Secure Multiparty Computation

Abstract

Protocols for secure multiparty computation (MPC) enable a set of parties to interact and compute a joint function of their private inputs while revealing nothing but the output. The potential applications for MPC are huge: privacy-preserving auctions, private DNA comparisons, private machine learning, threshold cryptography, and so on. This chapter reviews what MPC is, what problems it solves and how it is being currently used. MPC can be used to run machine learning models on data without revealing the model to the data owner, and without revealing the data to the model owner. MPC still requires great expertise to deploy, and additional research breakthroughs are needed to make secure computation practical on large data sets and for complex problems, and to make it easy to use for non-experts. The progress from the past few years, and the large amount of applied research now being generated paints a positive future for MPC in practice.