Date: 8 March 2024, Friday
Time: 4:30pm
Venue: YEUNG Y5-202, City University of Hong Kong
Zoom Meeting ID: 955 7075 2299
Password: 123456
Speaker: Mr Junyuan Fang, City University of Hong Kong
What Contributes More to the Robustness of Heterophilic Graph Neural Networks?
Mr Junyuan Fang, City University of Hong Kong
Abstract:In recent years, graph neural networks (GNNs) have gained significant attention due to their outstanding performance on graph-related tasks by utilizing neighborhood aggregation. However, traditional GNNs are primarily designed based on the homophily assumption, which means that they show poor performance on heterophilic networks where dissimilar nodes prefer to connect. To address this issue, several heterophilic GNNs have been proposed that employ techniques such as extending local neighbors and improving GNN architectures. From another perspective, recent studies have shown that, unlike homophilic GNNs, heterophilic GNNs exhibit higher robustness against graph adversarial attacks. In these attacks, the attackers try to inject small perturbations into the graph. Therefore, in this study, we delve into the core designs of heterophilic GNNs, including high-order neighbor, potential neighbor, ego-neighbor separation, and inter-layer combination designs. We further analyze the influence of these key designs on the robustness of GNNs. To compare the impact of different designs on baseline and real-world GNN models, we conducted comprehensive experiments. The findings of this work can serve as a reference for future studies aiming to design more robust and universal GNNs.
Speaker’s Bio:Junyuan Fang received the B.Eng. degree in software engineering from the Guangdong University of Technology, Guangzhou, China, in 2018, and the M.Eng. degree in software engineering from Sun Yat-sen University, Guangzhou, in 2020. He is currently pursuing the Ph.D. degree with the Department of Electrical Engineering, City University of Hong Kong. His current research interests include robustness analysis and optimization, applications of network science, and graph mining.
WEBINAR WEBSITE:
