Fast and Differentially Private Algorithms for Decentralized Collaborative Machine Learning

Aurélien Bellet 1 Rachid Guerraoui 2 Mahsa Taziki 2 Marc Tommasi 1, 3
1 MAGNET - Machine Learning in Information Networks
Inria Lille - Nord Europe, CRIStAL - Centre de Recherche en Informatique, Signal et Automatique de Lille (CRIStAL) - UMR 9189
Abstract : Consider a set of agents in a peer-to-peer communication network, where each agent has a personal dataset and a personal learning objective. The main question addressed in this paper is: how can agents collaborate to improve upon their locally learned model without leaking sensitive information about their data? Our first contribution is to reformulate this problem so that it can be solved by a block coordinate descent algorithm. We obtain an efficient and fully decentralized protocol working in an asynchronous fashion. Our second contribution is to make our algorithm differentially private to protect against the disclosure of any information about personal datasets. We prove convergence rates and exhibit the trade-off between utility and privacy. Our experiments show that our approach dramatically outperforms previous work in the non-private case, and that under privacy constraints we significantly improve over purely local models.
Complete list of metadatas

Cited literature [28 references]  Display  Hide  Download

https://hal.inria.fr/hal-01665410
Contributor : Aurélien Bellet <>
Submitted on : Friday, December 15, 2017 - 7:14:19 PM
Last modification on : Tuesday, September 10, 2019 - 11:32:02 AM

File

Decentralized_Personalized_CD_...
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01665410, version 1
  • ARXIV : 1705.08435

Citation

Aurélien Bellet, Rachid Guerraoui, Mahsa Taziki, Marc Tommasi. Fast and Differentially Private Algorithms for Decentralized Collaborative Machine Learning. [Research Report] INRIA Lille. 2017, pp.1-18. ⟨hal-01665410⟩

Share

Metrics

Record views

346

Files downloads

373