Asymptotically optimal and private statistical estimation

Adam Davison Smith

Research output: Chapter in Book/Report/Conference proceedingConference contribution

4 Scopus citations

Abstract

Differential privacy is a definition of "privacy" for statistical databases. The definition is simple, yet it implies strong semantics even in the presence of an adversary with arbitrary auxiliary information about the database. In this talk, we discuss recent work on measuring the utility of differentially private analyses via the traditional yardsticks of statistical inference. Specifically, we discuss two differentially private estimators that, given i.i.d. samples from a probability distribution, converge to the correct answer at the same rate as the optimal nonprivate estimator.

Original languageEnglish (US)
Title of host publicationCryptology and Network Security - 8th International Conference, CANS 2009, Proceedings
Pages53-57
Number of pages5
DOIs
StatePublished - Dec 14 2009
Event8th International Conference on Cryptology and Network Security, CANS 2009 - Kanazawa, Japan
Duration: Dec 12 2009Dec 14 2009

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume5888 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other8th International Conference on Cryptology and Network Security, CANS 2009
CountryJapan
CityKanazawa
Period12/12/0912/14/09

    Fingerprint

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Smith, A. D. (2009). Asymptotically optimal and private statistical estimation. In Cryptology and Network Security - 8th International Conference, CANS 2009, Proceedings (pp. 53-57). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5888 LNCS). https://doi.org/10.1007/978-3-642-10433-6_4