CARS 2023: Unlocking the potential of Fully Homomorphic Encryption

Shruthi Gorantala leads the practical FHE team at Google. Previously she worked on large-scale machine learning in Google Assistant and Google Maps. She received her Master’s in Computer Science from the University of Pennsylvania with a focus on searchable encryption and lattice-based cryptography with Nadia Heninger and Brett Hemenway. Abstract: Fully Homomorphic Encryption is a cryptographic primitive that has recently shown considerable advances toward being practical for real-world uses. This talk provides an overview of practical challenges in developing programs on encrypted data. It highlights the importance of Compilers with a focus on Intermediate representations (IR) at varying levels of abstraction to accelerate FHE research while taking advantage of decades of Compilers research. The talk also provides a broad perspective on the path to deploying FHE as a privacy enhancing technology, its practical challenges, and some open problems.