Input and Output Privacy-Preserving Linear Regression

Yoshinori AONO  Takuya HAYASHI  Le Trieu PHONG  Lihua WANG  

IEICE TRANSACTIONS on Information and Systems   Vol.E100-D   No.10   pp.2339-2347
Publication Date: 2017/10/01
Online ISSN: 1745-1361
DOI: 10.1587/transinf.2016INP0019
Type of Manuscript: Special Section PAPER (Special Section on Security, Privacy and Anonymity in Computation, Communication and Storage Systems)
Category: Privacy, anonymity, and fundamental theory
differential privacy,  homomorphic encryption,  LWE assumption,  linear regression,  

Full Text: PDF(614.7KB)>>
Buy this Article

We build a privacy-preserving system of linear regression protecting both input data secrecy and output privacy. Our system achieves those goals simultaneously via a novel combination of homomorphic encryption and differential privacy dedicated to linear regression and its variants (ridge, LASSO). Our system is proved scalable over cloud servers, and its efficiency is extensively checked by careful experiments.