Yupeng Zhang
Friday, August 25, 2017 at 12:00 PM in 373 Soda Hall
Title: SecureML: A System for Scalable Privacy-Preserving Machine Learning
Abstract: Machine learning is widely used in practice to produce predictive models for applications such as image processing, speech and text recognition. These models are more accurate when trained on large amount of data collected from different sources. However, the massive data collection raises privacy concerns. In this paper, we present new and efficient protocols for privacy preserving machine learning for linear regression, logistic regression and neural network training using the stochastic gradient descent method. Our protocols fall in the two-server model where data owners distribute their private data among two non-colluding servers who train various models on the joint data using secure two-party computation (2PC). We develop new techniques to support secure arithmetic operations on shared decimal numbers, and propose MPC-friendly alternatives to nonlinear functions such as sigmoid and softmax that are superior to prior work. We implement our system in C++. Our experiments validate that our protocols are several 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 first privacy preserving system for training neural networks. This is a joint work with Payman Mohassel.
Bio: Yupeng Zhang is a 4th year Ph.D. student at University of Maryland, working with Prof. Jonathan Katz and Prof. Charalampos (Babis) Papamanthou. His research interest is in the area of computer security, privacy and cryptography, with a specific focus on verifiable computation, searchable encryption and privacy-preserving machine learning. He is one of the recipients of Google PhD Fellowship 2017.