MPC in the real world: financial applications of multiparty computation
Multiparty computation (MPC) was first proposed in the 80s as a method for computing on encrypted data. A group of mistrustful parties can use MPC to calculate a joint function over their inputs while keeping them private. What was at first little more than a theoretical concept is now a full-fledged technology with multiple real-world deployments. However, the practicality of MPC depends heavily on the computation we want to perform. Among the applications that can benefit from MPC, several can be found within the financial sector. In this talk, we will dive into some of these applications, focusing on a recent work addressing the use case of dark pool trading, where pre-trade information remains hidden by default. We will examine the practical aspects and trade-offs to consider when designing algorithms with MPC, as well as how to quantify and mitigate information leakage from computation outputs.
