PPML: Privacy-preserving Machine Learning
PPML: This team will build a prototype for privacy-preserving machine learning to perform inference computations on the GPU without revealing sensitive data.
Advisors
Description
In this project, the team will work with privacy enhancing technologies to implement and deploy a privacy-preserving machine learning application. The project will involve working with tools from cryptography and machine learning to build a framework for performing inference computations on the GPU. This solution will allow a client holding a private input (e.g., medical data) to use a machine learning model owned by a server (which is proprietary), without the client input or the server’s model being revealed.
Relevant Technologies
- Machine learning
- Applied cryptography
- Graphics processing units
Prerequisite Knowledge
-
Prior experience with machine learning, cryptography, or GPUs is helpful, but not required.