Data Poisoning Attack against Unsupervised Node Embedding Methods

TL;DR

This paper introduces a data poisoning attack targeting unsupervised node embedding methods like DeepWalk and LINE, demonstrating how minimal graph modifications can significantly impair link prediction accuracy.

Contribution

It provides the first comprehensive analysis of attack strategies against unsupervised node embedding methods and offers efficient attack solutions with demonstrated transferability.

Findings

Attacks can significantly degrade link prediction performance.

Minimal graph modifications are sufficient for effective attacks.

Proposed attacks are transferable across different embedding methods.

Abstract

Unsupervised node embedding methods (e.g., DeepWalk, LINE, and node2vec) have attracted growing interests given their simplicity and effectiveness. However, although these methods have been proved effective in a variety of applications, none of the existing work has analyzed the robustness of them. This could be very risky if these methods are attacked by an adversarial party. In this paper, we take the task of link prediction as an example, which is one of the most fundamental problems for graph analysis, and introduce a data positioning attack to node embedding methods. We give a complete characterization of attacker's utilities and present efficient solutions to adversarial attacks for two popular node embedding methods: DeepWalk and LINE. We evaluate our proposed attack model on multiple real-world graphs. Experimental results show that our proposed model can significantly affect the results of link prediction by slightly changing the graph structures (e.g., adding or removing a few edges). We also show that our proposed model is very general and can be transferable across different embedding methods. Finally, we conduct a case study on a coauthor network to better understand our attack method.