We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Mathematics > Statistics Theory

Title: Local Privacy, Data Processing Inequalities, and Statistical Minimax Rates

Abstract: Working under a model of privacy in which data remains private even from the statistician, 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 depend on the privacy guarantees. When combined with standard minimax techniques, including the Le Cam, Fano, and Assouad methods, these inequalities allow for a precise characterization of statistical rates under local privacy constraints. We provide a treatment of several canonical families of problems: mean estimation, parameter estimation in fixed-design regression, multinomial probability estimation, 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.
Comments: 59 pages, 3 figures
Subjects: Statistics Theory (math.ST); Cryptography and Security (cs.CR); Information Theory (cs.IT)
Cite as: arXiv:1302.3203 [math.ST]
  (or arXiv:1302.3203v4 [math.ST] for this version)

Submission history

From: John Duchi [view email]
[v1] Wed, 13 Feb 2013 20:06:06 GMT (126kb)
[v2] Thu, 10 Oct 2013 18:00:24 GMT (151kb)
[v3] Thu, 9 Jan 2014 02:22:59 GMT (175kb)
[v4] Wed, 27 Aug 2014 04:40:22 GMT (179kb)

Link back to: arXiv, form interface, contact.