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

Statistics > Machine Learning

Title: Tight Differential Privacy for Discrete-Valued Mechanisms and for the Subsampled Gaussian Mechanism Using FFT

Abstract: We propose a numerical accountant for evaluating the tight $(\varepsilon,\delta)$-privacy loss for algorithms with discrete one dimensional output. The method is based on the privacy loss distribution formalism and it uses the recently introduced fast Fourier transform based accounting technique. We carry out an error analysis of the method in terms of moment bounds of the privacy loss distribution which leads to rigorous lower and upper bounds for the true $(\varepsilon,\delta)$-values. As an application, we present a novel approach to accurate privacy accounting of the subsampled Gaussian mechanism. This completes the previously proposed analysis by giving strict lower and upper bounds for the privacy parameters. We demonstrate the performance of the accountant on the binomial mechanism and show that our approach allows decreasing noise variance up to 75 percent at equal privacy compared to existing bounds in the literature. We also illustrate how to compute tight bounds for the exponential mechanism applied to counting queries.
Comments: 41 pages, 5 figures
Subjects: Machine Learning (stat.ML); Cryptography and Security (cs.CR); Machine Learning (cs.LG)
Journal reference: AISTATS (2021) 3358-3366
Cite as: arXiv:2006.07134 [stat.ML]
  (or arXiv:2006.07134v3 [stat.ML] for this version)

Submission history

From: Antti Koskela [view email]
[v1] Fri, 12 Jun 2020 12:46:42 GMT (83kb,D)
[v2] Sun, 1 Nov 2020 09:54:50 GMT (86kb,D)
[v3] Mon, 21 Jun 2021 11:49:45 GMT (720kb,D)

Link back to: arXiv, form interface, contact.