Papers

New on arXiv

Wed 24 Sep 2025

396 papers

Efficient On-Device Agents via Adaptive Context ManagementEfficient LLM inference and serving → Efficient reasoning and agent inference
MazeMate: An LLM-Powered Chatbot to Support Computational Thinking in Gamified Programming LearningLLM reasoning → Reasoning benchmarks and evaluation
OR-Toolformer: Modeling and Solving Operations Research Problems with Tool Augmented Large Language ModelsLLM reasoning → Mathematical reasoning
A Comparative Analysis of Sparse Autoencoder and Activation Difference in Language Model SteeringMechanistic interpretability → Steering and interpreting LLM reasoning
Detoxifying Large Language Models via Autoregressive Reward Guided Representation EditingLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks
SKYLENAGE Technical Report: Mathematical Reasoning and Contest-Innovation Benchmarks for Multi-Level Math EvaluationLLM reasoning → Mathematical reasoning
EpidemIQs: Prompt-to-Paper LLM Agents for Epidemic Modeling and AnalysisLLM agents and tool use → Social simulation and multi-agent debate
ToolBrain: A Flexible Reinforcement Learning Framework for Agentic ToolsLLM agents and tool use → Reinforcement learning for LLM agents
Learning to Lead Themselves: Agentic AI in MAS using MARLLLM agents and tool use → Multimodal and embodied agents
Blueprint-Bench: Comparing spatial intelligence of LLMs, agents and image modelsLLM reasoning → Spatial and embodied reasoning +1
AMLA: MUL by ADD in FlashAttention RescalingEfficient LLM inference and serving → KV cache optimization
LayoutAgent: A Vision-Language Agent Guided Compositional Diffusion for Spatial Layout PlanningLLM agents and tool use → Multimodal and embodied agents
RAR2: Retrieval-Augmented Medical Reasoning via Thought-Driven RetrievalLLM reasoning → Clinical and biomedical applications +1
JaiLIP: Jailbreaking Vision-Language Models via Loss Guided Image PerturbationLLM reasoning → Reasoning for agents, planning and code +2
SafeSteer: Adaptive Subspace Steering for Efficient Jailbreak Defense in Vision-Language ModelsLLM safety and security → Vision-language and text-to-image safety
MIXRAG : Mixture-of-Experts Retrieval-Augmented Generation for Textual Graph Understanding and Question AnsweringRetrieval-augmented generation → Knowledge-graph and structured-knowledge RAG +1
FS-DFM: Fast and Accurate Long Text Generation with Few-Step Diffusion Language ModelsDiffusion models → Discrete and text diffusion
Training Task Reasoning LLM Agents for Multi-turn Task Planning via Single-turn Reinforcement LearningLLM reasoning → Reinforcement learning for reasoning +1
MMG: Mutual Information Estimation via the MMSE Gap in DiffusionDiffusion models → Diffusion foundations and sampling
An LLM-based Agentic Framework for Accessible Network ControlLLM agents and tool use → General LLM agent frameworks and benchmarks
Dynamic Reasoning Chains through Depth-Specialized Mixture-of-Experts in Transformer ArchitecturesLLM reasoning → Test-time compute and efficient thinking +1
PIRF: Physics-Informed Reward Fine-Tuning for Diffusion ModelsDiffusion models → Diffusion and flow policies for robotics
SAMULE: Self-Learning Agents Enhanced by Multi-level ReflectionLLM agents and tool use → General LLM agent frameworks and benchmarks
Perspectra: Choosing Your Experts Enhances Critical Thinking in Multi-Agent Research IdeationLLM reasoning → Reasoning for agents, planning and code +1
Selective Progress-Aware Querying for Human-in-the-Loop Reinforcement LearningDeep reinforcement learning methods → Deep RL algorithms and representations
Adaptive Approach to Enhance Machine Learning Scheduling Algorithms During Runtime Using Reinforcement Learning in Metascheduling ApplicationsDeep reinforcement learning methods → RL for energy, networks and operations
A Recovery Theory for Diffusion Priors: Deterministic Analysis of the Implicit Prior AlgorithmDiffusion models → Inverse problems and image restoration
Complexity-Regularized Proximal Policy OptimizationDeep reinforcement learning methods → RL theory and sample complexity
MARS: toward more efficient multi-agent collaboration for LLM reasoningLLM reasoning → Reasoning for agents, planning and code +1
RadAgents: Multimodal Agentic Reasoning for Chest X-ray Interpretation with Radiologist-like WorkflowsLLM agents and tool use → Medical and scientific domain agents
Offline Goal-conditioned Reinforcement Learning with Quasimetric RepresentationsOffline RL and imitation learning from data → Diffusion and flow policies for offline RL
Seedream 4.0: Toward Next-generation Multimodal Image GenerationEfficient LLM inference and serving → Visual token pruning in multimodal models
FastEagle: Cascaded Drafting for Accelerating Speculative DecodingEfficient LLM inference and serving → Speculative decoding
Online-Optimized RAG for Tool Use and Function CallingRetrieval-augmented generation → RAG pipeline design and evaluation +1
SceneWeaver: All-in-One 3D Scene Synthesis with an Extensible and Self-Reflective AgentLLM agents and tool use → Multimodal and embodied agents
A Theory of Multi-Agent Generative Flow NetworksDiffusion models → Diffusion and flow policies for robotics
PhysCtrl: Generative Physics for Controllable and Physics-Grounded Video GenerationDiffusion models → Video generation
Language Models that Think, Chat BetterLLM reasoning → Reinforcement learning for reasoning +1
Uncovering Graph Reasoning in Decoder-only Transformers with Circuit TracingLLM reasoning → Prompting and chain-of-thought reasoning
Video models are zero-shot learners and reasonersLLM reasoning → Multimodal reasoning
RAG Security and Privacy: Formalizing the Threat Model and Attack SurfaceRetrieval-augmented generation → Privacy and security of RAG +1
SIM-CoT: Supervised Implicit Chain-of-ThoughtLLM reasoning → Prompting and chain-of-thought reasoning
A co-evolving agentic AI system for medical imaging analysisLLM agents and tool use → Medical and scientific domain agents
Scan-do Attitude: Towards Autonomous CT Protocol Management using a Large Language Model AgentLLM agents and tool use → Medical and scientific domain agents
Failure Modes of Maximum Entropy RLHFLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF
AnchDrive: Bootstrapping Diffusion Policies with Hybrid Trajectory Anchors for End-to-End DrivingDiffusion models → Diffusion and flow policies for robotics +1
Beyond Sharp Minima: Robust LLM Unlearning via Feedback-Guided Multi-Point OptimizationLLM safety and security → Machine unlearning
Play by the Type Rules: Inferring Constraints for LLM Functions in Declarative ProgramsLLM reasoning → Knowledge- and tool-augmented reasoning
Universal Camouflage Attack on Vision-Language Models for Autonomous DrivingMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness +1
Thinking Augmented Pre-trainingLLM reasoning → Reasoning benchmarks and evaluation
Automated Multi-Agent Workflows for RTL DesignLLM agents and tool use → Software engineering coding agents
Federation of Agents: A Semantics-Aware Communication Fabric for Large-Scale Agentic AILLM agents and tool use → General LLM agent frameworks and benchmarks
CyberSOCEval: Benchmarking LLMs Capabilities for Malware Analysis and Threat Intelligence ReasoningLLM reasoning → Reasoning for agents, planning and code
EchoBench: Benchmarking Sycophancy in Medical Large Vision-Language ModelsMultimodal LLMs / vision-language models → Medical vision-language models
KSDiff: Keyframe-Augmented Speech-Aware Dual-Path Diffusion for Facial AnimationDiffusion models → Audio, speech and music
A Simple Data Augmentation Strategy for Text-in-Image Scientific VQAMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Beyond Slater's Condition in Online CMDPs with Stochastic and Adversarial ConstraintsDeep reinforcement learning methods → RL theory and sample complexity
Discrete Diffusion for Reflective Vision-Language-Action Models in Autonomous DrivingMultimodal LLMs / vision-language models → Driving VLMs and open-vocabulary detection +3
PEPS: Quantum-Inspired Reinforcement Learning for Coherent Reasoning Traces in LLMsLLM reasoning → Reinforcement learning for reasoning
Steerable Adversarial Scenario Generation through Test-Time Preference AlignmentLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF
From Pheromones to Policies: Reinforcement Learning for Engineered Biological SwarmsDeep reinforcement learning methods → RL for robot navigation and UAV control
Unleashing the Potential of the Semantic Latent Space in Diffusion Models for Image DehazingDiffusion models → Inverse problems and image restoration
Queryable 3D Scene Representation: A Multi-Modal Framework for Semantic Reasoning and Robotic Task PlanningLLM reasoning → Spatial and embodied reasoning
SHMoAReg: Spark Deformable Image Registration via Spatial Heterogeneous Mixture of Experts and Attention HeadsMixture-of-experts → Vision mixture-of-experts
LLM Trainer: Automated Robotic Data Generation via Demonstration Augmentation using LLMsOffline RL and imitation learning from data → Robot manipulation imitation learning
MACD: Multi-Agent Clinical Diagnosis with Self-Learned Knowledge for LLMLLM agents and tool use → Medical and scientific domain agents
Discrete Diffusion for Generative Modeling of Text-Aligned Speech TokensDiffusion models → Audio, speech and music
MARG: MAstering Risky Gap Terrains for Legged Robots with Elevation MappingDeep reinforcement learning methods → RL for robot locomotion and manipulation
Embodied AI: From LLMs to World ModelsWorld models → World models for navigation
Learning Robust Penetration Testing Policies under Partial Observability: A systematic evaluationDeep reinforcement learning methods → Adversarial robustness and RL in finance
RAD: Towards Trustworthy Retrieval-Augmented Multi-modal Clinical DiagnosisRetrieval-augmented generation → Medical and biomedical RAG
OmniScene: Attention-Augmented Multimodal 4D Scene Understanding for Autonomous DrivingEnd-to-end autonomous driving → VLMs for autonomous driving +1
Learnable Sampler Distillation for Discrete Diffusion ModelsDiffusion models → Diffusion foundations and sampling
Beyond Language Barriers: Multi-Agent Coordination for Multi-Language Code GenerationLLM agents and tool use → Software engineering coding agents
Latent Iterative Refinement Flow: A Geometric Constrained Approach for Few-Shot GenerationDiffusion models → Flow matching, time series and science
PromptCoT 2.0: Scaling Prompt Synthesis for Large Language Model ReasoningLLM reasoning → Prompting and chain-of-thought reasoning
Future Policy Approximation for Offline Reinforcement Learning in LLM ReasoningLLM reasoning → Reinforcement learning for reasoning
D3Grasp: Diverse and Deformable Dexterous Grasping for General ObjectsDeep reinforcement learning methods → RL for robot locomotion and manipulation
Adaptive User Interest Modeling via Conditioned Denoising Diffusion For Click-Through Rate PredictionDiffusion models → Diffusion recommenders
Adaptive Guidance Semantically Enhanced via Multimodal LLM for Edge-Cloud Object DetectionMultimodal LLMs / vision-language models → Visual token pruning and efficiency
SpecMamba: Accelerating Mamba Inference on FPGA with Speculative DecodingEfficient LLM inference and serving → Speculative decoding
FreezeVLA: Action-Freezing Attacks against Vision-Language-Action ModelsVision-language-action models → Safety and adversarial attacks
Benchmarking Gaslighting Attacks Against Speech Large Language ModelsMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness
BoreaRL: A Multi-Objective Reinforcement Learning Environment for Climate-Adaptive Boreal Forest ManagementDeep reinforcement learning methods → RL for energy, networks and operations
ThinkFake: Reasoning in Multimodal Large Language Models for AI-Generated Image DetectionLLM reasoning → Multimodal reasoning +1
LatentGuard: Controllable Latent Steering for Robust Refusal of Attacks and Reliable Response GenerationLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
An Efficient Conditional Score-based Filter for High Dimensional Nonlinear Filtering ProblemsDiffusion models → Diffusion foundations and sampling
Efficient Speech Watermarking for Speech Synthesis via Progressive Knowledge DistillationKnowledge distillation → Speech and audio distillation
VCRL: Variance-based Curriculum Reinforcement Learning for Large Language ModelsLLM reasoning → Reinforcement learning for reasoning
Agentic Metacognition: Designing a "Self-Aware" Low-Code Agent for Failure Prediction and Human HandoffLLM agents and tool use → General LLM agent frameworks and benchmarks
Faster, Smaller, and Smarter: Task-Aware Expert Merging for Online MoE InferenceMixture-of-experts → Expert routing strategies
bi-GRPO: Bidirectional Optimization for Jailbreak Backdoor Injection on LLMsLLM reasoning → Reasoning for agents, planning and code +2
Frictional Q-LearningDeep reinforcement learning methods → RL theory and sample complexity
The Conductor and the Engine: A Path Towards Co-Designed ReasoningLLM reasoning → Reasoning for agents, planning and code +1
Logics-Parsing Technical ReportMultimodal LLMs / vision-language models → Document understanding and multimodal RAG
Beyond Human Demonstrations: Diffusion-Based Reinforcement Learning to Generate Data for VLA TrainingVision-language-action models → Reinforcement learning and post-training +1
UserRL: Training Interactive User-Centric Agent via Reinforcement LearningLLM agents and tool use → Reinforcement learning for LLM agents +1
TopoCut: Learning Multi-Step Cutting with Spectral Rewards and Discrete Diffusion PoliciesDiffusion models → Diffusion and flow policies for robotics
From Prompt to Progression: Taming Video Diffusion Models for Seamless Attribute TransitionDiffusion models → Video generation
Calibrated Reasoning: An Explanatory Verifier for Dynamic and Efficient Problem-SolvingLLM reasoning → Reinforcement learning for reasoning

All 396