Power Up! Robust Graph Convolutional Network via Graph Powering

We enhance adversarial robustness of GCNs by moving beyond spectral graph theory to robust graph theory. We introduce a novel convolution operator that is provably robust in the spectral domain and incorporate it into the GCN architecture to improve expressivity and interpretability. Extending the original graph to a sequence of graphs yields a robust training paradigm encouraging transferability across spatial and spectral characteristics. Extensive experiments show simultaneous gains in benign and adversarial settings.

Authors

Ming Jin

Heng Chang

Wenwu Zhu

Somayeh Sojoudi

Published

May 18, 2021