Link Prediction Adversarial Attack

TL;DR

This paper introduces the first formulation and attack method for link prediction adversarial attacks on graph auto-encoders, revealing their vulnerability and proposing potential uses in privacy protection and robustness evaluation.

Contribution

It defines the link prediction adversarial attack problem and proposes a novel iterative gradient attack method based on trained graph auto-encoders.

Findings

GAE is easily fooled with few link perturbations

Most deep and state-of-the-art link prediction algorithms are vulnerable

Adversarial attacks can be used for privacy protection and robustness assessment

Abstract

Deep neural network has shown remarkable performance in solving computer vision and some graph evolved tasks, such as node classification and link prediction. However, the vulnerability of deep model has also been revealed by carefully designed adversarial examples generated by various adversarial attack methods. With the wider application of deep model in complex network analysis, in this paper we define and formulate the link prediction adversarial attack problem and put forward a novel iterative gradient attack (IGA) based on the gradient information in trained graph auto-encoder (GAE). To our best knowledge, it is the first time link prediction adversarial attack problem is defined and attack method is brought up. Not surprisingly, GAE was easily fooled by adversarial network with only a few links perturbed on the clean network. By conducting comprehensive experiments on different real-world data sets, we can conclude that most deep model based and other state-of-art link prediction algorithms cannot escape the adversarial attack just like GAE. We can benefit the attack as an efficient privacy protection tool from link prediction unknown violation, on the other hand, link prediction attack can be a robustness evaluation metric for current link prediction algorithm in attack defensibility.