Cryptology ePrint Archive: Report 2022/146

Training Differentially Private Models with Secure Multiparty Computation

Sikha Pentyala and Davis Railsback and Ricardo Maia and Rafael Dowsley and David Melanson and Anderson Nascimento and Martine De Cock

Abstract: We address the problem of learning a machine learning model from training data that originates at multiple data owners, while providing formal privacy guarantees regarding the protection of each owner's data. Existing solutions based on Differential Privacy (DP) achieve this at the cost of a drop in accuracy. Solutions based on Secure Multiparty Computation (MPC) do not incur such accuracy loss but leak information when the trained model is made publicly available. We propose an MPC solution for training DP models. Our solution relies on an MPC protocol for model training, and an MPC protocol for perturbing the trained model coefficients with Laplace noise in a privacy-preserving manner. The resulting MPC+DP approach achieves higher accuracy than a pure DP approach, while providing the same formal privacy guarantees. Our work obtained first place in the iDASH2021 Track III competition on confidential computing for secure genome analysis.

Category / Keywords: cryptographic protocols / Secure Multiparty Computation, Differential Privacy, Logistic Regression

Date: received 9 Feb 2022

Contact author: sikha at uw edu, drail at uw edu, ricardo menezes at aluno unb br, rafael dowsley at monash edu, mence40 at uw edu, andclay at uw edu, mdecock at uw ed

Available format(s): PDF | BibTeX Citation

Version: 20220212:064349 (All versions of this report)

Short URL: ia.cr/2022/146