June 2022 Summaries

The FHE.org Conference 2022, the first annual event dedicated to fully homomorphic encryption, took place in Trondheim, Norway, alongside Eurocrypt 2022. Organized and sponsored by Zama, the conference was well-received, featuring a variety of presentations on homomorphic encryption and secure computation techniques, and attracting a large audience. Zama, a strong advocate for open-source and open science, is a founding member of the FHE.org community, aiming to make data sharing private by default. At the conference, Zama showcased several projects, including advancements in FHE instantiated with NTRU and LWE, hybrid attacks on LWE, and tools like Concrete ML for machine learning over encrypted data. The event highlighted the collaborative efforts of researchers and developers in advancing the field toward more secure data sharing solutions.

Zama's blog post series explores the intricacies of the Fully Homomorphic Encryption (FHE) scheme known as TFHE, specifically focusing on its implementation in Rust through TFHE-rs. The series covers various aspects, including ciphertext types, encodings, and operations like key switching and modulus switching, culminating in an in-depth examination of programmable bootstrapping. Bootstrapping, a pivotal operation in FHE, is dissected into steps involving modulus switching, blind rotation, and sample extraction to reduce noise and enable large-scale homomorphic computations. The post highlights applications of TFHE in neural networks, demonstrating its potential for secure, homomorphic evaluation of deep learning models. Readers are encouraged to engage with the open-source TFHE-rs project, and the post concludes by expressing gratitude to contributors and inviting the community to support and participate in ongoing FHE challenges.