Security and Privacy for Software Systems and Neural Network
COM1 Level 2
SR1, COM1-02-06
This is a distinguished talk as part of the NUS Computer Science Research Week 2019 (http://researchweek.comp.nus.edu.sg).
Abstract:
For the last several years, Google has been leading the development and real-world deployment of state-of-the-art, practical techniques for learning statistics and ML models with strong privacy guarantees for the data involved. This work has been motivated by a desire to simultaneously improve both the security and privacy guarantees of production systems. I'll give an overview of our work, and the practical techniques we've developed for training Deep Neural Networks with strong privacy guarantees. In particular, I'll cover recent results that show how local differential privacy guarantees can be strengthened by the addition of anonymity, and explain the motivation for that work. I'll also cover recent work on uncovering and measuring privacy problems due to unintended memorization in machine learning models.
The talk starts with a tutorial on the preliminaries and the theoretical foundations of this topic.
Biodata:
Ulfar Erlingsson is a Senior Staff Research Scientist in the Google Brain team, currently working primarily on privacy and security of deep learning systems. Previously, Ulfar has led computer security research at Google and been a researcher at Microsoft Research, Silicon Valley and Associate Professor at Reykjavik University. Ulfar was co-founder and CTO of the Internet security startup Green Border Technologies and Director of Privacy Protection at deCODE Genetics. Ulfar holds a PhD in computer science from Cornell University.