Unsupervised euclidean distance attack on network embedding

Abstract

Considering the wide application of network embedding methods in graph data mining, inspired by adversarial attacks in deep learning, a genetic algorithm-based Euclidean distance attack strategy is proposed to attack the network embedding method, thereby preventing structure information being discovered. EDA focuses on disturbing the Euclidean distance between a pair of nodes in the embedding space as much as possible through minimal modifications of the network structure. Since many downstream network algorithms, such as community detection and node classification, rely on the Euclidean distance between nodes to evaluate their similarity in the embedded space, EDA can be regarded as a general attack on various network algorithms. Different from traditional supervised attack strategies, EDA does not need labeling information, it is an unsupervised network embedding attack method. Experiments on a set of real networks demonstrate that the proposed EDA method can significantly reduce the performance of DeepWalk-based networking algorithms, i.e., community detection and node classification, and its performance is superior to several heuristic attack strategies.