Adversarial Attack on Graph Structured Data
This addresses the robustness issue in graph-based deep learning, which is an incremental contribution as it extends adversarial attack research from images and text to graphs.
The paper tackles the problem of adversarial attacks on graph-structured data, proposing reinforcement learning, genetic algorithm, and gradient-based methods to modify graph structures and fool Graph Neural Network models, showing that these models are vulnerable in both graph-level and node-level classification tasks.
Deep learning on graph structures has shown exciting results in various applications. However, few attentions have been paid to the robustness of such models, in contrast to numerous research work for image or text adversarial attack and defense. In this paper, we focus on the adversarial attacks that fool the model by modifying the combinatorial structure of data. We first propose a reinforcement learning based attack method that learns the generalizable attack policy, while only requiring prediction labels from the target classifier. Also, variants of genetic algorithms and gradient methods are presented in the scenario where prediction confidence or gradients are available. We use both synthetic and real-world data to show that, a family of Graph Neural Network models are vulnerable to these attacks, in both graph-level and node-level classification tasks. We also show such attacks can be used to diagnose the learned classifiers.