AI Thrust Seminar | Towards Automated Graph Machine Learning
Supporting the below United Nations Sustainable Development Goals:支持以下聯合國可持續發展目標：支持以下联合国可持续发展目标：
Graph learning has achieved great success in various interdisciplinary research and real applications, such as fraud detection, recommendation systems, drug discovery, protein structure-property prediction, etc. However, the neural network design for a given task needs experienced experts. Automated graph learning has been proposed to tackle this issue. However, most existing works focus more on the search algorithm, and pay less attention to the search space, leading to low model depth, inflexibility, and low scalability. Therefore, we first analyze the root cause that restricts the depth of the graph neural network (KDD22) through a large number of experimental evaluations. Based on this, we propose a deep and flexible graph neural architecture search space (ICML22). To tackle the scalability issue, we will introduce a scalable graph neural architecture search system PaSca (WWW 22 Best Student Paper).
Wentao Zhang is currently a postdoc research fellow working with Prof. Jian Tang at Montreal Institute for Learning Algorithms (Mila). Prior to that, he received his Ph.D. degree in computer science from Peking University in June 2022, supervised by Prof. Bin Cui. Besides, he has accumulated for 4 years industrial experience in Tencent and Apple Research.
Motivated by the industrial demand, Wentao’s research focuses on large-scale graph learning from three perspectives – data, model, and system. He has published 30+ papers, including 10+ first author papers in the top DB (SIGMOD, VLDB, ICDE), DM (KDD, WWW) and ML (ICML, NeurIPS, ICLR) venues. He is the contributor or designer of several system projects including Angel, SGL, MindWare, and OpenBox. His research works have been powering several billion-scale applications in Tencent, and some of them have been recognized by multiple prestigious awards, including the Outstanding Doctoral Dissertation Award, and the Best Student Paper Award at WWW’22. His homepage is https://zwt233.github.io