Quantifying Membership Privacy via Information Leakage

被引:20
作者
Saeidian, Sara [1 ]
Cervia, Giulia [2 ,3 ]
Oechtering, Tobias J. [1 ]
Skoglund, Mikael [1 ]
机构
[1] KTH Royal Inst Technol, Div Informat Sci & Engn, Sch Elect Engn & Comp Sci, S-10044 Stockholm, Sweden
[2] KTH Royal Inst Technol, Sch Elect Engn & Comp Sci, S-10044 Stockholm, Sweden
[3] Univ Lille, Ctr Digital Syst, IMT Lille Douai, Inst Mines Telecom, F-59000 Lille, France
关键词
Privacy; Differential privacy; Measurement; Training; Machine learning; Data models; Upper bound; Privacy-preserving machine learning; membership inference; maximal leakage; log-concave probability density;
D O I
10.1109/TIFS.2021.3073804
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
Machine learning models are known to memorize the unique properties of individual data points in a training set. This memorization capability can be exploited by several types of attacks to infer information about the training data, most notably, membership inference attacks. In this paper, we propose an approach based on information leakage for guaranteeing membership privacy. Specifically, we propose to use a conditional form of the notion of maximal leakage to quantify the information leaking about individual data entries in a dataset, i.e., the entrywise information leakage. We apply our privacy analysis to the Private Aggregation of Teacher Ensembles (PATE) framework for privacy-preserving classification of sensitive data and prove that the entrywise information leakage of its aggregation mechanism is Schur-concave when the injected noise has a log-concave probability density. The Schur-concavity of this leakage implies that increased consensus among teachers in labeling a query reduces its associated privacy cost. Finally, we derive upper bounds on the entrywise information leakage when the aggregation mechanism uses Laplace distributed noise.
引用
收藏
页码:3096 / 3108
页数:13
相关论文
共 50 条
[41]   Information Leakage in Embedding Models [J].
Song, Congzheng ;
Raghunathan, Ananth .
CCS '20: PROCEEDINGS OF THE 2020 ACM SIGSAC CONFERENCE ON COMPUTER AND COMMUNICATIONS SECURITY, 2020, :377-390
[42]   Quantifying privacy in multiagent planning [J].
van der Krogt, Roman .
MULTIAGENT AND GRID SYSTEMS, 2009, 5 (04) :451-469
[43]   On the Leakage of Personally Identifiable Information Via Online Social Networks [J].
Krishnamurthy, Balachander ;
Wills, Craig E. .
2ND ACM SIGCOMM WORKSHOP ON ONLINE SOCIAL NETWORKS (WOSN 09), 2009, :7-12
[44]   One Parameter Defense-Defending Against Data Inference Attacks via Differential Privacy [J].
Ye, Dayong ;
Shen, Sheng ;
Zhu, Tianqing ;
Liu, Bo ;
Zhou, Wanlei .
IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY, 2022, 17 :1466-1480
[45]   VisRecall: Quantifying Information Visualisation Recallability via Question Answering [J].
Wang, Yao ;
Jiao, Chuhan ;
Bace, Mihai ;
Bulling, Andreas .
IEEE TRANSACTIONS ON VISUALIZATION AND COMPUTER GRAPHICS, 2022, 28 (12) :4995-5005
[46]   Dealing with Privacy for Protecting Information [J].
Larrucea, Xabier ;
Santamaria, Izaskun .
SYSTEMS, SOFTWARE AND SERVICES PROCESS IMPROVEMENT, EUROSPI 2021, 2021, 1442 :518-530
[47]   Secure Metric Learning via Differential Pairwise Privacy [J].
Li, Jing ;
Pan, Yuangang ;
Sui, Yulei ;
Tsang, Ivor W. .
IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY, 2020, 15 :3640-3652
[48]   Defending CNN against privacy leakage in edge computing via binary neural networks [J].
Qiang, Weizhong ;
Liu, Renwan ;
Jin, Hai .
FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE, 2021, 125 :460-470
[49]   Privacy Leakage of SIFT Features via Deep Generative Model Based Image Reconstruction [J].
Wu, Haiwei ;
Zhou, Jiantao .
IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY, 2021, 16 :2973-2985
[50]   Privacy Enhancement Via Dummy Points in the Shuffle Model [J].
Li, Xiaochen ;
Liu, Weiran ;
Feng, Hanwen ;
Huang, Kunzhe ;
Hu, Yuke ;
Liu, Jinfei ;
Ren, Kui ;
Qin, Zhan .
IEEE TRANSACTIONS ON DEPENDABLE AND SECURE COMPUTING, 2024, 21 (03) :1001-1016