Document Type

Conference Paper

Publication Date

2025

DOI

10.1145/3746252.3760882

Publication Title

CIKM '25: Proceedings of the 34th ACM International Conference on Information and Knowledge Management

Pages

5530-5534

Conference Name

34th ACM International Conference on Information and Knowledge Management, November 10-14, 2025, Seoul, Republic of Korea

Abstract

Graph neural networks and graph transformers explicitly or implicitly rely on fundamental properties of the underlying graph, such as spectral properties and shortest-path distances. However, it is still not clear how these graph properties are vulnerable to adversarial attacks and what impacts this has on the downstream graph learning. Moreover, while graph sparsification has been used to improve computational cost of learning over graphs, its susceptibility to adversarial attacks has not been studied. In this paper, we study adversarial attacks on graph properties and graph sparsification and their impacts on downstream graph learning, paving the way for how to protect against these potential attacks. Our proposed methods are effective in attacking spectral properties, shortest distances, and graph sparsification as demonstrated in our experimental evaluation.

Rights

© 2025 Copyright held by the owner/authors.

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) License.

Original Publication Citation

Zhu, C., Gaines, B., Deng, J., & Bi, J. (2025). Adversarially attacking graph properties and sparsification in graph learning. In M. Cha, C. Park, N. Park, C. Yang, S. B. Roy, J. Li, J. Kamps, K. Shin, B. Hooi, & L. He (Eds.), CIKM '25: Proceedings of the 34th ACM International Conference on Information and Knowledge Management (pp. 5530-5534). Association for Computing Machinery. https://doi.org/10.1145/3746252.3760882

ORCID

0000-0002-5227-3575 (Zhu)

Download

Share

COinS