Matthew Jagielski

About Me

I am a fifth year PhD student advised by Alina Oprea and Cristina Nita-Rotaru, working as a member of the Network and Distributed Systems Security Lab (NDS2).

My research is broadly at the intersection between machine learning, security, and privacy. The goal of my research is to design training and deployment of machine learning that is secure from real world adversaries. I also study the design of machine learning algorithms that preserve the privacy of individuals in the training set. I rely on techniques drawn from machine learning, theoretical computer science, and security.

During summer '19, I was at Google Brain Privacy and Security, working with Nicolas Papernot on model extraction attacks. In summer '18, I worked at DoS and Abuse at Google, using machine learning to protect Google Cloud customers from DoS attacks.

In other news, I enjoy running, swimming, and biking. I'm also a retired Super Smash Brothers tournament competitor.

Selected Publications - see Google Scholar for full list

• Auditing Differentially Private Machine Learning - How Private is Private SGD?
  Matthew Jagielski, Jonathan Ullman, Alina Oprea
  NeurIPS 2020, TPDP 2020 Contributed Talk
  [Paper] [Code] [Poster] [3min talk]
• Subpopulation Data Poisoning Attacks
  Full - Matthew Jagielski, Giorgio Severi, Niklas Pousette Harger, Alina Oprea
  Full - Preprint
  Preliminary - Matthew Jagielski, Paul Hand, Alina Oprea
  Preliminary - NeurIPS 2019 Workshop on Robust AI in Financial Services
  [Full Paper][Preliminary Paper]
• High-Fidelity Extraction of Neural Network Models
  Matthew Jagielski, Nicholas Carlini, David Berthelot, Alex Kurakin, and Nicolas Papernot
  USENIX Security 2020
  [Paper] [Blog] [Talk]
• Threat Detection for Collaborative Adaptive Cruise Control in Connected Cars
  Matthew Jagielski, Nicholas Jones, Chung-Wei Lin, Cristina Nita-Rotaru, and Shinichi Shiraishi
  ACM WiSec 2018
  [Paper]
• Manipulating Machine Learning: Poisoning Attacks and Countermeasures for Regression Learning
  Matthew Jagielski, Alina Oprea, Chang Liu, Cristina Nita-Rotaru, and Bo Li
  IEEE S&P (Oakland) 2018
  [Code] [Paper] [Talk]