Secure Multiparty Computation For Data Sharing

Abstract

Secure Multiparty Computation (SMPC) enables multiple parties to collaborate and compute outcomes from their respective private data without sharing the data per se with one another. This document discusses applying the additive secret sharing method in order to reach such secure computation. Here, a portion of each datum is divided into random shares that are distributed between the members. No individual share reveals any information regarding the initial data. The operations, e.g., addition and multiplication, are performed directly on these shared values, and the resulting value can be achieved only once all shares are added together. This method maintains the data confidential yet useful computations can be executed. It is scalable, cost-effective, and highly secure against data leakage, and thus perfect for use in secure cloud computing, data analytics, and privacy-preserving machine learning.