In this episode, we chat with Flavio Bergamaschi from IBM research about Fully Homomorphic Encryption (FHE). FHE allows for computation on encrypted data. First developed in 2009 at IBM, this tech has long been the considered only theoretically possible. However, as we learn in the interview, there have been strides made in the last few years and we are starting to see FHE technology being used in some real world applications.
In this interview, we discuss the origin of the technology, what it is, how it can be combined with other cryptographic techniques such as MPC and ZKPs, and applications for FHE technology being explored today.
Here are some relevant links:
IBM's Open Source Homomorphic Encryption Library (HElib) v1.0.0
Towards a Homomorphic Machine Learning Big Data Pipeline for the Financial Services Sector
Homomorphic Training of 30,000 Logistic Regression Models
Homomorphically Securing AI at the Edge
Article on IBM's FHE work with a Brazilian bank
Episode 63: Proxy re-encryption and FHE with NuCypher
Nigel Smart on Multi-Party Computation: From Theory to Practice
Or directly here: