaptioningMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Voting Protocols as Coordination Mechanisms for Role-Constrained Multi-Agent Tutoring SystemsLLM agents and tool use → Educational agents and math problem solving
IEA: Amateur-Friendly Conversational Image Editing Agent via Three Stages of Multitask AlignmentLLM agents and tool use → Multimodal and embodied agents
Q-VGM: Q-Guided Value-Gradient Matching for Offline-to-Online RL of Flow-Matching VLA PoliciesVision-language-action models → Efficient VLA inference and adaptation +2
Rewrite to Translate, Translate to Reward: Reinforcement Learning for Source Rewriting in Machine TranslationDeep reinforcement learning methods → RL training of language and agent models
Customer-Agent: Overcoming Context Limitations in Ultra-Long Shopping Trajectories via Tool-Augmented Agents and RLVRLLM agents and tool use → Social simulation and multi-agent debate
VATS: Exploiting Implicit Authority in Error-Path Injection via Systematic MutationLLM agents and tool use → Agent threats, guardrails and governance +1
PAFO: Pareto Fairness Optimization for Personalized Reward ModelingLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
PRISM: PRior-guided Imagination Sampling in world ModelsWorld models → Latent predictive world models
Defending Against Malicious Finetuning by Scaling Train-time Adversarial AttacksLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks
RecurGuard: Runtime Monitoring for Reasoning-Token Consumption AttacksLLM reasoning → Reasoning for agents, planning and code
DisCo: World Models with Discrete Camera Motion ControlWorld models → Video-generation world models and simulators
Zero-Shot Learning in Industrial Scenarios: New Large-Scale Benchmark, Challenges and BaselineMultimodal LLMs / vision-language models → Chart understanding and anomaly detection
Shared Latent Structures Enable Unified Backdoor Detection and Mitigation in LLMsLLM safety and security → Backdoor and data poisoning attacks +1
ChronoPhyBench: Do MLLMs Truly Understand the World or Merely Exploit Language Priors?LLM reasoning → Multimodal reasoning +1
Unification of Closed-Open Industrial Detection Scenarios: New Large-Scale Benchmarks,Challenges and BaselinesMultimodal LLMs / vision-language models → Chart understanding and anomaly detection
The Easy, the Hard, and the Learnable: Confidence and Difficulty-Adaptive Policy Optimization for LLM ReasoningLLM reasoning → Reinforcement learning for reasoning
EduMirror: Modeling Educational Social Dynamics with Value-driven Multi-agent SimulationLLM agents and tool use → Educational agents and math problem solving
Poise: Position-Aware One-Instruction Skill Injection for Silent Execution on LLM AgentsLLM agents and tool use → Agent threats, guardrails and governance
Collective Hallucination in Multi-Agent LLMs:Modeling and DefenseLLM agents and tool use → Medical and scientific domain agents
SGTO-MAS: Secure Gorilla Troops Optimization for Multi-Agent LLM SystemsLLM agents and tool use → General LLM agent frameworks and benchmarks
Hallucination Cascade: Analyzing Error Propagation in Multi-Agent LLM SystemsLLM agents and tool use → Medical and scientific domain agents
Decoupling Semantics and Logic: A Training-Free Coarse-to-Fine Pipeline for Video Retrieval-Augmented GenerationRetrieval-augmented generation → Multimodal RAG
Combining Reinforcement Learning with Arc-search Interior-Point Method for Path PlanningDeep reinforcement learning methods → RL for robot navigation and UAV control
MemToolAgent: Leveraging Memory for Tool Using Agents Based on Environment and User FeedbackLLM agents and tool use → Agent memory systems
Expert-validated STEM QA
Self- and Other-Labels Induce Bidirectional Bias in LLM Judges
From Regulatory Approvals to Patents: Cross-Domain Linking for Cardiovascular Device Traceability
KFTD: Koopman-Fourier Time-Differentiable Network for Continuous Ocean Spatiotemporal Forecasting
SPDM: Geometry-Modulated State Space Modeling with Manifold Constraints for Time Series Forecasting
IDP-Bench: Benchmarking ability of LLMs to protect personal information in interdependent privacy contexts
Predictive Coding with Bayesian Priors via Proximal Gradients
SoK: Reconstruction Attacks on Synthetic Tabular Data (Insights from Winning the NIST CRC)
Programmable Silicon Retina on Pixel Processor Array
An Information-Theoretic Definition for Open-Ended Learning
Self-Supervised Vision Transformers for CBCT-Based Detection of Temporomandibular Joint Osteoarthritis
Generative Frontier Planning for Adaptive Peer-Referral Recruitment under Covariate-Dependent Arrivals
Tensorizing Engram: Sharing Latents Across N-Gram Embeddings is Beneficial in LLMs

All 237