Abstract:
Secure multi-party computation (MPC) allows a set of parties to jointly compute a function on their private inputs, whilst leaking no further information on their inputs other than what can be learnt from the output. Secure computation techniques have many applications towards reducing the privacy and security risks from the vast amounts of personal data being stored today, as well as enabling new applications, by allowing data processing without the underlying plaintext being kept in any single place.

In this talk I will start with an introduction to the problem of secure computation, and a short survey of the progress that has been made in improving its efficiency in recent years. This will be focused around my research in designing improved building blocks and protocols. I will illustrate this with a few highlights based on my work on efficient oblivious transfer protocols, general-purpose MPC in the preprocessing model, and new applications of techniques from lattice-based cryptography. Along the way I will also discuss future research directions in these areas, which are needed to lower the cost and expand the scope of applications where secure computation can be used, to make the most out of this promising technology.