Papers

New on arXiv

Fri 27 Jun 2025

267 papers

Toward the Autonomous AI Doctor: Quantitative Benchmarking of an Autonomous Agentic AI Versus Board-Certified Clinicians in a Real World SettingLLM agents and tool use → Medical and scientific domain agents
Frequency-Aligned Knowledge Distillation for Lightweight Spatiotemporal ForecastingKnowledge distillation → General knowledge distillation methods
A Large Language Model-Empowered Agent for Reliable and Robust Structural AnalysisLLM agents and tool use → Multimodal and embodied agents
Smooth-Distill: A Self-distillation Framework for Multitask Learning with Wearable Sensor DataKnowledge distillation → Distillation for medical imaging
VERA: Variational Inference Framework for Jailbreaking Large Language ModelsLLM safety and security → Jailbreaking
Knowledge-Guided Multi-Agent Framework for Automated Requirements Development: A VisionLLM agents and tool use → Software engineering coding agents
URSA: The Universal Research and Scientific AgentLLM agents and tool use → Web, search, and research agents
Evaluating Hybrid Retrieval Augmented Generation using Dynamic Test Sets: LiveRAG ChallengeRetrieval-augmented generation → RAG robustness and adaptive retrieval
Layer Importance for Mathematical Reasoning is Forged in Pre-Training and Invariant after Post-TrainingLLM reasoning → Mathematical reasoning +1
CaO2: Rectifying Inconsistencies in Diffusion-Based Dataset DistillationDiffusion models → Diffusion foundations and sampling
ReCo: Reminder Composition Mitigates Hallucinations in Vision-Language ModelsMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
Seg-R1: Segmentation Can Be Surprisingly Simple with Reinforcement LearningLLM reasoning → Multimodal reasoning +2
Bootstrapping Human-Like Planning via LLMsSynthetic data and self-improvement for LLMs → Self-evolving LLM agents
RExBench: Can coding agents autonomously implement AI research extensions?LLM agents and tool use → Web, search, and research agents
BrainMT: A Hybrid Mamba-Transformer Architecture for Modeling Long-Range Dependencies in Functional MRI DataState-space sequence models (S4, Mamba) → Physiological signal modeling
The Hidden Link Between RLHF and Contrastive LearningLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF
Unifying Biomedical Vision-Language Expertise: Towards a Generalist Foundation Model via Multi-CLIP Knowledge DistillationKnowledge distillation → Multimodal distillation
Exploration Behavior of Untrained PoliciesDeep reinforcement learning methods → Deep RL algorithms and representations
Adjoint Schr"odinger Bridge SamplerDiffusion models → Diffusion foundations and sampling
MetaCipher: A Time-Persistent and Universal Multi-Agent Framework for Cipher-Based Jailbreak Attacks for LLMsLLM agents and tool use → Agent threats, guardrails and governance +1
MiCo: Multi-image Contrast for Reinforcement Visual ReasoningLLM reasoning → Multimodal reasoning
ARMOR: Robust Reinforcement Learning-based Control for UAVs under Physical AttacksDeep reinforcement learning methods → RL for robot navigation and UAV control
HyperCLOVA X THINK Technical ReportLLM reasoning → Reasoning benchmarks and evaluation
Exploration from a Primal-Dual Lens: Value-Incentivized Actor-Critic Methods for Sample-Efficient Online RLDeep reinforcement learning methods → RL theory and sample complexity
HazeMatching: Dehazing Light Microscopy Images with Guided Conditional Flow MatchingDiffusion models → Inverse problems and image restoration
QuickSilver -- Speeding up LLM Inference through Dynamic Token Halting, KV Skipping, Contextual Token Fusion, and Adaptive Matryoshka QuantizationEfficient LLM inference and serving → Early exit and edge inference
Test-Time Consistency in Vision Language ModelsMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Can Video Large Multimodal Models Think Like Doubters-or Double-Down: A Study on Defeasible Video EntailmentLLM reasoning → Multimodal reasoning +1
OptScale: Probabilistic Optimality for Inference-time ScalingLLM reasoning → Test-time compute and efficient thinking
Exploiting Vision Language Model for Training-Free 3D Point Cloud OOD Detection via Graph Score PropagationMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
Reinforcement Learning with Physics-Informed Symbolic Program Priors for Zero-Shot Wireless Indoor NavigationDeep reinforcement learning methods → RL for robot locomotion and manipulation
DiffSoundStream: Efficient Speech Tokenization via Diffusion DecodingDiffusion models → Audio, speech and music
Embodied AI Agents: Modeling the WorldWorld models → World models for navigation
Optimal Estimation of Watermark Proportions in Hybrid AI-Human TextsLLM safety and security → Watermarking generated content and models
OutDreamer: Video Outpainting with a Diffusion TransformerDiffusion models → 3D, point cloud and scene generation
Score-Based Model for Low-Rank Tensor RecoveryDiffusion models → Diffusion foundations and sampling
Rethinking Visual Token Reduction in LVLMs Under Cross-Modal MisalignmentLLM reasoning → Multimodal reasoning +2
JointRank: Rank Large Set with Single PassRetrieval-augmented generation → Generation and context integration
EAMamba: Efficient All-Around Vision State Space Model for Image RestorationState-space sequence models (S4, Mamba) → Mamba for image restoration
4D-VLA: Spatiotemporal Vision-Language-Action Pretraining with Cross-Scene CalibrationVision-language-action models → 3D and spatial-aware VLA
UiS-IAI@LiveRAG: Retrieval-Augmented Information Nugget-Based Generation of ResponsesRetrieval-augmented generation → RAG pipeline design and evaluation
A Reinforcement Learning Framework for Some Singular Stochastic Control ProblemsDeep reinforcement learning methods → RL theory and sample complexity
EFRame: Deeper Reasoning via Exploration-Filter-Replay Reinforcement Learning FrameworkLLM reasoning → Reinforcement learning for reasoning
Exploring Modularity of Agentic Systems for Drug DiscoveryLLM agents and tool use → Medical and scientific domain agents
MPipeMoE: Memory Efficient MoE for Pre-trained Models with Adaptive Pipeline ParallelismMixture-of-experts → Efficient MoE training and inference systems
Visual Structures Helps Visual Reasoning: Addressing the Binding Problem in VLMsLLM reasoning → Spatial and embodied reasoning +1
Q-Frame: Query-aware Frame Selection and Multi-Resolution Adaptation for Video-LLMsMultimodal LLMs / vision-language models → Video understanding with MLLMs
Reward Balancing Revisited: Enhancing Offline Reinforcement Learning for Recommender SystemsWorld models → Latent dynamics models and model-based RL
Reasoning in machine vision by learning fast and slow thinkingLLM reasoning → Multimodal reasoning
Query as Test: An Intelligent Driving Test and Data Storage Method for Integrated Cockpit-Vehicle-Road ScenariosEnd-to-end autonomous driving → VLMs for autonomous driving
Lost at the Beginning of ReasoningLLM reasoning → Test-time compute and efficient thinking
Hyper-modal Imputation Diffusion Embedding with Dual-Distillation for Federated Multimodal Knowledge Graph CompletionLLM reasoning → Knowledge- and tool-augmented reasoning
Noise-Inspired Diffusion Model for Generalizable Low-Dose CT ReconstructionDiffusion models → Medical imaging and segmentation
TROFI: Trajectory-Ranked Offline Inverse Reinforcement LearningOffline RL and imitation learning from data → Inverse reinforcement learning
RoboEnvision: A Long-Horizon Video Generation Model for Multi-Task Robot ManipulationDiffusion models → Diffusion and flow policies for robotics
StableCodec: Taming One-Step Diffusion for Extreme Image CompressionDiffusion models → Diffusion for compression and wireless
SceneDiffuser++: City-Scale Traffic Simulation via a Generative World ModelWorld models → World models for autonomous driving
Don't Trust Generative Agents to Mimic Communication on Social Networks Unless You Benchmarked their Empirical RealismLLM agents and tool use → Social simulation and multi-agent debate
Advancing Jailbreak Strategies: A Hybrid Approach to Exploiting LLM Vulnerabilities and Bypassing Modern DefensesLLM safety and security → Jailbreaking
More Vulnerable than You Think: On the Stability of Tool-Integrated LLM AgentsLLM reasoning → Reasoning for agents, planning and code +1
Optimal Return-to-Go Guided Decision Transformer for Auto-Bidding in AdvertisementOffline RL and imitation learning from data → Decision transformers and sequence modeling
CAL-RAG: Retrieval-Augmented Multi-Agent Generation for Content-Aware Layout DesignRetrieval-augmented generation → Multimodal RAG +1
ARAG: Agentic Retrieval Augmented Generation for Personalized RecommendationRetrieval-augmented generation → RAG for recommendation +1
SPAZER: Spatial-Semantic Progressive Reasoning Agent for Zero-shot 3D Visual GroundingLLM reasoning → Spatial and embodied reasoning
RAUM-Net: Regional Attention and Uncertainty-aware Mamba NetworkState-space sequence models (S4, Mamba) → Vision Mamba backbones
Advancements and Challenges in Continual Reinforcement Learning: A Comprehensive ReviewDeep reinforcement learning methods → Deep RL algorithms and representations
Do Vision-Language Models Have Internal World Models? Towards an Atomic EvaluationWorld models → World models for language agents +1
WildSpeech-Bench: Benchmarking End-to-End SpeechLLMs in the WildMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
On the Feasibility of Poisoning Text-to-Image AI Models via Adversarial MislabelingMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness
Grounding-Aware Token Pruning: Recovering from Drastic Performance Drops in Visual Grounding Caused by PruningMultimodal LLMs / vision-language models → Visual token pruning and efficiency +1
A Survey of Continual Reinforcement LearningDeep reinforcement learning methods → Deep RL algorithms and representations
DeepOmni: Towards Seamless and Smart Speech Interaction with Adaptive Modality-Specific MoEMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning +1
Remote Sensing Large Vision-Language Model: Semantic-augmented Multi-level Alignment and Semantic-aware Expert ModelingMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
LLaVA-Scissor: Token Compression with Semantic Connected Components for Video LLMsMultimodal LLMs / vision-language models → Video understanding with MLLMs +1
Adversarial Threats in Quantum Machine Learning: A Survey of Attacks and DefensesLLM safety and security → Machine unlearning
Image2Net: Datasets, Benchmark and Hybrid Framework to Convert Analog Circuit Diagrams into Netlists
The Attribution Crisis in LLM Search Results
Control Synthesis in Partially Observable Environments for Complex Perception-Related Objectives
GameTileNet: A Semantic Dataset for Low-Resolution Game Art in Procedural Content Generation
Towards a Comparative Framework for Compositional AI Models
XxaCT-NN: Structure Agnostic Multimodal Learning for Materials Science
M3: Conversational LLMs Simplify Secure Clinical Data Access, Understanding, and Analysis
Long-Sequence Memory with Temporal Kernels and Dense Hopfield Functionals
Can AI be Consentful?
Mitigating Semantic Collapse in Generative Personalization with Test-Time Embedding Adjustment
Assessing the feasibility of Large Language Models for detecting micro-behaviors in team interactions during space missions
3D Shape Generation: A Survey
Bayesian Invariance Modeling of Multi-Environment Data
DistShap: Scalable GNN Explanations with Distributed Shapley Values
Learning Stochastic Multiscale Models
Interact2Vec -- An efficient neural network-based model for simultaneously learning users and items embeddings in recommender systems
Cost-effective Reduced-Order Modeling via Bayesian Active Learning
Diversity by Design: Addressing Mode Collapse Improves scRNA-seq Perturbation Modeling on Well-Calibrated Metrics
A hierarchical Vovk-Azoury-Warmuth forecaster with discounting for online regression in RKHS
Temperature Matters: Enhancing Watermark Robustness Against Paraphrasing Attacks
Modeling Hierarchical Spaces: A Review and Unified Framework for Surrogate-Based Architecture Design
Deep Hedging to Manage Tail Risk
Ludax: A GPU-Accelerated Domain Specific Language for Board Games
Learning Individual Reproductive Behavior from Aggregate Fertility Rates via Neural Posterior Estimation
A User-Centric, Privacy-Preserving, and Verifiable Ecosystem for Personal Data Management and Utilization

All 267