Papers

New on arXiv

Thu 10 Jul 2025

327 papers

SemRAG: Semantic Knowledge-Augmented RAG for Improved Question-AnsweringRetrieval-augmented generation → Knowledge-graph and structured-knowledge RAG
Self-Improving Language Models for Evolutionary Program Synthesis: A Case Study on ARC-AGISynthetic data and self-improvement for LLMs → Self-evolving LLM agents
Response Wide Shut? Surprising Observations in Basic Vision Language Model CapabilitiesMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
AirScape: An Aerial Generative World Model with Motion ControllabilityWorld models → World models for navigation
ODIA: Oriented Distillation for Inline Acceleration of LLM-based Function CallingLLM agents and tool use → Tool learning and function calling
Contrastive Language-Image Pre-Training Model based Semantic Communication Performance OptimizationDeep reinforcement learning methods → RL for energy, networks and operations
Simple Mechanistic Explanations for Out-Of-Context ReasoningLLM reasoning → Test-time compute and efficient thinking +1
From Curiosity to Competence: How World Models Interact with the Dynamics of ExplorationWorld models → Latent dynamics models and model-based RL +1
Reasoning and Behavioral Equilibria in LLM-Nash Games: From Mindsets to ActionsLLM reasoning → Reasoning for agents, planning and code
Toward a Dynamic Stackelberg Game-Theoretic Framework for Agentic AI Defense Against LLM JailbreakingLLM agents and tool use → Agent threats, guardrails and governance +1
Deep Reinforcement Learning in Applied Control: Challenges, Analysis, and InsightsDeep reinforcement learning methods → Deep RL algorithms and representations
CTRLS: Chain-of-Thought Reasoning via Latent State-TransitionLLM reasoning → Test-time compute and efficient thinking
KP-A: A Unified Network Knowledge Plane for Catalyzing Agentic Network IntelligenceLLM agents and tool use → Agentic AI for wireless networks
Adaptive Diffusion Denoised Smoothing : Certified Robustness via Randomized Smoothing with Differentially Private Guided Denoising DiffusionDiffusion models → Robustness, privacy and adversarial aspects
A Unified Framework for Adversary-Aware Differential Privacy BoundsLLM safety and security → Membership inference attacks
ALCo-FM: Adaptive Long-Context Foundation Model for Accident PredictionLLM reasoning → Chain-of-thought reasoning for driving
Code with Me or for Me? How Increasing AI Automation Transforms Developer WorkflowsLLM agents and tool use → Software engineering coding agents
Compactor: Calibrated Query-Agnostic KV Cache Compression with Approximate Leverage ScoresEfficient LLM inference and serving → KV cache optimization
Audio Flamingo 3: Advancing Audio Intelligence with Fully Open Large Audio Language ModelsLLM reasoning → Audio-language model reasoning
Imitation Learning for Obstacle Avoidance Using End-to-End CNN-Based Sensor FusionOffline RL and imitation learning from data → Language-conditioned policies and VLAs
Audit, Alignment, and Optimization of LM-Powered Subroutines with Application to Public Comment ProcessingLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
GRASP: Generic Reasoning And SPARQL Generation across Knowledge GraphsLLM reasoning → Knowledge- and tool-augmented reasoning
Predicting Flow Dynamics using Diffusion ModelsDiffusion models → Flow matching, time series and science
Quantile Reward Policy Optimization: Alignment with Pointwise Regression and Exact Partition FunctionsLLM reasoning → Reinforcement learning for reasoning +1
PUMA: Layer-Pruned Language Model for Efficient Unified Multimodal Retrieval with Modality-Adaptive LearningKnowledge distillation → Multimodal distillation
The relative importance of being GaussianDiffusion models → Diffusion foundations and sampling
TableReasoner: Advancing Table Reasoning Framework with Large Language ModelsLLM reasoning → Knowledge- and tool-augmented reasoning
Krul: Efficient State Restoration for Multi-turn Conversations with Dynamic Cross-layer KV SharingEfficient LLM inference and serving → KV cache optimization
Impact of Pretraining Word Co-occurrence on Compositional Generalization in Multimodal ModelsMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
PyVision: Agentic Vision with Dynamic ToolingLLM agents and tool use → Multimodal and embodied agents
Multi-Granular Spatio-Temporal Token Merging for Training-Free Acceleration of Video LLMsEfficient LLM inference and serving → Visual token pruning in multimodal models
Automating Expert-Level Medical Reasoning Evaluation of Large Language ModelsLLM reasoning → Clinical and biomedical applications
EXPO: Stable Reinforcement Learning with Expressive PoliciesDeep reinforcement learning methods → Deep RL algorithms and representations
Common Data Properties Limit Object-Attribute Binding in CLIPMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
OST-Bench: Evaluating the Capabilities of MLLMs in Online Spatio-temporal Scene UnderstandingMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Geometry Forcing: Marrying Video Diffusion and 3D Representation for Consistent World ModelingWorld models → Video-generation world models and simulators +1
Why is Your Language Model a Poor Implicit Reward Model?LLM alignment / RLHF / preference optimisation → Reward modeling and RLHF
Martian World Model: Controllable Video Synthesis with Physically Accurate 3D ReconstructionsWorld models → Video-generation world models and simulators
Defending Against Prompt Injection With a Few DefensiveTokensLLM safety and security → Prompt injection and agent security
Reinforcement Learning with Action ChunkingOffline RL and imitation learning from data → Offline RL algorithms +1
Scaling RL to Long VideosLLM reasoning → Multimodal reasoning +2
MIRIX: Multi-Agent Memory System for LLM-Based AgentsLLM agents and tool use → Agent memory systems
Reconstructing Template-Memorized Images from Natural PromptsDiffusion models → Robustness, privacy and adversarial aspects
SAGE: A Visual Language Model for Anomaly Detection via Fact Enhancement and Entropy-aware AlignmentMultimodal LLMs / vision-language models → Chart understanding and anomaly detection +1
Agentic Retrieval of Topics and Insights from Earnings CallsLLM agents and tool use → Financial trading and forecasting agents
MIRA: A Novel Framework for Fusing Modalities in Medical RAGMultimodal LLMs / vision-language models → Medical vision-language models +1
The Trust Fabric: Decentralized Interoperability and Economic Coordination for the Agentic WebLLM agents and tool use → Agent threats, guardrails and governance
Single-Step Latent Diffusion for Underwater Image RestorationDiffusion models → Inverse problems and image restoration
Mitigating Watermark Forgery in Generative Models via Randomized Key SelectionLLM safety and security → Watermarking generated content and models
DPO Unchained: Your Training Algorithm is Secretly Disentangled in Human Choice Theory (and its Loss' Convexity is Dispensable)LLM alignment / RLHF / preference optimisation → LLM preference optimization methods
"So, Tell Me About Your Policy...": Distillation of interpretable policies from Deep Reinforcement Learning agentsDeep reinforcement learning methods → Interpretable and logic-guided RL
From Ambiguity to Accuracy: The Transformative Effect of Coreference Resolution on Retrieval-Augmented Generation systemsRetrieval-augmented generation → RAG pipeline design and evaluation
MoSE: Skill-by-Skill Mixture-of-Experts Learning for Embodied Autonomous MachinesMixture-of-experts → MoE in reinforcement learning and robotics
SURPRISE3D: A Dataset for Spatial Understanding and Reasoning in Complex 3D ScenesLLM reasoning → Spatial and embodied reasoning
BEAVER: Building Environments with Assessable Variation for Evaluating Multi-Objective Reinforcement LearningDeep reinforcement learning methods → RL for energy, networks and operations
One-Point Contraction: Erasing Representational Separability toward Irreversible Deep ForgettingLLM safety and security → Machine unlearning
GuardVal: Dynamic Large Language Model Jailbreak Evaluation for Comprehensive Safety TestingLLM safety and security → Jailbreaking
Energy-Guided Decoding for Object Hallucination MitigationMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
Not All Preferences are What You Need for Post-Training: Selective Alignment Strategy for Preference OptimizationLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
Stable Preference Optimization: A Bilevel Approach to Catastrophic Preference ShiftLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
One Object, Multiple Lies: A Benchmark for Cross-task Adversarial Attack on Unified Vision-Language ModelsMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness
KeyKnowledgeRAG (K^2RAG): An Enhanced RAG method for improved LLM question-answering capabilitiesRetrieval-augmented generation → RAG robustness and adaptive retrieval
From Domain Documents to Requirements: Retrieval-Augmented Generation in the Space IndustryRetrieval-augmented generation → RAG pipeline design and evaluation
Tree-Mamba: A Tree-Aware Mamba for Underwater Monocular Depth EstimationState-space sequence models (S4, Mamba) → Mamba for image restoration
Rationale-Enhanced Decoding for Multi-modal Chain-of-ThoughtLLM reasoning → Multimodal reasoning +1
FloorplanQA: A Benchmark for Spatial Reasoning in LLMs using Structured RepresentationsLLM reasoning → Spatial and embodied reasoning
FrugalRAG: Less is More in RL Finetuning for Multi-Hop Question AnsweringLLM reasoning → Knowledge- and tool-augmented reasoning +1
Exploring the Limits of Model Compression in LLMs: A Knowledge Distillation Study on QA TasksKnowledge distillation → Knowledge distillation for LLMs
ViLU: Learning Vision-Language Uncertainties for Failure PredictionMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
LOSC: LiDAR Open-voc Segmentation ConsolidatorMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Stable-Hair v2: Real-World Hair Transfer via Multiple-View Diffusion ModelDiffusion models → 3D, point cloud and scene generation
Your Absorbing Discrete Diffusion Secretly Models the Bayesian PosteriorDiffusion models → Discrete and text diffusion
Beyond the Linear Separability Ceiling: Aligning Representations in VLMsMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Single-to-mix Modality Alignment with Multimodal Large Language Model for Document Image Machine TranslationMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
The Synergy Dilemma of Long-CoT SFT and RL: Investigating Post-Training Techniques for Reasoning VLMsLLM reasoning → Multimodal reasoning +1
ArchiveGPT: A human-centered evaluation of using a vision language model for image cataloguingMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
The Cross-Lingual Cost: Retrieval Biases in RAG over Arabic-English CorporaRetrieval-augmented generation → Generation and context integration
Divergence Minimization Preference Optimization for Diffusion Model AlignmentDiffusion models → Diffusion and flow policies for robotics +1
Toward Real-World Chinese Psychological Support Dialogues: CPsDD Dataset and a Co-Evolving Multi-Agent SystemLLM agents and tool use → Medical and scientific domain agents
Hallucination Stations: On Some Basic Limitations of Transformer-Based Language ModelsLLM agents and tool use → Medical and scientific domain agents
Teaching LLM to Reason: Reinforcement Learning from Algorithmic Problems without CodeLLM reasoning → Reinforcement learning for reasoning
Degradation-Agnostic Statistical Facial Feature Transformation for Blind Face Restoration in Adverse Weather ConditionsDiffusion models → Inverse problems and image restoration
RLEP: Reinforcement Learning with Experience Replay for LLM ReasoningLLM reasoning → Reinforcement learning for reasoning
StarDojo: Benchmarking Open-Ended Behaviors of Agentic Multimodal LLMs in Production-Living Simulations with Stardew ValleyLLM agents and tool use → Reinforcement learning for LLM agents
SAND: Boosting LLM Agents with Self-Taught Action DeliberationLLM agents and tool use → General LLM agent frameworks and benchmarks +1
Towards Interpretable Time Series Foundation ModelsLLM reasoning → Time series reasoning
Neural networks leverage nominally quantum and post-quantum representationsWorld models → World models for language agents
DrugMCTS: a drug repurposing framework combining multi-agent, RAG and Monte Carlo Tree SearchLLM agents and tool use → Medical and scientific domain agents
Corvid: Improving Multimodal Large Language Models Towards Chain-of-Thought ReasoningLLM reasoning → Multimodal reasoning +1
May I have your Attention? Breaking Fine-Tuning based Prompt Injection Defenses using Architecture-Aware AttacksLLM safety and security → Prompt injection and agent security
EscherNet++: Simultaneous Amodal Completion and Scalable View Synthesis through Masked Fine-Tuning and Enhanced Feed-Forward 3D ReconstructionDiffusion models → 3D, point cloud and scene generation
KVFlow: Efficient Prefix Caching for Accelerating LLM-Based Multi-Agent WorkflowsLLM agents and tool use → Software engineering coding agents +1
Bradley-Terry and Multi-Objective Reward Modeling Are ComplementaryLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF
Zero-Shot Context Generalization in Reinforcement Learning from Few Training ContextsDeep reinforcement learning methods → Deep RL algorithms and representations
Real-World Receptivity to Adaptive Mental Health Interventions: Findings from an In-the-Wild Study
Modelling Program Spaces in Program Synthesis with Constraints
A Formal Rebuttal of "The Blockchain Trilemma: A Formal Proof of the Inherent Trade-Offs Among Decentralization, Security, and Scalability"
Task-Focused Consolidation with Spaced Recall: Making Neural Networks Learn like College Students
Interpretable inverse design of optical multilayer thin films based on extended neural adjoint and regression activation mapping
A Regression-Based Share Market Prediction Model for Bangladesh

All 327