Skip to main content

Research Repository

Advanced Search

Encrypted accelerated least squares regression

Esperança, P.M.; Aslett, L.J.M.; Holmes, C.C.

Encrypted accelerated least squares regression Thumbnail


Authors

P.M. Esperança

C.C. Holmes



Contributors

Aarti Singh
Editor

Jerry Zhu
Editor

Abstract

Information that is stored in an encrypted format is, by definition, usually not amenable to statistical analysis or machine learning methods. In this paper we present detailed analysis of coordinate and accelerated gradient descent algorithms which are capable of fitting least squares and penalised ridge regression models, using data encrypted under a fully homomorphic encryption scheme. Gradient descent is shown to dominate in terms of encrypted computational speed, and theoretical results are proven to give parameter bounds which ensure correctness of decryption. The characteristics of encrypted computation are empirically shown to favour a non-standard acceleration technique. This demonstrates the possibility of approximating conventional statistical regression methods using encrypted data without compromising privacy.

Citation

Esperança, P., Aslett, L., & Holmes, C. (2017). Encrypted accelerated least squares regression. In A. Singh, & J. Zhu (Eds.), Proceedings of the 20th International Conference on Artificial Intelligence and Statistics, PMLR 54 (334-343)

Conference Name The 20th International Conference on Artificial Intelligence and Statistics
Conference Location Fort Lauderdale, Florida
Acceptance Date Jan 25, 2017
Online Publication Date Apr 20, 2017
Publication Date Jan 1, 2017
Deposit Date Apr 24, 2017
Publicly Available Date May 4, 2017
Volume 54
Pages 334-343
Series Title Proceedings of machine learning research
Series ISSN 2640-3498
Book Title Proceedings of the 20th International Conference on Artificial Intelligence and Statistics, PMLR 54.
Public URL https://durham-repository.worktribe.com/output/1146354
Publisher URL http://proceedings.mlr.press/v54/esperanca17a/esperanca17a.pdf

Files

Published Conference Proceeding (416 Kb)
PDF

Copyright Statement
Copyright 2017 by the author(s).




You might also like



Downloadable Citations