👋About Me

Hi, I am Xing Lei (雷星), a Ph.D. student in Artificial Intelligence at the School of Artificial Intelligence, Xi'an Jiaotong University (2020–present). I received my M.S. in Software Engineering from Xi'an Jiaotong University (2019) and my B.S. in Computer Science from Chang'an University (2016).

🔭 My research interests lie in offline goal-conditioned reinforcement learning, sample-efficient reinforcement learning, and sequence modeling with Transformers and Mamba for decision making — with a focus on trajectory stitching, goal representations, and history-aware architectures.

✉️ Welcome to reach out for any discussion and collaboration!

🔥News

  • 2026.07 NEWOne paper (DAGR) is released on arXiv — state-conditioned goal representations via difference-aware goal cross-attention.
  • 2026.07 NEWOne paper (NFTR) is released on arXiv — normalizing-flow subgoal policies with triangle-slack reweighting for offline goal-conditioned RL.
  • 2026.05 One paper (QHyer) is accepted by ICML 2026.
  • 2025.08 One paper (GCHR) is released on arXiv — goal-conditioned hindsight regularization for sample-efficient RL.
  • 2025.02 One paper (MGDA) is accepted by AAAI 2025.
  • 2020.09 Started my Ph.D. journey at the School of Artificial Intelligence, Xi'an Jiaotong University.
  • 2019.05 One paper is accepted by ICIP 2019.

📝Publications

* Equal contribution. Full list available on Google Scholar.

Published

  • QHyer: Q-conditioned Hybrid Attention-mamba Transformer for Offline Goal-conditioned RL
    QHyer
    ICML 2026
    Xing Lei, Jincheng Wang, Xuetao Zhang, Donglin Wang
    arXiv
  • MGDA: Model-based Goal Data Augmentation for Offline Goal-conditioned Weighted Supervised Learning
    MGDA
    AAAI 2025
    Xing Lei, Xuetao Zhang, Donglin Wang
    Paper arXiv
  • Exploring Hardware Friendly Bottleneck Architecture in CNN for Embedded Computing Systems
    Hardware Friendly CNN
    ICIP 2019
    Xing Lei, Longjun Liu, Zhiheng Zhou, Hongbin Sun, Nanning Zheng
    Paper arXiv

Preprints & Under Submission

  • DAGR: State-Conditioned Goal Representations via Difference-Aware Goal Cross-Attention
    DAGR
    arXiv 2026
    Xing Lei, Wenyan Yang, Xuetao Zhang, Donglin Wang
    arXiv Code
  • NFTR: From Provable Mode-Averaging to Geodesic Subgoal Selection in Offline Goal-Conditioned RL
    NFTR
    arXiv 2026
    Erdemt Bao*, Xing Lei*, Jun Chen
    arXiv Code
  • GCHR: Goal-Conditioned Hindsight Regularization for Sample-Efficient Reinforcement Learning
    GCHR
    arXiv 2025
    Xing Lei, Wenyan Yang, Kaiqiang Ke, Shentao Yang, Xuetao Zhang, Joni Pajarinen, Donglin Wang
    arXiv Code
  • Closing the Gap between TD Learning and Supervised Learning with Q-Conditioned Maximization
    Q-Conditioned Maximization
    arXiv 2025
    Xing Lei, Zifeng Zhuang, Shentao Yang, Sheng Xu, Yunhao Luo, Fei Shen, Wenyan Yang, Xuetao Zhang, Donglin Wang
    arXiv
  • Q-WSL: Optimizing Goal-Conditioned RL with Weighted Supervised Learning via Dynamic Programming
    Q-WSL
    arXiv 2024
    Xing Lei, Zifeng Zhuang, Xuetao Zhang, Donglin Wang
    arXiv

🎓Academic Service

  • Conference Reviewer
    • NeurIPS 2026
    • ICLR 2026
    • ICLR 2025

🏅Honors & Awards

  • 2013Outstanding Student Cadre (优秀学生干部), Chang'an University