SAM: Query-efficient Adversarial Attacks against Graph Neural Networks

Zhang, C; Zhang, S; Yu, JJQ; Yu, S

SAM: Query-efficient Adversarial Attacks against Graph Neural Networks

Zhang, C

Zhang, S Yu, JJQ Yu, S

Permalink

Publisher:: ASSOC COMPUTING MACHINERY
Publication Type:: Journal Article
Citation:: ACM Transactions on Privacy and Security, 2023, 26, (4)
Issue Date:: 2023-11-13

In Progress

	Filename	Description	Size
	SAM- Query-efficient Adversarial Attacks against Graph Neural Networks.pdf	Published version	1.46 MB	Adobe PDF	View/Open

Copyright Clearance Process

Recently Added
In Progress
Open Access

This item is being processed and is not currently available.

Full metadata record

Field	Value	Language
dc.contributor.author	Zhang, C https://orcid.org/0000-0002-2352-0485
dc.contributor.author	Zhang, S
dc.contributor.author	Yu, JJQ
dc.contributor.author	Yu, S https://orcid.org/0000-0003-4485-6743
dc.date.accessioned	2024-03-22T04:07:14Z
dc.date.available	2024-03-22T04:07:14Z
dc.date.issued	2023-11-13
dc.identifier.citation	ACM Transactions on Privacy and Security, 2023, 26, (4)
dc.identifier.issn	2471-2566
dc.identifier.issn	2471-2574
dc.identifier.uri	http://hdl.handle.net/10453/177016
dc.description.abstract	Recent studies indicate that Graph Neural Networks (GNNs) are vulnerable to adversarial attacks. Particularly, adversarially perturbing the graph structure, e.g., flipping edges, can lead to salient degeneration of GNNs' accuracy. In general, efficiency and stealthiness are two significant metrics to evaluate an attack method in practical use. However, most prevailing graph structure-based attack methods are query intensive, which impacts their practical use. Furthermore, while the stealthiness of perturbations has been discussed in previous studies, the majority of them focus on the attack scenario targeting a single node. To fill the research gap, we present a global attack method against GNNs, Saturation adversarial Attack with Meta-gradient, in this article. We first propose an enhanced meta-learning-based optimization method to obtain useful gradient information concerning graph structural perturbations. Then, leveraging the notion of saturation attack, we devise an effective algorithm to determine the perturbations based on the derived meta-gradients. Meanwhile, to ensure stealthiness, we introduce a similarity constraint to suppress the number of perturbed edges. Thorough experiments demonstrate that our method can effectively depreciate the accuracy of GNNs with a small number of queries. While achieving a higher misclassification rate, we also show that the perturbations developed by our method are not noticeable.
dc.language	English
dc.publisher	ASSOC COMPUTING MACHINERY
dc.relation	http://purl.org/au-research/grants/arc/DP200101374
dc.relation.ispartof	ACM Transactions on Privacy and Security
dc.relation.isbasedon	10.1145/3611307
dc.rights	info:eu-repo/semantics/restrictedAccess
dc.title	SAM: Query-efficient Adversarial Attacks against Graph Neural Networks
dc.type	Journal Article
utslib.citation.volume	26
pubs.organisational-group	University of Technology Sydney
pubs.organisational-group	University of Technology Sydney/Faculty of Engineering and Information Technology
pubs.organisational-group	University of Technology Sydney/Faculty of Engineering and Information Technology/School of Computer Science
pubs.organisational-group	University of Technology Sydney/Strength - CCSP - Centre for Cyber Security and Privacy
utslib.copyright.status	in_progress	*
dc.date.updated	2024-03-22T04:07:13Z
pubs.issue	4
pubs.publication-status	Published
pubs.volume	26
utslib.citation.issue	4

Abstract:

Recent studies indicate that Graph Neural Networks (GNNs) are vulnerable to adversarial attacks. Particularly, adversarially perturbing the graph structure, e.g., flipping edges, can lead to salient degeneration of GNNs' accuracy. In general, efficiency and stealthiness are two significant metrics to evaluate an attack method in practical use. However, most prevailing graph structure-based attack methods are query intensive, which impacts their practical use. Furthermore, while the stealthiness of perturbations has been discussed in previous studies, the majority of them focus on the attack scenario targeting a single node. To fill the research gap, we present a global attack method against GNNs, Saturation adversarial Attack with Meta-gradient, in this article. We first propose an enhanced meta-learning-based optimization method to obtain useful gradient information concerning graph structural perturbations. Then, leveraging the notion of saturation attack, we devise an effective algorithm to determine the perturbations based on the derived meta-gradients. Meanwhile, to ensure stealthiness, we introduce a similarity constraint to suppress the number of perturbed edges. Thorough experiments demonstrate that our method can effectively depreciate the accuracy of GNNs with a small number of queries. While achieving a higher misclassification rate, we also show that the perturbations developed by our method are not noticeable.

Please use this identifier to cite or link to this item:

http://hdl.handle.net/10453/177016