Analyzing Adversarial Vulnerabilities of Graph Lottery Tickets

• Published in: Proceedings of the ... IEEE International Conference on Acoustics, Speech and Signal Processing (1998) pp. 7830 - 7834
• Main Authors: Chowdhury, Subhajit Dutta, Ni, Zhiyu, Peng, Qingyuan, Kundu, Souvik, Nuzzo, Pierluigi
• Format: Conference Proceeding
• Language: English
• Published: IEEE 14.04.2024
• Subjects: Acoustics; Adversarial attacks; graph lottery tickets; Graph neural networks; Perturbation methods; Resilience; Signal processing; Speech processing; structure perturbation; Task analysis
• ISSN: 2379-190X
• DOI: 10.1109/ICASSP48485.2024.10448491

Graph neural networks (GNNs) have displayed significant potential in various graph-based learning tasks. However, the computational demands of deploying GNNs on large-scale graphs can grow exponentially. A recent method, termed unified graph sparsification (UGS), shows that there exists a pair consisting of a subgraph and a sparse subnetwork, called graph lottery ticket (GLT), that can effectively speed up GNN inference. However, despite their advantages, the performance of GLTs against adversarial structure perturbations remains largely unexplored. In this paper, we investigate the resilience of GLTs against different structure perturbation attacks under the poisoning attack setting. The evaluation results show that the GLTs identified by UGS are vulnerable and exhibit a large drop in classification accuracy for the adversarially perturbed graphs. We then propose a new technique for defending UGS that leverages self-training to find GLTs that are more resilient and can achieve better performance than plain UGS.