Distributed Point Functions meet Secure Multiparty Computation

Abstract:

Secure Multiparty Computation allows two or more computing parties to collaboratively compute some function on their (secret) inputs while keeping their inputs secret. MPC has a wide range of applications for privacy-protecting systems. However, unfortunately, the efficiency of MPC protocols is still somewhat limited. Distributed Point Functions (DPFs) are cryptographic primitives allowing us to share a standard-basis vector among two or more parties succinctly. In this talk, we present novel techniques which use DPFs to construct efficient MPC protocols. In this presentation, we will discuss some of the novel hitherto unobserved properties of DPFs. We expand these observations to produce the most efficient MPC protocols to evaluate piecewise polynomial functions, also known as splines. This method’s scalability and efficiency allow splines to be used for extremely high accuracy approximation of non-linear functions in MPC.