Papers

New on arXiv

Thu 6 Mar 2025

391 papers

TS-RAG: Retrieval-Augmented Generation based Time Series Foundation Models are Stronger Zero-Shot ForecasterRetrieval-augmented generation → Time series and agricultural RAG
The day-ahead scenario generation method for new energy based on an improved conditional generative diffusion modelDiffusion models → Flow matching, time series and science
Refined Policy Distillation: From VLA Generalists to RL ExpertsVision-language-action models → Reinforcement learning and post-training
Provably Correct Automata Embeddings for Optimal Automata-Conditioned Reinforcement LearningDeep reinforcement learning methods → Interpretable and logic-guided RL
Bridging the AI Adoption Gap: Designing an Interactive Pedagogical Agent for Higher Education InstructorsLLM agents and tool use → Educational agents and math problem solving
Collapse of Dense Retrievers: Short, Early, and Literal Biases Outranking Factual EvidenceRetrieval-augmented generation → RAG robustness and adaptive retrieval
Continual Pre-training of MoEs: How robust is your router?Mixture-of-experts → Expert routing strategies
Lessons learned from field demonstrations of model predictive control and reinforcement learning for residential and commercial HVAC: A reviewDeep reinforcement learning methods → RL for energy, networks and operations
Safety is Not Only About Refusal: Reasoning-Enhanced Fine-tuning for Interpretable LLM SafetyLLM reasoning → Reasoning for agents, planning and code +3
Quantifying and Modeling Driving Styles in Trajectory ForecastingEnd-to-end autonomous driving → Imitation-learning end-to-end planners
Wanda++: Pruning Large Language Models via Regional GradientsEfficient LLM inference and serving → Early exit and edge inference
LVLM-Compress-Bench: Benchmarking the Broader Impact of Large Vision-Language Model CompressionMultimodal LLMs / vision-language models → Visual token pruning and efficiency +1
Energy-Weighted Flow Matching for Offline Reinforcement LearningDiffusion models → Diffusion and flow policies for robotics +1
Beyond RAG: Task-Aware KV Cache Compression for Comprehensive Knowledge ReasoningLLM reasoning → Test-time compute and efficient thinking +1
DB-Explore: Automated Database Exploration and Instruction Synthesis for Text-to-SQLSynthetic data and self-improvement for LLMs → Synthetic reasoning data and bootstrapping
SafeArena: Evaluating the Safety of Autonomous Web AgentsLLM agents and tool use → Agent threats, guardrails and governance +1
Spectral Informed Mamba for Robust Point Cloud ProcessingState-space sequence models (S4, Mamba) → Point clouds, detection and remote sensing
Curiosity-Driven Imagination: Discovering Plan Operators and Learning Associated Policies for Open-World AdaptationDeep reinforcement learning methods → Deep RL algorithms and representations
HILGEN: Hierarchically-Informed Data Generation for Biomedical NER Using Knowledgebases and Large Language ModelsSynthetic data and self-improvement for LLMs → Medical synthetic data and agents
FirePlace: Geometric Refinements of LLM Common Sense Reasoning for 3D Object PlacementLLM reasoning → Spatial and embodied reasoning
Fine-Tuning Florence2 for Enhanced Object Detection in Un-constructed Environments: Vision-Language Model ApproachMultimodal LLMs / vision-language models → Driving VLMs and open-vocabulary detection
Maximizing Signal in Human-Model Preference AlignmentLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
Adapt3R: Adaptive 3D Scene Representation for Domain Transfer in Imitation LearningOffline RL and imitation learning from data → Robot manipulation imitation learning
TinyR1-32B-Preview: Boosting Accuracy with Branch-Merge DistillationLLM reasoning → Test-time compute and efficient thinking +1
Toward Lightweight and Fast Decoders for Diffusion Models in Image and Video GenerationDiffusion models → Diffusion foundations and sampling
E4: Energy-Efficient DNN Inference for Edge Video Analytics Via Early-Exit and DVFSEfficient LLM inference and serving → Early exit and edge inference
SHAPE : Self-Improved Visual Preference Alignment by Iteratively Generating Holistic WinnerMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning +2
M2S: Multi-turn to Single-turn jailbreak in Red Teaming for LLMsLLM safety and security → Jailbreaking
L1: Controlling How Long A Reasoning Model Thinks With Reinforcement LearningLLM reasoning → Test-time compute and efficient thinking
UIPE: Enhancing LLM Unlearning by Removing Knowledge Related to Forgetting TargetsLLM safety and security → Machine unlearning
Quantifying the Reasoning Abilities of LLMs on Real-world Clinical CasesLLM reasoning → Clinical and biomedical applications
Teach YOLO to Remember: A Self-Distillation Approach for Continual Object DetectionKnowledge distillation → Detection and dense prediction distillation
DIMSUM: Discourse in Mathematical Reasoning as a Supervision ModuleLLM reasoning → Mathematical reasoning
Multi-Agent Inverse Q-Learning from DemonstrationsLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF +1
Implicit Cross-Lingual Rewarding for Efficient Multilingual Preference AlignmentLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
Mark Your LLM: Detecting the Misuse of Open-Source Large Language Models via WatermarkingLLM safety and security → Watermarking generated content and models
START: Self-taught Reasoner with ToolsLLM reasoning → Test-time compute and efficient thinking
Better Process Supervision with Bi-directional Rewarding SignalsLLM reasoning → Reinforcement learning for reasoning
The Best of Both Worlds: Integrating Language Models and Diffusion Models for Video GenerationDiffusion models → Video generation
LLM Applications: Current Paradigms and the Next FrontierLLM agents and tool use → Agent threats, guardrails and governance
A Benchmark for Multi-Lingual Vision-Language Learning in Remote Sensing Image CaptioningMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Compositional Causal Reasoning Evaluation in Language ModelsLLM reasoning → Causal reasoning in LLMs
Benchmarking Reasoning Robustness in Large Language ModelsLLM reasoning → Reasoning benchmarks and evaluation
An Empirical Study on Eliciting and Improving R1-like Reasoning ModelsLLM reasoning → Reasoning benchmarks and evaluation
Keeping Yourself is Important in Downstream Tuning Multimodal Large Language ModelLLM reasoning → Multimodal reasoning +1
SOLAR: Scalable Optimization of Large-scale Architecture for ReasoningLLM reasoning → Prompting and chain-of-thought reasoning
Dynamic Pricing for On-Demand DNN Inference in the Edge-AI MarketEfficient LLM inference and serving → Early exit and edge inference
Generalized Interpolating Discrete DiffusionDiffusion models → Discrete and text diffusion
ToolFuzz -- Automated Agent Tool TestingLLM reasoning → Reasoning for agents, planning and code +1
Know Thy Judge: On the Robustness Meta-Evaluation of LLM Safety JudgesLLM safety and security → Safety alignment, refusal and guardrails
DAST: Difficulty-Adaptive Slow-Thinking for Large Reasoning ModelsLLM reasoning → Test-time compute and efficient thinking
PALo: Learning Posture-Aware Locomotion for Quadruped RobotsDeep reinforcement learning methods → RL for robot locomotion and manipulation
Question-Aware Gaussian Experts for Audio-Visual Question AnsweringLLM reasoning → Audio-language model reasoning +1
TPC: Cross-Temporal Prediction Connection for Vision-Language Model Hallucination ReductionMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
ToFu: Visual Tokens Reduction via Fusion for Multi-modal, Multi-patch, Multi-image TaskMultimodal LLMs / vision-language models → Visual token pruning and efficiency
Activation Space Interventions Can Be Transferred Between Large Language ModelsLLM safety and security → Safety alignment, refusal and guardrails
Revisiting the Othello World Model HypothesisWorld models → World models for language agents
From Idea to CAD: A Language Model-Driven Multi-Agent System for Collaborative DesignLLM agents and tool use → Multimodal and embodied agents
Learning Transformer-based World Models with Contrastive Predictive CodingWorld models → Latent dynamics models and model-based RL
Wider or Deeper? Scaling LLM Inference-Time Compute with Adaptive Branching Tree SearchLLM reasoning → Test-time compute and efficient thinking
Semantic Parallelism: Redefining Efficient MoE Inference via Model-Data Co-SchedulingMixture-of-experts → Efficient MoE training and inference systems +1
AgentSafe: Safeguarding Large Language Model-based Multi-agent Systems via Hierarchical Data ManagementLLM agents and tool use → Agent threats, guardrails and governance
More Documents, Same Length: Isolating the Challenge of Multiple Documents in RAGRetrieval-augmented generation → RAG pipeline design and evaluation
TRACT: Regression-Aware Fine-tuning Meets Chain-of-Thought Reasoning for LLM-as-a-JudgeLLM reasoning → Reasoning benchmarks and evaluation
HelpSteer3: Human-Annotated Feedback and Edit Data to Empower Inference-Time Scaling in Open-Ended General-Domain TasksLLM reasoning → Test-time compute and efficient thinking
Adding Alignment Control to Language ModelsLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
LEDiT: Your Length-Extrapolatable Diffusion Transformer without Positional EncodingDiffusion models → Diffusion foundations and sampling
In-depth Analysis of Graph-based RAG in a Unified FrameworkRetrieval-augmented generation → Knowledge-graph and structured-knowledge RAG
Towards Autonomous Reinforcement Learning for Real-World Robotic Manipulation with Large Language ModelsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Knowledge Retention for Continual Model-Based Reinforcement LearningWorld models → Latent dynamics models and model-based RL +1
ADOR: A Design Exploration Framework for LLM Serving with Enhanced Latency and ThroughputEfficient LLM inference and serving → Serving systems and request scheduling
An Egocentric Vision-Language Model based Portable Real-time Smart AssistantMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
How to Mitigate Overfitting in Weak-to-strong Generalization?LLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
DiffPO: Diffusion-styled Preference Optimization for Efficient Inference-Time Alignment of Large Language ModelsLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
Synthetic Data is an Elegant GIFT for Continual Vision-Language ModelsMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP +1
FuseChat-3.0: Preference Optimization Meets Heterogeneous Model FusionLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
Quantum-Inspired Reinforcement Learning in the Presence of Epistemic AmbivalenceDeep reinforcement learning methods → Quantum reinforcement learning
Energy-Guided Optimization for Personalized Image Editing with Pretrained Text-to-Image Diffusion ModelsDiffusion models → Image generation, text-to-image and editing
Measuring temporal effects of agent knowledge by date-controlled tool useLLM agents and tool use → Web, search, and research agents
The Role of Visual Modality in Multimodal Mathematical Reasoning: Challenges and InsightsLLM reasoning → Multimodal reasoning
VLA Model-Expert Collaboration for Bi-directional Manipulation LearningVision-language-action models → Generalist VLA models and benchmarks
Semantic Retrieval Augmented Contrastive Learning for Sequential RecommendationRetrieval-augmented generation → RAG for recommendation
KidneyTalk-open: No-code Deployment of a Private Large Language Model with Medical Documentation-Enhanced Knowledge Database for Kidney DiseaseLLM reasoning → Clinical and biomedical applications
Dynamic Benchmarking of Reasoning Capabilities in Code Large Language Models Under Data ContaminationLLM reasoning → Reasoning for agents, planning and code
DM-Adapter: Domain-Aware Mixture-of-Adapters for Text-Based Person RetrievalMixture-of-experts → Vision mixture-of-experts
MTS: A Deep Reinforcement Learning Portfolio Management Framework with Time-Awareness and Short-SellingDeep reinforcement learning methods → Adversarial robustness and RL in finance
STORM: Token-Efficient Long Video Understanding for Multimodal LLMsMultimodal LLMs / vision-language models → Video understanding with MLLMs +1
Diff-Reg v2: Diffusion-Based Matching Matrix Estimation for Image Matching and 3D RegistrationDiffusion models → 3D, point cloud and scene generation
Disparities in LLM Reasoning Accuracy and Explanations: A Case Study on African American EnglishLLM reasoning → Reasoning benchmarks and evaluation
Chart-HQA: A Benchmark for Hypothetical Question Answering in ChartsMultimodal LLMs / vision-language models → Chart understanding and anomaly detection
Pok'eChamp: an Expert-level Minimax Language AgentLLM agents and tool use → Reinforcement learning for LLM agents
SED2AM: Solving Multi-Trip Time-Dependent Vehicle Routing Problem using Deep Reinforcement LearningDeep reinforcement learning methods → RL for games and combinatorial optimization
Can We Optimize Deep RL Policy Weights as Trajectory Modeling?Deep reinforcement learning methods → Deep RL algorithms and representations
EVE: Towards End-to-End Video Subtitle Extraction with Vision-Language ModelsMultimodal LLMs / vision-language models → Video understanding with MLLMs
RA-DP: Rapid Adaptive Diffusion Policy for Training-Free High-frequency Robotics ReplanningDiffusion models → Diffusion and flow policies for robotics
Underlying Semantic Diffusion for Effective and Efficient In-Context LearningDiffusion models → Image generation, text-to-image and editing
Robust Data Watermarking in Language Models by Injecting Fictitious KnowledgeLLM safety and security → Watermarking generated content and models
TextDoctor: Unified Document Image Inpainting via Patch Pyramid Diffusion ModelsDiffusion models → 3D, point cloud and scene generation
Dexterous Grasping with Real-World Robotic Reinforcement LearningDeep reinforcement learning methods → RL for robot locomotion and manipulation
Robotic Compliant Object Prying Using Diffusion Policy Guided by Vision and Force ObservationsDiffusion models → Diffusion and flow policies for robotics

All 391