Detecting Topology Attacks against Graph Neural Networks


Graph neural networks (GNNs) have been widely used in many real applications, and recent studies have revealed their vulnerabilities against topology attacks. To address this issue, existing efforts have mainly been dedicated to improving the robustness of GNNs, while little attention is paid to the detection of such attacks. In this work, we study the victim node detection problem under topology attacks against GNNs. Our approach is built upon the observation rooted in the intrinsic message passing nature of GNNs, and thus applicable to a variety of them. That is, the neighborhood of a victim node tends to have two competing group forces, pushing the node classification results towards the original label and the targeted label, respectively. Based on this observation, we propose to detect victim nodes by deliberately designing an effective measurement of the neighborhood variance for each node. Extensive experimental results on four real-world datasets and five existing topology attacks show the effectiveness and and efficiency of the proposed detection approach..

In the CIKM 2022 Workshop on Trust Worthy Learning on Graphs.