Papers

New on arXiv

Tue 28 May 2024

425 papers

Model-Based Diffusion for Trajectory OptimizationDiffusion models → Diffusion and flow policies for robotics
Interpretable DRL-based Maneuver Decision of UCAV DogfightDeep reinforcement learning methods → RL for robot navigation and UAV control
Personalized Steering of Large Language Models: Versatile Steering Vectors Through Bi-directional Preference OptimizationLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment +1
Improved Generation of Adversarial Examples Against Safety-aligned LLMsLLM safety and security → Jailbreaking
Black-Box Detection of Language Model WatermarksLLM safety and security → Watermarking generated content and models
Imitating from auxiliary imperfect demonstrations via Adversarial Density Weighted RegressionOffline RL and imitation learning from data → Imitation learning methods
PureEBM: Universal Poison Purification via Mid-Run Dynamics of Energy-Based ModelsLLM safety and security → Backdoor and data poisoning attacks
Mitigating Object Hallucination in MLLMs via Data-augmented Phrase-level AlignmentMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
Lisa: Lazy Safety Alignment for Large Language Models against Harmful Fine-tuning AttackLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
ConSiDERS-The-Human Evaluation Framework: Rethinking Human Evaluation for Generative Large Language ModelsLLM reasoning → Reasoning benchmarks and evaluation
A Theoretical Understanding of Self-Correction through In-context AlignmentLLM reasoning → Test-time compute and efficient thinking
Hardware-Aware Parallel Prompt Decoding for Memory-Efficient Acceleration of LLM InferenceEfficient LLM inference and serving → Speculative decoding
PureGen: Universal Data Purification for Train-Time Poison Defense via Generative Model DynamicsDiffusion models → Robustness, privacy and adversarial aspects +1
DTR-Bench: An in silico Environment and Benchmark Platform for Reinforcement Learning Based Dynamic Treatment RegimeDeep reinforcement learning methods → RL for clinical treatment and surgery
OpenConvoy: Universal Platform for Real-World Testing of Cooperative Driving SystemsEnd-to-end autonomous driving → Cooperative and V2X driving
No Dtrain: Model-Agnostic Counterfactual Explanations Using Reinforcement LearningDeep reinforcement learning methods → Interpretable and logic-guided RL
Reinforcement Learning in Dynamic Treatment Regimes Needs Critical ReexaminationOffline RL and imitation learning from data → Safe offline RL +1
Low-Rank Few-Shot Adaptation of Vision-Language ModelsMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
Learning diverse attacks on large language models for robust red-teaming and safety tuningLLM safety and security → Prompt injection and agent security
Cardiovascular Disease Detection from Multi-View Chest X-rays with BI-MambaState-space sequence models (S4, Mamba) → Mamba for medical image analysis
Aligning in a Compact Space: Contrastive Knowledge Distillation between Heterogeneous ArchitecturesKnowledge distillation → General knowledge distillation methods
Offline-Boosted Actor-Critic: Adaptively Blending Optimal Historical Behaviors in Deep Off-Policy RLDeep reinforcement learning methods → Deep RL algorithms and representations
Atlas3D: Physically Constrained Self-Supporting Text-to-3D for Simulation and FabricationDiffusion models → 3D, point cloud and scene generation
Understanding Transformer Reasoning Capabilities via Graph AlgorithmsLLM reasoning → Prompting and chain-of-thought reasoning
Improved Emotional Alignment of AI and Humans: Human Ratings of Emotions Expressed by Stable Diffusion v1, DALL-E 2, and DALL-E 3Diffusion models → Audio, speech and music
SoundCTM: Unifying Score-based and Consistency Models for Full-band Text-to-Sound GenerationDiffusion models → Audio, speech and music
DiG: Scalable and Efficient Diffusion Models with Gated Linear AttentionDiffusion models → Video generation
Hierarchical World Models as Visual Whole-Body Humanoid ControllersWorld models → Humanoid and legged locomotion world models
Don't Forget to Connect! Improving RAG with Graph-based RerankingRetrieval-augmented generation → Knowledge-graph and structured-knowledge RAG
Phased Consistency ModelsDiffusion models → Diffusion foundations and sampling
Thai Winograd Schemas: A Benchmark for Thai Commonsense ReasoningLLM reasoning → Reasoning benchmarks and evaluation
Is a 3D-Tokenized LLM the Key to Reliable Autonomous Driving?LLM reasoning → Chain-of-thought reasoning for driving +2
MMCTAgent: Multi-modal Critical Thinking Agent Framework for Complex Visual ReasoningLLM reasoning → Multimodal reasoning +2
Faithful Logical Reasoning via Symbolic Chain-of-ThoughtLLM reasoning → Prompting and chain-of-thought reasoning
Frustratingly Easy Test-Time Adaptation of Vision-Language ModelsMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
VITON-DiT: Learning In-the-Wild Video Try-On from Human Dance Videos via Diffusion TransformersDiffusion models → 3D, point cloud and scene generation
Highway Reinforcement LearningDeep reinforcement learning methods → RL theory and sample complexity
CT-based brain ventricle segmentation via diffusion Schr"odinger Bridge without target domain ground truthsDiffusion models → Medical imaging and segmentation
Extreme Value Monte Carlo Tree Search for Classical PlanningDeep reinforcement learning methods → RL theory and sample complexity
Recurrent Natural Policy Gradient for POMDPsDeep reinforcement learning methods → RL theory and sample complexity
A Human-Like Reasoning Framework for Multi-Phases Planning Task with Large Language ModelsLLM reasoning → Reasoning for agents, planning and code +1
In-Context Symmetries: Self-Supervised Learning through Contextual World ModelsWorld models → Latent predictive world models
Mutation-Bias Learning in GamesDeep reinforcement learning methods → RL for games and combinatorial optimization
AlignIQL: Policy Alignment in Implicit Q-Learning through Constrained OptimizationOffline RL and imitation learning from data → Offline RL algorithms
Safe Reinforcement Learning in Black-Box Environments via Adaptive ShieldingDeep reinforcement learning methods → Safe and constrained RL
Defending Large Language Models Against Jailbreak Attacks via Layer-specific EditingLLM safety and security → Jailbreaking
VividPose: Advancing Stable Video Diffusion for Realistic Human Image AnimationDiffusion models → Video generation
EG4D: Explicit Generation of 4D Object without Score DistillationDiffusion models → 3D, point cloud and scene generation
An agent design with goal reaching guarantees for enhancement of learningDeep reinforcement learning methods → Deep RL algorithms and representations
MockLLM: A Multi-Agent Behavior Collaboration Framework for Online Job Seeking and RecruitingLLM agents and tool use → Social simulation and multi-agent debate
ATM: Adversarial Tuning Multi-agent System Makes a Robust Retrieval-Augmented GeneratorRetrieval-augmented generation → Agentic RAG, agents and memory +1
A Unified Temporal Knowledge Graph Reasoning Model Towards Interpolation and ExtrapolationLLM reasoning → Knowledge- and tool-augmented reasoning
A Pontryagin Perspective on Reinforcement LearningDeep reinforcement learning methods → RL theory and sample complexity
LLM experiments with simulation: Large Language Model Multi-Agent System for Simulation Model Parametrization in Digital TwinsLLM agents and tool use → General LLM agent frameworks and benchmarks
FlowSDF: Flow Matching for Medical Image Segmentation Using Distance TransformsDiffusion models → Medical imaging and segmentation
HarmoDT: Harmony Multi-Task Decision Transformer for Offline Reinforcement LearningOffline RL and imitation learning from data → Decision transformers and sequence modeling
Towards Dialogues for Joint Human-AI Reasoning and Value AlignmentLLM reasoning → Reasoning benchmarks and evaluation +1
Fast-FedUL: A Training-Free Federated Unlearning with Provable Skew ResilienceLLM safety and security → Federated unlearning and poisoning
Are Images Indistinguishable to Humans Also Indistinguishable to Classifiers?Diffusion models → Medical imaging and segmentation
Coupled Mamba: Enhanced Multi-modal Fusion with Coupled State Space ModelState-space sequence models (S4, Mamba) → Vision Mamba backbones
MAVIN: Multi-Action Video Generation with Diffusion Models via Transition Video InfillingDiffusion models → Video generation
Cross-Context Backdoor Attacks against Graph Prompt LearningLLM safety and security → Backdoor and data poisoning attacks
Reinforced Model Predictive Control via Trust-Region Quasi-Newton Policy OptimizationDeep reinforcement learning methods → RL for autonomous driving and traffic
Aligning to Thousands of Preferences via System Message GeneralizationLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
Yuan 2.0-M32: Mixture of Experts with Attention RouterMixture-of-experts → Expert routing strategies
Knowledge Circuits in Pretrained TransformersMechanistic interpretability → Circuits and mechanistic analysis
AttenCraft: Attention-guided Disentanglement of Multiple Concepts for Text-to-Image CustomizationDiffusion models → Image generation, text-to-image and editing
Unified Preference Optimization: Language Model Alignment Beyond the Preference FrontierLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
Efficient Time Series Processing for Transformers and State-Space Models through Token MergingEfficient LLM inference and serving → Token pruning and merging for ViTs
Self-Guiding Exploration for Combinatorial ProblemsLLM reasoning → Mathematical reasoning
World Models for General Surgical GraspingWorld models → Medical and surgical world models
Tool Learning with Large Language Models: A SurveyLLM agents and tool use → Tool learning and function calling
Proof of Quality: A Costless Paradigm for Trustless Generative AI Model Inference on BlockchainsDiffusion models → Discrete and text diffusion
Online Merging Optimizers for Boosting Rewards and Mitigating Tax in AlignmentLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
White-box Multimodal Jailbreaks Against Large Vision-Language ModelsMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness +1
Arithmetic Reasoning with LLM: Prolog Generation & PermutationLLM reasoning → Mathematical reasoning
SLMRec: Distilling Large Language Models into Small for Sequential RecommendationKnowledge distillation → Distillation for recommender systems
Improving Discrete Diffusion Models via Structured Preferential GenerationDiffusion models → Discrete and text diffusion
Getting More Juice Out of the SFT Data: Reward Learning from Human Demonstration Improves SFT for LLM AlignmentLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF +1
Diffusion Rejection SamplingDiffusion models → Diffusion foundations and sampling
An Information Theoretic Evaluation Metric For Strong UnlearningLLM safety and security → Machine unlearning
MixDQ: Memory-Efficient Few-Step Text-to-Image Diffusion Models with Metric-Decoupled Mixed Precision QuantizationDiffusion models → Diffusion foundations and sampling
Seeing the Image: Prioritizing Visual Correlation by Contrastive AlignmentMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
MMDisCo: Multi-Modal Discriminator-Guided Cooperative Diffusion for Joint Audio and Video GenerationDiffusion models → Audio, speech and music
Benchmarks Underestimate the Readiness of Multi-lingual Dialogue AgentsLLM agents and tool use → General LLM agent frameworks and benchmarks
Enabling Generative Design Tools with LLM Agents for Mechanical Computation Devices: A Case StudyLLM agents and tool use → Multimodal and embodied agents
Mollification Effects of Policy Gradient MethodsDeep reinforcement learning methods → RL theory and sample complexity
LDMol: A Text-to-Molecule Diffusion Model with Structurally Informative Latent Space Surpasses AR ModelsDiffusion models → Flow matching, time series and science
Diffusion Model Patching via Mixture-of-PromptsDiffusion models → Diffusion foundations and sampling
RITUAL: Random Image Transformations as a Universal Anti-hallucination Lever in Large Vision Language ModelsMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
Don't Miss the Forest for the Trees: Attentional Vision Calibration for Large Vision Language ModelsMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
Visual Anchors Are Strong Information Aggregators For Multimodal Large Language ModelMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
Adaptive Horizon Actor-Critic for Policy Learning in Contact-Rich Differentiable SimulationDeep reinforcement learning methods → RL theory and sample complexity
Rethinking Pruning for Backdoor Mitigation: An Optimization PerspectiveLLM safety and security → Backdoor and data poisoning attacks
LoRA-Switch: Boosting the Efficiency of Dynamic LLM Adapters via System-Algorithm Co-designMixture-of-experts → Mixture of LoRA experts
ClavaDDPM: Multi-relational Data Synthesis with Cluster-guided Diffusion ModelsDiffusion models → Flow matching, time series and science +1
AI Alignment with Changing and Influenceable Reward FunctionsLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
Media Insights Engine for Advanced Media Analysis: A Case Study of a Computer Vision Innovation for Pet Health Diagnosis
The Cost of Arbitrariness for Individuals: Examining the Legal and Technical Challenges of Model Multiplicity
Use of Boosting Algorithms in Household-Level Poverty Measurement: A Machine Learning Approach to Predict and Classify Household Wealth Quintiles in the Philippines

All 425