Zvika Brakerski
Stanford University

Fully Homomorphic Encryption


Fully homomorphic encryption (FHE) is a method of encryption that
supports computation on encrypted data: For any efficient function f,
it is possible to convert Enc(x) => Enc(f(x)) using only public
information. FHE enables a multitude of cryptographic applications,
such as outsourcing computation to third parties without compromising
data privacy, and is thus a sought-after tool for cloud applications.

Developing an FHE scheme has been a goal in cryptography since the
late 70s, but the first candidate scheme was only introduced in 2009,
in Gentry's breakthrough work. Unfortunately, despite the great
theoretical implications, the first schemes were quite complicated to
explain and implement, and relied on fairly strong computational
assumptions. Recently we showed how to base FHE on the more standard
learning with errors (LWE) assumption, yielding constructions with
simplified presentation and improved efficiency and security. In the
talk I will explain how to construct such a "new-generation" FHE scheme.