Papers

New on arXiv

Mon 4 Aug 2025

408 papers

XFacta: Contemporary, Real-World Dataset and Evaluation for Multimodal Misinformation Detection with Multimodal LLMsMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness
MetAdv: A Unified and Interactive Adversarial Testing Platform for Autonomous DrivingEnd-to-end autonomous driving → Adversarial scenarios and safety testing
LMAR: Language Model Augmented Retriever for Domain-specific Knowledge IndexingRetrieval-augmented generation → Generation and context integration
Fine-Tuning Vision-Language Models for Markdown Conversion of Financial Tables in Malaysian Audited Financial ReportsMultimodal LLMs / vision-language models → Chart understanding and anomaly detection
Large Reasoning Models Are Autonomous Jailbreak AgentsLLM reasoning → Reasoning for agents, planning and code +1
Defend LLMs Through Self-ConsciousnessLLM safety and security → Prompt injection and agent security
Polymath: A Self-Optimizing Agent with Dynamic Hierarchical WorkflowLLM agents and tool use → Software engineering coding agents
AMD-Mamba: A Phenotype-Aware Multi-Modal Framework for Robust AMD PrognosisState-space sequence models (S4, Mamba) → Mamba for medical image analysis
MedBLINK: Probing Basic Perception in Multimodal Language Models for MedicineMultimodal LLMs / vision-language models → Medical vision-language models
X-Actor: Emotional and Expressive Long-Range Portrait Acting from AudioDiffusion models → Audio, speech and music
AQUAH: Automatic Quantification and Unified Agent in HydrologyLLM agents and tool use → Multimodal and embodied agents
Realizing Scaling Laws in Recommender Systems: A Foundation-Expert Paradigm for Hyperscale Model DeploymentKnowledge distillation → Distillation for recommender systems
GrandJury: A Collaborative Machine Learning Model Evaluation Protocol for Dynamic Quality RubricsLLM reasoning → Reasoning benchmarks and evaluation
PentestJudge: Judging Agent Behavior Against Operational RequirementsLLM agents and tool use → Agent threats, guardrails and governance
Context-aware Risk Assessment and Its Application in Autonomous DrivingEnd-to-end autonomous driving → Adversarial scenarios and safety testing
Following Route Instructions using Large Vision-Language Models: A Comparison between Low-level and Panoramic Action SpacesMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Enhancing Japanese Large Language Models with Reasoning VectorsLLM reasoning → Reasoning benchmarks and evaluation
Communicating Plans, Not Percepts: Scalable Multi-Agent Coordination with Embodied World ModelsWorld models → World models for language agents
RDDPM: Robust Denoising Diffusion Probabilistic Model for Unsupervised Anomaly SegmentationDiffusion models → Flow matching, time series and science
VisuCraft: Enhancing Large Vision-Language Models for Complex Visual-Guided Creative Content Generation via Structured Information ExtractionMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
REFLECT: Rectified Flows for Efficient Brain Anomaly Correction TransportDiffusion models → Medical imaging and segmentation
Coherent Multimodal Reasoning with Iterative Self-Evaluation for Vision-Language ModelsLLM reasoning → Multimodal reasoning +1
CauKer: Classification Time Series Foundation Models Can Be Pretrained on Synthetic DataSynthetic data and self-improvement for LLMs → Model collapse from synthetic data
Highlight & Summarize: RAG without the jailbreaksRetrieval-augmented generation → Privacy and security of RAG +1
Modeling Annotator Disagreement with Demographic-Aware Experts and Synthetic PerspectivesMixture-of-experts → MoE large language models
A Multi-Agent System for Complex Reasoning in Radiology Visual Question AnsweringLLM reasoning → Clinical and biomedical applications +1
Defending Against Knowledge Poisoning Attacks During Retrieval-Augmented GenerationRetrieval-augmented generation → Privacy and security of RAG
TIC-GRPO: Provable and Efficient Optimization for Reinforcement Learning from Human FeedbackLLM reasoning → Reinforcement learning for reasoning +1
Elucidating the Role of Feature Normalization in IJEPAKnowledge distillation → General knowledge distillation methods
DreamVVT: Mastering Realistic Video Virtual Try-On in the Wild via a Stage-Wise Diffusion Transformer FrameworkDiffusion models → 3D, point cloud and scene generation
Adaptive Knowledge Distillation for Device-Directed Speech DetectionKnowledge distillation → Speech and audio distillation
Cognitive Loop via In-Situ Optimization: Self-Adaptive Reasoning for ScienceLLM reasoning → Knowledge- and tool-augmented reasoning
Raw Data Matters: Enhancing Prompt Tuning by Internal Augmentation on Vision-Language ModelsMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
MedVLThinker: Simple Baselines for Multimodal Medical ReasoningLLM reasoning → Clinical and biomedical applications
Evaluating Variance in Visual Question Answering BenchmarksMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
D2PPO: Diffusion Policy Policy Optimization with Dispersive LossDiffusion models → Diffusion and flow policies for robotics
What Is Your AI Agent Buying? Evaluation, Biases, Model Dependence, & Emerging Implications for Agentic E-CommerceLLM agents and tool use → Social simulation and multi-agent debate
HyCodePolicy: Hybrid Language Controllers for Multimodal Monitoring and Decision in Embodied AgentsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models +1
Alleviating Attention Hacking in Discriminative Reward Modeling through Interaction DistillationLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF
ReMoMask: Retrieval-Augmented Masked Motion GenerationRetrieval-augmented generation → Multimodal RAG
Parameter-Efficient Routed Fine-Tuning: Mixture-of-Experts Demands Mixture of Adaptation ModulesMixture-of-experts → Expert routing strategies
CAMA: Enhancing Mathematical Reasoning in Large Language Models with Causal KnowledgeLLM reasoning → Causal reasoning in LLMs
Sparse-dLLM: Accelerating Diffusion LLMs with Dynamic Cache EvictionLLM reasoning → Test-time compute and efficient thinking +1
MonoDream: Monocular Vision-Language Navigation with Panoramic DreamingVision-language-action models → Generalist VLA models and benchmarks
Contextual Graph Transformer: A Small Language Model for Enhanced Engineering Document Information ExtractionRetrieval-augmented generation → Knowledge-graph and structured-knowledge RAG
Accurate and Interpretable Postmenstrual Age Prediction via Multimodal Large Language ModelMultimodal LLMs / vision-language models → Medical vision-language models
Transportation Cyber Incident Awareness through Generative AI-Based Incident Analysis and Retrieval-Augmented Question-Answering SystemsRetrieval-augmented generation → Privacy and security of RAG
Engagement Prediction of Short Videos with Large Multimodal ModelsMultimodal LLMs / vision-language models → Video understanding with MLLMs
Test-time Prompt InterventionLLM reasoning → Test-time compute and efficient thinking
R3A: Reinforced Reasoning for Relevance Assessment for RAG in User-Generated Content PlatformsLLM reasoning → Knowledge- and tool-augmented reasoning +1
Federated Graph UnlearningLLM safety and security → Federated unlearning and poisoning
Toward Using Machine Learning as a Shape Quality Metric for Liver Point Cloud GenerationDiffusion models → Medical imaging and segmentation
PoseGuard: Pose-Guided Generation with Safety GuardrailsLLM safety and security → Vision-language and text-to-image safety
Computationally efficient Gauss-Newton reinforcement learning for model predictive controlDeep reinforcement learning methods → RL for autonomous driving and traffic
Beyond Chunks and Graphs: Retrieval-Augmented Generation through Triplet-Driven ThinkingLLM reasoning → Knowledge- and tool-augmented reasoning +1
Multimodal Large Language Models for End-to-End Affective Computing: Benchmarking and Boosting with Generative Knowledge PromptingMultimodal LLMs / vision-language models → Multimodal emotion understanding
Superior resilience to poisoning and amenability to unlearning in quantum machine learningLLM safety and security → Machine unlearning
Modality Bias in LVLMs: Analyzing and Mitigating Object Hallucination via Attention LensLLM reasoning → Hallucination detection and mitigation +1
CompressKV: Semantic Retrieval Heads Know What Tokens are Not Important Before GenerationEfficient LLM inference and serving → KV cache optimization
Inference-time Scaling for Diffusion-based Audio Super-resolutionDiffusion models → Audio, speech and music
Talking Surveys: How Photorealistic Embodied Conversational Agents Shape Response Quality, Engagement, and SatisfactionLLM agents and tool use → Social simulation and multi-agent debate
Uni-Layout: Integrating Human Feedback in Unified Layout Generation and EvaluationLLM alignment / RLHF / preference optimisation → Diffusion and multimodal model alignment
Optimal Transport for Rectified Flow Image Editing: Unifying Inversion-Based and Direct MethodsDiffusion models → Image generation, text-to-image and editing
Traffic-R1: Reinforced LLMs Bring Human-Like Reasoning to Traffic Signal Control SystemsLLM reasoning → Chain-of-thought reasoning for driving
Agentic Personalized Fashion Recommendation in the Age of Generative AI: Challenges, Opportunities, and EvaluationLLM agents and tool use → Agentic recommender systems
VITRIX-CLIPIN: Enhancing Fine-Grained Visual Understanding in CLIP via Instruction Editing Data and Long CaptionsMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
CAMERA: Multi-Matrix Joint Compression for MoE Models via Micro-Expert Redundancy AnalysisMixture-of-experts → MoE large language models
VeOmni: Scaling Any Modality Model Training with Model-Centric Distributed Recipe ZooMixture-of-experts → Multimodal MoE models
Data Security in Large Language Models: Risks, Defense, and DirectionsLLM safety and security → Prompt injection and agent security
CAPO: Towards Enhancing LLM Reasoning through Generative Credit AssignmentLLM reasoning → Reinforcement learning for reasoning +1
Knowing When Not to Answer: Lightweight KB-Aligned OOD Detection for Safe RAGRetrieval-augmented generation → Privacy and security of RAG
CellForge: Agentic Design of Virtual Cell ModelsLLM agents and tool use → Medical and scientific domain agents
Decomposing the Entropy-Performance Exchange: The Missing Keys to Unlocking Effective Reinforcement LearningLLM reasoning → Reinforcement learning for reasoning
Patho-AgenticRAG: Towards Multimodal Agentic Retrieval-Augmented Generation for Pathology VLMs via Reinforcement LearningMultimodal LLMs / vision-language models → Medical vision-language models +2
Forecasting When to Forecast: Accelerating Diffusion Models with Confidence-Gated TaylorDiffusion models → Diffusion foundations and sampling
CO-RFT: Efficient Fine-Tuning of Vision-Language-Action Models through Chunked Offline Reinforcement LearningVision-language-action models → Reinforcement learning and post-training +1
Population-Free Pareto Tracking for Sample-Efficient Multi-Policy MORLDeep reinforcement learning methods → RL for games and combinatorial optimization
Constrained Reinforcement Learning for Unstable Point-Feet Bipedal Locomotion Applied to the Bolt RobotDeep reinforcement learning methods → RL for robot locomotion and manipulation
Seed Diffusion: A Large-Scale Diffusion Language Model with High-Speed InferenceDiffusion models → Discrete and text diffusion
Content-Aware Mamba for Learned Image CompressionState-space sequence models (S4, Mamba) → Vision Mamba backbones
FedVLA: Federated Vision-Language-Action Learning with Dual Gating Mixture-of-Experts for Robotic ManipulationVision-language-action models → Generalist VLA models and benchmarks +1
Retrieval-augmented Decoding for Improving Truthfulness in Open-ended GenerationRetrieval-augmented generation → RAG robustness and adaptive retrieval
Reconsidering Overthinking: Penalizing Internal and External Redundancy in CoT ReasoningLLM reasoning → Test-time compute and efficient thinking
Hidden in the Noise: Unveiling Backdoors in Audio LLMs Alignment through Latent Acoustic Pattern TriggersLLM safety and security → Audio jailbreaks and speech backdoors
Subject or Style: Adaptive and Training-Free Mixture of LoRAsMixture-of-experts → Mixture of LoRA experts
PIGDreamer: Privileged Information Guided World Models for Safe Partially Observable Reinforcement LearningWorld models → Latent dynamics models and model-based RL +1
DreamPainter: Image Background Inpainting for E-commerce ScenariosDiffusion models → 3D, point cloud and scene generation
AttriCtrl: Fine-Grained Control of Aesthetic Attribute Intensity in Diffusion ModelsDiffusion models → Image generation, text-to-image and editing
Beyond the Trade-off: Self-Supervised Reinforcement Learning for Reasoning Models' Instruction FollowingLLM reasoning → Reinforcement learning for reasoning
AURORA:Augmented Understanding via Structured Reasoning and Reinforcement Learning for Reference Audio-Visual SegmentationLLM reasoning → Audio-language model reasoning
Large-Scale Model Enabled Semantic Communication Based on Robust Knowledge DistillationKnowledge distillation → General knowledge distillation methods
FedLAD: A Linear Algebra Based Data Poisoning Defence for Federated LearningLLM safety and security → Federated unlearning and poisoning
Free-MoRef: Instantly Multiplexing Context Perception Capabilities of Video-MLLMs within Single InferenceMultimodal LLMs / vision-language models → Video understanding with MLLMs
VDEGaussian: Video Diffusion Enhanced 4D Gaussian Splatting for Dynamic Urban Scenes ModelingDiffusion models → 3D, point cloud and scene generation
A Survey on AgentOps: Categorization, Challenges, and Future DirectionsLLM reasoning → Reasoning for agents, planning and code +1
Don't Overthink It: A Survey of Efficient R1-style Large Reasoning ModelsLLM reasoning → Test-time compute and efficient thinking
DeflareMamba: Hierarchical Vision Mamba for Contextually Consistent Lens Flare RemovalState-space sequence models (S4, Mamba) → Mamba for image restoration
Attractive Metadata Attack: Inducing LLM Agents to Invoke Malicious ToolsLLM agents and tool use → Agent threats, guardrails and governance
AutoLoRA: Automatic LoRA Retrieval and Fine-Grained Gated Fusion for Text-to-Image GenerationDiffusion models → Image generation, text-to-image and editing
REACT-KD: Region-Aware Cross-modal Topological Knowledge Distillation for Interpretable Medical Image ClassificationKnowledge distillation → Distillation for medical imaging

All 408