报告题目：Graph Neural Networks
Dong Xu is Curators’ Distinguished Professor in the Department of Electrical Engineering and Computer Science, with appointments in the Christopher S. Bond Life Sciences Center and the Informatics Institute at the University of Missouri-Columbia. He obtained his Ph.D. from the University of Illinois, Urbana-Champaign in 1995 and did two years of postdoctoral work at the US National Cancer Institute. He was a Staff Scientist at Oak Ridge National Laboratory until 2003 before joining the University of Missouri, where he served as Department Chair of Computer Science during 2007-2016 and Director of Information Technology Program during 2017-2020. Over the past 30 years, he has conducted research in many areas of computational biology and bioinformatics, including single-cell data analysis, protein structure prediction and modeling, protein post-translational modifications, protein localization prediction, computational systems biology, biological information systems, and bioinformatics applications in human, microbes, and plants. His research since 2012 has focused on the interface between bioinformatics and deep learning. He has published more than 400 papers with more than 21,000 citations and an H-index of 74, according to Google Scholar. He was elected to the rank of American Association for the Advancement of Science (AAAS) Fellow in 2015 and American Institute for Medical and Biological Engineering (AIMBE) Fellow in 2020.
Graph neural network (GNN) has achieved great success in many areas recently. GNN learns representations of the topological structures and distribution patterns within data in addition to the features of individual data points, which makes it applicable to many research problems with excellent performance. In this talk, I will first introduce some basic shallow graph embedding methods, such as locally linear embedding, random walk embedding, DeepWalk, Node2Vec, and LINE. I will then explain Graph Convolutional Network (GCN) by covering Laplacian embedding, Chebyshev polynomial approximation (ChebNet), and spectral convolution. I will also discuss several GCN variants, including GraphSAGE, gated graph neural networks, message-passing neural networks, graph attention networks, and heterogeneous graph transformer. I will discuss various GNN tasks, such as node classification, edge inference, and graph property prediction, as well as a major GNN implementation issue, i.e., over-smoothing. Finally, I will show some dynamic GNNs, including spectral temporal GNN and neural relational inference. I will use some of our own GNN applications in bioinformatics as examples, including property prediction of small molecules and drugs, electronic medical record study, single-cell data annotation, and molecular dynamics simulation analysis.