Secure MPC
Secure MPC
Welcome to Secure multi-party computation. Please watch the video below, in order to get an overview on our system.
Why MPC?
Consider a set of parties who do not trust each other, nor the channels by which they communicate.
Still, the parties wish to correctly compute some common function on their local inputs in the purpose of creating statistics, draw broader conclusions, train models of machine learning etc.
Our System
Our software uses secure MPC in order to sort the users' data.
It takes as input a list of different numbers from each user, and sorts the union of these lists.
Later, it keeps the output for further computations.
During the whole process, the real data is never revealed, only in a shared-secret form. This way we keep the users' data private.
Credits
Our system is based on an algorithm developed by Dr. Elette Boyle, Dr. Niv Gilboa & Prof. Yuval Ishai, found in the following paper: Secure Computation with Preprocessing via Function Secret Sharing.
We used an implementation of the algorithm written in
found in the following GitHub repository: github.com/OpenMined/sycret