Minimax Optimal Procedures for Locally Private Estimation

被引:262
作者
Duchi, John C. [1 ,2 ]
Jordan, Michael I. [3 ,4 ]
Wainwright, Martin J. [3 ,4 ]
机构
[1] Stanford Univ, Dept Stat, Stanford, CA 94305 USA
[2] Stanford Univ, Dept Elect Engn, Stanford, CA 94305 USA
[3] Univ Calif Berkeley, Dept Stat, Berkeley, CA 94720 USA
[4] Univ Calif Berkeley, Dept Elect Engn & Comp Sci, Berkeley, CA 94720 USA
基金
美国国家科学基金会;
关键词
Differential privacy; Minimax bounds; Optimal estimators; STOCHASTIC-APPROXIMATION; CONVERGENCE; RATES; INFORMATION; CENSORSHIP; DISCLOSURE; FRAMEWORK; NOISY; RISK;
D O I
10.1080/01621459.2017.1389735
中图分类号
O21 [概率论与数理统计]; C8 [统计学];
学科分类号
020208 ; 070103 ; 0714 ;
摘要
Working under a model of privacy in which data remain private even from the statistician, we study the tradeoff between privacy guarantees and the risk of the resulting statistical estimators. We develop private versions of classical information-theoretical bounds, in particular those due to Le Cam, Fano, and Assouad. These inequalities allow for a precise characterization of statistical rates under local privacy constraints and the development of provably (minimax) optimal estimation procedures. We provide a treatment of several canonical families of problems: mean estimation and median estimation, generalized linear models, and nonparametric density estimation. For all of these families, we provide lower and upper bounds that match up to constant factors, and exhibit new (optimal) privacy-preserving mechanisms and computationally efficient estimators that achieve the bounds. Additionally, we present a variety of experimental results for estimation problems involving sensitive data, including salaries, censored blog posts and articles, and drug abuse; these experiments demonstrate the importance of deriving optimal procedures. Supplementary materials for this article are available online.
引用
收藏
页码:182 / 201
页数:20
相关论文
共 55 条
[1]  
[Anonymous], 1990, Entropy and Information Theory
[2]  
[Anonymous], 2013, ARXIV13046133
[3]  
[Anonymous], 1990, Proceedings of the 1990 Symposium on Information Theory and its Applications (ISITA-90)
[4]  
[Anonymous], 134760 SMA SUBST AB
[5]  
[Anonymous], 1986, IMS Lecture Notes Monograph Series
[6]  
[Anonymous], ANN REP EMPL COMP 20
[7]  
[Anonymous], 2012, J PRIV CONFID
[8]   On the Fundamental Limits of Adaptive Sensing [J].
Arias-Castro, Ery ;
Candes, Emmanuel J. ;
Davenport, Mark A. .
IEEE TRANSACTIONS ON INFORMATION THEORY, 2013, 59 (01) :472-481
[9]  
Assouad B. Yu., 1997, Festschrift for Lucien Le Cam, P423
[10]  
ASSOUAD P, 1983, CR ACAD SCI I-MATH, V296, P1021