Adversarial Attacks on Graph Neural Networks via Meta Learning

Daniel Zügner, Stephan Günnemann

Abstract: Deep learning models for graphs have advanced the state of the art on many tasks. Despite their recent success, little is known about their robustness. We investigate training time attacks on graph neural networks for node classification that perturb the discrete graph structure. Our core principle is to use meta-gradients to solve the bilevel problem underlying training-time attacks, essentially treating the graph as a hyperparameter to optimize. Our experiments show that small graph perturbations consistently lead to a strong decrease in performance for graph convolutional networks, and even transfer to unsupervised embeddings. Remarkably, the perturbations created by our algorithm can misguide the graph neural networks such that they perform worse than a simple baseline that ignores all relational information. Our attacks do not assume any knowledge about or access to the target classifiers.

Keywords: graph mining, adversarial attacks, meta learning, graph neural networks, node classification

TL;DR: We use meta-gradients to attack the training procedure of deep neural networks for graphs.

Code: [ danielzuegner/gnn-meta-attack](https://github.com/danielzuegner/gnn-meta-attack)

Data: [Citeseer](https://paperswithcode.com/dataset/citeseer), [Cora](https://paperswithcode.com/dataset/cora)