Differentially private Bayesian programming

Gilles Barthe; Gian Pietro Farina; Andy Gordon; Emilio Jesús Gallego Arias; Marco Gaboardi; Justin Hsu; Pierre Yves Strub

doi:10.1145/2976749.2978371

Differentially private Bayesian programming

Gilles Barthe
, Gian Pietro Farina
, Andy Gordon
, Emilio Jesús Gallego Arias
, Marco Gaboardi
, Justin Hsu
, Pierre Yves Strub

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

We present PrivInfer, an expressive framework for writing and verifying differentially private Bayesian machine learning algorithms. Programs in PrivInfer are written in a rich functional probabilistic programming language with constructs for performing Bayesian inference. Then, differential privacy of programs is established using a relational refinement type system, in which refinements on probability types are indexed by a metric on distributions. Our framework leverages recent developments in Bayesian inference, probabilistic programming languages, and in relational refinement types. We demonstrate the expressiveness of PrivInfer by verifying privacy for several examples of private Bayesian inference.

Original language	English
Title of host publication	CCS 2016 - Proceedings of the 2016 ACM SIGSAC Conference on Computer and Communications Security
Publisher	Association for Computing Machinery
Pages	68-79
Number of pages	12
ISBN (Electronic)	9781450341394
DOIs	https://doi.org/10.1145/2976749.2978371
Publication status	Published - 24 Oct 2016
Externally published	Yes
Event	23rd ACM Conference on Computer and Communications Security, CCS 2016 - Vienna, Austria Duration: 24 Oct 2016 → 28 Oct 2016

Publication series

Name	Proceedings of the ACM Conference on Computer and Communications Security
Volume	24-28-October-2016
ISSN (Print)	1543-7221

Conference

Conference	23rd ACM Conference on Computer and Communications Security, CCS 2016
Country/Territory	Austria
City	Vienna
Period	24/10/16 → 28/10/16

Access to Document

10.1145/2976749.2978371

Cite this

Barthe, G., Farina, G. P., Gordon, A., Arias, E. J. G., Gaboardi, M., Hsu, J., & Strub, P. Y. (2016). Differentially private Bayesian programming. In CCS 2016 - Proceedings of the 2016 ACM SIGSAC Conference on Computer and Communications Security (pp. 68-79). (Proceedings of the ACM Conference on Computer and Communications Security; Vol. 24-28-October-2016). Association for Computing Machinery. https://doi.org/10.1145/2976749.2978371

@inproceedings{607313660d0e44b49248151e2fe3b637,

title = "Differentially private Bayesian programming",

abstract = "We present PrivInfer, an expressive framework for writing and verifying differentially private Bayesian machine learning algorithms. Programs in PrivInfer are written in a rich functional probabilistic programming language with constructs for performing Bayesian inference. Then, differential privacy of programs is established using a relational refinement type system, in which refinements on probability types are indexed by a metric on distributions. Our framework leverages recent developments in Bayesian inference, probabilistic programming languages, and in relational refinement types. We demonstrate the expressiveness of PrivInfer by verifying privacy for several examples of private Bayesian inference.",

author = "Gilles Barthe and Farina, \{Gian Pietro\} and Andy Gordon and Arias, \{Emilio Jes{\'u}s Gallego\} and Marco Gaboardi and Justin Hsu and Strub, \{Pierre Yves\}",

note = "Publisher Copyright: {\textcopyright} 2016 Copyright held by the owner/author(s). Publication rights licensed to ACM.; 23rd ACM Conference on Computer and Communications Security, CCS 2016 ; Conference date: 24-10-2016 Through 28-10-2016",

year = "2016",

month = oct,

day = "24",

doi = "10.1145/2976749.2978371",

language = "English",

series = "Proceedings of the ACM Conference on Computer and Communications Security",

publisher = "Association for Computing Machinery",

pages = "68--79",

booktitle = "CCS 2016 - Proceedings of the 2016 ACM SIGSAC Conference on Computer and Communications Security",

}

Barthe, G, Farina, GP, Gordon, A, Arias, EJG, Gaboardi, M, Hsu, J & Strub, PY 2016, Differentially private Bayesian programming. in CCS 2016 - Proceedings of the 2016 ACM SIGSAC Conference on Computer and Communications Security. Proceedings of the ACM Conference on Computer and Communications Security, vol. 24-28-October-2016, Association for Computing Machinery, pp. 68-79, 23rd ACM Conference on Computer and Communications Security, CCS 2016, Vienna, Austria, 24/10/16. https://doi.org/10.1145/2976749.2978371

Differentially private Bayesian programming. / Barthe, Gilles; Farina, Gian Pietro; Gordon, Andy et al.
CCS 2016 - Proceedings of the 2016 ACM SIGSAC Conference on Computer and Communications Security. Association for Computing Machinery, 2016. p. 68-79 (Proceedings of the ACM Conference on Computer and Communications Security; Vol. 24-28-October-2016).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Differentially private Bayesian programming

AU - Barthe, Gilles

AU - Farina, Gian Pietro

AU - Gordon, Andy

AU - Arias, Emilio Jesús Gallego

AU - Gaboardi, Marco

AU - Hsu, Justin

AU - Strub, Pierre Yves

PY - 2016/10/24

Y1 - 2016/10/24

N2 - We present PrivInfer, an expressive framework for writing and verifying differentially private Bayesian machine learning algorithms. Programs in PrivInfer are written in a rich functional probabilistic programming language with constructs for performing Bayesian inference. Then, differential privacy of programs is established using a relational refinement type system, in which refinements on probability types are indexed by a metric on distributions. Our framework leverages recent developments in Bayesian inference, probabilistic programming languages, and in relational refinement types. We demonstrate the expressiveness of PrivInfer by verifying privacy for several examples of private Bayesian inference.

AB - We present PrivInfer, an expressive framework for writing and verifying differentially private Bayesian machine learning algorithms. Programs in PrivInfer are written in a rich functional probabilistic programming language with constructs for performing Bayesian inference. Then, differential privacy of programs is established using a relational refinement type system, in which refinements on probability types are indexed by a metric on distributions. Our framework leverages recent developments in Bayesian inference, probabilistic programming languages, and in relational refinement types. We demonstrate the expressiveness of PrivInfer by verifying privacy for several examples of private Bayesian inference.

UR - https://www.scopus.com/pages/publications/84995467622

U2 - 10.1145/2976749.2978371

DO - 10.1145/2976749.2978371

M3 - Conference contribution

AN - SCOPUS:84995467622

T3 - Proceedings of the ACM Conference on Computer and Communications Security

SP - 68

EP - 79

BT - CCS 2016 - Proceedings of the 2016 ACM SIGSAC Conference on Computer and Communications Security

PB - Association for Computing Machinery

T2 - 23rd ACM Conference on Computer and Communications Security, CCS 2016

Y2 - 24 October 2016 through 28 October 2016

ER -

Barthe G, Farina GP, Gordon A, Arias EJG, Gaboardi M, Hsu J et al. Differentially private Bayesian programming. In CCS 2016 - Proceedings of the 2016 ACM SIGSAC Conference on Computer and Communications Security. Association for Computing Machinery. 2016. p. 68-79. (Proceedings of the ACM Conference on Computer and Communications Security). doi: 10.1145/2976749.2978371

Differentially private Bayesian programming

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this