Privacy Preserving Machine Learning with Fully Homomorphic Encryption

Explore the cutting-edge intersection of data privacy and machine learning in this 57-minute Google TechTalk presented by Jordan Frery. Dive into the world of Fully Homomorphic Encryption (FHE) and its application in preserving privacy during Machine Learning operations, particularly crucial for sectors like healthcare and finance. Learn about Concrete ML, an open-source library that makes practical FHE for ML possible, and discover how it enables secure inference on encrypted data across various models. Examine the process of FHE training and its potential for using encrypted data from multiple sources without compromising individual privacy. Investigate the synergies between FHE and Federated Learning, and how their integration enhances privacy-preserving features throughout the ML pipeline. Finally, delve into the application of FHE in generative AI and the development of Hybrid FHE models, exploring solutions that balance intellectual property protection, user privacy, and computational performance in modern AI applications.

Privacy Preserving ML with Fully Homomorphic Encryption