Local Privacy and Statistical Minimax Rates

被引:726
作者
Duchi, John C. [1 ]
Jordan, Michael I. [1 ]
Wainwright, Martin J. [1 ]
机构
[1] Univ Calif Berkeley, Dept Elect Engn & Comp Sci, Berkeley, CA 94720 USA
来源
2013 IEEE 54TH ANNUAL SYMPOSIUM ON FOUNDATIONS OF COMPUTER SCIENCE (FOCS) | 2013年
关键词
differential privacy; minimax rates; estimation; DIFFERENTIAL PRIVACY; NOISE;
D O I
10.1109/FOCS.2013.53
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
Working under local differential privacy-a model of privacy in which data remains private even from the statistician or learner-we study the tradeoff between privacy guarantees and the utility of the resulting statistical estimators. We prove bounds on information-theoretic quantities, including mutual information and Kullback-Leibler divergence, that influence estimation rates as a function of the amount of privacy preserved. When combined with minimax techniques such as Le Cam's and Fano's methods, these inequalities allow for a precise characterization of statistical rates under local privacy constraints. In this paper, we provide a treatment of two canonical problem families: mean estimation in location family models and convex risk minimization. For these families, we provide lower and upper bounds for estimation of population quantities that match up to constant factors, giving privacy-preserving mechanisms and computationally efficient estimators that achieve the bounds.
引用
收藏
页码:429 / 438
页数:10
相关论文
共 28 条
[1]   Information-Theoretic Lower Bounds on the Oracle Complexity of Stochastic Convex Optimization [J].
Agarwal, Alekh ;
Bartlett, Peter L. ;
Ravikumar, Pradeep ;
Wainwright, Martin J. .
IEEE TRANSACTIONS ON INFORMATION THEORY, 2012, 58 (05) :3235-3249
[2]  
[Anonymous], 2006, Elements of Information Theory
[3]  
[Anonymous], 2009, P 41 ANN ACM S THEOR
[4]  
Barak B., 2007, P 26 ACM S PRINC DAT
[5]  
Beimel A, 2008, LECT NOTES COMPUT SC, V5157, P451, DOI 10.1007/978-3-540-85174-5_25
[6]  
Chaudhuri Kamalika, 2012, P 29 INT C MACH LEAR
[7]  
De A., 2012, P 9 THEOR CRYPT C
[8]  
Duchi J. C., 2012, ADV NEURAL INFORM PR, V25
[9]  
Duchi J. C., 2013, ARXIV13023203MATHST
[10]  
DUNCAN GT, 1986, J AM STAT ASSOC, V81, P10, DOI 10.2307/2287959