Yang Yu

A Study on PAVE Specification for Learnware

ADM-v2: Pursuing Full-Horizon Roll-out in Dynamics Models for Offline Policy Learning and Evaluation

Hierarchical Value-Decomposed Offline Reinforcement Learning for Whole-Body Control

Off-Policy Token Clipped Supervised Fine-Tuning Yields a Robust Cold-Start

On the Representation Degradation in Vision-Language-Action Models

EMFuse: Energy-based Model Fusion for Decision Making

MiniOpt: Reasoning to Model and Solve General Optimization Problems with Limited Resources

MOBA: Model-Based Offline Reinforcement Learning with Adaptive Contextual Penalties

Provably Efficient Policy-Reward Co-Pretraining for Adversarial Imitation Learning

Planning with Unified Multimodal Models

Proactive Cost Generation for Offline Safe Reinforcement Learning Without Unsafe Data

Learning Disentangled Multi-Agent World Model for Decentralized Control

ReLAM: Learning Anticipation Model for Rewarding Visual Robotic Manipulation

ImagineBench: Evaluating Reinforcement Learning with Large Language Model Rollouts

Long CoT In-Context Learning Can Empower Pre-trained LLMs

Focus-Then-Reuse: Fast Adaptation in Visual Perturbation Environments

Adaptable Safe Policy Learning from Multi-task Data with Constraint Prioritized Decision Transformer

LLM-Assisted Semantically Diverse Teammate Generation for Efficient Multi-agent Coordination

Efficient Multi-agent Offline Coordination via Diffusion-based Trajectory Stitching

Uncertainty-Sensitive Privileged Learning

On the Optimization Landscape of Low Rank Adaptation Methods for Large Language Models

Behavior-Regularized Diffusion Policy Optimization for Offline Reinforcement Learning

Semantic Temporal Abstraction via Vision-Language Model Guidance for Efficient Reinforcement Learning

Q-Adapter: Customizing Pre-trained LLMs to New Preferences with Forgetting Mitigation

Any-step Dynamics Model Improves Future Predictions for Online and Offline Reinforcement Learning

SOO-Bench: Benchmarks for Evaluating the Stability of Offline Black-Box Optimization

Safe Multi-task Pretraining with Constraint Prioritized Decision Transformer

Improving Reward Model Generalization from Adversarial Process Enhanced Preferences

LLMOPT: Learning to Define and Solve General Optimization Problems from Scratch

Controlling Large Language Model with Latent Action

Multi-Agent Imitation by Learning and Sampling from Factorized Soft Q-Function

Learning to Reuse Policies in State Evolvable Environments

Learning View-invariant World Models for Visual Robotic Manipulation

Hindsight Preference Learning for Offline Preference-based Reinforcement Learning

Haland: Human-AI Coordination via Policy Generation from Language-guided Diffusion

Boosting Offline Multi-Objective Reinforcement Learning via Preference Conditioned Diffusion Models

Learning Generalizable Environment Models via Discovering Superposed Causal Relationships

Diffusion-Guided Safe Policy Optimization From Cost-Label-Free Offline Dataset

Whale-X: Learning Scalable Embodied World Models with Enhanced Generalizability