When Fairness Meets Privacy: Exploring Privacy Threats in Fair Binary Classifiers via Membership Inference Attacks

Tian, H; Zhang, G; Liu, B; Zhu, T; Ding, M; Zhou, W

When Fairness Meets Privacy: Exploring Privacy Threats in Fair Binary Classifiers via Membership Inference Attacks

Tian, H Zhang, G Liu, B

Zhu, T Ding, M Zhou, W

Permalink

Publisher:: International Joint Conferences on Artificial Intelligence
Publication Type:: Conference Proceeding
Citation:: Proceedings of the Thirty-ThirdInternational Joint Conference on Artificial Intelligence, 2024, pp. 512-520
Issue Date:: 2024-08-01

Open Access

Copyright Clearance Process

Recently Added
In Progress
Open Access

This item is open access.

Adobe PDF

Download Accepted versionAdobe PDF (8.39 MB)

View on publisher's site

View statistics

Full metadata record

Field	Value	Language
dc.contributor.author	Tian, H
dc.contributor.author	Zhang, G
dc.contributor.author	Liu, B https://orcid.org/0000-0002-3603-6617
dc.contributor.author	Zhu, T
dc.contributor.author	Ding, M
dc.contributor.author	Zhou, W
dc.date	2023-08-19
dc.date.accessioned	2024-08-21T03:10:31Z
dc.date.available	2024-08-21T03:10:31Z
dc.date.issued	2024-08-01
dc.identifier.citation	Proceedings of the Thirty-ThirdInternational Joint Conference on Artificial Intelligence, 2024, pp. 512-520
dc.identifier.uri	http://hdl.handle.net/10453/180429
dc.description.abstract	<jats:p>While in-processing fairness approaches show promise in mitigating bias predictions, their potential impact on privacy leakage remains under-explored. We aim to address this gap by assessing the privacy risks of fairness-enhanced binary classifiers with membership inference attacks (MIAs). Surprisingly, our results reveal that these fairness interventions exhibit increased resilience against existing attacks, indicating that enhancing fairness does not necessarily lead to privacy compromises. However, we find current attack methods are ineffective as they typically degrade into simple threshold models with limited attack effectiveness. Following this observation, we discover a novel threat dubbed Fairness Discrepancy Membership Inference Attacks (FD-MIA) that exploits prediction discrepancies between fair and biased models. This attack reveals more potent vulnerabilities and poses significant privacy risks to model privacy. Extensive experiments across multiple datasets, attack methods, and representative fairness approaches confirm our findings and demonstrate the efficacy of the proposed attack method. Our study exposes the overlooked privacy threats in fairness studies, advocating for thorough evaluations of potential security vulnerabilities before model deployments.</jats:p>
dc.language	en
dc.publisher	International Joint Conferences on Artificial Intelligence
dc.relation	http://purl.org/au-research/grants/arc/DP230100246
dc.relation	http://purl.org/au-research/grants/arc/LP220200808
dc.relation.ispartof	Proceedings of the Thirty-ThirdInternational Joint Conference on Artificial Intelligence
dc.relation.ispartof	Proceedings of the Thirty-ThirdInternational Joint Conference on Artificial Intelligence
dc.relation.isbasedon	10.24963/ijcai.2024/57
dc.rights	info:eu-repo/semantics/openAccess
dc.title	When Fairness Meets Privacy: Exploring Privacy Threats in Fair Binary Classifiers via Membership Inference Attacks
dc.type	Conference Proceeding
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/Strength - AAII - Australian Artificial Intelligence Institute
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	open_access	*
dc.date.updated	2024-08-21T03:10:29Z
pubs.finish-date	2023-08-25
pubs.publication-status	Published
pubs.start-date	2023-08-19

Abstract:

While in-processing fairness approaches show promise in mitigating bias predictions, their potential impact on privacy leakage remains under-explored. We aim to address this gap by assessing the privacy risks of fairness-enhanced binary classifiers with membership inference attacks (MIAs). Surprisingly, our results reveal that these fairness interventions exhibit increased resilience against existing attacks, indicating that enhancing fairness does not necessarily lead to privacy compromises. However, we find current attack methods are ineffective as they typically degrade into simple threshold models with limited attack effectiveness. Following this observation, we discover a novel threat dubbed Fairness Discrepancy Membership Inference Attacks (FD-MIA) that exploits prediction discrepancies between fair and biased models. This attack reveals more potent vulnerabilities and poses significant privacy risks to model privacy. Extensive experiments across multiple datasets, attack methods, and representative fairness approaches confirm our findings and demonstrate the efficacy of the proposed attack method. Our study exposes the overlooked privacy threats in fairness studies, advocating for thorough evaluations of potential security vulnerabilities before model deployments.

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

http://hdl.handle.net/10453/180429