avatar

Da-Wei Zhou

avatar
Ph.D. Candidate
LAMDA Group, Nanjing University
zhoudw(at)lamda.nju.edu.cn

avatar avatar

About Me

I am a final-year Ph.D. candidate in the LAMDA Group, Nanjing University (NJU), advised by Prof. De-Chuan Zhan and Han-Jia Ye. Prior to attending NJU, I obtained my B.Sc. degree from Hunan University in June 2018. From October 2022 to October 2023, I had the invaluable experience of being a visiting scholar at MMLab@NTU, working closely with Prof. Ziwei Liu.

🤚 Open to collaborations. Feel free to contact me if you have interests in my work.

Research Interests

My research interests include machine learning and data mining. More specifically, I focus on:

  1. Class-Incremental Learning: learn from evolutive streams to adapt to new class data without forgetting existing knowledge.
    Open-Set Recognition: not only accurately classify the seen classes but also effectively deal with unseen ones during testing.
    Open-World Learning: enable the learning system to automatically and incrementally detect and learn new classes.

News

  • [2023-09] One paper about few-shot class-incremental learning is accepted to NeurIPS 2023. See you in New Orleans!
  • [2023-09] We have released PILOT toolbox for class-incremental learning with pre-trained models (technical report).
  • [2023-09] One paper about contextualizing meta-learning is accepted to TPAMI.
  • [2023-08] We are hosting the tuotorial (slides) on continual learning at IJCAI 2023.
  • [2023-05] One paper about class-incremental learning with VLM is uploaded to arXiv.
  • [2023-03] One paper about class-incremental learning with PTM is uploaded to arXiv.
  • [2023-02] One survey about class-incremental learning is uploaded to arXiv.
  • [2023-01] Two papers about class-incremental learning are accepted to ICLR 2023 (one spotlight).
  • [2022-10] Our toolbox for class-incremental learning (PyCIL) is accepted to SCIS.
  • [2022-08] One paper about few-shot class-incremental learning is accepted to TPAMI.
  • [2022-07] One paper about class-incremental learning is accepted to ECCV 2022.
  • [2022-03] One paper about few-shot class-incremental learning is accepted to CVPR 2022.
  • [2021-12] One paper about active incremental learning is accepted to TKDE.
  • [2021-08] One paper about open-world learning is accepted to TNNLS.
  • [2021-07] One paper about class-incremental learning is accepted to ACM MM 2021.
  • [2021-03] One oral paper about open-set recognition is accepted to CVPR 2021.
  • Selected Publications

    For more details, please view the full publication page or Google Scholar profile.

    Preprints

    1. Preprint
      Learning without Forgetting for Vision-Language Models
      Da-Wei Zhou, Yuanhan Zhang, Jingyi Ning, Han-Jia Ye, De-Chuan Zhan, Ziwei Liu
      arXiv:2305.19270. 2023.
      [arXiv]
    2. Preprint
      Revisiting Class-Incremental Learning with Pre-Trained Models: Generalizability and Adaptivity are All You Need
      Da-Wei Zhou, Han-Jia Ye, De-Chuan Zhan, Ziwei Liu
      arXiv:2303.07338. 2023.
      [arXiv] [Code]
    3. Preprint
      Deep Class-Incremental Learning: A Survey
      Da-Wei Zhou, Qi-Wei Wang, Zhi-Hong Qi, Han-Jia Ye, De-Chuan Zhan, Ziwei Liu
      arXiv:2302.03648. 2023.
      [arXiv] [Code] [中文解读]
    4. Preprint
      PILOT: A Pre-Trained Model-Based Continual Learning Toolbox
      Hai-Long Sun, Da-Wei Zhou, Han-Jia Ye, De-Chuan Zhan
      arXiv:2309.07117. 2023.
      [arXiv] [Code]

    Conference Paper

    1. NeurIPS
      Few-Shot Class-Incremental Learning via Training-Free Prototype Calibration
      Qi-Wei Wang, Da-Wei Zhou, Yi-Kai Zhang, De-Chuan Zhan, Han-Jia Ye
      Thirty-seventh Conference on Neural Information Processing Systems. NeurIPS 2023.
      [Paper]
    2. ICLR
      A Model or 603 Exemplars: Towards Memory-Efficient Class-Incremental Learning
      Da-Wei Zhou, Qi-Wei Wang, Han-Jia Ye, De-Chuan Zhan
      Eleventh International Conference on Learning Representations. ICLR 2023. Spotlight Presentation
      [Paper] [arXiv] [Project Page] [Code]
    3. ICLR
      BEEF: Bi-Compatible Class-Incremental Learning via Energy-Based Expansion and Fusion
      Fu-Yun Wang, Da-Wei Zhou, Liu Liu, Yatao Bian, Han-Jia Ye, De-Chuan Zhan, Peilin Zhao
      Eleventh International Conference on Learning Representations. ICLR 2023.
      [Paper] [Code]
    4. CVPR
      Forward Compatible Few-Shot Class-Incremental Learning
      Da-Wei Zhou, Fu-Yun Wang, Han-Jia Ye, Liang Ma, Shiliang Pu, De-Chuan Zhan
      IEEE Conference on Computer Vision and Pattern Recognition. CVPR 2022.
      [Paper] [Supplementary] [arXiv] [Project Page] [Code]
    5. ECCV
      FOSTER: Feature Boosting and Compression for Class-Incremental Learning
      Fu-Yun Wang, Da-Wei Zhou, Han-Jia Ye, De-Chuan Zhan
      European Conference on Computer Vision. ECCV 2022.
      [Paper] [Supplementary] [arXiv] [Code]
    6. ACM MM
      Co-Transport for Class-Incremental Learning
      Da-Wei Zhou, Han-Jia Ye, De-Chuan Zhan
      The 29th ACM International Conference on Multimedia. ACM MM 2021.
      [Paper] [arXiv] [Project Page] [Code]
    7. CVPR
      Learning Placeholders for Open-Set Recognition
      Da-Wei Zhou, Han-Jia Ye, De-Chuan Zhan
      IEEE Conference on Computer Vision and Pattern Recognition. CVPR 2021. Oral Presentation
      [Paper] [Supplementary] [arXiv] [Project Page] [Code]
    8. KDD
      Adaptive Deep Models for Incremental Learning: Considering Capacity Scalability and Sustainability
      Yang Yang, Da-Wei Zhou, De-Chuan Zhan, Hui Xiong, Yuan Jiang.
      The 25th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. KDD 2019. Oral Presentation
      [Paper] [Video] [Code]

    Journal Article

    1. TPAMI
      Few-Shot Class-Incremental Learning by Sampling Multi-Phase Tasks
      Da-Wei Zhou, Han-Jia Ye, Liang Ma, Di Xie, Shiliang Pu, De-Chuan Zhan
      IEEE Transactions on Pattern Analysis and Machine Intelligence. TPAMI.
      [Paper] [arXiv] [Project Page] [Code]
    2. TPAMI
      Contextualizing Meta-Learning via Learning to Decompose
      Han-Jia Ye, Da-Wei Zhou, Lanqing Hong, Zhenguo Li, Xiu-Shen Wei, De-Chuan Zhan
      IEEE Transactions on Pattern Analysis and Machine Intelligence. TPAMI.
      [Paper] [arXiv] [Code]
    3. SCIS
      PyCIL: A Python Toolbox for Class-Incremental Learning
      Da-Wei Zhou*, Fu-Yun Wang*, Han-Jia Ye, De-Chuan Zhan
      (* indicates equal contributions)
      SCIENCE CHINA Information Sciences. SCIS.
      [Paper] [arXiv] [Code] [Media] [中文解读]
    4. TKDE
      Cost-Effective Incremental Deep Model: Matching Model Capacity with the Least Sampling
      Yang Yang*, Da-Wei Zhou*, De-Chuan Zhan, Hui Xiong, Yuan Jiang, Jian Yang (* indicates equal contributions)
      IEEE Transactions on Knowledge and Data Engineering. TKDE.
      [Paper] [Code]
    5. TNNLS
      Learning to Classify with Incremental New Class
      Da-Wei Zhou, Yang Yang, De-Chuan Zhan
      IEEE Transactions on Neural Networks and Learning Systems. TNNLS.
      [Paper]

    Correspondence

    Please read this page before asking for source code/pre-trained models/data, etc.

    Email: zhoudw [at] lamda.nju.edu.cn
    Office: Room A201, Yifu Building, Xianlin Campus of Nanjing University
    Address: National Key Laboratory for Novel Software Technology, Nanjing University, Xianlin Campus Mailbox 603, 163 Xianlin Avenue, Qixia District, Nanjing 210023, China


    Many thanks to Yaoyao for sharing this great theme.