Adversarial Defense Framework for Graph Neural Network
This addresses the problem of adversarial attacks on GNNs for users in graph-based applications, representing an incremental improvement with novel methods for known bottlenecks.
The paper tackles the vulnerability of graph neural networks (GNNs) to adversarial attacks by proposing DefNet, a defense framework that improves robustness through strategies like dual-stage aggregation and adversarial contrastive learning, with experiments on three datasets showing effectiveness across GNN variants.
Graph neural network (GNN), as a powerful representation learning model on graph data, attracts much attention across various disciplines. However, recent studies show that GNN is vulnerable to adversarial attacks. How to make GNN more robust? What are the key vulnerabilities in GNN? How to address the vulnerabilities and defense GNN against the adversarial attacks? In this paper, we propose DefNet, an effective adversarial defense framework for GNNs. In particular, we first investigate the latent vulnerabilities in every layer of GNNs and propose corresponding strategies including dual-stage aggregation and bottleneck perceptron. Then, to cope with the scarcity of training data, we propose an adversarial contrastive learning method to train the GNN in a conditional GAN manner by leveraging the high-level graph representation. Extensive experiments on three public datasets demonstrate the effectiveness of DefNet in improving the robustness of popular GNN variants, such as Graph Convolutional Network and GraphSAGE, under various types of adversarial attacks.