Papers

New on arXiv

Sun 25 May 2025

280 papers

AI Pedagogy: Dialogic Social Learning for Artificial AgentsLLM agents and tool use → Educational agents and math problem solving
Scalable Generation of Spatial Transcriptomics from Histology Images via Whole-Slide Flow MatchingDiffusion models → Medical imaging and segmentation
Rethinking the Understanding Ability across LLMs through Mutual InformationLLM reasoning → Reasoning benchmarks and evaluation
Nine Ways to Break Copyright Law and Why Our LLM Won't: A Fair Use Aligned Generation FrameworkLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks
Position: Collaborative Agentic AI Needs Interoperability Across EcosystemsLLM agents and tool use → General LLM agent frameworks and benchmarks
SeRL: Self-Play Reinforcement Learning for Large Language Models with Limited DataLLM reasoning → Reinforcement learning for reasoning +1
Adaptive Diffusion Guidance via Stochastic Optimal ControlDiffusion models → Diffusion foundations and sampling
RADEP: A Resilient Adaptive Defense Framework Against Model Extraction AttacksLLM safety and security → Watermarking generated content and models
Consistency-based Abductive Reasoning over Perceptual Errors of Multiple Pre-trained Models in Novel EnvironmentsLLM reasoning → Reasoning benchmarks and evaluation
ChartLens: Fine-grained Visual Attribution in ChartsMultimodal LLMs / vision-language models → Chart understanding and anomaly detection
Towards Humanoid Robot Autonomy: A Dynamic Architecture Integrating Continuous thought Machines (CTM) and Model Context Protocol (MCP)LLM agents and tool use → Multimodal and embodied agents
Prompting Decision Transformers for Zero-Shot Reach-Avoid PoliciesOffline RL and imitation learning from data → Decision transformers and sequence modeling +1
Paying Alignment Tax with Contrastive LearningLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
Concept Reachability in Diffusion Models: Beyond Dataset ConstraintsDiffusion models → Image generation, text-to-image and editing
DocMMIR: A Framework for Document Multi-modal Information RetrievalMultimodal LLMs / vision-language models → Document understanding and multimodal RAG
Retrieval-Augmented Generation for Service Discovery: Chunking Strategies and BenchmarkingRetrieval-augmented generation → RAG pipeline design and evaluation
A Novel Zero-Trust Identity Framework for Agentic AI: Decentralized Authentication and Fine-Grained Access ControlLLM agents and tool use → Agent threats, guardrails and governance
SituatedThinker: Grounding LLM Reasoning with Real-World through Situated ThinkingLLM reasoning → Reasoning for agents, planning and code
TextDiffuser-RL: Efficient and Robust Text Layout Optimization for High-Fidelity Text-to-Image SynthesisDiffusion models → Image generation, text-to-image and editing
Hypercube-Based Retrieval-Augmented Generation for Scientific Question-AnsweringRetrieval-augmented generation → Knowledge-graph and structured-knowledge RAG
RankLLM: A Python Package for Reranking with LLMsMixture-of-experts → MoE large language models
A Snapshot of Influence: A Local Data Attribution Framework for Online Reinforcement LearningDeep reinforcement learning methods → Deep RL algorithms and representations
Agentic Information Theory: Ergodicity and Intrinsic Semantics of Information ProcessesLLM agents and tool use → General LLM agent frameworks and benchmarks
Enhancing Text-to-Image Diffusion Transformer via Split-Text ConditioningDiffusion models → Image generation, text-to-image and editing
ALRPHFS: Adversarially Learned Risk Patterns with Hierarchical Fast & Slow Reasoning for Robust Agent DefenseLLM agents and tool use → Agent threats, guardrails and governance
Towards Large Reasoning Models for AgricultureLLM reasoning → Clinical and biomedical applications
VTool-R1: VLMs Learn to Think with Images via Reinforcement Learning on Multimodal Tool UseLLM reasoning → Multimodal reasoning +1
DeepResearchGym: A Free, Transparent, and Reproducible Evaluation Sandbox for Deep ResearchLLM agents and tool use → Web, search, and research agents
PATS: Process-Level Adaptive Thinking Mode SwitchingLLM reasoning → Test-time compute and efficient thinking
Improving Value Estimation Critically Enhances Vanilla Policy GradientDeep reinforcement learning methods → RL theory and sample complexity
To CoT or To Loop? A Formal Comparison Between Chain-of-Thought and Looped TransformersLLM reasoning → Prompting and chain-of-thought reasoning
Deformable Attentive Visual Enhancement for Referring Segmentation Using Vision-Language ModelMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
ActiveDPO: Active Direct Preference Optimization for Sample-Efficient AlignmentLLM reasoning → Reinforcement learning for reasoning +1
DriveX: Omni Scene Modeling for Learning Generalizable World Knowledge in Autonomous DrivingWorld models → World models for autonomous driving +1
Efficient Policy Optimization in Robust Constrained MDPs with Iteration Complexity GuaranteesDeep reinforcement learning methods → RL theory and sample complexity
Sensorimotor Self-Recognition in Multimodal Large Language Model-Driven RobotsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
CoreMatching: A Co-adaptive Sparse Inference Framework with Token and Neuron Pruning for Comprehensive Acceleration of Vision-Language ModelsMultimodal LLMs / vision-language models → Visual token pruning and efficiency
GUARDIAN: Safeguarding LLM Multi-Agent Collaborations with Temporal Graph ModelingLLM agents and tool use → Agent threats, guardrails and governance
LLaDA 1.5: Variance-Reduced Preference Optimization for Large Language Diffusion ModelsDiffusion models → Diffusion and flow policies for robotics +2
Where Paths Collide: A Comprehensive Survey of Classic and Learning-Based Multi-Agent PathfindingLLM agents and tool use → General LLM agent frameworks and benchmarks
The Overthinker's DIET: Cutting Token Calories with DIfficulty-AwarE TrainingLLM reasoning → Test-time compute and efficient thinking +1
Improving Medical Reasoning with Curriculum-Aware Reinforcement LearningLLM reasoning → Clinical and biomedical applications
Towards Understanding the Mechanisms of Classifier-Free GuidanceDiffusion models → Diffusion foundations and sampling
OptiMindTune: A Multi-Agent Framework for Intelligent Hyperparameter OptimizationLLM agents and tool use → General LLM agent frameworks and benchmarks
DREAM: Drafting with Refined Target Features and Entropy-Adaptive Cross-Attention Fusion for Multimodal Speculative DecodingMultimodal LLMs / vision-language models → Visual token pruning and efficiency +1
Structuring the Unstructured: A Multi-Agent System for Extracting and Querying Financial KPIs and GuidanceLLM agents and tool use → Financial trading and forecasting agents
CoCA: Step-level Reward for Free in RL-based T2I Diffusion Model Fine-tuningDiffusion models → Diffusion and flow policies for robotics +2
CardioCoT: Hierarchical Reasoning for Multimodal Survival AnalysisLLM reasoning → Clinical and biomedical applications
I2MoE: Interpretable Multimodal Interaction-aware Mixture-of-ExpertsMixture-of-experts → Multimodal MoE models
POQD: Performance-Oriented Query Decomposer for Multi-vector retrievalRetrieval-augmented generation → RAG pipeline design and evaluation
LIMOPro: Reasoning Refinement for Efficient and Effective Test-time ScalingLLM reasoning → Test-time compute and efficient thinking
When Two LLMs Debate, Both Think They'll WinLLM reasoning → Reasoning for agents, planning and code
Assistant-Guided Mitigation of Teacher Preference Bias in LLM-as-a-JudgeLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
Investigating Pedagogical Teacher and Student LLM Agents: Genetic Adaptation Meets Retrieval Augmented Generation Across Learning StyleRetrieval-augmented generation → Agentic RAG, agents and memory +1
JEDI: The Force of Jensen-Shannon Divergence in Disentangling Diffusion ModelsDiffusion models → Diffusion foundations and sampling
SRDiffusion: Accelerate Video Diffusion Inference via Sketching-Rendering CooperationDiffusion models → Video generation
Shifting AI Efficiency From Model-Centric to Data-Centric CompressionMultimodal LLMs / vision-language models → Visual token pruning and efficiency
Incentivizing High-Quality Human Annotations with Golden QuestionsLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
MMATH: A Multilingual Benchmark for Mathematical ReasoningLLM reasoning → Mathematical reasoning
RTime-QA: A Benchmark for Atomic Temporal Event Understanding in Large Multi-modal ModelsMultimodal LLMs / vision-language models → Video understanding with MLLMs
Exploring Magnitude Preservation and Rotation Modulation in Diffusion TransformersDiffusion models → Diffusion foundations and sampling
CreatiDesign: A Unified Multi-Conditional Diffusion Transformer for Creative Graphic DesignDiffusion models → Image generation, text-to-image and editing
Self-Critique Guided Iterative Reasoning for Multi-hop Question AnsweringLLM reasoning → Knowledge- and tool-augmented reasoning
Remote Sensing Image Classification with Decoupled Knowledge DistillationKnowledge distillation → Detection and dense prediction distillation
CCHall: A Novel Benchmark for Joint Cross-Lingual and Cross-Modal Hallucinations Detection in Large Language ModelsLLM reasoning → Hallucination detection and mitigation
Latent Mamba Operator for Partial Differential EquationsState-space sequence models (S4, Mamba) → Time series forecasting with SSMs
Agentic Visualization: Extracting Agent-based Design Patterns from Visualization SystemsLLM agents and tool use → Multimodal and embodied agents
ASPO: Adaptive Sentence-Level Preference Optimization for Fine-Grained Multimodal ReasoningLLM reasoning → Multimodal reasoning +1
SeePhys: Does Seeing Help Thinking? -- Benchmarking Vision-Based Physics ReasoningLLM reasoning → Spatial and embodied reasoning
ScreenExplorer: Training a Vision-Language Model for Diverse Exploration in Open GUI WorldMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
SATORI-R1: Incentivizing Multimodal Reasoning through Explicit Visual AnchoringLLM reasoning → Multimodal reasoning
Reinforced Latent Reasoning for LLM-based RecommendationLLM reasoning → LLM reasoning for recommendation
ReadBench: Measuring the Dense Text Visual Reading Ability of Vision-Language ModelsMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
Plug-and-Play Context Feature Reuse for Efficient Masked GenerationEfficient LLM inference and serving → Fast autoregressive visual generation
ReFineVLA: Reasoning-Aware Teacher-Guided Transfer Fine-TuningVision-language-action models → Reasoning and chain-of-thought in VLA
ChartSketcher: Reasoning with Multimodal Feedback and Reflection for Chart UnderstandingLLM reasoning → Multimodal reasoning
Universal Reasoner: A Single, Composable Plug-and-Play Reasoner for Frozen LLMsLLM reasoning → Test-time compute and efficient thinking
Training-free Stylized Text-to-Image Generation with Fast InferenceDiffusion models → Image generation, text-to-image and editing
Distributionally Robust Deep Q-LearningDeep reinforcement learning methods → RL theory and sample complexity
An Embarrassingly Simple Defense Against LLM Abliteration AttacksLLM safety and security → Safety alignment, refusal and guardrails
Structured Reinforcement Learning for Combinatorial Decision-MakingDeep reinforcement learning methods → RL for games and combinatorial optimization
When Models Don't Collapse: On the Consistency of Iterative MLESynthetic data and self-improvement for LLMs → Model collapse from synthetic data
Medical Large Vision Language Models with Multi-Image Visual AbilityMultimodal LLMs / vision-language models → Medical vision-language models
InfoChartQA: A Benchmark for Multimodal Question Answering on Infographic ChartsMultimodal LLMs / vision-language models → Chart understanding and anomaly detection
WorldEval: World Model as Real-World Robot Policies EvaluatorWorld models → Embodied and robot world models +1
Can Multimodal Large Language Models Understand Spatial Relations?Multimodal LLMs / vision-language models → Embodied and 3D vision-language models
Semi-pessimistic Reinforcement LearningOffline RL and imitation learning from data → Offline RL algorithms +1
VerIPO: Cultivating Long Reasoning in Video-LLMs via Verifier-Gudied Iterative Policy OptimizationLLM reasoning → Multimodal reasoning +1
STRICT: Stress Test of Rendering Images Containing TextDiffusion models → Image generation, text-to-image and editing
Dynamic Optimization and Safety Indicator Injection for Jailbreaking Text-to-Image Models with Multimodal Safety FiltersLLM safety and security → Vision-language and text-to-image safety
AI4Math: A Native Spanish Benchmark for University-Level Mathematical Reasoning in Large Language ModelsLLM reasoning → Mathematical reasoning
FastMamba: A High-Speed and Efficient Mamba Accelerator on FPGA with Accurate QuantizationState-space sequence models (S4, Mamba) → Core SSM architectures for language
Hierarchical Mamba Meets Hyperbolic Geometry: A New Paradigm for Structured Language EmbeddingsState-space sequence models (S4, Mamba) → Core SSM architectures for language
MGD3: Mode-Guided Dataset Distillation using Diffusion ModelsDiffusion models → Diffusion foundations and sampling
System-1.5 Reasoning: Traversal in Language and Latent Spaces with Dynamic ShortcutsLLM reasoning → Prompting and chain-of-thought reasoning
Weaver: Interweaving SQL and LLM for Table ReasoningLLM reasoning → Knowledge- and tool-augmented reasoning
Co-PatcheR: Collaborative Software Patching with Component(s)-specific Small Reasoning ModelsLLM reasoning → Reasoning for agents, planning and code
Online Knowledge Distillation with Reward GuidanceLLM alignment / RLHF / preference optimisation → LLM preference optimization methods +1
Exact Expressive Power of Transformers with PaddingLLM reasoning → Prompting and chain-of-thought reasoning
OpenHOI: Open-World Hand-Object Interaction Synthesis with Multimodal Large Language ModelMultimodal LLMs / vision-language models → Embodied and 3D vision-language models

All 280