Papers

New on arXiv

Sun 1 Jun 2025

274 papers

Optimizing RAG Pipelines for Arabic: A Systematic Analysis of Core ComponentsRetrieval-augmented generation → RAG pipeline design and evaluation
HASHIRU: Hierarchical Agent System for Hybrid Intelligent Resource UtilizationLLM agents and tool use → General LLM agent frameworks and benchmarks
HADA: Human-AI Agent Decision Alignment ArchitectureLLM agents and tool use → General LLM agent frameworks and benchmarks
A Graph-Retrieval-Augmented Generation Framework Enhances Decision-Making in the Circular EconomyRetrieval-augmented generation → Agentic RAG, agents and memory
Language-Driven Coordination and Learning in Multi-Agent Simulation EnvironmentsLLM agents and tool use → Reinforcement learning for LLM agents
SafeSteer: Interpretable Safety Steering with Refusal-Evasion in LLMsMechanistic interpretability → Steering and probing for LLM safety
Infinity Parser: Layout Aware Reinforcement Learning for Scanned Document ParsingDeep reinforcement learning methods → RL training of language and agent models
Unlabeled Data Improves Fine-Grained Image Zero-shot Classification with Multimodal LLMsMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
The Measurement Imbalance in Agentic AI Evaluation Undermines Industry Productivity ClaimsLLM agents and tool use → General LLM agent frameworks and benchmarks
Decoupled Hierarchical Reinforcement Learning with State Abstraction for Discrete GridsDeep reinforcement learning methods → Deep RL algorithms and representations
Improving LLM Agents with Reinforcement Learning on Cryptographic CTF ChallengesLLM agents and tool use → Agent threats, guardrails and governance
Compress, Gather, and Recompute: REFORMing Long-Context Processing in TransformersEfficient LLM inference and serving → Early exit and edge inference
Mamba Drafters for Speculative DecodingState-space sequence models (S4, Mamba) → Core SSM architectures for language +1
Self-Supervised Multi-View Representation Learning using Vision-Language Model for 3D/4D Facial Expression RecognitionMultimodal LLMs / vision-language models → Multimodal emotion understanding
Incorporating Hierarchical Semantics in Sparse Autoencoder ArchitecturesMechanistic interpretability → SAE features and their applications
OG-VLA: Orthographic Image Generation for 3D-Aware Vision-Language Action ModelVision-language-action models → 3D and spatial-aware VLA
Culturally-Grounded Chain-of-Thought (CG-CoT):Enhancing LLM Performance on Culturally-Specific Tasks in Low-Resource LanguagesLLM reasoning → Reasoning benchmarks and evaluation
HoMeR: Learning In-the-Wild Mobile Manipulation via Hybrid Imitation and Whole-Body ControlOffline RL and imitation learning from data → Robot manipulation imitation learning
Doubly Robust Alignment for Large Language ModelsLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
Humanoid World Models: Open World Foundation Models for Humanoid RoboticsWorld models → Humanoid and legged locomotion world models
Accelerated Learning with Linear Temporal Logic using Differentiable SimulationDeep reinforcement learning methods → Interpretable and logic-guided RL
Efficient Regression-Based Training of Normalizing Flows for Boltzmann GeneratorsDiffusion models → Flow matching, time series and science
Slow Feature Analysis on Markov Chains from Goal-Directed BehaviorDeep reinforcement learning methods → RL theory and sample complexity
FlowMo: Variance-Based Flow Guidance for Coherent Motion in Video GenerationDiffusion models → Video generation
PARROT: Synergizing Mamba and Attention-based SSL Pre-Trained Models via Parallel Branch Hadamard Optimal Transport for Speech Emotion RecognitionState-space sequence models (S4, Mamba) → Speech and audio SSMs
From Words to Waves: Analyzing Concept Formation in Speech and Text-Based Foundation ModelsLLM reasoning → Reasoning benchmarks and evaluation
Neuro-Symbolic Generative Diffusion Models for Physically Grounded, Robust, and Safe GenerationDiffusion models → Flow matching, time series and science
ChemAU: Harness the Reasoning of LLMs in Chemical Research with Adaptive Uncertainty EstimationLLM reasoning → Clinical and biomedical applications
Is Random Attention Sufficient for Sequence Modeling? Disentangling Trainable Components in the TransformerLLM reasoning → Test-time compute and efficient thinking
Reconsidering LLM Uncertainty Estimation Methods in the WildLLM safety and security → Safety alignment, refusal and guardrails
DeepVerse: 4D Autoregressive Video Generation as a World ModelWorld models → Video-generation world models and simulators
Task-Related Token Compression in Multimodal Large Language Models from an Explainability PerspectiveMultimodal LLMs / vision-language models → Visual token pruning and efficiency +1
SuperRL: Reinforcement Learning with Supervision to Boost Language Model ReasoningLLM reasoning → Reinforcement learning for reasoning +1
Learning What Matters: Prioritized Concept Learning via Relative Error-driven Sample SelectionMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
Generative diffusion posterior sampling for informative likelihoodsDiffusion models → Diffusion foundations and sampling
The Coming Crisis of Multi-Agent Misalignment: AI Alignment Must Be a Dynamic and Social ProcessLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
GThinker: Towards General Multimodal Reasoning via Cue-Guided RethinkingLLM reasoning → Multimodal reasoning +1
Fighting Fire with Fire (F3): A Training-free and Efficient Visual Adversarial Example Purification Method in LVLMsMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness
AI4Contracts: LLM & RAG-Powered Encoding of Financial Derivative ContractsRetrieval-augmented generation → Legal and financial RAG
SealQA: Raising the Bar for Reasoning in Search-Augmented Language ModelsLLM reasoning → Knowledge- and tool-augmented reasoning
MCP-Zero: Active Tool Discovery for Autonomous LLM AgentsLLM agents and tool use → Tool learning and function calling
Simple Prompt Injection Attacks Can Leak Personal Data Observed by LLM Agents During Task ExecutionLLM agents and tool use → Agent threats, guardrails and governance +1
A Robust widetildeO(1/√T) Rate for Unprojected TD Learning with Linear Function ApproximationDeep reinforcement learning methods → RL theory and sample complexity
ECP-Mamba: An Efficient Multi-scale Self-supervised Contrastive Learning Method with State Space Model for PolSAR Image ClassificationState-space sequence models (S4, Mamba) → Point clouds, detection and remote sensing
Self-supervised ControlNet with Spatio-Temporal Mamba for Real-world Video Super-resolutionDiffusion models → Inverse problems and image restoration +1
ReFlow-VC: Zero-shot Voice Conversion Based on Rectified Flow and Speaker Feature OptimizationDiffusion models → Audio, speech and music
NavBench: Probing Multimodal Large Language Models for Embodied NavigationMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Modality Translation and Registration of MR and Ultrasound Images Using Diffusion ModelsDiffusion models → Medical imaging and segmentation
Optimistic critics can empower small actorsDeep reinforcement learning methods → RL theory and sample complexity
Rhythm Controllable and Efficient Zero-Shot Voice Conversion via Shortcut Flow MatchingDiffusion models → Audio, speech and music
Autoregressive Images Watermarking through Lexical Biasing: An Approach Resistant to Regeneration AttackLLM safety and security → Watermarking generated content and models
Temporal In-Context Fine-Tuning with Temporal Reasoning for Versatile Control of Video Diffusion ModelsLLM reasoning → Multimodal reasoning +1
Robust and Safe Multi-Agent Reinforcement Learning with Communication for Autonomous Vehicles: From Simulation to HardwareEnd-to-end autonomous driving → Cooperative and V2X driving
FLEx: Personalized Federated Learning for Mixture-of-Experts LLMs via Expert GraftingMixture-of-experts → Federated mixture-of-experts
Reinforcement Learning with Random Time HorizonsDeep reinforcement learning methods → RL theory and sample complexity
Aligning VLM Assistants with Personalized Situated CognitionMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning +1
Deep Temporal Reasoning in Video Language Models: A Cross-Linguistic Evaluation of Action Duration and Completion through Perfect TimesLLM reasoning → Multimodal reasoning
Q-learning with Posterior SamplingDeep reinforcement learning methods → RL theory and sample complexity
PCoreSet: Effective Active Learning through Knowledge Distillation from Vision-Language ModelsMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP +1
HMPC-assisted Adversarial Inverse Reinforcement Learning for Smart Home Energy ManagementOffline RL and imitation learning from data → Inverse reinforcement learning +1
State-Covering Trajectory Stitching for Diffusion PlannersDiffusion models → Diffusion and flow policies for robotics +1
Position: Agent Should Invoke External Tools ONLY When Epistemically NecessaryLLM agents and tool use → Tool learning and function calling
CoVoMix2: Advancing Zero-Shot Dialogue Generation with Fully Non-Autoregressive Flow MatchingDiffusion models → Audio, speech and music
Improve MLLM Benchmark Efficiency through InterviewMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
Not Every Token Needs Forgetting: Selective Unlearning to Limit Change in Utility in Large Language Model UnlearningLLM safety and security → Machine unlearning
What's Missing in Vision-Language Models? Probing Their Struggles with Causal Order ReasoningLLM reasoning → Multimodal reasoning +1
Local Manifold Approximation and Projection for Manifold-Aware Diffusion PlanningDiffusion models → Diffusion and flow policies for robotics
FourierFlow: Frequency-aware Flow Matching for Generative Turbulence ModelingDiffusion models → Flow matching, time series and science
Can AI Master Econometrics? Evidence from Econometrics AI Agent on Expert-Level TasksLLM agents and tool use → Financial trading and forecasting agents
MedBookVQA: A Systematic and Comprehensive Medical Benchmark Derived from Open-Access BookMultimodal LLMs / vision-language models → Medical vision-language models
Generalization in VAE and Diffusion Models: A Unified Information-Theoretic AnalysisDiffusion models → Diffusion foundations and sampling
Speech UnlearningLLM safety and security → Machine unlearning
Generalizable LLM Learning of Graph Synthetic Data with Post-training AlignmentLLM reasoning → Knowledge- and tool-augmented reasoning +2
Toward Structured Knowledge Reasoning: Contrastive Retrieval-Augmented Generation on ExperienceLLM reasoning → Knowledge- and tool-augmented reasoning +1
SynPO: Synergizing Descriptiveness and Preference Optimization for Video Detailed CaptioningLLM alignment / RLHF / preference optimisation → Diffusion and multimodal model alignment
SkyReels-Audio: Omni Audio-Conditioned Talking Portraits in Video Diffusion TransformersDiffusion models → Audio, speech and music
QuantFace: Efficient Quantization for Face RestorationDiffusion models → Inverse problems and image restoration
DriveMind: A Dual Visual Language Model-based Reinforcement Learning Framework for Autonomous DrivingMultimodal LLMs / vision-language models → Driving VLMs and open-vocabulary detection +1
Generalized Linear Markov Decision ProcessOffline RL and imitation learning from data → Offline RL algorithms +1
GuessBench: Sensemaking Multimodal Creativity in the WildMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Enhancing LLM Reasoning for Time Series Classification by Tailored Thinking and Fused DecisionLLM reasoning → Time series reasoning
Fast or Slow? Integrating Fast Intuition and Deliberate Thinking for Enhancing Visual Question AnsweringLLM reasoning → Multimodal reasoning +1
HSCR: Hierarchical Self-Contrastive Rewarding for Aligning Medical Vision Language ModelsMultimodal LLMs / vision-language models → Medical vision-language models +1
Adversarial Reinforcement Learning: A Duality-Based Approach To Solving Optimal Control ProblemsDeep reinforcement learning methods → Adversarial robustness and RL in finance
Action Dependency Graphs for Globally Optimal Coordinated Reinforcement LearningDeep reinforcement learning methods → RL for games and combinatorial optimization
Closing the Gap between TD Learning and Supervised Learning with Q-Conditioned MaximizationOffline RL and imitation learning from data → Offline RL algorithms +1
RARE: Retrieval-Aware Robustness Evaluation for Retrieval-Augmented Generation SystemsRetrieval-augmented generation → RAG robustness and adaptive retrieval
GeoChain: Multimodal Chain-of-Thought for Geographic ReasoningLLM reasoning → Spatial and embodied reasoning +1
KG-TRACES: Enhancing Large Language Models with Knowledge Graph-constrained Trajectory Reasoning and Attribution SupervisionLLM reasoning → Knowledge- and tool-augmented reasoning
Jailbreak-R1: Exploring the Jailbreak Capabilities of LLMs via Reinforcement LearningLLM safety and security → Prompt injection and agent security
CoP: Agentic Red-teaming for Large Language Models using Composition of PrinciplesLLM agents and tool use → Agent threats, guardrails and governance +1
LIFT the Veil for the Truth: Principal Weights Emerge after Rank Reduction for Reasoning-Focused Supervised Fine-TuningLLM reasoning → Test-time compute and efficient thinking
NeuronSeek: On Stability and Expressivity of Task-driven Neurons
Adaptive Two Sided Laplace Transforms: A Learnable, Interpretable, and Scalable Replacement for Self-Attention
Bound by semanticity: universal laws governing the generalization-identification tradeoff
PFMBench: Protein Foundation Model Benchmark
Multimodal Fusion with Semi-Supervised Learning Minimizes Annotation Quantity for Modeling Videoconference Conversation Experience
Deontological Keyword Bias: The Impact of Modal Expressions on Normative Judgments of Language Models
Unified Game Moderation: Soft-Prompting and LLM-Assisted Label Transfer for Resource-Efficient Toxicity Detection
LD-RPMNet: Near-Sensor Diagnosis for Railway Point Machines

All 274