Seminar: Deep Graph Representation Learning: Scalability and Efficiency
Kaixiong Zhou
Ph.D. Student, Department of Computer Science
Rice University
Wednesday, February 15, 2023
11:00 AM - 12:00 PM
via Zoom
Abstract
The large-scale graph data is ubiquitous in science and industry, including social networks, knowledge graphs, and biochemical molecules. Graph neural networks (GNNs) have emerged as de facto standard models to analyze the node features and graph topology. Despite the prominent effectiveness obtained in recent progress, GNNs are notoriously challenged throughout machine learning cycle, including the low model depth scalability, limited data processing scalability, and inefficient learning paradigms. In this talk, we will discuss some recent results in improving the scalability and efficiency of GNN models. These results are (i) answering how to develop deep graph neural networks from theories, tools, and benchmark platform; (ii) designing scalable graph machine learning algorithms to save memory usage and computation time in the large-scale graph data; and (iii) proposing graph prompt learning and quantum graph neural networks to enhance modeling efficiency in massive data.
Biography
Kaixiong Zhou is a fifth-year Ph.D. student in the Department of Computer Science at Rice University, advised by Professor Xia "Ben" Hu. His reseraches focus on the large-scale graph machine learning, particularly in deep graph analysis, efficient graph representation learning, graph quantum computing, and the science applications in biochemical informatics. With his collaborators, his contributions include several benchmark platforms and automated models for scalable graph analytics, and an initial work on quantum graph neural networks. Prior to joining Rice University, he obtained B.S. and M.S. degrees in electrical engineering from Sun Yat-Sen University and University of Science and Technology of China, respectively.