Robust Mid-Pass Filtering Graph Convolutional Networks
This addresses the problem of adversarial robustness in graph neural networks for researchers and practitioners, offering a simple, effective defense method that is not data- or attack-specific, though it is incremental as it builds on existing GCN paradigms.
The paper tackles the vulnerability of Graph Convolutional Networks (GCNs) to adversarial attacks by proposing Mid-GCN, a mid-pass filtering method that improves robustness without extra training overhead, achieving higher node classification accuracy on six benchmark datasets under various attacks.
Graph convolutional networks (GCNs) are currently the most promising paradigm for dealing with graph-structure data, while recent studies have also shown that GCNs are vulnerable to adversarial attacks. Thus developing GCN models that are robust to such attacks become a hot research topic. However, the structural purification learning-based or robustness constraints-based defense GCN methods are usually designed for specific data or attacks, and introduce additional objective that is not for classification. Extra training overhead is also required in their design. To address these challenges, we conduct in-depth explorations on mid-frequency signals on graphs and propose a simple yet effective Mid-pass filter GCN (Mid-GCN). Theoretical analyses guarantee the robustness of signals through the mid-pass filter, and we also shed light on the properties of different frequency signals under adversarial attacks. Extensive experiments on six benchmark graph data further verify the effectiveness of our designed Mid-GCN in node classification accuracy compared to state-of-the-art GCNs under various adversarial attack strategies.