论文标题
图形分类对拓扑攻击的认证鲁棒性，并随机平滑
Certified Robustness of Graph Classification against Topology Attack with Randomized Smoothing
论文作者
论文摘要
图分类在不同领域具有实际应用。最近的研究表明，由于图数据的非i.i.d性质，基于图的机器学习模型特别容易受到对抗扰动的影响。通过在图中添加或删除少量边缘，对手可以大大更改图形分类模型预测的图形标签。在这项工作中，我们建议建立一个具有认证鲁棒性保证的平滑图分类模型。我们已经证明，所得的图形分类模型将在$ L_0 $有限的对抗扰动下输出相同的图表。我们还评估了基于图形卷积网络（GCN）的多级图分类模型下的方法的有效性。
Graph classification has practical applications in diverse fields. Recent studies show that graph-based machine learning models are especially vulnerable to adversarial perturbations due to the non i.i.d nature of graph data. By adding or deleting a small number of edges in the graph, adversaries could greatly change the graph label predicted by a graph classification model. In this work, we propose to build a smoothed graph classification model with certified robustness guarantee. We have proven that the resulting graph classification model would output the same prediction for a graph under $l_0$ bounded adversarial perturbation. We also evaluate the effectiveness of our approach under graph convolutional network (GCN) based multi-class graph classification model.
