Papers

New on arXiv

Fri 12 Dec 2025

287 papers

CHIME: Chiplet-based Heterogeneous Near-Memory Acceleration for Edge Multimodal LLM InferenceMultimodal LLMs / vision-language models → Visual token pruning and efficiency
TeleMem: Building Long-Term and Multimodal Memory for Agentic AILLM agents and tool use → Agent memory systems
LeafTutor: An AI Agent for Programming Assignment TutoringLLM agents and tool use → Educational agents and math problem solving
Data-Chain Backdoor: Do You Trust Diffusion Models as Generative Data Supplier?Diffusion models → Robustness, privacy and adversarial aspects +1
LOOPRAG: Enhancing Loop Transformation Optimization with Retrieval-Augmented Large Language ModelsRetrieval-augmented generation → RAG for code
Shapley-based Data Valuation for LLM Alignment via Sequential Preference OptimizationLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
VERAFI: Verified Agentic Financial Intelligence through Neurosymbolic Policy GenerationRetrieval-augmented generation → Legal and financial RAG +1
Persistent Backdoor Attacks under Continual Fine-Tuning of LLMsLLM safety and security → Backdoor and data poisoning attacks
NoveltyRank: A Retrieval-Augmented Framework for Conceptual Novelty Estimation in AI ResearchRetrieval-augmented generation → Agentic RAG, agents and memory
VEGAS: Mitigating Hallucinations in Large Vision-Language Models via Vision-Encoder Attention Guided Adaptive SteeringMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
CLOAK: Contrastive Guidance for Latent Diffusion-Based Data ObfuscationDiffusion models → Robustness, privacy and adversarial aspects
FloodSQL-Bench: A Retrieval-Augmented Benchmark for Geospatially-Grounded Text-to-SQLLLM reasoning → Knowledge- and tool-augmented reasoning +1
RePack then Refine: Efficient Diffusion Transformer with Vision Foundation ModelDiffusion models → Diffusion foundations and sampling
BAgger: Backwards Aggregation for Mitigating Drift in Autoregressive Video Diffusion ModelsWorld models → Video-generation world models and simulators +1
Rethinking Jailbreak Detection of Large Vision Language Models with Representational Contrastive ScoringLLM reasoning → Reasoning for agents, planning and code +2
The Instability of Safety: How Random Seeds and Temperature Expose Inconsistent LLM Refusal BehaviorLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
CreativeVR: Diffusion-Prior-Guided Approach for Structure and Motion Restoration in Generative and Real VideosDiffusion models → Inverse problems and image restoration
Goal Reaching with Eikonal-Constrained Hierarchical Quasimetric Reinforcement LearningDeep reinforcement learning methods → Deep RL algorithms and representations
Benchmarking Contextual Understanding for In-Car Conversational SystemsLLM reasoning → Reasoning benchmarks and evaluation
Hold Onto That Thought: Assessing KV Cache Compression On ReasoningLLM reasoning → Test-time compute and efficient thinking +1
Direct Confidence Alignment: Aligning Verbalized Confidence with Internal Confidence In Large Language ModelsLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
V-REX: Benchmarking Exploratory Visual Reasoning via Chain-of-QuestionsLLM reasoning → Multimodal reasoning +1
Policy Gradient Algorithms for Age-of-Information Cost MinimizationDeep reinforcement learning methods → RL for energy, networks and operations
Neural Chameleons: Language Models Can Learn to Hide Their Thoughts from Unseen Activation MonitorsLLM reasoning → Prompting and chain-of-thought reasoning
How AI Agents Follow the Herd of AI? Network Effects, History, and Machine OptimismLLM agents and tool use → Social simulation and multi-agent debate
AGAPI-Agents: An Open-Access Agentic AI Platform for Accelerated Materials Design on AtomGPT.orgLLM agents and tool use → Medical and scientific domain agents
The Agentic Regulator: Risks for AI in Finance and a Proposed Agent-based Framework for GovernanceLLM agents and tool use → Financial trading and forecasting agents
Evolutionary Reinforcement Learning based AI tutor for Socratic Interdisciplinary InstructionDeep reinforcement learning methods → RL training of language and agent models
AnchorDream: Repurposing Video Diffusion for Embodiment-Aware Robot Data SynthesisWorld models → Embodied and robot world models +2
Structure From Tracking: Distilling Structure-Preserving Motion for Video GenerationDiffusion models → Video generation
BLURR: A Boosted Low-Resource Inference for Vision-Language-Action ModelsVision-language-action models → Efficient VLA inference and adaptation
SVG-T2I: Scaling Up Text-to-Image Latent Diffusion Model Without Variational AutoencoderDiffusion models → Image generation, text-to-image and editing
Speculative Decoding Speed-of-Light: Optimal Lower Bounds via Branching Random WalksEfficient LLM inference and serving → Speculative decoding
EditMGT: Unleashing Potentials of Masked Generative Transformers in Image EditingDiffusion models → Image generation, text-to-image and editing
MedAI: Evaluating TxAgent's Therapeutic Agentic Reasoning in the NeurIPS CURE-Bench CompetitionRetrieval-augmented generation → Medical and biomedical RAG +1
Cross-modal Context-aware Learning for Visual Prompt Guided Multimodal Image Understanding in Remote SensingLLM reasoning → Spatial and embodied reasoning
FactorPortrait: Controllable Portrait Animation via Disentangled Expression, Pose, and ViewpointDiffusion models → Audio, speech and music
Bounding Hallucinations: Merlin-Arthur Protocols for Mutual-Information Bounds in Language ModelsRetrieval-augmented generation → RAG robustness and adaptive retrieval
Using GUI Agent for Electronic Design AutomationLLM agents and tool use → GUI and mobile agents
UniBYD: A Unified Framework for Learning Robotic Manipulation Across Embodiments Beyond Imitation of Human DemonstrationsOffline RL and imitation learning from data → Robot manipulation imitation learning +1
A Study of Library Usage in Agent-Authored Pull RequestsLLM agents and tool use → Software engineering coding agents
Atomic Action Slicing: Planner-Aligned Options for Generalist VLA AgentsVision-language-action models → Generalist VLA models and benchmarks
DentalGPT: Incentivizing Multimodal Complex Reasoning in DentistryLLM reasoning → Clinical and biomedical applications +1
PD-Swap: Prefill-Decode Logic Swapping for End-to-End LLM Inference on Edge FPGAs via Dynamic Partial ReconfigurationEfficient LLM inference and serving → Serving systems and request scheduling
xGR: Efficient Generative Recommendation Serving at ScaleEfficient LLM inference and serving → KV cache optimization
Reconstruction as a Bridge for Event-Based Visual Question AnsweringMultimodal LLMs / vision-language models → Video understanding with MLLMs
EmeraldMind: A Knowledge Graph-Augmented Framework for Greenwashing DetectionLLM agents and tool use → Web, search, and research agents
TSkel-Mamba: Temporal Dynamic Modeling via State Space Model for Human Skeleton-based Action RecognitionState-space sequence models (S4, Mamba) → Video, motion and generative Mamba models
Mistake Notebook Learning: Batch-Clustered Failures for Training-Free Agent AdaptationLLM agents and tool use → General LLM agent frameworks and benchmarks
Rethinking Expert Trajectory Utilization in LLM Post-training for Mathematical ReasoningLLM reasoning → Reinforcement learning for reasoning
Exploring MLLM-Diffusion Information Transfer with MetaCanvasMultimodal LLMs / vision-language models → Image editing and quality assessment
Flowception: Temporally Expansive Flow Matching for Video GenerationDiffusion models → Video generation
AgentBalance: Backbone-then-Topology Design for Cost-Effective Multi-Agent Systems under Budget ConstraintsLLM agents and tool use → General LLM agent frameworks and benchmarks
Towards Trustworthy Multi-Turn LLM Agents via Behavioral GuidanceLLM reasoning → Reasoning for agents, planning and code +1
Minimal Clips, Maximum Salience: Long Video Summarization via Key Moment ExtractionMultimodal LLMs / vision-language models → Video understanding with MLLMs
AutoFSM: A Multi-agent Framework for FSM Code Generation with IR and SystemC-Based TestingLLM agents and tool use → Software engineering coding agents
FlowDC: Flow-Based Decoupling-Decay for Complex Image EditingDiffusion models → Image generation, text-to-image and editing
Mitigating the Safety Alignment Tax with Null-Space Constrained Policy OptimizationLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks
An Anatomy of Vision-Language-Action Models: From Modules to Milestones and ChallengesVision-language-action models → Generalist VLA models and benchmarks
Incremental Validation of Automated Driving Functions using Generic Volumes in Micro- Operational Design DomainsEnd-to-end autonomous driving → Adversarial scenarios and safety testing
PhraseVAE and PhraseLDM: Latent Diffusion for Full-Song Multitrack Symbolic Music GenerationDiffusion models → Audio, speech and music
Symmetry-Aware Steering of Equivariant Diffusion Policies: Benefits and LimitsDiffusion models → Diffusion and flow policies for robotics
DAPO: Design Structure-Aware Pass Ordering in High-Level Synthesis with Graph Contrastive and Reinforcement LearningDeep reinforcement learning methods → RL for energy, networks and operations
UFVideo: Towards Unified Fine-Grained Video Cooperative Understanding with Large Language ModelsMultimodal LLMs / vision-language models → Video understanding with MLLMs
Robust MLLM Unlearning via Visual Knowledge DistillationLLM safety and security → Machine unlearning +1
CAPTURE: A Benchmark and Evaluation for LVLMs in CAPTCHA ResolvingLLM reasoning → Multimodal reasoning +1
Benchmarking the Generality of Vision-Language-Action ModelsVision-language-action models → Generalist VLA models and benchmarks
Unifying Dynamic Tool Creation and Cross-Task Experience Sharing through Cognitive Memory ArchitectureLLM agents and tool use → Agent memory systems
AdaSD: Adaptive Speculative Decoding for Efficient Language Model InferenceEfficient LLM inference and serving → Speculative decoding
When Actions Teach You to Think: Reasoning-Action Synergy via Reinforcement Learning in Conversational AgentsLLM reasoning → Reinforcement learning for reasoning +1
Towards Logic-Aware Manipulation: A Knowledge Primitive for VLM-Based Assistants in Smart ManufacturingMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
TriFlow: A Progressive Multi-Agent Framework for Intelligent Trip PlanningLLM agents and tool use → General LLM agent frameworks and benchmarks
A-LAMP: Agentic LLM-Based Framework for Automated MDP Modeling and Policy GenerationLLM agents and tool use → Reinforcement learning for LLM agents
Features Emerge as Discrete States: The First Application of SAEs to 3D RepresentationsMechanistic interpretability → Sparse autoencoder training and evaluation
Multi-Objective Reinforcement Learning for Large-Scale Mixed Traffic ControlDeep reinforcement learning methods → RL for autonomous driving and traffic
REST: Diffusion-based Real-time End-to-end Streaming Talking Head Generation via ID-Context Caching and Asynchronous Streaming DistillationDiffusion models → Audio, speech and music
FutureX: Enhance End-to-End Autonomous Driving via Latent Chain-of-Thought World ModelWorld models → World models for autonomous driving +2
VFMF: World Modeling by Forecasting Vision Foundation Model FeaturesWorld models → Video-generation world models and simulators
Adaptive Soft Rolling KV Freeze with Entropy-Guided Recovery: Sublinear Memory Growth for Efficient LLM InferenceEfficient LLM inference and serving → KV cache optimization
Seeing to Act, Prompting to Specify: A Bayesian Factorization of Vision Language Action PolicyMultimodal LLMs / vision-language models → Embodied and 3D vision-language models +1
SmokeBench: Evaluating Multimodal Large Language Models for Wildfire Smoke DetectionMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
FutureWeaver: Planning Test-Time Compute for Multi-Agent Systems with Modularized CollaborationLLM reasoning → Reasoning for agents, planning and code +1
AutoRefiner: Improving Autoregressive Video Diffusion Models via Reflective Refinement Over the Stochastic Sampling PathDiffusion models → Video generation
Beyond Memorization: Selective Learning for Copyright-Safe Diffusion Model TrainingDiffusion models → Robustness, privacy and adversarial aspects
Operation Veja: Fixing Fundamental Concepts Missing from Modern Roleplaying Training Paradigms
TrueBrief: Faithful Summarization through Small Language Models
Forecasting N-Body Dynamics: A Comparative Study of Neural Ordinary Differential Equations and Universal Differential Equations
Flow Gym: A framework for the development, benchmarking, training, and deployment of flow-field quantification methods
Generative AI for Analysts
A 96pJ/Frame/Pixel and 61pJ/Event Anti-UAV System with Hybrid Object Tracking Modes
PHANTOM: Progressive High-fidelity Adversarial Network for Threat Object Modeling
Bridging Data and Physics: A Graph Neural Network-Based Hybrid Twin Framework
AdaGradSelect: An adaptive gradient-guided layer selection method for efficient fine-tuning of SLMs
Cross-Sample Augmented Test-Time Adaptation for Personalized Intraoperative Hypotension Prediction
Machine Learning Framework for Thrombosis Risk Prediction in Rotary Blood Pumps
A Tutorial on Dimensionless Learning: Geometric Interpretation and the Effect of Noise
Semantic-Constrained Federated Aggregation: Convergence Theory and Privacy-Utility Bounds for Knowledge-Enhanced Distributed Learning
Yantra AI -- An intelligence platform which interacts with manufacturing operations
Twin Restricted Kernel Machines for Multiview Classification
Scientific feedback reveals productive regimes beyond training conventions

All 287