  .  . .  .  . .  .  . .  .  . .  .  . .
   .       .       .       .       .     .
     .  .    .  :8:SS X @t    . .    .
.       .  8:S @88888@@S@;8.S    .    .
   .  .    .t@8Xt.:t@@@88@tX%@@ .   .   .
  .    .  .t@8@% 88   t@8@8XX%8       .
    .     :%8@ S:XS;@  8Xt@@@8@8 . .     .
  .   . . tX :   .    :8X888@%8      .  .
    .     S@@:     .  8%88888;8S
  .    .  t%t;XSX8t: .%S:88 888X;  . . .
     .   .@8S8t8t XX :8tX8Xt% X88        .
  .    . XXX8%SX;X8..tt;X%:S 888t. .  .
    .    .%@88S:   XS8X@8XS8 8S.:       .
  .   .   SX888X8: 8 @8888 88S88..   .
    .   . .88888@8@8%@ 88%888 88.  .   .
  .       .888888888tt8 8888 88@  .      .
     . .  .:888888@8888@8:8888%...   .
  .      . :@88888.@8@X%;8 8 X     .   .
    .  .    ;::8888 8%888S888   .
  .       . ..;. 88X88@8888; .    . .  .
    . .  .  .   :8@8%8@88;..   .     .
  .          . :8S@%@@8SX@S      .      .
     . .  .   :88X8@88:8XX8. . .   . .
  .         .:888888 8X8X88@      .    .

Wang-Zhou DAI

Can be pronounced as "won joe dye"

Ph.D.
Associate Professor
School of Intelligence Science and Technology
Nanjing University, Suzhou Campus

About

My research interest is in machine learning, a sub-field of artificial intelligence. Currently, I am interested in combining sub-symbolic machine learning and logic-based symbolic machine learning. (CV)

Research

Leveraging the power of logic reasoning in machine learning is my main focus right now. Popular machine learning techniques, such as Deep Neural Network and Statistical Learning, are good at mapping noisy sub-symbolic data (e.g. images) into symbols (e.g., labels, clusters, etc.); While symbolic machine learning techniques, such as Inductive Logic Programming and Statistical Relational Learning, are good at modelling complex (e.g., recursive) relationships in symbolic data. The two sub-areas in AI have been developed separately throughout the most of the history, resulting a huge gap between machine perception and reasoning. I am trying in various aspects to bridge the two islands, aiming at building ultra-strong machine learning systems that are human understandable, sample-efficient and applicaple to physical-world tasks.

Publications Codes

News

Jul. 2023. A minimal example of Abductive Learning using Jupyter Notebook is available on Github, welcome to try it!
Jun. 2023. A paper about how to utilise Knowledge Graphs for abductive reasoning under the Abductive Learning framework is going to be presented in IJCAI’23.
Jan. 2023. A paper about how to discover novel concepts/predicates from raw data under the Abductive Learning framework has been accepted by AAAI’23.

Teaching

90111101. Discrete Mathematics, 2023 Spr.

Academic Service

Conference (Senior) PC Member: IJCAI, AAAI, ICML, NeurIPS, ICLR, IJCLR, AISTATS, ICDM, KDD, ECAI, PAKDD, ICPR, CIKM;
Journal Reviewer: IEEE TKDE, JMLR, ACM TKDD.

Talks

Abductive Learning: Connecting Machine Learning and Logical Reasoning, @Dagstuhl Seminar “Approaches and Applications of Inductive Programming”, May 11 2021.
Combining Machine Learning and Logical Reasoning, @ Hisilicon, Huawei Shanghai Research Center, Dec 28 2020. (slides)
Abductive Learning: Connecting Machine Learning and Logical Reasoning, @ Huawei Autonomous Driving Network Forum, Nov 27 2020. (Slides)
Introduction to Abductive Learning, @ Nanjing University of Science and Technology, Nov 24 2020.
Abductive Knowledge Induction from Raw Data, @ Samsung AI Research Cambridge, Nov 16 2020.
Bridging Machine Learning and Logical Reasoning by Abductive Learning. Imperial @ NeurIPS 2019, Data Science Institute, Imperial College London. Nov 28th 2019.
Abductive Learning: Towards Bridging Machine Learning and Logical Reasoning, @ London Machine Learning Meetup, Aug 28 2019. (Slides)
- Machine Learning seminar @ City, University of London, May 17 2019.
Research on Integrating First-Order Logical Domain Knowledge with Machine Learning, @ School of Mechanical Engineering, Xiangtan University, Apr 8 2019.