Robust Graph Convolutional Networks Against Adversarial Attacks

By Dingyuan Zhu, Ziwei Zhang, Peng Cui, Wenwu Zhu. Knowledge Discovery and Data Mining, 2019

559 citations recorded by the source corpus.

Graph Convolutional Networks (GCNs) are an emerging type of neural network model on graphs which have achieved state-of-the-art performance in the task of node classification. However, recent studies show that GCNs are vulnerable to adversarial attacks, i.e. small deliberate perturbations in graph structures and node attributes, which poses great challenges for applying GCNs to real world applications. How to enhance the robustness of GCNs remains a critical open problem. To address this problem, we propose Robust GCN (RGCN), a novel model that "fortifies'' GCNs against adversarial attacks. Specifically, instead of representing nodes as vectors, our method adopts Gaussian distributions as the hidden representations of nodes in each convolutional layer. In this way, when the graph is attacked, our model can automatically absorb the effects of adversarial changes in the variances of the Gaussian distributions. Moreover, to remedy the propagation of adversarial attacks in GCNs, we propose a variance-based attention mechanism, i.e. assigning different weights to node neighborhoods according to their variances when performing convolutions. Extensive experimental results demonstrate that our proposed method can effectively improve the robustness of GCNs. On three benchmark graphs, our RGCN consistently shows a substantial gain in node classification accuracy compared with state-of-the-art GCNs against various adversarial attack strategies.

Subfields: Artificial Intelligence, Materials Chemistry

View Paper: https://www.semanticscholar.org/paper/f5252075bb34666863cd01cc82c2d941d4ffe6c6

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Philip S. Yu
2. Marinka Zitnik
3. Jie Tang
4. Yuxiao Dong
5. Lichao Sun
6. Jiliang Tang
7. Kai Chen
8. Yao Ma

In Materials Chemistry

1. Philip S. Yu
2. Bo Li
3. Bo Dai
4. Jiliang Tang
5. Yuxiao Dong
6. Marinka Zitnik
7. Jie Tang
8. Suhang Wang

Citations in the corpus

• Finding Global Homophily in Graph Neural Networks When Meeting Heterophily (2022, background): “, 2019), the adversarial attack and defense (Dai et al., 2018; Zhu et al., 2019), and the model explainability (Vu & Thai, 2020; Shan et al.”
• Graph Structure Learning for Robust Graph Neural Networks (2020, methodology): “• RGCN [38]: RGCN models node representations as gaussian distributions to absorb effects of adversarial attacks.”
• GNNGuard: Defending Graph Neural Networks against Adversarial Attacks (2020, methodology): “For example, RobustGCN [17] is a novel GNN that adopts Gaussian distributions as the hidden representations of nodes in each convolutional layer to absorb the effect of an attack.”
• Deep Learning on Graphs: A Survey (2018, unknown): “attacks as shown in Section 7.2, enhancing the robustness of the existing methods is another important issue. Some pioneering works regarding interpretability and robustness can be found in [161] and [162], [163], respectively. 8.4 Summary The above survey shows that deep learning on graphs is a pro”
• Graph Neural Networks in Recommender Systems: A Survey (2020, background): “In the field of GNN, there are emerging efforts on graph adversarial learning to enhance the robustness [55, 186, 210].”
• Graph-Refined Convolutional Network for Multimedia Recommendation with Implicit Feedback (2020, unknown): “GCNs against structure perturbations [3, 43].”
• Revisiting Skeleton-based Action Recognition (2021, background): “Despite the great success of GCN in skeleton-based action recognition, it is also limited in robustness [40] and scalability.”
• Graph Information Bottleneck (2020, unknown): “Accordingly, several defense models are proposed to increase the robustness by injecting random noise in the representations [33], removing suspicious and uninformative edges [34], low-rank approximation of the adjacency matrix [35], additional hinge loss for certified robustness [36].”
• Graph Random Neural Networks for Semi-Supervised Learning on Graphs (2020, background): “Second, GNNs are often not robust to graph attacks [48, 45], due to the deterministic propagation adopted in most of them.”
• Graph Neural Networks for Graphs With Heterophily: A Survey (2022, unknown): “All datasets can be divided in three categories: standard datasets, large-scale datasets and high-quality datasets.”
• Subgraph Federated Learning with Missing Neighbor Generation (2021, background): “, 2021) or adding/removing edges on the graph (Zhu et al., 2019).”
• Adversarial Attack and Defense on Graph Data: A Survey (2018, methodology): “Citation Network Node/Link Citeseer [87] 3,327 4,732 3,703 6 [145], [28], [23], [106], [92], [8], [147], [12], [129], [120], [125], [17], [127], [40], [140], [89], [22], [144], [105] , [57], [148], [9], [52], [53], [107], [36], [134], [96], [20], [54], [29], [56], [31], [76], [77], [146], [59], [49]”