Papers

New on arXiv

Fri 11 Sep 2026

378 papers

CHOREO: Every Humanoid Skill as a TrajectoryDeep reinforcement learning methods → RL for robot locomotion and manipulation
Bi-MoDe: Bilateral Control-based Imitation Learning via Modifier-Conditioned Decoding for Modulation of Execution Speed and Contact IntensityOffline RL and imitation learning from data → Robot manipulation imitation learning
Anatomy-Change-Aware Bidirectional Selective State-Space Memory for Clinically Deployed Thoracic Radiotherapy Auto-ContouringState-space sequence models (S4, Mamba) → Mamba for medical image analysis
StepAudio 3 Music Technical ReportDiffusion models → Audio, speech and music
Same Patient, Different Order: Action-Level Reliability of Clinical LLM Agents Under Repeated RunsLLM agents and tool use → Medical and scientific domain agents
FLoKD: Adaptive Knowledge Distillation for Federated Low-Rank LLM over Wireless NetworksKnowledge distillation → Federated, robust and private distillation
How User-AI Mistreatment Occurs and Matters in Conversational Systems?LLM safety and security → Safety alignment, refusal and guardrails
A Hybrid Agentic AI Framework for Intelligent Supply Chain AnalyticsLLM agents and tool use → Social simulation and multi-agent debate
Carbon-Aware Routing for Function Calling in Edge-Cloud LLM SystemsLLM agents and tool use → General LLM agent frameworks and benchmarks
AutoTailor: Automatic, User-Aligned Capability Selection and Adaptation for Web AgentsLLM agents and tool use → Web, search, and research agents
Asclepius: An Adaptive Harness for Long-Horizon Clinical AgentsLLM agents and tool use → Medical and scientific domain agents
Dynamic HBM Repartitioning for Multi-Turn MoE ServingMixture-of-experts → Efficient MoE training and inference systems +1
Harmfulness Propagation Dynamics: Layer-wise Trajectories of Adversarial Intent in Large Language ModelsLLM safety and security → Safety alignment, refusal and guardrails
Generative Interpretability via Scalable Neuro-Symbolic ModelsLLM agents and tool use → General LLM agent frameworks and benchmarks
From Token Probabilities to Semantic Constraints: Towards Declarative Probabilistic Evaluation of Language ModelsLLM reasoning → Reasoning benchmarks and evaluation
Constraint-Grounded Reinforcement Learning for Variable Impedance Control in Contact-Rich Robotic InsertionDeep reinforcement learning methods → RL for robot locomotion and manipulation
Token Efficient Task Execution via Application Behavior Modeling for Web AgentsLLM agents and tool use → Web, search, and research agents
Pre-retrieval Query Clustering for Adaptive Top-k Document Retrieval in RAG SystemsRetrieval-augmented generation → RAG pipeline design and evaluation
Mixture-of-Experts Language Models Can Be Strong and Efficient RetrieversMixture-of-experts → MoE large language models
BadEngram: Backdoor Attack on Gated Memory Components in LLMsLLM safety and security → Backdoor and data poisoning attacks
OrchSLM: Probing the Dynamics of Small Language Model OrchestrationLLM agents and tool use → General LLM agent frameworks and benchmarks
Root-Cause Attribution Is a Search Problem: Continual Search for Long-Horizon Agent FailuresLLM agents and tool use → Software engineering coding agents
Rethinking Software Development as a Self-Adaptive Socio-Technical SystemLLM agents and tool use → Software engineering coding agents
TimeThink: Eliciting Compositional Reasoning in Timeseries Large Language ModelsLLM reasoning → Time series reasoning +1
Hindsight Bias in Clinical Temporal Reasoning: How Future Data Exposure Affects Large Language Model JudgmentLLM reasoning → Clinical and biomedical applications
Toward Self-Adaptive Physical AI: Can LLM Agents Manage Long-Horizon Physical Tasks?LLM agents and tool use → General LLM agent frameworks and benchmarks
ReCAST: Reward Credit Assignment across Timesteps for Online Diffusion ReinforcementDiffusion models → Diffusion and flow policies for robotics
Vibe Patenting: Evaluating LLM Judges for Professional Patent-Drafting AgentsLLM agents and tool use → Web, search, and research agents
Task-Aware Federated Fine-Tuning for MoE-based Large Language ModelsMixture-of-experts → Federated mixture-of-experts
RFCLLM: Evaluating LLMs' Reasoning Ability of Network Protocol State MachinesLLM reasoning → Reasoning for agents, planning and code
ZGCM-1: A Fully Open and Extremely Efficient Foundation Model for Math and Agentic SearchLLM agents and tool use → Educational agents and math problem solving
SkillAtlas: An Attack Trace Library for Agent SkillsLLM agents and tool use → Agent threats, guardrails and governance
Bridging Thought and Action: Taming Long-Horizon Instability in Open-Source LLM Agents with a MetaTool-Enhanced ROS FrameworkLLM reasoning → Spatial and embodied reasoning +1
The Agentic Company OS: Substrate Inversion for Sustained Enterprise Agent DeploymentLLM agents and tool use → General LLM agent frameworks and benchmarks
Rethinking Heterogeneous System Disaggregation for Subquadratic AttentionLLM reasoning → Test-time compute and efficient thinking
Embodied-BenchForge: A Closed-Loop Agentic Workflow for Embodied Benchmark ConstructionLLM agents and tool use → Agent skill learning and evolution
CanvasAnneal: Curriculum Reinforcement Learning for Diffusion Language ModelsLLM reasoning → Diffusion reasoning and test-time scaling +1
Expert-Space Exploration in MoE Reinforcement LearningMixture-of-experts → MoE in reinforcement learning and robotics +1
Dynin-Robotics: Omnimodal Unified Diffusion Vision-Language-Action ModelVision-language-action models → Efficient VLA inference and adaptation
A Unified and Constrained View of Regularization-Based Robust Reinforcement LearningDeep reinforcement learning methods → Adversarial robustness and RL in finance
MCRL2: Multi-resource Cross-attention-based Representation Learning-augmented Reinforcement Learning for Cloud Microservice SchedulingDeep reinforcement learning methods → RL for energy, networks and operations
Diffusion Models and Concept FormationDiffusion models → Diffusion foundations and sampling
Robust Policy Optimization via Adversarial Importance SamplingDeep reinforcement learning methods → Adversarial robustness and RL in finance
Groupoid-Based Internal State Representations for Reinforcement Learning with Local SymmetriesDeep reinforcement learning methods → RL for robot locomotion and manipulation
Label-Guided Knowledge Distillation for 3D-CNNs in Action RecognitionKnowledge distillation → General knowledge distillation methods
Comfort by Construction: Adaptive, Comfort-Bounded Action Spaces for Learned Driving PoliciesEnd-to-end autonomous driving → Reinforcement learning for driving
How Good Are Frontier Models at Physics? Expert Re-Grading Reveals Broken Evaluations and Near-Saturation of Leading BenchmarksLLM reasoning → Spatial and embodied reasoning
Tasks over Application Manuals: Revealing Gaps in Long-Horizon Procedural Reasoning for Language ModelsLLM reasoning → Reasoning benchmarks and evaluation
SeqMoE: Toward Full-Load Performance via Predictive and Graph-Compatible MoE OffloadingMixture-of-experts → Efficient MoE training and inference systems
Fast and Faithful: Principled Conditional Flow Matching for Inverse ProblemsDiffusion models → Flow matching, time series and science
Forging Tree-Ring: Reproducing and Instrumenting Black-Box Semantic Watermark ForgeryDiffusion models → Robustness, privacy and adversarial aspects
Offline Reinforcement Learning for Wind Farm Control: A Wind Tunnel Study under Dynamic Wind DirectionsOffline RL and imitation learning from data → Offline RL algorithms
Physical-State-Guided Diffusion Sampling for Full-Waveform InversionDiffusion models → Flow matching, time series and science
Behavior Quotient Learning for Low-Rank Adaptation of LLM AgentsLLM agents and tool use → Reinforcement learning for LLM agents
Evaluating Context Segmentation in Locally Deployable SLMs for Cybersecurity CTF TasksLLM safety and security → Prompt injection and agent security
SCDM: Spatial-Contextual Disentanglement Mamba via Differential Inference for Efficient Image ClassificationState-space sequence models (S4, Mamba) → Mamba for medical image analysis
Scaling Clinical Judgment to Evaluate Medical AILLM reasoning → Clinical and biomedical applications
Online Video Agent Harness for Long Video UnderstandingLLM agents and tool use → Multimodal and embodied agents
K-Bench: A Benchmark for LLM Unlearning in Agentic DeploymentsLLM agents and tool use → Agent memory systems +1
Cognition on Graph: Navigating Massive Knowledge Space via Cognitive Cycles and Bidirectional Graph-Text SynergyRetrieval-augmented generation → Knowledge-graph and structured-knowledge RAG
DiffSynth-Music: Audio-Conditioned KV-Cache Adapters for Controllable Music GenerationDiffusion models → Audio, speech and music
Curriculum-Based Adversarial Heterogeneous Agent Reinforcement Learning for Autonomous Quad-Copter Landing in Maritime SettingsDeep reinforcement learning methods → RL for robot navigation and UAV control
GraphAHA: Graph-Based Adaptive Search with Heterogeneous Actions for Test-Time Code GenerationLLM reasoning → Reasoning for agents, planning and code
SCQ: Stabilizing Conservative Q-Learning with Sigmoid-Bounded EntropyOffline RL and imitation learning from data → Offline RL algorithms
The Mechanics of a Swarm: A Reproducible External Reconstruction of an Unintended Agent-Coordination Episode on a Third-Party WikiLLM agents and tool use → General LLM agent frameworks and benchmarks
Assisted Spatial Cognition Through Vision-Language ModelsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
What Drives Recovery in Agentic Text-to-Cypher? LAST-CQ: An LLM Agent Self-Refinement FrameworkLLM agents and tool use → Web, search, and research agents
Skill Issue: Lessons from Optimizing Repository SKILLs for Coding AgentsLLM agents and tool use → Software engineering coding agents
GRACE: Adaptive Concept Erasure with Geometry-Guided Retention in Diffusion ModelsDiffusion models → Robustness, privacy and adversarial aspects
Improving Imitation Learning Efficiency for Manipulation through Geometric Prior PretrainingOffline RL and imitation learning from data → Robot manipulation imitation learning
ExpertHTR: Unified Handwritten Text Recognition with Multi-Task Learning and Sparse Mixture-of-ExpertsMixture-of-experts → Multimodal MoE models
Size Doesn't Matter: Material-State Reinforcement Learning for Excavator Transferable Soil ManipulationDeep reinforcement learning methods → RL for robot navigation and UAV control
What Did the MLLM Hear? Token-Level Spectro-Temporal Grounding for Audio MLLM ExplainabilityMultimodal LLMs / vision-language models → Video understanding with MLLMs
Driving Context-guided Model Predictive Planning and Control for Autonomous Car Racing at the Limit and BeyondEnd-to-end autonomous driving → Reinforcement learning for driving
Separating Engineering Reasoning from DEXPI Serialization in LLM-Based Greenfield Surface-Process Design: A Three-Case Study for Underground Gas StorageLLM reasoning → Clinical and biomedical applications
LifeMem: Enabling Lifelong Experience Reuse for LLM AgentsLLM agents and tool use → Agent memory systems
Distortion of AI Alignment Revisited: RLHF is a Decent Utilitarian AlignerLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF
Online Material Estimation for Conditioned Diffusion Policy in Shaping Deformable Linear ObjectsDiffusion models → Diffusion and flow policies for robotics +1
Correlation-Guided Fast Machine Unlearning via Hessian AnalysisLLM safety and security → Machine unlearning
Direct Preference Density Alignment for Conversational Audio EqualizationLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
Beyond Generation and Accuracy: Diagnosing and Enhancing Visual Chain-of-Thought for Geometry Problem SolvingLLM reasoning → Spatial and embodied reasoning
Where Decoder Cosine Similarity Fails for SAE Feature Flow DiscoveryMechanistic interpretability → Sparse autoencoder training and evaluation
Reproducing and Evaluating the Generalizability of Subliminal Learning in Open-Weight ModelsKnowledge distillation → On-policy and self-distillation for LLMs
SCOPE-OPSD: Fisher-Conditioned Privileged Subspaces for On-Policy Self-DistillationKnowledge distillation → On-policy and self-distillation for LLMs
From Collaboration to Capability: Internalizing Routed LLM Experts into Compact ReasonersLLM reasoning → Reinforcement learning for reasoning
A Retrieval-Augmented Automated Stakeholder for Requirements Elicitation Education: A Comparative StudyRetrieval-augmented generation → RAG pipeline design and evaluation
Calibrated Ambiguity in Multimodal Language Models: Humans reach for cultural references, while models describe the pictureMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
An Ultra-Widefield Swept-Source OCTA Dataset and a Polar-Gated Mamba Network for Retinal Vessel SegmentationState-space sequence models (S4, Mamba) → Mamba for medical image analysis
RoofLang: Enabling AI-Driven Architecting of LLM Inference SystemsEfficient LLM inference and serving → Serving systems and request scheduling
Quality-Constrained Routing over a Fixed Pool of Quantized Mixture-of-Experts InstancesMixture-of-experts → Expert routing strategies
Earth-Agent-Pro: Towards Real-World Full-Chain Earth Observation with AgentsLLM agents and tool use → Multimodal and embodied agents
A Splitting Method for SDE Terminal-Law EstimationDiffusion models → Diffusion foundations and sampling
ChitraMiti: Benchmarking Visual Grounding and Modality Reliance in Bengali Geometric ReasoningLLM reasoning → Spatial and embodied reasoning +1
LettuceVisSim: A Simulator That Generates Lettuce Image Time-series for Vision-Based Reinforcement LearningDeep reinforcement learning methods → RL for energy, networks and operations
Information Specialization and Constrained Synthesis in Multi-Agent LLM Forecasting: A Prospective Live-Study of the 2026 FIFA World CupLLM agents and tool use → Financial trading and forecasting agents
TripPattern: A Pattern-based Text Watermarking Method for Large Language ModelsLLM safety and security → Watermarking generated content and models
AMDKernelVault: Large-Scale Datasets and Agentic Training for AMD GPU Kernel OptimizationLLM agents and tool use → Software engineering coding agents
EvoRS: On-Policy Self-Evolution of Reward Systems for Open-Ended Reinforcement LearningLLM reasoning → Reinforcement learning for reasoning
Bridging Vision Foundation Model Priors with CLIP for Spatial-aware Few-shot Anomaly Detection in Medical ImagesMultimodal LLMs / vision-language models → Chart understanding and anomaly detection
Do LLMs Trust the Accuser or the Accusation? Measuring Belief Shifts in WerewolfLLM agents and tool use → Social simulation and multi-agent debate

All 378