 | 姓名 | 郅大鹏 |
| 职务/职称 | 讲师 | |
| 研究方向 | 可信人工智能 | |
| 联系方式 | Zhi.dapeng@163.com |
个人简介：
郅大鹏，男，讲师，硕士研究生导师。博士毕业于华东师范大学软件工程专业。主要研究方向为可信人工智能，包括深度神经网络鲁棒性验证与训练以及强化学习系统的可靠性训练与验证。相关成果发表于 AAAI、CAV、NeurIPS、Neural Networks等顶级期刊及会议。
主授课程：《程序设计基础》、《数据分析》等
论文发表
[1] Shi Peng, Si Liu, Dapeng Zhi, Peixin Wang, Min Zhang: MaskCtrl: Training mask networks as self-explainable and performant controllers via deep reinforcement learning. Neural Networks 203: 109107 (2026)
[2] Dapeng Zhi, Peixin Wang, Min Zhang: Formal Verification of Neural Network-Controlled Systems via Proof Certificates. SETSS 2025: 59-85
[3] Shi Peng, Si Liu, Dapeng Zhi, Peixin Wang, Chenyang Xu, Cheng Chen, Min Zhang: ATA: An Abstract-Train-Abstract approach for explanation-friendly deep reinforcement learning. Neural Networks 190: 107749 (2025)
[4] Dapeng Zhi, Peixin Wang, Cheng Chen, Min Zhang. Robustness Verification of Deep Reinforcement Learning Based Control Systems Using Reward Martingales [C] Thirty-Eighth AAAI Conference on Artificial Intelligence (AAAI), 2024.
[5] Dapeng Zhi, Peixin Wang, Si Liu, C.-H. Luke Ong, Min Zhang. Unifying Qualitative and Quantitative Safety Verification of DNN-Controlled Systems [C] Computer Aided Verification - 36th International Conference (CAV), 2024.
[6] Jiaxu Tian, Dapeng Zhi, Si Liu, Peixin Wang, Guy Katz, Min Zhang. Taming Reachability Analysis of DNN-Controlled Systems via Abstraction-Based Training [C] Verification, Model Checking, and Abstract Interpretation - 25th International Conference (VMCAI), 2024.
[7] Jiaxu Tian, Dapeng Zhi, Si Liu, Peixin Wang, Cheng Chen, Min Zhang. Boosting Verification of Deep Reinforcement Learning via Piece-Wise Linear Decision Neural Networks [C] Annual Conference on Neural Information Processing Systems (NeurIPS), 2023.
[8] Peng Jin, Jiaxu Tian, Dapeng Zhi, Xuejun Wen, Min Zhang. Trainify: A CEGARDriven Training and Verification Framework for Safe Deep Reinforcement Learning [C] Computer Aided Verification - 34th International Conference (CAV), 2022.
研究课题
1. 主持多项企业委托项目。
