Chenjia Bai

Align-Then-stEer: Adapting the Vision-Language Action Models through Unified Latent Guidance

CoNavBench: Collaborative Long-Horizon Vision-Language Navigation Benchmark

Cross-domain Offline Policy Adaptation with Dynamics- and Value-aligned Data Filtering

Online Preference Alignment for Language Models via Count-based Exploration

Information-Theoretic Reward Decomposition for Generalizable RLHF

KungfuBot: Physics-Based Humanoid Whole-Body Control for Learning Highly-Dynamic Skills

Task-Agnostic Pre-training and Task-Guided Fine-tuning for Versatile Diffusion Planner

Discriminator-Guided Embodied Planning for LLM Agent

HumanoidGen: Data Generation for Bimanual Dexterous Manipulation via LLM Reasoning

Adversarial Locomotion and Motion Imitation for Humanoid Policy Learning

Exponential Topology-enabled Scalable Communication in Multi-agent Reinforcement Learning

Towards Reliable LLM-based Robots Planning via Combined Uncertainty Estimation

Revisiting Multi-Agent World Modeling from a Diffusion-Inspired Perspective

Decentralized Transformers with Centralized Aggregation are Sample-Efficient Multi-Agent World Models

VLP: Vision-Language Preference Learning for Embodied Manipulation

Towards Efficient LLM Grounding for Embodied Multi-Agent Collaboration

Unsupervised Reinforcement Learning by Maximizing Skill Density Deviation

Task-agnostic Pre-training and Task-guided Fine-tuning for Versatile Diffusion Planner

A Provable Quantile Regression Adapter via Transfer Learning