Complex Network Effects on the Robustness of Graph Convolutional Networks

TL;DR

This paper proposes training data selection strategies based on network characteristics to enhance the robustness of graph convolutional networks against targeted poisoning attacks, outperforming existing defenses.

Contribution

It introduces two novel training data selection methods leveraging network structure to significantly improve GCN robustness against attacks.

Findings

Changing training data makes networks harder to attack.

Attackers need 2-4 times more perturbations to succeed.

Proposed methods outperform random training and complement existing defenses.

Abstract

Vertex classification -- the problem of identifying the class labels of nodes in a graph -- has applicability in a wide variety of domains. Examples include classifying subject areas of papers in citation networks or roles of machines in a computer network. Vertex classification using graph convolutional networks is susceptible to targeted poisoning attacks, in which both graph structure and node attributes can be changed in an attempt to misclassify a target node. This vulnerability decreases users' confidence in the learning method and can prevent adoption in high-stakes contexts. Defenses have also been proposed, focused on filtering edges before creating the model or aggregating information from neighbors more robustly. This paper considers an alternative: we leverage network characteristics in the training data selection process to improve robustness of vertex classifiers. We propose two alternative methods of selecting training data: (1) to select the highest-degree nodes and (2) to iteratively select the node with the most neighbors minimally connected to the training set. In the datasets on which the original attack was demonstrated, we show that changing the training set can make the network much harder to attack. To maintain a given probability of attack success, the adversary must use far more perturbations; often a factor of 2--4 over the random training baseline. These training set selection methods often work in conjunction with the best recently published defenses to provide even greater robustness. While increasing the amount of randomly selected training data sometimes results in a more robust classifier, the proposed methods increase robustness substantially more. We also run a simulation study in which we demonstrate conditions under which each of the two methods outperforms the other, controlling for the graph topology, homophily of the labels, and node attributes.