Privacy in AI
Kristin discusses the groundbreaking work on cryptonets, which demonstrated the feasibility of evaluating deep neural networks on homomorphically encrypted data. She highlights the challenges posed by common activation functions not being in polynomial form, and reflects on how this initial research laid the foundation for ongoing advancements in private AI. The conversation reveals the evolution of this field and its significance in the broader context of secure machine learning.In this clip
From this podcast
The Gradient
Kristin Lauter: Private AI, Homomorphic Encryption, and AI for Cryptography
Related Questions