Privacy Preserving ML

Summary

Pedagogic activities for companies wanting to adopt Privacy Preserving Machine Learning (PPML).

The story

Privacy is increasingly becoming a concern with vast amounts of data and advanced ML/AI algorithms. There are several ways in which ML can help preserve privacy. In this project at CeADAR, we created a series of short videos covering the different ways like anonymization, synthetic data generation and differential privacy which can be used to make data more private. We also created a short course on PPML accessible to CeADAR’s members. Below we list the pedagogic videos we created during the project.

  • Technical discussion with a few industry experts working on PPML
  • Overview of PPML

  • Some use cases for PPML

  • Synthetic Data Generation for privacy protection

  • Differential Privacy: Part 1

  • Does anonymization alone ensure privacy?

  • Using Synthetic Data Vault library

  • Differential Privacy: Part 2