Privacy under hard distortion constraints

Jiachun Liao, Oliver Kosut, Lalitha Sankar, Flavio P. Calmon

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

Abstract

We study the problem of data disclosure with privacy guarantees, wherein the utility of the disclosed data is ensured via a hard distortion constraint. Unlike average distortion, hard distortion provides a deterministic guarantee of fidelity. For the privacy measure, we use a tunable information leakage measure, namely maximal α-leakage (α ∈ [1, ∞]), and formulate the privacy-utility tradeoff problem. The resulting solution highlights that under a hard distortion constraint, the nature of the solution remains unchanged for both local and non-local privacy requirements. More precisely, we show that both the optimal mechanism and the optimal tradeoff are invariant for any α > 1; i.e., the tunable leakage measure only behaves as either of the two extrema, i.e., mutual information for α = 1 and maximal leakage for α = ∞.

Original languageEnglish (US)
Title of host publication2018 IEEE Information Theory Workshop, ITW 2018
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781538635995
DOIs
StatePublished - Jan 15 2019
Event2018 IEEE Information Theory Workshop, ITW 2018 - Guangzhou, China
Duration: Nov 25 2018Nov 29 2018

Publication series

Name2018 IEEE Information Theory Workshop, ITW 2018

Conference

Conference2018 IEEE Information Theory Workshop, ITW 2018
CountryChina
CityGuangzhou
Period11/25/1811/29/18

Keywords

  • F-divergence
  • Hard distortion
  • Maximal α-leakage
  • Privacy-utility tradeoff

ASJC Scopus subject areas

  • Information Systems

Cite this

Liao, J., Kosut, O., Sankar, L., & Calmon, F. P. (2019). Privacy under hard distortion constraints. In 2018 IEEE Information Theory Workshop, ITW 2018 [8613385] (2018 IEEE Information Theory Workshop, ITW 2018). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ITW.2018.8613385

Privacy under hard distortion constraints. / Liao, Jiachun; Kosut, Oliver; Sankar, Lalitha; Calmon, Flavio P.

2018 IEEE Information Theory Workshop, ITW 2018. Institute of Electrical and Electronics Engineers Inc., 2019. 8613385 (2018 IEEE Information Theory Workshop, ITW 2018).

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

Liao, J, Kosut, O, Sankar, L & Calmon, FP 2019, Privacy under hard distortion constraints. in 2018 IEEE Information Theory Workshop, ITW 2018., 8613385, 2018 IEEE Information Theory Workshop, ITW 2018, Institute of Electrical and Electronics Engineers Inc., 2018 IEEE Information Theory Workshop, ITW 2018, Guangzhou, China, 11/25/18. https://doi.org/10.1109/ITW.2018.8613385
Liao J, Kosut O, Sankar L, Calmon FP. Privacy under hard distortion constraints. In 2018 IEEE Information Theory Workshop, ITW 2018. Institute of Electrical and Electronics Engineers Inc. 2019. 8613385. (2018 IEEE Information Theory Workshop, ITW 2018). https://doi.org/10.1109/ITW.2018.8613385
Liao, Jiachun ; Kosut, Oliver ; Sankar, Lalitha ; Calmon, Flavio P. / Privacy under hard distortion constraints. 2018 IEEE Information Theory Workshop, ITW 2018. Institute of Electrical and Electronics Engineers Inc., 2019. (2018 IEEE Information Theory Workshop, ITW 2018).
@inproceedings{325f71bc5753470db548d7ac93634b0b,
title = "Privacy under hard distortion constraints",
abstract = "We study the problem of data disclosure with privacy guarantees, wherein the utility of the disclosed data is ensured via a hard distortion constraint. Unlike average distortion, hard distortion provides a deterministic guarantee of fidelity. For the privacy measure, we use a tunable information leakage measure, namely maximal α-leakage (α ∈ [1, ∞]), and formulate the privacy-utility tradeoff problem. The resulting solution highlights that under a hard distortion constraint, the nature of the solution remains unchanged for both local and non-local privacy requirements. More precisely, we show that both the optimal mechanism and the optimal tradeoff are invariant for any α > 1; i.e., the tunable leakage measure only behaves as either of the two extrema, i.e., mutual information for α = 1 and maximal leakage for α = ∞.",
keywords = "F-divergence, Hard distortion, Maximal α-leakage, Privacy-utility tradeoff",
author = "Jiachun Liao and Oliver Kosut and Lalitha Sankar and Calmon, {Flavio P.}",
year = "2019",
month = "1",
day = "15",
doi = "10.1109/ITW.2018.8613385",
language = "English (US)",
series = "2018 IEEE Information Theory Workshop, ITW 2018",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
booktitle = "2018 IEEE Information Theory Workshop, ITW 2018",

}

TY - GEN

T1 - Privacy under hard distortion constraints

AU - Liao, Jiachun

AU - Kosut, Oliver

AU - Sankar, Lalitha

AU - Calmon, Flavio P.

PY - 2019/1/15

Y1 - 2019/1/15

N2 - We study the problem of data disclosure with privacy guarantees, wherein the utility of the disclosed data is ensured via a hard distortion constraint. Unlike average distortion, hard distortion provides a deterministic guarantee of fidelity. For the privacy measure, we use a tunable information leakage measure, namely maximal α-leakage (α ∈ [1, ∞]), and formulate the privacy-utility tradeoff problem. The resulting solution highlights that under a hard distortion constraint, the nature of the solution remains unchanged for both local and non-local privacy requirements. More precisely, we show that both the optimal mechanism and the optimal tradeoff are invariant for any α > 1; i.e., the tunable leakage measure only behaves as either of the two extrema, i.e., mutual information for α = 1 and maximal leakage for α = ∞.

AB - We study the problem of data disclosure with privacy guarantees, wherein the utility of the disclosed data is ensured via a hard distortion constraint. Unlike average distortion, hard distortion provides a deterministic guarantee of fidelity. For the privacy measure, we use a tunable information leakage measure, namely maximal α-leakage (α ∈ [1, ∞]), and formulate the privacy-utility tradeoff problem. The resulting solution highlights that under a hard distortion constraint, the nature of the solution remains unchanged for both local and non-local privacy requirements. More precisely, we show that both the optimal mechanism and the optimal tradeoff are invariant for any α > 1; i.e., the tunable leakage measure only behaves as either of the two extrema, i.e., mutual information for α = 1 and maximal leakage for α = ∞.

KW - F-divergence

KW - Hard distortion

KW - Maximal α-leakage

KW - Privacy-utility tradeoff

UR - http://www.scopus.com/inward/record.url?scp=85062071748&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85062071748&partnerID=8YFLogxK

U2 - 10.1109/ITW.2018.8613385

DO - 10.1109/ITW.2018.8613385

M3 - Conference contribution

T3 - 2018 IEEE Information Theory Workshop, ITW 2018

BT - 2018 IEEE Information Theory Workshop, ITW 2018

PB - Institute of Electrical and Electronics Engineers Inc.

ER -