Enhanced Graph Convolutional Networks using Local Structural Information

Over the past few years, Graph Convolutional Networks (GCNs) have achieved state-of-the-art performance in machine learning tasks on graph data and have been widely applied to many real-world applications across different fields, such as traffic prediction, user behavior analysis, and fraud detection. However, networks in the real world are often with heterogeneous degree distributions, such as power-law. This means that the function of nodes with various degrees can vary significantly, with high-degree nodes playing a crucial role in information spread and other spreading phenomena such as message passing. Current GCNs do not consider the roles of nodes with different degrees during the training process, which could influence the performance of GCNs. In this project, we aim at constructing an enhanced GCNs to involve the degree distribution information of the graph to achieve a better performance compared to the current GCNs.

Faculty Supervisor:

Yuanzhu Chen

Student:

Zhihao Dong

Partner:

Verafin Inc.

Discipline:

Computer science

Sector:

Information and cultural industries

University:

Memorial University of Newfoundland