Obfuscating the Dataset: Impacts and Applications

Yu, G; Wang, X; Sun, C; Yu, P; Ni, W; Liu, RP

Obfuscating the Dataset: Impacts and Applications

Yu, G Wang, X

Sun, C Yu, P Ni, W

Liu, RP

Permalink

Publisher:: ASSOC COMPUTING MACHINERY
Publication Type:: Journal Article
Citation:: ACM Transactions on Intelligent Systems and Technology, 2023, 14, (5)
Issue Date:: 2023-09-30

Closed Access

	Filename	Description	Size
	3597936.pdf	Published version	3.1 MB	Adobe PDF	View/Open

Copyright Clearance Process

Recently Added
In Progress
Closed Access

This item is closed access and not available.

Full metadata record

Field	Value	Language
dc.contributor.author	Yu, G
dc.contributor.author	Wang, X https://orcid.org/0000-0001-9439-6437
dc.contributor.author	Sun, C
dc.contributor.author	Yu, P
dc.contributor.author	Ni, W https://orcid.org/0000-0002-4933-594X
dc.contributor.author	Liu, RP
dc.date.accessioned	2024-04-12T04:58:01Z
dc.date.available	2024-04-12T04:58:01Z
dc.date.issued	2023-09-30
dc.identifier.citation	ACM Transactions on Intelligent Systems and Technology, 2023, 14, (5)
dc.identifier.issn	2157-6904
dc.identifier.issn	2157-6912
dc.identifier.uri	http://hdl.handle.net/10453/177832
dc.description.abstract	Obfuscating a dataset by adding random noises to protect the privacy of sensitive samples in the training dataset is crucial to prevent data leakage to untrusted parties when dataset sharing is essential. We conduct comprehensive experiments to investigate how the dataset obfuscation can affect the resultant model weights - in terms of the model accuracy, ℓ2-distance-based model distance, and level of data privacy - and discuss the potential applications with the proposed Privacy, Utility, and Distinguishability (PUD)-triangle diagram to visualize the requirement preferences. Our experiments are based on the popular MNIST and CIFAR-10 datasets under both independent and identically distributed (IID) and non-IID settings. Significant results include a tradeoff between the model accuracy and privacy level and a tradeoff between the model difference and privacy level. The results indicate broad application prospects for training outsourcing and guarding against attacks in federated learning both of which have been increasingly attractive in many areas, particularly learning in edge computing.
dc.language	English
dc.publisher	ASSOC COMPUTING MACHINERY
dc.relation.ispartof	ACM Transactions on Intelligent Systems and Technology
dc.relation.isbasedon	10.1145/3597936
dc.rights	info:eu-repo/semantics/closedAccess
dc.subject	0801 Artificial Intelligence and Image Processing, 0806 Information Systems
dc.subject.classification	4602 Artificial intelligence
dc.subject.classification	4611 Machine learning
dc.title	Obfuscating the Dataset: Impacts and Applications
dc.type	Journal Article
utslib.citation.volume	14
utslib.for	0801 Artificial Intelligence and Image Processing
utslib.for	0806 Information Systems
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 - GBDTC - Global Big Data Technologies
pubs.organisational-group	University of Technology Sydney/Faculty of Engineering and Information Technology/School of Electrical and Data Engineering
pubs.organisational-group	University of Technology Sydney/Strength - CCSP - Centre for Cyber Security and Privacy
utslib.copyright.status	closed_access	*
dc.date.updated	2024-04-12T04:58:00Z
pubs.issue	5
pubs.publication-status	Published
pubs.volume	14
utslib.citation.issue	5

Abstract:

Obfuscating a dataset by adding random noises to protect the privacy of sensitive samples in the training dataset is crucial to prevent data leakage to untrusted parties when dataset sharing is essential. We conduct comprehensive experiments to investigate how the dataset obfuscation can affect the resultant model weights - in terms of the model accuracy, ℓ2-distance-based model distance, and level of data privacy - and discuss the potential applications with the proposed Privacy, Utility, and Distinguishability (PUD)-triangle diagram to visualize the requirement preferences. Our experiments are based on the popular MNIST and CIFAR-10 datasets under both independent and identically distributed (IID) and non-IID settings. Significant results include a tradeoff between the model accuracy and privacy level and a tradeoff between the model difference and privacy level. The results indicate broad application prospects for training outsourcing and guarding against attacks in federated learning both of which have been increasingly attractive in many areas, particularly learning in edge computing.

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

http://hdl.handle.net/10453/177832