SecureML: A System for Scalable Privacy-Preserving …

secureml : A System for Scalable privacy -PreservingMachine LearningPayman Mohassel Yupeng Zhang AbstractMachine learning is widely used in practice to produce predictive models for applicationssuch as image processing, speech and text recognition. These models are more accurate whentrained on large amount of data collected from different sources. However, the massive datacollection raises privacy this paper, we present new and efficient protocols for privacy preserving machine learningfor linear regression, logistic regression and neural network training using the stochastic gradientdescent method. Our protocols fall in the two-server model where data owners distribute theirprivate data among two non-colluding servers who train various models on the joint data usingsecure two-party computation (2PC).

orders of magnitude faster than the state of the art implementations for privacy preserving linear and logistic regressions, and scale to millions of data samples with thousands of features. We also implement the rst privacy preserving system for training neural networks. 1 Introduction

Loading..

Tags:

  System, Privacy, Preserving, Scalable, Secureml, A system for scalable privacy preserving, Privacy preserving, Privacy preserving system

Information