White Paper

Hardware Acceleration of Fully Homomorphic Encryption: Making Privacy Preserving Machine Learning Practical

In this paper, we discuss the challenge of protecting sensitive data while allowing organizations to collaborate. FHE is a potential solution, enabling arbitrary computations over encrypted data without using the secret key. We particularly focus on the important and rapidly expanding field of machine learning (ML). Bootstrapping is almost always required in FHE for ML. An extremely compute-intensive procedure with a high memory footprint, bootstrapping is the performance chokepoint on regular computers. 

Download the white paper to learn more about making privacy-preserving machine learning practical.

Fill out the form to download now