About me

I am an Assistant Researcher at the Institute of Automation, Chinese Academy of Sciences. Previously, I received my Ph.D. from the School of Computer Science and Engineering (SCSE) , Beihang University, Beijing, China, under the supervision of Prof. Wenjun Wu.

Research

My research focuses on controlling multi-robot systems through multi-agent reinforcement learning (MARL). To address the practical training challenges of reinforcement learning (RL), I integrate domain expertise with RL to improve sample efficiency and convergence. This strategy not only mitigates the practical training constraints of RL but also expands the applicability of MARL in multi-robot operations. In addition, I am interested in Imitation Learning, Embodied AI based on Large Language Models (LLM), and Robust MARL.

Updates

[2025.3] One paper as the corresponding author accepted by IEEE TMC.

[2024.01] One first-authored paper on multi-robot accepted by ICRA 2024.

[2023.12] One co-authored paper on robust MARL accepted by ICLR 2024.

[2023.12] One first-authored paper on physics-informed MARL accepted by AAAI 2024.

[2023.12] One co-authored paper on reinforcement learning accepted by ICDE 2024.

[2023.07] One first-authored paper on physics-informed MARL accepted by ECAI 2023.

Publications

  1. Leveraging Partial Symmetry for Multi-Agent Reinforcement Learning
    Xin Yu, Rongye Shi, Yongkai Tian,Simin Li,Shuhao Liao, Wenjun Wu
    AAAI 2024 (CCF A) $\vert$ pdf $\vert$ Project page
  2. Symmetry-Informed MARL: A Decentralized and Cooperative UAV Swarm Control Approach for Communication Coverage
    Rongye Shi, Xin Yu✝, Yandong Wang, Yongkai Tian, Zhenyu Liu, Wenjun Wu, Xiaoping Zhang, Manuela M. Veloso (✝ Corresponding Author)
    IEEE Transactions on Mobile Computing, 2025. (CCF-A)
  3. AdaptAUG: Adaptive Data Augmentation Framework for Multi-Agent Reinforcement Learning
    Xin Yu,Yongkai Tian, Li Wang, Pu Feng, Wenjun Wu, Rongye Shi
    ICRA 2024 (CCF B) $\vert$ pdf $\vert$ Project page
  4. ESP: Exploiting Symmetry Prior for Multi-Agent Reinforcement Learning
    Xin Yu, Rongye Shi, Pu Feng, Yongkai Tian, Jie Luo, Wenjun Wu
    ECAI 2023 (CCF B) $\vert$ pdf $\vert$ Project page
  5. Swarm inverse reinforcement learning for biological systems
    Xin Yu, Pu Feng, Yongkai Tian, Wenjun Wu
    BIBM 2021 (CCF B) $\vert$ pdf
  6. 面向群体共识机制的逆强化学习辨识方法
    于鑫,吴文峻,罗杰,李未
    中国科学:技术科学, 2023, 53(2): 258-267.
  7. Byzantine Robust Cooperative Multi-Agent Reinforcement Learning as a Bayesian Game
    Simin Li, Jun Guo, Jingqiao Xiu, Xu Yu, Jiakai Wang, Aishan Liu, Yaodong Yang, Xianglong Liu
    ICLR 2024
  8. Safe and Efficient Multi-Agent Collision Avoidance with Physics-Informed Reinforcement Learning
    Pu Feng, Rongye Shi, Size Wang, Junkang Liang, Xin Yu, Simin Li, Wenjun Wu
    IEEE Robotics and Automation Letters
  9. MACT: Multi-agent Collision Avoidance with Continuous Transition Reinforcement Learning via Mixup
    Pu Feng, Xin Yu, Junkang Liang, Wenjun Wu, Yongkai Tian.
    ICSI $\vert$ pdf
  10. GraphRARE: Reinforcement Learning Enhanced Graph Neural Network with Relative Entropy
    Tianhao Peng , Wenjun Wu, Haitao Yuan, Zhifeng Bao , Zhao Pengrui , Xin Yu , Xuetao Lin , Yu Liang , Yanjun Pu
    ICDE 2024 (CCF A) $\vert$ pdf

Work in Progress

  1. CLGA: A Collaborative LLM Framework for Dynamic Goal Assignment in Multi-Robot Systems
    Xin Yu, Haoyuan Li, Yandong Wang, Simin Li, Rongye Shi, Wenjun Wu
    (Submitted to IROS 2025)

  2. Exploiting Spatio-Temporal Symmetry for Multi-Agent Reinforcement Learning
    Xin Yu, Rongye Shi, Yongkai Tian, Li Wang, Tianhao Peng, Simin Li, Pu Feng, Wenjun Wu

  3. Decentralized Multi-Agent Reinforcement Learning with Equivariant Graph Neural Network
    Xin Yu, Yongkai Tian, Yandong Wang, Shuhao Liao, Pu Feng, Tianhao Peng, Rongye Shi, Wenjun Wu

  4. Integrating Permutation Invariance into Inverse Reinforcement Learning: A New Approach to Collective Behavior Analysis
    Xin Yu, Yongkai Tian, Haoyuan Li, Tianhao Peng, Rongye Shi, Wenjun Wu

Book

I am one of the authors of the book ”AIDevOps: Principles and Practices of Development, Operation and Maintenance of Intelligent Microservices”.