Papers

New on arXiv

Fri 24 Oct 2025

378 papers

From Failure Modes to Reliability Awareness in Generative and Agentic AI SystemLLM agents and tool use → Agent threats, guardrails and governance
VisCoder2: Building Multi-Language Visualization Coding AgentsLLM agents and tool use → Multimodal and embodied agents
Noise is All You Need: Solving Linear Inverse Problems by Noise Combination Sampling with Diffusion ModelsDiffusion models → Inverse problems and image restoration
Beyond Pairwise: Empowering LLM Alignment With Ranked Choice ModelingLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
NUM2EVENT: Interpretable Event Reasoning from Numerical time-seriesLLM reasoning → Time series reasoning
Chain of Execution Supervision Promotes General Reasoning in Large Language ModelsLLM reasoning → Reasoning for agents, planning and code
QuArch: A Benchmark for Evaluating LLM Reasoning in Computer ArchitectureLLM reasoning → Reasoning for agents, planning and code
Jailbreak Mimicry: Automated Discovery of Narrative-Based Jailbreaks for Large Language ModelsLLM safety and security → Jailbreaking
Compositional Bias Control in Large Language Models: Preference Learning Fails, Supervision SucceedsLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
Agentic Reinforcement Learning for Real-World Code RepairLLM agents and tool use → Software engineering coding agents
Capturing Gaze Shifts for Guidance: Cross-Modal Fusion Enhancement for VLM Hallucination MitigationMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
A Vision Toward Energy-Efficient Domain-Specific Artificial Intelligence Models and AgentsLLM agents and tool use → General LLM agent frameworks and benchmarks
VLM-SlideEval: Evaluating VLMs on Structured Comprehension and Perturbation Sensitivity in PPTMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Predictive Coding Enhances Meta-RL To Achieve Interpretable Bayes-Optimal Belief Representation Under Partial ObservabilityDeep reinforcement learning methods → Deep RL algorithms and representations
LLM-AR: LLM-powered Automated Reasoning FrameworkLLM reasoning → Reasoning benchmarks and evaluation
Online Optimization for Offline Safe Reinforcement LearningOffline RL and imitation learning from data → Safe offline RL
Do You Trust the Process?: Modeling Institutional Trust for Community Adoption of Reinforcement Learning PoliciesDeep reinforcement learning methods → RL for games and combinatorial optimization
Toward Understanding the Transferability of Adversarial Suffixes in Large Language ModelsLLM safety and security → Jailbreaking
FlowOpt: Fast Optimization Through Whole Flow Processes for Training-Free EditingDiffusion models → Image generation, text-to-image and editing
OpenPhone: Mobile Agentic Foundation ModelsLLM agents and tool use → GUI and mobile agents
Optimal Detection for Language Watermarks with Pseudorandom CollisionLLM safety and security → Watermarking generated content and models
LiteDiffDiffusion models → Medical imaging and segmentation
Is Temporal Difference Learning the Gold Standard for Stitching in RL?Deep reinforcement learning methods → RL theory and sample complexity
FeaGPT: an End-to-End agentic-AI for Finite Element AnalysisLLM agents and tool use → Multimodal and embodied agents
Two-Steps Diffusion Policy for Robotic Manipulation via Genetic DenoisingDiffusion models → Diffusion and flow policies for robotics
Sprint: Sparse-Dense Residual Fusion for Efficient Diffusion TransformersDiffusion models → Diffusion foundations and sampling +1
Uncovering the Persuasive Fingerprint of LLMs in Jailbreaking AttacksLLM safety and security → Jailbreaking
Beyond Reasoning Gains: Mitigating General-Capability Forgetting in Large Reasoning ModelsLLM reasoning → Reinforcement learning for reasoning +1
Parallel Sampling from Masked Diffusion Models via Conditional Independence TestingDiffusion models → Discrete and text diffusion
δ-STEAL: LLM Stealing Attack with Local Differential PrivacyLLM safety and security → Watermarking generated content and models
Adversarial D'ej`a Vu: Jailbreak Dictionary Learning for Stronger Generalization to Unseen AttacksLLM safety and security → Jailbreaking
Structure-Aware Cooperative Ensemble Evolutionary Optimization on Combinatorial Problems with Multimodal Large Language ModelsMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
TOM-SWE: User Mental Modeling For Software Engineering AgentsLLM agents and tool use → Software engineering coding agents
Software Engineering Agents for Embodied Controller Generation : A Study in Minigrid EnvironmentsLLM agents and tool use → Software engineering coding agents
The Principles of Diffusion ModelsDiffusion models → Diffusion foundations and sampling
Computational Hardness of Reinforcement Learning with Partial qπ-RealizabilityDeep reinforcement learning methods → RL theory and sample complexity
Visual Diffusion Models are Geometric SolversDiffusion models → 3D, point cloud and scene generation
BachVid: Training-Free Video Generation with Consistent Background and CharacterDiffusion models → Video generation
Mechanistic Interpretability for Neural TSP SolversMechanistic interpretability → Circuits and mechanistic analysis
FlowSynth: Instrument Generation Through Distributional Flow Matching and Test-Time SearchDiffusion models → Audio, speech and music
A Dynamic Knowledge Distillation Method Based on the Gompertz CurveKnowledge distillation → General knowledge distillation methods
DEEDEE: Fast and Scalable Out-of-Distribution Dynamics DetectionDeep reinforcement learning methods → Deep RL algorithms and representations
Few-Shot Knowledge Distillation of LLMs With Counterfactual ExplanationsKnowledge distillation → General knowledge distillation methods
The Universal Landscape of Human ReasoningLLM reasoning → Reasoning benchmarks and evaluation
DeepAgent: A General Reasoning Agent with Scalable ToolsetsLLM reasoning → Reasoning for agents, planning and code +1
Epipolar Geometry Improves Video Generation ModelsDiffusion models → 3D, point cloud and scene generation
Huxley-G"odel Machine: Human-Level Coding Agent Development by an Approximation of the Optimal Self-Improving MachineLLM agents and tool use → Software engineering coding agents +1
Enhancing Tactile-based Reinforcement Learning for Robotic ControlDeep reinforcement learning methods → RL for robot locomotion and manipulation
Generalised Flow Maps for Few-Step Generative Modelling on Riemannian ManifoldsDiffusion models → Flow matching, time series and science
Modest-Align: Data-Efficient Alignment for Vision-Language ModelsMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
RETuning: Upgrading Inference-Time Scaling for Stock Movement Prediction with Large Language ModelsLLM reasoning → Reasoning for agents, planning and code
Principled RL for Flow Matching Emerges from the Chunk-level Policy OptimizationLLM reasoning → Reinforcement learning for reasoning +1
Scalable Vision-Language-Action Model Pretraining for Robotic Manipulation with Real-Life Human Activity VideosVision-language-action models → Human-video and latent-action pretraining
ColorEcosystem: Powering Personalized, Standardized, and Trustworthy Agentic Service in massive-agent EcosystemLLM agents and tool use → Software engineering coding agents
Co-Sight: Enhancing LLM-Based Agents via Conflict-Aware Meta-Verification and Trustworthy Reasoning with Structured FactsLLM reasoning → Reasoning for agents, planning and code +1
InterpDetect: Interpretable Signals for Detecting Hallucinations in Retrieval-Augmented GenerationRetrieval-augmented generation → Hallucination detection in RAG
EU-Agent-Bench: Measuring Illegal Behavior of LLM Agents Under EU LawLLM agents and tool use → Social simulation and multi-agent debate
Head Pursuit: Probing Attention Specialization in Multimodal TransformersMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
Towards a Golden Classifier-Free Guidance Path via Foresight Fixed Point IterationsDiffusion models → Diffusion foundations and sampling
MRO: Enhancing Reasoning in Diffusion Language Models via Multi-Reward OptimizationLLM reasoning → Diffusion reasoning and test-time scaling +1
Unified token representations for sequential decision modelsOffline RL and imitation learning from data → Decision transformers and sequence modeling +1
PhysWorld: From Real Videos to World Models of Deformable Objects via Physics-Aware Demonstration SynthesisWorld models → Latent predictive world models
Causality Meets Locality: Provably Generalizable and Scalable Policy Learning for Networked SystemsDeep reinforcement learning methods → Deep RL algorithms and representations
Vision Language Models for Dynamic Human Activity Recognition in Healthcare SettingsMultimodal LLMs / vision-language models → Medical vision-language models
Boosting Accuracy and Efficiency of Budget Forcing in LLMs via Reinforcement Learning for Mathematical ReasoningLLM reasoning → Reinforcement learning for reasoning
HIKMA: Human-Inspired Knowledge by Machine Agents through a Multi-Agent Framework for Semi-Autonomous Scientific ConferencesLLM agents and tool use → Web, search, and research agents
BADiff: Bandwidth Adaptive Diffusion ModelDiffusion models → Diffusion foundations and sampling
FairImagen: Post-Processing for Bias Mitigation in Text-to-Image ModelsDiffusion models → Image generation, text-to-image and editing
Compositional Monte Carlo Tree Diffusion for Extendable PlanningDiffusion models → Diffusion and flow policies for robotics
Multi-turn Training with Basic Human Feedback Helps Little on LLM ReasoningLLM reasoning → Reinforcement learning for reasoning +1
CXRAgent: Director-Orchestrated Multi-Stage Reasoning for Chest X-Ray InterpretationLLM reasoning → Clinical and biomedical applications
VL-SAE: Interpreting and Enhancing Vision-Language Alignment with a Unified Concept SetLLM reasoning → Multimodal reasoning +2
Leverage Unlearning to Sanitize LLMsLLM safety and security → Machine unlearning
FineRS: Fine-grained Reasoning and Segmentation of Small Objects with Reinforcement LearningLLM reasoning → Multimodal reasoning +1
PARL: Prompt-based Agents for Reinforcement LearningLLM agents and tool use → Reinforcement learning for LLM agents +1
When Models Outthink Their Safety: Unveiling and Mitigating Self-Jailbreak in Large Reasoning ModelsLLM reasoning → Reasoning for agents, planning and code +1
Topology Sculptor, Shape Refiner: Discrete Diffusion Model for High-Fidelity 3D Meshes GenerationDiffusion models → 3D, point cloud and scene generation
PINN Balls: Scaling Second-Order Methods for PINNs with Domain Decomposition and Adaptive SamplingMixture-of-experts → MoE theory and gating functions
VoiceAgentEval: A Dual-Dimensional Benchmark for Expert-Level Intelligent Voice-Agent Evaluation of Xbench's Professional-Aligned SeriesLLM agents and tool use → General LLM agent frameworks and benchmarks
AgentBound: Securing Execution Boundaries of AI AgentsLLM agents and tool use → Agent threats, guardrails and governance
How Hard is it to Confuse a World Model?World models → Latent dynamics models and model-based RL +1
DispatchMAS: Fusing taxonomy and artificial intelligence agents for emergency medical servicesLLM agents and tool use → Medical and scientific domain agents
Enhanced MLLM Black-Box Jailbreaking Attacks and DefensesLLM safety and security → Vision-language and text-to-image safety
On the flow matching interpretabilityDiffusion models → Flow matching, time series and science
Adaptive Graph Mixture of Residual Experts: Unsupervised Learning on Diverse Graphs with Heterogeneous SpecializationMixture-of-experts → MoE for time series and graphs
The Trojan Example: Jailbreaking LLMs through Template Filling and Unsafety ReasoningLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
Adjacent Words, Divergent Intents: Jailbreaking Large Language Models via Task ConcurrencyLLM safety and security → Jailbreaking
KBE-DME: Dynamic Multimodal Evaluation via Knowledge Enhanced Benchmark EvolutionMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
Memory-Free Continual Learning with Null Space Adaptation for Zero-Shot Vision-Language ModelsMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
Blockwise Flow Matching: Improving Flow Matching Models For Efficient High-Quality GenerationDiffusion models → Flow matching, time series and science
Uncertainty-Aware Multi-Objective Reinforcement Learning-Guided Diffusion Models for 3D De Novo Molecular DesignDiffusion models → Flow matching, time series and science +1
String Seed of Thought: Prompting LLMs for Distribution-Faithful and Diverse GenerationLLM reasoning → Prompting and chain-of-thought reasoning
How to Auto-optimize Prompts for Domain Tasks? Adaptive Prompting and Reasoning through Evolutionary Domain Knowledge AdaptationLLM reasoning → Prompting and chain-of-thought reasoning
NeuroGenPoisoning: Neuron-Guided Attacks on Retrieval-Augmented Generation of LLM via Genetic Optimization of External KnowledgeRetrieval-augmented generation → Privacy and security of RAG
Enhanced Evolutionary Multi-Objective Deep Reinforcement Learning for Reliable and Efficient Wireless Rechargeable Sensor NetworksDeep reinforcement learning methods → RL for energy, networks and operations
NoisyGRPO: Incentivizing Multimodal CoT Reasoning via Noise Injection and Bayesian EstimationLLM reasoning → Multimodal reasoning +1
DAO-AI: Evaluating Collective Decision-Making through Agentic AI in Decentralized GovernanceLLM agents and tool use → Financial trading and forecasting agents
Controllable-LPMoE: Adapting to Challenging Object Segmentation via Dynamic Local Priors from Mixture-of-ExpertsMixture-of-experts → Vision mixture-of-experts
PhysVLM-AVR: Active Visual Reasoning for Multimodal Large Language Models in Physical EnvironmentsLLM reasoning → Multimodal reasoning +1
Confounding Robust Deep Reinforcement Learning: A Causal ApproachDeep reinforcement learning methods → Deep RL algorithms and representations

All 378