Papers

New on arXiv

Thu 17 Oct 2024

441 papers

IGOR: Image-GOal Representations are the Atomic Control Units for Foundation Models in Embodied AIWorld models → Embodied and robot world models
Stochastic Flow Matching for Resolving Small-Scale PhysicsDiffusion models → Flow matching, time series and science
ControlAgent: Automating Control System Design via Novel Integration of LLM Agents and Domain ExpertiseLLM agents and tool use → General LLM agent frameworks and benchmarks
LocateBench: Evaluating the Locating Ability of Vision Language ModelsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Holistic Utility Preference Learning for Listwise AlignmentLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
CFTS-GAN: Continual Few-Shot Teacher Student for Generative Adversarial NetworksKnowledge distillation → General knowledge distillation methods
Eliciting Uncertainty in Chain-of-Thought to Mitigate Bias against Forecasting Harmful User BehaviorsLLM reasoning → Test-time compute and efficient thinking
CAKD: A Correlation-Aware Knowledge Distillation Framework Based on Decoupling Kullback-Leibler DivergenceKnowledge distillation → General knowledge distillation methods
MMAD-Purify: A Precision-Optimized Framework for Efficient and Scalable Multi-Modal AttacksDiffusion models → Robustness, privacy and adversarial aspects
Reward-free World Models for Online Imitation LearningWorld models → Embodied and robot world models +2
Efficient Vision-Language Models by Summarizing Visual Tokens into Compact RegistersMultimodal LLMs / vision-language models → Visual token pruning and efficiency
Rethinking Optimal Transport in Offline Reinforcement LearningOffline RL and imitation learning from data → Offline RL algorithms
Provable Benefits of Complex Parameterizations for Structured State Space ModelsState-space sequence models (S4, Mamba) → Core SSM architectures for language
Towards Cross-Cultural Machine Translation with Retrieval-Augmented Generation from Multilingual Knowledge GraphsRetrieval-augmented generation → Generation and context integration
Style-Compress: An LLM-Based Prompt Compression Framework Considering Task-Specific StylesEfficient LLM inference and serving → Prompt and context compression
From Barriers to Tactics: A Behavioral Science-Informed Agentic Workflow for Personalized Nutrition CoachingLLM agents and tool use → Medical and scientific domain agents
WARPD: World model Assisted Reactive Policy DiffusionWorld models → Embodied and robot world models +2
Sliding Puzzles Gym: A Scalable Benchmark for State Representation in Visual Reinforcement LearningDeep reinforcement learning methods → RL for robot locomotion and manipulation
Language Conditioned Multi-Finger Dexterous Manipulation Enabled by Physical Compliance and Switching of ControllersVision-language-action models → Human-video and latent-action pretraining
Personalized Adaptation via In-Context Preference LearningLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
Adversarial Inception Backdoor Attacks against Reinforcement LearningDeep reinforcement learning methods → Adversarial robustness and RL in finance
RiTeK: A Dataset for Large Language Models Complex Reasoning over Textual Knowledge Graphs in MedicineLLM reasoning → Clinical and biomedical applications
Debiasing Large Vision-Language Models by Ablating Protected Attribute RepresentationsMultimodal LLMs / vision-language models → Social bias and fairness in VLMs
Sparse Mixture-of-Experts for Compositional Generalization: Empirical Evidence and Theoretical Foundations of Optimal SparsityMixture-of-experts → MoE theory and gating functions
Approximating Auction Equilibria with Reinforcement LearningDeep reinforcement learning methods → Adversarial robustness and RL in finance
On Diffusion Models for Multi-Agent Partial Observability: Shared Attractors, Error Bounds, and Composite FlowDiffusion models → Diffusion and flow policies for robotics
Automatically Interpreting Millions of Features in Large Language ModelsMechanistic interpretability → SAE features and their applications
FiTv2: Scalable and Improved Flexible Vision Transformer for Diffusion ModelDiffusion models → Diffusion foundations and sampling
LLM Agent Honeypot: Monitoring AI Hacking Agents in the WildLLM agents and tool use → Agent threats, guardrails and governance
FTSmartAudit: A Knowledge Distillation-Enhanced Framework for Automated Smart Contract Auditing Using Fine-Tuned LLMsKnowledge distillation → Knowledge distillation for LLMs
γ-MoD: Exploring Mixture-of-Depth Adaptation for Multimodal Large Language ModelsMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
How Numerical Precision Affects Arithmetical Reasoning Capabilities of LLMsLLM reasoning → Mathematical reasoning
Diffusing States and Matching Scores: A New Framework for Imitation LearningDiffusion models → Diffusion and flow policies for robotics +1
Can MLLMs Understand the Deep Implication Behind Chinese Images?Multimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Influence Functions for Scalable Data Attribution in Diffusion ModelsDiffusion models → Diffusion foundations and sampling
LightTransfer: Your Long-Context LLM is Secretly a Hybrid Model with Effortless AdaptationLLM reasoning → Test-time compute and efficient thinking
D-FINE: Redefine Regression Task in DETRs as Fine-grained Distribution RefinementKnowledge distillation → Detection and dense prediction distillation
Accelerating Codec-based Speech Synthesis with Multi-Token Prediction and Speculative DecodingEfficient LLM inference and serving → Speculative decoding
ORSO: Accelerating Reward Design via Online Reward Selection and Policy OptimizationDeep reinforcement learning methods → Deep RL algorithms and representations
A Common Pitfall of Margin-based Language Model Alignment: Gradient EntanglementLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
AgentOccam: A Simple Yet Strong Baseline for LLM-Based Web AgentsLLM agents and tool use → Web, search, and research agents
Harnessing Webpage UIs for Text-Rich Visual UnderstandingMultimodal LLMs / vision-language models → Chart understanding and anomaly detection
Guided Reinforcement Learning for Robust Multi-Contact Loco-ManipulationDeep reinforcement learning methods → RL for robot locomotion and manipulation
Steering Your Generalists: Improving Robotic Foundation Models via Value GuidanceVision-language-action models → Reinforcement learning and post-training
De-mark: Watermark Removal in Large Language ModelsLLM safety and security → Watermarking generated content and models
Improving Consistency in Diffusion Models for Image Super-ResolutionDiffusion models → Inverse problems and image restoration
A Watermark for Order-Agnostic Language ModelsLLM safety and security → Watermarking generated content and models
Arbitrarily-Conditioned Multi-Functional Diffusion for Multi-Physics EmulationDiffusion models → Flow matching, time series and science
PopAlign: Diversifying Contrasting Patterns for a More Comprehensive AlignmentLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
Is Prior-Free Black-Box Non-Stationary Reinforcement Learning Feasible?Deep reinforcement learning methods → RL theory and sample complexity
Probing the Latent Hierarchical Structure of Data via Diffusion ModelsDiffusion models → Diffusion foundations and sampling
Theory on Score-Mismatched Diffusion Models and Zero-Shot Conditional SamplersDiffusion models → Diffusion foundations and sampling
Instance-dependent Convergence Theory for Diffusion ModelsDiffusion models → Diffusion foundations and sampling
Improving Multi-modal Large Language Model through Boosting Vision CapabilitiesMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
MIRAGE-Bench: Automatic Multilingual Benchmark Arena for Retrieval-Augmented Generation SystemsRetrieval-augmented generation → RAG pipeline design and evaluation
On the Role of Attention Heads in Large Language Model SafetyLLM safety and security → Safety alignment, refusal and guardrails
Unconstrained Model Merging for Enhanced LLM ReasoningLLM reasoning → Test-time compute and efficient thinking
Exploring the Design Space of Visual Context Representation in Video MLLMsMultimodal LLMs / vision-language models → Video understanding with MLLMs
Jailbreaking LLM-Controlled RobotsLLM safety and security → Prompt injection and agent security
Diffusion Curriculum: Synthetic-to-Real Data Curriculum via Image-Guided DiffusionDiffusion models → Diffusion foundations and sampling
HEALTH-PARIKSHA: Assessing RAG Models for Health Chatbots in Real-World Multilingual SettingsRetrieval-augmented generation → Medical and biomedical RAG
VL-GLUE: A Suite of Fundamental yet Challenging Visuo-Linguistic Reasoning TasksLLM reasoning → Multimodal reasoning
Fine-Tuning Discrete Diffusion Models via Reward Optimization with Applications to DNA and Protein DesignDiffusion models → Flow matching, time series and science +1
An Active Learning Framework for Inclusive Generation by Large Language ModelsKnowledge distillation → Knowledge distillation for LLMs
A Comparative Study on Reasoning Patterns of OpenAI's o1 ModelLLM reasoning → Test-time compute and efficient thinking
H2OVL-Mississippi Vision Language Models Technical ReportMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
MeNTi: Bridging Medical Calculator and LLM Agent with Nested Tool CallingLLM agents and tool use → Medical and scientific domain agents
Text-guided multi-property molecular optimization with a diffusion language modelDiffusion models → Flow matching, time series and science
Preference Aligned Diffusion Planner for Quadrupedal Locomotion ControlDiffusion models → Diffusion and flow policies for robotics
DriveDreamer4D: World Models Are Effective Data Machines for 4D Driving Scene RepresentationWorld models → World models for autonomous driving +1
RemoteDet-Mamba: A Hybrid Mamba-CNN Network for Multi-modal Object Detection in Remote Sensing ImagesState-space sequence models (S4, Mamba) → Point clouds, detection and remote sensing
Can Medical Vision-Language Pre-training Succeed with Purely Synthetic Data?Multimodal LLMs / vision-language models → Medical vision-language models +1
Contracting With a Reinforcement Learning Agent by Playing Trick or TreatDeep reinforcement learning methods → RL theory and sample complexity
GeoCoder: Solving Geometry Problems by Generating Modular Code through Vision-Language ModelsLLM reasoning → Spatial and embodied reasoning +1
RAG-DDR: Optimizing Retrieval-Augmented Generation Using Differentiable Data RewardsRetrieval-augmented generation → Agentic RAG, agents and memory
Integrating Large Language Models and Reinforcement Learning for Non-Linear ReasoningLLM reasoning → Reinforcement learning for reasoning
Deep Reinforcement Learning for Online Optimal Execution StrategiesDeep reinforcement learning methods → Adversarial robustness and RL in finance
Novelty-based Sample Reuse for Continuous Robotics ControlDeep reinforcement learning methods → Deep RL algorithms and representations
Truncating Trajectories in Monte Carlo Policy Evaluation: an Adaptive ApproachDeep reinforcement learning methods → RL theory and sample complexity
Progressive Mixed-Precision Decoding for Efficient LLM InferenceEfficient LLM inference and serving → KV cache quantization
Solving Prior Distribution Mismatch in Diffusion Models via Optimal TransportDiffusion models → Diffusion foundations and sampling
Generate and Instantiate What You Prefer: Text-Guided Diffusion for Sequential RecommendationDiffusion models → Diffusion recommenders
Think Thrice Before You Act: Progressive Thought Refinement in Large Language ModelsLLM reasoning → Test-time compute and efficient thinking
MagicTailor: Component-Controllable Personalization in Text-to-Image Diffusion ModelsDiffusion models → Image generation, text-to-image and editing
RAP: Retrieval-Augmented Personalization for Multimodal Large Language ModelsMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning +1
LAR-ECHR: A New Legal Argument Reasoning Task and Dataset for Cases of the European Court of Human RightsLLM reasoning → Reasoning for agents, planning and code
Cerberus: Efficient Inference with Adaptive Parallel Decoding and Sequential Knowledge EnhancementEfficient LLM inference and serving → Speculative decoding
Probing-RAG: Self-Probing to Guide Language Models in Selective Document RetrievalRetrieval-augmented generation → RAG pipeline design and evaluation
SSD-TS: Exploring the Potential of Linear State Space Models for Diffusion Models in Time Series ImputationDiffusion models → Flow matching, time series and science
BiasJailbreak:Analyzing Ethical Biases and Jailbreak Vulnerabilities in Large Language ModelsLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
Comparing the Utility, Preference, and Performance of Course Material Search Functionality and Retrieval-Augmented Generation Large Language Model (RAG-LLM) AI Chatbots in Information-Seeking TasksRetrieval-augmented generation → Agentic RAG, agents and memory
Arc-Length-Based Warping for Robot Skill Synthesis from Multiple DemonstrationsOffline RL and imitation learning from data → Robot learning from demonstration
Mitigating Hallucinations in Large Vision-Language Models via Summary-Guided DecodingMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
Precipitation Nowcasting Using Diffusion Transformer with Causal AttentionDiffusion models → Flow matching, time series and science
Advancing Large Language Model Attribution through Self-ImprovingLLM reasoning → Test-time compute and efficient thinking +1
SBI-RAG: Enhancing Math Word Problem Solving for Students through Schema-Based Instruction and Retrieval-Augmented GenerationRetrieval-augmented generation → Agentic RAG, agents and memory
SeerAttention: Learning Intrinsic Sparse Attention in Your LLMsMixture-of-experts → MoE large language models
Breaking Chains: Unraveling the Links in Multi-Hop Knowledge UnlearningLLM safety and security → Machine unlearning
FRAG: Toward Federated Vector Database Management for Collaborative and Secure Retrieval-Augmented GenerationRetrieval-augmented generation → Privacy and security of RAG
Constraint Decoupled Latent Diffusion for Protein BackmappingDiffusion models → Flow matching, time series and science

All 441