(There will be Live Q&A at end of each talk on Zoom)

Practical applications of ML via cloud-based or machine-learning-as-a-service platforms pose a range of security and privacy challenges. There are a number of technical approaches being studied including: homomorphic encryption, secure multi-party computation, federated learning, on-device computation, and differential privacy. This tutorial will dive into some of the important areas that are shaping the future of how we interpret our models and build AI with security and privacy in mind. We will cover the major challenges and walk through some solutions. The material will be presented in the following talks:

• Introduction to Privacy Preserving Machine Learning - Geeta Chauhan
• Secure Computation using CrypTen (https://crypten.ai/); - Laurens van der Maaten
• Training models differentially private at scale using Opacus (https://ai.facebook.com/blog/introducing-opacus-a-high-speed-library-for-training-pytorch-models-with-differential-privacy/); - Davide Testuggine
• Training models across multiple organizations privately with federated learning and PySyft from OpenMined (https://www.openmined.org/) - Andrew Trask