Private learning and sanitization: Pure vs. approximate differential privacy

Amos Beimel, Kobbi Nissim, Uri Stemmer

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

77 Scopus citations

Abstract

We compare the sample complexity of private learning and sanitization tasks under pure ε-differential privacy [Dwork, McSherry, Nissim, and Smith TCC 2006] and approximate (ε,δ)-differential privacy [Dwork, Kenthapadi, McSherry, Mironov, and Naor EUROCRYPT 2006]. We show that the sample complexity of these tasks under approximate differential privacy can be significantly lower than that under pure differential privacy.

Original languageEnglish
Title of host publicationApproximation, Randomization, and Combinatorial Optimization
Subtitle of host publicationAlgorithms and Techniques - 16th International Workshop, APPROX 2013 and 17th International Workshop, RANDOM 2013, Proceedings
Pages363-378
Number of pages16
DOIs
StatePublished - 2013
Externally publishedYes
Event16th International Workshop on Approximation Algorithms for Combinatorial Optimization Problems, APPROX 2013 and the 17th International Workshop on Randomization and Computation, RANDOM 2013 - Berkeley, CA, United States
Duration: 21 Aug 201323 Aug 2013

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8096 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference16th International Workshop on Approximation Algorithms for Combinatorial Optimization Problems, APPROX 2013 and the 17th International Workshop on Randomization and Computation, RANDOM 2013
Country/TerritoryUnited States
CityBerkeley, CA
Period21/08/1323/08/13

Funding

FundersFunder number
Frankel Center for Computer Science at Ben-Gurion University
Google, Inc.
National Science FoundationCNS-1237235
Israel Science Foundation2761/12, 938/09

    Keywords

    • Differential Privacy
    • Private Learning
    • Sanitization

    Fingerprint

    Dive into the research topics of 'Private learning and sanitization: Pure vs. approximate differential privacy'. Together they form a unique fingerprint.

    Cite this