Papers

New on arXiv

Wed 3 Dec 2025

375 papers

RAGVUE: A Diagnostic View for Explainable and Automated Evaluation of Retrieval-Augmented GenerationRetrieval-augmented generation → RAG pipeline design and evaluation
D-STEER - Preference Alignment Techniques Learn to Behave, not to Believe -- Beneath the Surface, DPO as Steering Vector Perturbation in Activation SpaceLLM alignment / RLHF / preference optimisation → LLM preference optimization methods +1
On the Design of One-step Diffusion via Shortcutting Flow PathsDiffusion models → Diffusion foundations and sampling
Cognitive Mirrors: Exploring the Diverse Functional Roles of Attention Heads in LLM ReasoningLLM reasoning → Reasoning benchmarks and evaluation
Peek-a-Boo Reasoning: Contrastive Region Masking in MLLMsLLM reasoning → Multimodal reasoning +1
EFDiT: Efficient Fine-grained Image Generation Using Diffusion Transformer ModelsDiffusion models → Diffusion foundations and sampling
Data-regularized Reinforcement Learning for Diffusion Models at ScaleDiffusion models → Diffusion and flow policies for robotics +1
A Retrieval-Augmented Generation Approach to Extracting Algorithmic Logic from Neural NetworksRetrieval-augmented generation → RAG pipeline design and evaluation
DAComp: Benchmarking Data Agents across the Full Data Intelligence LifecycleLLM agents and tool use → Web, search, and research agents
ResponsibleRobotBench: Benchmarking Responsible Robot Manipulation using Multi-modal Large Language ModelsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Evaluating Long-Context Reasoning in LLM-Based WebAgentsLLM reasoning → Reasoning for agents, planning and code
How (Mis)calibrated is your Federated CLIP and what to do about it?Multimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
Towards better dense rewards in Reinforcement Learning ApplicationsLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF +1
Plug-and-Play Image Restoration with Flow Matching: A Continuous ViewpointDiffusion models → Inverse problems and image restoration
Driving Beyond Privilege: Distilling Dense-Reward Knowledge into Sparse-Reward PoliciesWorld models → World models for autonomous driving +1
Hey GPT-OSS, Looks Like You Got It -- Now Walk Me Through It! An Assessment of the Reasoning Language Models Chain of Thought Mechanism for Digital ForensicsLLM reasoning → Reasoning for agents, planning and code
MVRoom: Controllable 3D Indoor Scene Generation with Multi-View Diffusion ModelsDiffusion models → 3D, point cloud and scene generation
Toward Virtuous Reinforcement Learning: A Critique and RoadmapDeep reinforcement learning methods → RL training of language and agent models
6 Fingers, 1 Kidney: Natural Adversarial Medical Images Reveal Critical Weaknesses of Vision-Language ModelsMultimodal LLMs / vision-language models → Medical vision-language models
Addressing Logical Fallacies In Scientific Reasoning From Large Language Models: Towards a Dual-Inference Training FrameworkLLM reasoning → Reasoning benchmarks and evaluation
ReasonX: MLLM-Guided Intrinsic Image DecompositionLLM reasoning → Multimodal reasoning
MoReGen: Multi-Agent Motion-Reasoning Engine for Code-based Text-to-Video SynthesisLLM reasoning → Spatial and embodied reasoning +1
On Group Relative Policy Optimization Collapse in Agent Search: The Lazy Likelihood-DisplacementLLM reasoning → Reinforcement learning for reasoning +1
Balancing Safety and Helpfulness in Healthcare AI Assistants through Iterative Preference AlignmentLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks
RippleBench: Capturing Ripple Effects Using Existing Knowledge RepositoriesLLM safety and security → Machine unlearning
Decoding Large Language Diffusion Models with Foreseeing MovementDiffusion models → Discrete and text diffusion
Don't Trust Your Upstream: Exploiting LLM Multi-Agent System via Topology-Guided Adversarial PropagationLLM agents and tool use → Agent threats, guardrails and governance
SkillFactory: Self-Distillation For Learning Cognitive BehaviorsLLM reasoning → Reinforcement learning for reasoning +1
SpaceTools: Tool-Augmented Spatial Reasoning via Double Interactive RLLLM reasoning → Spatial and embodied reasoning +1
MarkTune: Improving the Quality-Detectability Trade-off in Open-Weight LLM WatermarkingLLM safety and security → Watermarking generated content and models
RELIC: Interactive Video World Model with Long-Horizon MemoryWorld models → Video-generation world models and simulators
Fast & Efficient Normalizing Flows and Applications of Image Generative ModelsDiffusion models → Flow matching, time series and science
jina-vlm: Small Multilingual Vision Language ModelMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Divide, then Ground: Adapting Frame Selection to Query Types for Long-Form Video UnderstandingMultimodal LLMs / vision-language models → Video understanding with MLLMs
Highly Efficient Test-Time Scaling for T2I Diffusion Models with Text Embedding PerturbationLLM reasoning → Diffusion reasoning and test-time scaling +1
Value-Guided Iterative Refinement and the DIQ-H Benchmark for Evaluating VLM RobustnessMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
BlurDM: A Blur Diffusion Model for Image DeblurringDiffusion models → Inverse problems and image restoration
Guided Flow Policy: Learning from High-Value Actions in Offline Reinforcement LearningOffline RL and imitation learning from data → Diffusion and flow policies for offline RL
Training for Identity, Inference for Controllability: A Unified Approach to Tuning-Free Face PersonalizationDiffusion models → Image generation, text-to-image and editing
TempR1: Improving Temporal Understanding of MLLMs via Temporal-Aware Multi-Task Reinforcement LearningMultimodal LLMs / vision-language models → Video understanding with MLLMs
Benchmark for Planning and Control with Large Language Model Agents: Blocksworld with Model Context ProtocolLLM agents and tool use → General LLM agent frameworks and benchmarks
Driving is a Game: Combining Planning and Prediction with Bayesian Iterative Best ResponseEnd-to-end autonomous driving → Imitation-learning end-to-end planners
OD-MoE: On-Demand Expert Loading for Cacheless Edge-Distributed MoE InferenceMixture-of-experts → Efficient MoE training and inference systems
A Theoretical Framework for Auxiliary-Loss-Free Load Balancing of Sparse Mixture-of-Experts in Large-Scale AI ModelsMixture-of-experts → Expert routing strategies
Hierarchical Vision Language Action Model Using Success and Failure DemonstrationsVision-language-action models → Failure recovery and robustness
Autonomous Reinforcement Learning Robot Control with Intel's Loihi 2 Neuromorphic HardwareDeep reinforcement learning methods → Spiking and neuromorphic RL
Zero-Shot Video Translation and Editing with Frame Spatial-Temporal CorrespondenceDiffusion models → Video generation
Digital Twin-based Control Co-Design of Full Vehicle Active Suspensions via Deep Reinforcement LearningDeep reinforcement learning methods → RL for autonomous driving and traffic
A Hierarchical Tree-based approach for creating Configurable and Static Deep Research Agent (Static-DRA)Retrieval-augmented generation → Agentic RAG, agents and memory +1
A Modular Architecture Design for Autonomous Driving Racing in Controlled EnvironmentsEnd-to-end autonomous driving → Imitation-learning end-to-end planners
OmniDexVLG: Learning Dexterous Grasp Generation from Vision Language Model-Guided Grasp Semantics, Taxonomy and Functional AffordanceMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Reconstructing KV Caches with Cross-layer Fusion For Enhanced TransformersEfficient LLM inference and serving → KV cache optimization
Traffic Image Restoration under Adverse Weather via Frequency-Aware MambaState-space sequence models (S4, Mamba) → Mamba for image restoration
CoDA: From Text-to-Image Diffusion Models to Training-Free Dataset DistillationDiffusion models → Diffusion foundations and sampling
Training and Evaluation of Guideline-Based Medical Reasoning in LLMsLLM reasoning → Clinical and biomedical applications
Deep Reinforcement Learning for Dynamic Algorithm Configuration: A Case Study on Optimizing OneMax with the (1+(λ,λ))-GADeep reinforcement learning methods → Deep RL algorithms and representations
MPCFormer: A physics-informed data-driven approach for explainable socially-aware autonomous drivingEnd-to-end autonomous driving → Cooperative and V2X driving
AdaptVision: Efficient Vision-Language Models via Adaptive Visual AcquisitionMultimodal LLMs / vision-language models → Visual token pruning and efficiency
Omni-AutoThink: Adaptive Multimodal Reasoning via Reinforcement LearningLLM reasoning → Multimodal reasoning
Safety Reinforced Model Predictive Control (SRMPC): Improving MPC with Reinforcement Learning for Motion Planning in Autonomous DrivingEnd-to-end autonomous driving → Reinforcement learning for driving +1
In-Context Representation HijackingLLM safety and security → Jailbreaking
Sample-Efficient Model-Free Policy Gradient Methods for Stochastic LQR via Robust Linear RegressionDeep reinforcement learning methods → RL for games and combinatorial optimization
Fully Unsupervised Self-debiasing of Text-to-Image Diffusion ModelsDiffusion models → Image generation, text-to-image and editing
Thinking with Programming Vision: Towards a Unified View for Thinking with ImagesLLM reasoning → Multimodal reasoning +1
Crossing the Sim2Real Gap Between Simulation and Ground Testing to Space Deployment of Autonomous Free-flyer ControlDeep reinforcement learning methods → RL for robot navigation and UAV control
PosA-VLA: Enhancing Action Generation via Pose-Conditioned Anchor AttentionVision-language-action models → Generalist VLA models and benchmarks
ContactRL: Safe Reinforcement Learning based Motion Planning for Contact based Human Robot CollaborationDeep reinforcement learning methods → RL for robot locomotion and manipulation
Overcoming State Inertia: Minimally Invasive Temporal Alignment for Evolving ContextsLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
SRPG: Semantically Reconstructed Privacy Guard for Zero-Trust Privacy in Educational Multi-Agent SystemsLLM agents and tool use → Agent threats, guardrails and governance
ConvRot: Rotation-Based Plug-and-Play 4-bit Quantization for Diffusion TransformersDiffusion models → Diffusion foundations and sampling
Colon-X: Advancing Intelligent Colonoscopy toward Clinical ReasoningLLM reasoning → Clinical and biomedical applications
Dynamically Scaled Activation SteeringMechanistic interpretability → Activation steering methods
Optical Context Compression Is Just (Bad) AutoencodingEfficient LLM inference and serving → Visual token pruning in multimodal models
The promising potential of vision language models for the generation of textual weather forecastsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
KVNAND: Efficient On-Device Large Language Model Inference Using DRAM-Free In-Flash ComputingEfficient LLM inference and serving → Serving systems and request scheduling
EnCompass: Enhancing Agent Programming with Search Over Program Execution PathsLLM agents and tool use → Software engineering coding agents
Synthetic Cognitive Walkthrough: Aligning Large Language Model Performance with Human Cognitive WalkthroughLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
GAOT: Generating Articulated Objects Through Text-Guided Diffusion ModelsDiffusion models → Image generation, text-to-image and editing
Towards Irreversible Machine Unlearning for Diffusion ModelsDiffusion models → Robustness, privacy and adversarial aspects +1
State Space Models for Bioacoustics: A Comparative Evaluation with TransformersState-space sequence models (S4, Mamba) → Speech and audio SSMs
CartoMapQA: A Fundamental Benchmark Dataset Evaluating Vision-Language Models on Cartographic Map UnderstandingMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
RoboScape-R: Unified Reward-Observation World Models for Generalizable Robotics Training via RLWorld models → Latent dynamics models and model-based RL
PARC: An Autonomous Self-Reflective Coding Agent for Robust Execution of Long-Horizon TasksLLM agents and tool use → Medical and scientific domain agents
V-ITI: Mitigating Hallucinations in Multimodal Large Language Models via Visual Inference-Time InterventionMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
AdaPower: Specializing World Foundation Models for Predictive ManipulationWorld models → Embodied and robot world models
Rethinking Prompt Design for Inference-time Scaling in Text-to-Visual GenerationLLM reasoning → Multimodal reasoning
Adaptive sampling using variational autoencoder and reinforcement learningDeep reinforcement learning methods → RL with diffusion and flow models
Exploiting Domain Properties in Language-Driven Domain Generalization for Semantic SegmentationMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
Understanding LLM Reasoning for Abstractive SummarizationLLM reasoning → Reasoning benchmarks and evaluation
EEA: Exploration-Exploitation Agent for Long Video UnderstandingLLM agents and tool use → Multimodal and embodied agents
Fairness-Aware Fine-Tuning of Vision-Language Models for Medical Glaucoma DiagnosisMultimodal LLMs / vision-language models → Medical vision-language models
ATHENA: Agentic Team for Hierarchical Evolutionary Numerical AlgorithmsLLM agents and tool use → Web, search, and research agents
AsymPuzl: An Asymmetric Puzzle for multi-agent cooperationLLM agents and tool use → General LLM agent frameworks and benchmarks
Text-Printed Image: Bridging the Image-Text Modality Gap for Text-centric Training of Large Vision-Language ModelsMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Think Before You Drive: World Model-Inspired Multimodal Grounding for Autonomous VehiclesWorld models → World models for autonomous driving +2
GeoVideo: Introducing Geometric Regularization into Video Generation ModelDiffusion models → 3D, point cloud and scene generation
GalaxyDiT: Efficient Video Generation with Guidance Alignment and Adaptive Proxy in Diffusion TransformersDiffusion models → Video generation
KeyPointDiffuser: Unsupervised 3D Keypoint Learning via Latent Diffusion ModelsDiffusion models → 3D, point cloud and scene generation
Multi-Aspect Knowledge-Enhanced Medical Vision-Language Pretraining with Multi-Agent Data GenerationMultimodal LLMs / vision-language models → Medical vision-language models +1
Multimodal Reinforcement Learning with Adaptive Verifier for AI AgentsLLM reasoning → Multimodal reasoning +2

All 375