A research overview on when memorization helps, when it hurts, and how it can be controlled in deep learning.
I am a senior research scientist at Google DeepMind, based in Toronto, an adjunct professor in the McGill University School of Computer Science, and an associate industry member of Mila, the Quebec AI Institute. Prior to joining Google, I led the Trustworthy AI program at Element AI / ServiceNow. My research combines theoretical and empirical approaches to understanding deep learning, with a focus on generalization and network compression.
I obtained my Ph.D. in machine learning from the University of Cambridge, under the supervision of Zoubin Ghahramani. I studied Mathematics at the University of Warwick and read Part III in Mathematics at the University of Cambridge, receiving a Masters of Advanced Study (MASt) in Applied Mathematics.
See my curriculum vitae for more details.
In 2020, I was a member of the Institute for Advanced Study in Princeton, N.J., participating in the Special Year on Optimization, Statistics, and Theoretical Machine Learning. During Winter 2017, I was a long-term participant in the Foundations of Machine Learning program at the Simons Institute for the Theory of Computing at the University of Berkeley. I returned to Simons as a fellow in Summer 2019 for the Foundations of Deep Learning program. I will be returning to Simons again in 2022 for the Learning and Games program.
PhD, Machine Learning, 2019
University of Cambridge
MASt, Applied Mathematics, 2013
University of Cambridge
BSc, Mathematics, 2012
University of Warwick
A research overview on when memorization helps, when it hurts, and how it can be controlled in deep learning.