IP Protection and Privacy in LLM: Leveraging Fully Homomorphic Encryption

Presenters:
Benoit Chevallier-Mames, VP of Cloud and Machine Learning, Zama
Jordan Frery, Research Scientist, Zama
Large Language Models (LLMs) are increasingly utilized in various applications. However, there's a dilemma between safeguarding the model owner's assets and ensuring the user's data privacy. This session introduces a hybrid method that employs Fully Homomorphic Encryption to address both these concerns. Presenting a live demonstration of the approach, highlighting its practicality and efficiency.
www.rsaconference.com/usa