Matthew Jagielski

Google DeepMind
Email: (my last name)@google.com
Github
Google Scholar
me.jpg
News
[Dec 2023] Our paper Privacy Auditing in One (1) Training Run received an outstanding paper award at NeurIPS 2023!

[June - Sept 2023] I enjoyed hosting Karan Chadha as a student researcher, together with Nicolas Papernot! Stay tuned for his work, and hire him - he's on the job market!

[Aug 2023] Our paper Tight Auditing of Differentially Private Machine Learning won a best paper award at USENIX Security 2023!

[July 2023] Our paper "Extracting Training Data from Large Language Models" won runner up for the Caspar Bowden award at PETS 2023!

[June 2023] Lishan Yang and I cochaired the DSML 2023 workshop, colocated with DSN 2023 in Porto, Portugal! Thank you to everyone involved, especially our attendees, keynote speakers (Paolo Rech and Andrew Paverd) and our steering committee!

About Me
I am a research scientist at Google DeepMind, working on Andreas Terzis's team. I work on security, privacy, and memorization in machine learning systems. This includes directions like privacy auditing, memorization in generative models, data poisoning, and model stealing.

I received my PhD from Northeastern University, where I was fortunate to be advised by Alina Oprea and Cristina Nita-Rotaru, as a member of the Network and Distributed Systems Security Lab (NDS2).

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
  • Measuring Forgetting of Memorized Training Examples
    Matthew Jagielski, Om Thakkar, Florian Tramèr, Daphne Ippolito, Katherine Lee, Nicholas Carlini, Eric Wallace, Shuang Song, Abhradeep Thakurta, Nicolas Papernot, Chiyuan Zhang
    ICLR 2023
    [Paper]
  • Extracting Training Data from Large Language Models
    Nicholas Carlini, Florian Tramer, Eric Wallace, Matthew Jagielski, Ariel Herbert-Voss, Katherine Lee, Adam Roberts, Tom Brown, Dawn Song, Ulfar Erlingsson, Alina Oprea, Colin Raffel
    USENIX Security 2021
    [Paper]
  • 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]
  • High-Fidelity Extraction of Neural Network Models
    Matthew Jagielski, Nicholas Carlini, David Berthelot, Alex Kurakin, and Nicolas Papernot
    USENIX Security 2020
    [Paper] [Blog] [Talk]
  • 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]