PROGCRYPTO 2023

Abstract

By default noisy homomorphic encryption schemes, such as the ones based on the Ring-Learning-With-Errors problem, can only evaluate a limited number of operations before decryption is required. Bootstrapping is the ability for such a scheme to evaluate its own decryption circuit under encryption. This enables these schemes to perform an unlimited (or practically unlimited) number of operations without requiring intermediate decryption. This talk will give an introduction to the concept of bootstrapping for approximate homomorphic encryption, a homomorphic encryption scheme that enables encrypted fixed point arithmetic over the complex field. The aim of the talk is for the participant to understand how the process of bootstrapping for such a scheme works and what the current state of the art is. Prior knowledge of this specific scheme is not required, but some familiarity with the concept of Ring-Learning-With-Errors and polynomial rings is recommended.

Jean-Philippe Bossuat

Cryptography Research Scientist & Software Specialist

My research interests include applied lattice based cryptography, privacy preserving machine learning and secure analytics.