研究兴趣

专注于贝叶斯深度学习领域，主要研究内容包括概率方法、近似推理、生成模型、深度神经网络。致力于用深度学习和概率方法来突破统计推断的界限。我和合作者开发的概率深度模型和推理算法已经证明了它们在解决跨领域复杂任务方面的有效性。这些领域包括计算机视觉、自然语言处理、文本分析、图像处理、时间序列建模。具体地：

• 结构化贝叶斯深度学习（基于结构化先验的贝叶斯深度学习算法）

• 概率深度生成模型（面向可控生成的生成模型，解耦与组合泛化，大模型Post-training的概率方法）

• 概率深度表示学习模型 （可解释，可泛化，鲁棒的概率特征学习方法）

欢迎优秀本科生（西交，西工大，西电）实习，尤其是致力于出国读PHD的本科生，将面向机器学习领域前沿提供有效的科研指导与计算资源支持。

联系实习时请提供能够反映个人学习能力的材料（编程基础、数学能力、英语能力）（请与我提前邮件联系 zbduan@xjtu.edu.cn）

教育与工作经历

• 2025/03 - 至今，西安交通大学，数学与统计学院，助理教授

• 2025/03 - 至今，西安交通大学，数学与统计学院，博士后（合作导师：徐宗本院士）

• 2019/09 - 2024/12，西安电子科技大学，信号与信息处理，博士学位（导师：陈渤教授）

• 2015/09 - 2019/07，西安电子科技大学，通信工程，学士学位

发表论文

1. Zhibin Duan, Tiansheng Wen, Muyao Wang, Wenchao Chen, Hao Zhang, Bo Chen, Fei Meng, Hongwei Liu, and Mingyuan Zhou, A Non-negative Deep VAE: the Generalized Gamma Belief Network, to appear in IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2026

2. Zhibin Duan, Xinyue Hu, Bo Chen, Chaojie Wang, Xuefei Cao, Fei Meng, and Mingyuan Zhou, Rethinking Topic Modeling with Information Bottleneck Principle, to appear in IEEE Transactions on Neural Networks and Learning Systems (TNNLS), 2026.

3. Xinyue Hu*, Zhibin Duan*, Xinyang Liu, Yuxin Li, Bo Chen, Chaojie Wang, Yilin He, Hongwei Liu, Xuefei Cao, and Mingyuan Zhou, Disentangled Generative Graph Representation Learning, to appear in IEEE Transactions on Neural Networks and Learning Systems (TNNLS), 2026.（共同一作）

4.  Xinyue Hu⁺, Zhibin Duan⁺, Bo Chen, and Mingyuan Zhou. "Enhancing Uncertainty Estimation and Interpretability with Bayesian Non-negative Decision Layer." In The Thirteenth International Conference on Learning Representations. (ICLR 2025,共同一作 ).

5. Zhibin Duan, Zhiyi Lv, Chaojie Wang, Bo Chen, Bo An, Mingyuan Zhou: Few-shot Generation via Recalling Brain-Inspired Episodic-Semantic Memory. Advances in Neural Information Processing Systems, 36, 7443-7461. (NeurIPS 2023). 

6. Zhibin Duan, Xinyang Liu, Yudi Su, Yishi Xu, Bo Chen, Mingyuan Zhou: Bayesian Progressive Deep Topic Model with Knowledge Informed Textual Data Coarsening Process. In International Conference on Machine Learning, pp. 8731-8746. PMLR, 2023. (ICML 2023).

7.  Zhibin Duan, Yishi Xu, Jianqiao Sun, Bo Chen, Wenchao Chen,  Chaojie  Wang, Mingyuan Zhou: Bayesian Deep Embedding Topic Meta-Learner. In International Conference on Machine Learning, pp. 5659-5670. PMLR, 2022. (ICML 2022).

8.  Zhibin Duan, Yishi Xu, Bo Chen, Dongsheng Wang, Chaojie Wang, Mingyuan Zhou: TopicNet: Semantic Graph-Guided Topic Discovery. Advances in Neural Information Processing Systems 34 (2021): 547-559. (NeurIPS 2021).

9. Zhibin Duan, Dongsheng Wang, Bo Chen, Chaojie Wang, Wenchao Chen, Yewen Li, Jie Ren, Mingyuan Zhou: Sawtooth Factorial Topic Embeddings Guided Gamma Belief Network. In International Conference on Machine Learning, pp. 2903-2913. PMLR, 2021. (ICML 2021).

10. Zhibin Duan, Hao Zhang, Chaojie Wang, Zhengjue Wang, Bo Chen, Mingyuan Zhou: EnsLM: Ensemble Language Model for Data Diversity by Semantic Clustering. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers), pp. 2954-2967. 2021. (ACL 2021,口头报告).

11.  Zhengjue Wang⁺, Zhibin Duan⁺, Hao Zhang, Chaojie Wang, Long Tian, Bo Chen, and Mingyuan Zhou. Friendly topic assistant for transformer based abstractive summarization. In Proceedings of the 2020 conference on empirical methods in natural language processing , pp. 485-497. 2020. (EMNLP 2020,共同一作).

12. Xu, Yishi, Jianqiao Sun, Yudi Su, Xinyang Liu, Zhibin Duan, Bo Chen, and Mingyuan Zhou. Context-guided embedding adaptation for effective topic modeling in low-resource regimes. Advances in Neural Information Processing Systems 36 (2023): 79959-79979. (NeurIPS 2023).

13.  Yishi Xu, Dongsheng Wang, Bo Chen, Ruiying Lu, Zhibin Duan, and Mingyuan Zhou. Hyperminer: Topic taxonomy mining with hyperbolic embedding. Advances in Neural Information Processing Systems 35 (2022): 31557-31570.  (NeurIPS 2022).

14.  Wang, Dongsheng, Yi Xu, Miaoge Li, Zhibin Duan, Chaojie Wang, Bo Chen, and Mingyuan Zhou. Knowledge-aware Bayesian deep topic model. Advances in Neural Information Processing Systems 35 (2022): 14331-14344. (NeurIPS 2022).

15. Yewen Li, Chaojie Wang, Zhibin Duan, Dongsheng Wang, Bo Chen, Bo An, Mingyuan Zhou: Alleviating "Posterior Collapse" in Deep Topic Models via Policy Gradient. Advances in Neural Information Processing Systems 35 (2022): 22562-22575. (NeurIPS 2022）

16.  Wenchao Chen, Long Tian, Bo Chen, Liang Dai, Zhibin Duan, and Mingyuan Zhou. Deep variational graph convolutional recurrent network for multivariate time series anomaly detection. In International conference on machine learning, pp. 3621-3633. PMLR; (ICML 2022).

17. Chaojie Wang, Bo Chen, Zhibin Duan, Wenchao Chen, Hao Zhang, Mingyuan Zhou: Generative Text Convolutional Neural Network for Hierarchical Document Representation Learning. IEEE Transactions on Pattern Analysis and Machine Intelligence 45, no. 4 (2022): 4586-4604. 

合作团队
     

西电 陈渤 教授 

德州大学奥斯汀分校 终身教授 Mingyuan Zhou

合作学生

博士生

胡欣悦 （西电）

硕士生

刘昕洋（西电，毕业时间：2024，PhD at University of Texas at Austin）

本科生

吕帜一 （西电，毕业时间：2024，PhD at Nanyang Technological University） 

李晔文（西电，毕业时间：2022，PhD at Nanyang Technological University） 

任杰（西电，毕业时间：2022，PhD at University of Edinburgh）