Biography


I am a Research Scientist at Institute of Artificial Intelligence (TeleAI), China Telecom, specialized in the cutting-edge field of Embodied AI and Reinforcement Learning (RL). Our group is dedicated to develop embodied technologies encompassing perception, planning, locomotion, manipulation, and promoting the industrial application of embodied AI. Our group thrives under the leadership of Prof. Xuelong Li, who serves as the dean of TeleAI. Previously, I was a Researcher at Shanghai AI Laboratory, affiliated with IPEC group. My research interests include diffusion/transformer policy, LLM-driven planning, world model, preference learning, RL/MPC-based locomotion, dexterous manipulation, representation learning, sim-to-real, multi-agent collaboration, as well as real-world applications for robot arm, dexterous hand, quadruped robot, and humanoid robot.

I holds a Ph.D. degree in Computer Science from Harbin Institute of Technology (HIT), advised by Prof. Peng Liu. I am fortunate to have been collaborated with many fantastic researchers. I was a joint PhD student at University of Toronto and Vector Institute, working with Prof. Animesh Garg. I also used to be an intern at Huawei Noah’s Ark Lab (advised by Prof. Jianye Hao), Tencent Robotics X (advised by Dr. Lei Han), and Alibaba. I received my Bachelor’s degree and Master’s degree in Computer Science from HIT.

团队招收具身智能方向全职研究人员、实习生、联培博士生，具体详见链接.

Interests
  • Deep Reinforcement Learning
  • Embodied AI
  • Foundation Model for Decision Making
Education

  • PhD in Computer Science, 2017-2022

    Harbin Institute of Technology

  • Joint PhD Program, 2021-2022

    University of Toronto

Book

强化学习：前沿算法与应用

白辰甲，赵英男，郝建业，刘鹏，王震. 《强化学习：前沿算法与应用》. 机械工业出版社. 2023

[京东购买链接] | [淘宝购买链接] | [目录页] | [ 媒体报道：深度强化学习实验室 | RL China | 机器之心 ] | [ 勘误 ]
强化学习：前沿算法与应用

Publications

"✉" denotes corresponding author
Quickly discover relevant content by filtering publications.
大模型驱动的具身智能：发展与挑战
中国科学：信息科学
We give a comprehensive survey for embodied AI driven by large-scale models.
Chenjia Bai , Huazhe Xu , Xuelong Li✉
PDF Cite 公众号
大模型驱动的具身智能：发展与挑战
Task-agnostic Pre-training and Task-guided Fine-tuning for Versatile Diffusion Planner.
under review
We develop a versatile diffusion planner that can leverage large-scale inferior data that contains task-agnostic sub-optimal trajectories, with the ability to fast adapt to specific tasks.
Chenyou Fan , Chenjia Bai✉ , Zhao Shan , Haoran He , Yang Zhang , Zhen Wang
PDF Cite
Task-agnostic Pre-training and Task-guided Fine-tuning for Versatile Diffusion Planner.
Forward KL Regularized Preference Optimization for Aligning Diffusion Policies.
under review
We propose Forward KL regularized Preference optimization for aligning Diffusion policies to align the diffusion policy with preferences, learning to align the policy output with human intents in various tasks.
Zhao Shan , Chenyou Fan , Shuang Qiu , Jiyuan Shi , Chenjia Bai✉
PDF Cite
Forward KL Regularized Preference Optimization for Aligning Diffusion Policies.
Decentralized Transformers with Centralized Aggregation are Sample-Efficient Multi-Agent World Models.
under review
we propose a novel world model for MARL that learns decentralized local dynamics for scalability, combined with a centralized representation aggregation from all agents.
Yang Zhang , Chenjia Bai✉ , Bin Zhao , Junchi Yan , Xiu Li , Xuelong Li
PDF Cite
Decentralized Transformers with Centralized Aggregation are Sample-Efficient Multi-Agent World Models.
Towards Efficient LLM Grounding for Embodied Multi-Agent Collaboration.
under review
We propose a novel framework for multi-agent collaboration that introduces Reinforced Advantage feedback (ReAd) for efficient self-refinement of plans.
Yang Zhang , Shixin Yang , Chenjia Bai✉ , Fei Wu , Xiu Li , Xuelong Li , Zhen Wang
PDF Cite Project 公众号
Towards Efficient LLM Grounding for Embodied Multi-Agent Collaboration.
Provably Efficient Information-Directed Sampling Algorithms for Multi-Agent Reinforcement Learning.
under review
This work designs and analyzes a novel set of algorithms for multi-agent reinforcement learning (MARL) based on the principle of information-directed sampling (IDS).
Qiaosheng Zhang , Chenjia Bai , Shuyu Hu , Zhen Wang✉ , Xuelong Li✉
PDF Cite
Provably Efficient Information-Directed Sampling Algorithms for Multi-Agent Reinforcement Learning.
ODRL: A Benchmark for Off-Dynamics Reinforcement Learning.
In Neural Information Processing Systems (NeurIPS), Datasets and Benchmarks Track, 2024
We introduce ODRL, the first benchmark tailored for evaluating off-dynamics RL methods where one needs to transfer policies across different domains with dynamics mismatch.
Jiafei Lyu , Kang Xu , Jiacheng Xu , Mengbei Yan , Jing-Wen Yang , Zongzhang Zhang , Chenjia Bai , Zongqing Lu , Xiu Li
PDF Cite Code
ODRL: A Benchmark for Off-Dynamics Reinforcement Learning.
Regularized Conditional Diffusion Model for Multi-Task Preference Alignment.
In Neural Information Processing Systems (NeurIPS), 2024
We adopt multi-task preferences as a unified condition for both single- and multi-task decision-making, and propose preference representations aligned with preference labels.
Xudong Yu , Chenjia Bai✉ , Haoran He , Changhong Wang , Xuelong Li
PDF Cite
Regularized Conditional Diffusion Model for Multi-Task Preference Alignment.
Learning an Actionable Discrete Diffusion Policy via Large-Scale Actionless Video Pre-Training.
In Neural Information Processing Systems (NeurIPS), 2024
We introduce a novel framework that leverages a unified discrete diffusion to combine generative pre-training on human videos and policy fine-tuning on a small number of action-labeled robot videos.
Haoran He , Chenjia Bai✉ , Ling Pan , Weinan Zhang , Bin Zhao , Xuelong Li
PDF Cite Project
Learning an Actionable Discrete Diffusion Policy via Large-Scale Actionless Video Pre-Training.
Bridging the Sim-to-Real Gap from the Information Bottleneck Perspective.
In Annual Conference on Robot Learning (CORL), 2024     Oral
We propose a novel single-stage privileged knowledge distillation method called the Historical Information Bottleneck (HIB) to narrow the sim-to-real gap for legged locomotion.
Haoran He , Peilin Wu , Chenjia Bai , Hang Lai , Lingxiao Wang , Ling Pan, , Xiaolin Hu , Weinan Zhang✉
PDF Cite
Bridging the Sim-to-Real Gap from the Information Bottleneck Perspective.
SelfBC: Self Behavior Cloning for Offline Reinforcement Learning .
In European Conference on Artificial Intelligence (ECAI), 2024
We propose a novel dynamic policy constraint that restricts the learned policy on the samples generated by the exponentional moving average of previously learned policies for offline RL.
Shirong Liu , Chenjia Bai , Zixian Guo , Hao Zhang , Gaurav Sharma , Yang Liu✉
PDF Cite
SelfBC: Self Behavior Cloning for Offline Reinforcement Learning .
Constrained Ensemble Exploration for Unsupervised Skill Discovery.
In International Conference on Machine Learning (ICML), 2024
We propose a novel unsupervised RL framework via an ensemble of skills, where each skill performs partition exploration based on the state prototypes.
Chenjia Bai , Rushuai Yang , Qiaosheng Zhang , Kang Xu , Yi Chen , Ting Xiao , Xuelong Li
PDF Cite Code
Constrained Ensemble Exploration for Unsupervised Skill Discovery.
Contrastive Representation for Data Filtering in Cross-Domain Offline Reinforcement Learning.
In International Conference on Machine Learning (ICML), 2024
We propose a novel representation-based approach to measure the domain gap, where the representation is learned through a contrastive objective by sampling transitions from different domains.
Xiaoyu Wen , Chenjia Bai✉ , Kang Xu , Xudong Yu , Yang Zhang , Xuelong Li , Zhen Wang
PDF Cite Code
Contrastive Representation for Data Filtering in Cross-Domain Offline Reinforcement Learning.
SAM-E: Leveraging Visual Foundation Model with Sequence Imitation for Embodied Manipulation.
In International Conference on Machine Learning (ICML), 2024
We propose SAM-E, a novel architecture for robot manipulation by leveraging a vision-foundation model for generalizable scene understanding and sequence imitation for long-term action reasoning.
Junjie Zhang , Chenjia Bai✉ , Haoran He , Zhigang Wang , Bin Zhao , Xiu Li , Xuelong Li
PDF Cite Code Project 公众号
SAM-E: Leveraging Visual Foundation Model with Sequence Imitation for Embodied Manipulation.
How Does Goal Relabeling Improve Sample Efficiency?
In International Conference on Machine Learning (ICML), 2024
We construct an example to show the information-theoretical improvement in sample efficiency achieved by goal relabeling and develop an RL algorithm called GOALIVE.
Sirui Zheng , Chenjia Bai , Zhuoran Yang , Zhaoran Wang
PDF Cite
How Does Goal Relabeling Improve Sample Efficiency?
Cross-Domain Policy Adaptation by Capturing Representation Mismatch.
In International Conference on Machine Learning (ICML), 2024
We consider dynamics adaptation settings where there exists dynamics mismatch between the source domain and the target domain, and one can get access to sufficient source domain data, while can only have limited interactions with the target domain.
Jiafei Lyu , Chenjia Bai , Jing-Wen Yang , Zongqing Lu , Xiu Li
PDF Cite Code
Cross-Domain Policy Adaptation by Capturing Representation Mismatch.
Skill Matters: Dynamic Skill Learning for Multi-Agent Cooperative Reinforcement Learning.
Neural Networks (under review), 2024
We propose a novel Dynamic Skill Learning (DSL) framework to enable more effective adaptation and collaboration in complex tasks.
Tong Li , Chenjia Bai✉ , Kang Xu , Chen Chu , Peican Zhu , Zhen Wang✉
PDF Cite
Skill Matters: Dynamic Skill Learning for Multi-Agent Cooperative Reinforcement Learning.
Robust Quadrupedal Locomotion via Risk-Averse Policy Learning.
In IEEE International Conference on Robotics and Automation (ICRA), 2024     Oral
We consider a novel risk-sensitive perspective to enhance the robustness of legged locomotion.
Jiyuan Shi , Chenjia Bai✉ , Haoran He , Lei Han , Dong Wang , Bin Zhao , Mingguo Zhao , Xiu Li , Xuelong Li
PDF Cite Project
Robust Quadrupedal Locomotion via Risk-Averse Policy Learning.
OVD-Explorer: Optimism should not be the Sole Pursuit of Exploration in Noisy Environments.
In AAAI Conference on Artificial Intelligence (AAAI), 2024
We propose Optimistic Value Distribution Explorer (OVD-Explorer) to achieve a noise-aware optimistic exploration for continuous control.
Jinyi Liu , Zhi Wang , Yan Zheng , Jianye Hao , Chenjia Bai , Junjie Ye , Zhen Wang , Et Al.
PDF Cite
OVD-Explorer: Optimism should not be the Sole Pursuit of Exploration in Noisy Environments.
Towards Robust Offline-to-Online Reinforcement Learning via Uncertainty and Smoothness.
Journal of Artificial Intelligence Research (under review), 2023
We propose the Robust Offline-to-Online (RO2O) algorithm, designed to enhance offline policies through uncertainty and smoothness, and to mitigate the performance drop in online adaptation.
Xiaoyu Wen , Xudong Yu , Rui Yang , Chenjia Bai✉ , Zhen Wang
PDF Cite
Towards Robust Offline-to-Online Reinforcement Learning via Uncertainty and Smoothness.
Diffusion Model is an Effective Planner and Data Synthesizer for Multi-Task Reinforcement Learning.
In Neural Information Processing Systems (NeurIPS), 2023
We aim to investigate the effectiveness of a single diffusion model in modeling large-scale multi-task offline data, which can be challenging due to diverse and multimodal data distribution.
Haoran He , Chenjia Bai✉ , Kang Xu , Zhuoran Yang , Weinan Zhang , Dong Wang , Bin Zhao , Xuelong Li
PDF Cite
Diffusion Model is an Effective Planner and Data Synthesizer for Multi-Task Reinforcement Learning.
Cross-Domain Policy Adaptation via Value-Guided Data Filtering.
In Neural Information Processing Systems (NeurIPS), 2023
We reveal the limitations of these methods and explore the problem from the value difference perspective via a novel insight on the value consistency across domains.
Kang Xu , Chenjia Bai✉ , Xiaoteng Ma , Dong Wang , Bin Zhao , Zhen Wang , Xuelong Li , Wei Li
PDF Cite
Cross-Domain Policy Adaptation via Value-Guided Data Filtering.
On the Value of Myopic Behavior in Policy Reuse.
IEEE Transactions on Pattern Analysis and Machine Intelligence. 2023 (under review)
We present a framework called Selective Myopic bEhavior Control~(SMEC), which results from the insight that the short-term behaviors of prior policies are sharable across tasks.
Kang Xu , Chenjia Bai✉ , Shuang Qiu , Haoran He , Bin Zhao , Zhen Wang , Wei Li , Xuelong Li
PDF Cite
On the Value of Myopic Behavior in Policy Reuse.
Ensemble Successor Representations for Task Generalization in Offline-to-Online Reinforcement Learning.
In SCIENCE CHINA Information Sciences, 2023
Our work builds upon the investigation of successor representations for task generalization in online RL and extends the framework to incorporate offline-to-online learning.
Changhong Wang , Xudong Yu , Chenjia Bai , Qiaosheng Zhang , Zhen Wang✉
PDF Cite
Ensemble Successor Representations for Task Generalization in Offline-to-Online Reinforcement Learning.
Pessimistic Value Iteration for Multi-Task Data Sharing in Offline Reinforcement Learning.
In Artificial Intelligence (AIJ), 2023
We propose an uncertainty-based MTDS approach that shares the entire dataset without data selection.
Chenjia Bai , Lingxiao Wang , Jianye Hao , Zhuoran Yang , Bin Zhao , Zhen Wang✉ , Xuelong Li✉
PDF Cite Code 公众号
Pessimistic Value Iteration for Multi-Task Data Sharing in Offline Reinforcement Learning.
Behavior Contrastive Learning for Unsupervised Skill Discovery.
In International Conference on Machine Learning (ICML), 2023
We propose a novel unsupervised skill discovery method through contrastive learning among behaviors, which makes the agent produce similar behaviors for the same skill and diverse behaviors for different skills.
Rushuai Yang , Chenjia Bai✉ , Hongyi Guo , Siyuan Li , Bin Zhao , Zhen Wang , Peng Liu , Xuelong Li
PDF Cite Code
Behavior Contrastive Learning for Unsupervised Skill Discovery.
Diverse Randomized Value Functions: A Provably Pessimistic Approach for Offline Reinforcement Learning.
In Information Sciences, 2023
We introduce a novel strategy employing diverse randomized value functions to estimate the posterior distribution of Q-values.
Xudong Yu , Chenjia Bai✉ , Hongyi Guo , Changhong Wang✉ , Zhen Wang
PDF Cite
Diverse Randomized Value Functions: A Provably Pessimistic Approach for Offline Reinforcement Learning.
False Correlation Reduction for Offline Reinforcement Learning.
IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2023
We propose falSe COrrelation REduction (SCORE) for offline RL, a practically effective and theoretically provable algorithm.
Zhihong Deng , Zuyue Fu , Lingxiao Wang , Zhuoran Yang , Chenjia Bai , Tianyi Zhou , Jing Jiang
PDF Cite
False Correlation Reduction for Offline Reinforcement Learning.
RORL: Robust Offline Reinforcement Learning via Conservative Smoothing.
In Neural Information Processing Systems (NeurIPS), 2022     Spotlight
We propose Robust Offline Reinforcement Learning (RORL) with a novel conservative smoothing technique.
Rui Yang✉ , Chenjia Bai✉ , Xiaoteng Ma , Zhaoran Wang , Chongjie Zhang , Lei Han
PDF Cite
RORL: Robust Offline Reinforcement Learning via Conservative Smoothing.
Self-Supervised Imitation for Offline Reinforcement Learning with Hindsight Relabeling.
IEEE Transactions on Systems, Man, and Cybernetics: Systems. 2022
We present an offline RL algorithm that combines hindsight relabeling and supervised regression to predict actions without oracle information.
Xudong Yu , Chenjia Bai , Changhong Wang , Dengxiu Yu , C. L. Philip Chen , Zhen Wang✉
PDF Cite
Self-Supervised Imitation for Offline Reinforcement Learning with Hindsight Relabeling.
Contrastive UCB: Provably Efficient Contrastive Self-Supervised Learning in Online Reinforcement Learning.
In International Conference on Machine Learning (ICML), 2022     Spotlight
We study how RL can be empowered by contrastive learning in a class of Markov decision processes (MDPs) and Markov games (MGs) with low-rank transitions. For both models, we propose to extract the correct feature representations of the low-rank model by minimizing a contrastive loss.
Shuang Qiu , Lingxiao Wang , Chenjia Bai , Zhuoran Yang , Zhaoran Wang
PDF Cite Code
Monotonic Quantile Network for Worst-Case Offline Reinforcement Learning.
IEEE Transactions on Neural Networks and Learning Systems, 2022
We propose monotonic quantile network (MQN) with conservative quantile regression (CQR) for risk-averse policy learning.
Chenjia Bai , Ting Xiao , Zhoufan Zhu , Lingxiao Wang , Fan Zhou , Peng Liu
PDF Cite
Monotonic Quantile Network for Worst-Case Offline Reinforcement Learning.
Exploration in Deep Reinforcement Learning: From Single-Agent to Multiagent Domain.
IEEE Transactions on Neural Networks and Learning Systems, 2022
We conduct a comprehensive survey on existing exploration methods for both single-agent RL and multiagent RL.
Jianye Hao , Tianpei Yang , Hongyao Tang , Chenjia Bai , Jinyi Liu , Zhaopeng Meng , Peng Liu , Zhen Wang
PDF Cite
Exploration in Deep Reinforcement Learning: From Single-Agent to Multiagent Domain.
Pessimistic Bootstrapping for Uncertainty-Driven Offline Reinforcement Learning.
International Conference on Learning Representations (ICLR), 2022     Spotlight
We propose Pessimistic Bootstrapping for offline RL (PBRL), a purely uncertainty-driven offline algorithm without explicit policy constraints.
Chenjia Bai , Lingxiao Wang , Zhuoran Yang , Zhihong Deng , Animesh Garg , Peng Liu , Zhaoran Wang
PDF Cite Code
Pessimistic Bootstrapping for Uncertainty-Driven Offline Reinforcement Learning.
Dynamic Bottleneck for Robust Self-Supervised Exploration.
In Neural Information Processing Systems (NeurIPS), 2021
We propose a Dynamic Bottleneck (DB) model, which attains a dynamics-relevant representation based on the information-bottleneck principle.
Chenjia Bai , Lingxiao Wang , Lei Han , Animesh Garg , Jianye Hao , Peng Liu , Zhaoran Wang
PDF Cite Code
Dynamic Bottleneck for Robust Self-Supervised Exploration.
Principled Exploration via Optimistic Bootstrapping and Backward Induction.
In International Conference on Machine Learning (ICML), 2021     Spotlight
We propose a principled exploration method for DRL through Optimistic Bootstrapping and Backward Induction (OB2I).
Chenjia Bai , Lingxiao Wang , Lei Han , Jianye Hao , Animesh Garg , Peng Liu , Zhaoran Wang
PDF Cite Code
Principled Exploration via Optimistic Bootstrapping and Backward Induction.
Addressing Hindsight Bias in Multi-Goal Reinforcement Learning.
IEEE Transactions on Cybernetics, 2021
We analyze the hindsight bias due to this use of hindsight goals and propose the bias-corrected HER (BHER), an efficient algorithm that corrects the hindsight bias in training.
Chenjia Bai , Lingxiao Wang , Yixin Wang , Rui Zhao , Chenyao Bai , Peng Liu
PDF Cite Code
Addressing Hindsight Bias in Multi-Goal Reinforcement Learning.
Variational Dynamic for Self-Supervised Exploration in Deep Reinforcement Learning.
IEEE Transactions on Neural Networks and Learning Systems, 2021
We propose a variational dynamic model based on the conditional variational inference to model the multimodality and stochasticity.
Chenjia Bai , Peng Liu , Kaiyu Liu , Lingxiao Wang , Yingnan Zhao , Lei Han
PDF Cite Code Project
Variational Dynamic for Self-Supervised Exploration in Deep Reinforcement Learning.
Generating Attentive Goals for Prioritized Hindsight Reinforcement Learning.
Knowledge-Based Systems (KBS), 2020
We propose a novel prioritized hindsight model for multi-goal RL in which the agent is provided with more valuable goals, as measured by the expected temporal-difference (TD) error.
Peng Liu , Chenjia Bai , Yingnan Zhao , Chenyao Bai , Wei Zhao , Xianglong Tang
PDF Cite
Generating Attentive Goals for Prioritized Hindsight Reinforcement Learning.
Obtaining Accurate Estimated Action Values in Categorical Distributional Reinforcement Learning.
Knowledge-Based Systems (KBS), 2020
This paper describes a method of obtaining more accurate estimated action values for CDRL using adaptive bounds.
Yingnan Zhao , Peng Liu , Chenjia Bai , Wei Zhao , Xianglong Tang
PDF Cite
Obtaining Accurate Estimated Action Values in Categorical Distributional Reinforcement Learning.
Active Sampling for Deep Q-learning Based on TD-error Adaptive Correction.
Journal of Computer Research and Development (in Chinese), 2019
We propose an active sampling method based on TD-error adaptive correction in order to solve sample efficiency problem in deep Q-learning.
Chenjia Bai , Peng Liu , Wei Zhao , Xianglong Tang
PDF Cite
Active Sampling for Deep Q-learning Based on TD-error Adaptive Correction.

Talks

Service

  • Senior Program Committee Member (SPC) / Area Chair (AC) of AAMAS (2024 - 2025)
  • Program Committee Members (PC) / Conference Reviewer of NeurIPS (2021 - 2024)
  • Program Committee Members (PC) / Conference Reviewer of ICLR (2021 - 2024)
  • Program Committee Members (PC) / Conference Reviewer of ICML (2022 - 2024)
  • Program Committee Members (PC) / Conference Reviewer of AAAI (2021 - 2024)
  • Program Committee Members (PC) / Conference Reviewer of ECAI (2023 - 2024)
  • Journal Reviewer: IEEE Trans. Cybernetics, IEEE Trans. TNNLS, IEEE Trans. TETCI, IEEE Trans. Intelligent Vehicles

Experience

 
 
 
 
 
Research Scientist
TeleAI, China Telecom
2024 – Present China
 
 
 
 
 
Researcher
Shanghai AI Laboratory
2022 – 2024 China
 
 
 
 
 
Joint PhD Student
University of Toronto
2021 – 2022 Canada
 
 
 
 
 
PhD Student
Harbin Institute of Technology
2017 – 2022 China