Matt J. Kusner
About
Appointed Canada CIFAR AI Chair – 2025
Matt J. Kusner is an associate professor at École Polytechnique de Montréal and a Core Academic Member at Mila – Quebec Artificial Intelligence Institute. His research interests are in the fundamentals of machine learning evaluation and machine learning for society. Before joining Polytechnique Montréal, he was an associate professor at University College London, and earlier an associate professor at the University of Oxford and a tutorial fellow at Jesus College.
He has given talks for the Federal Reserve Banks of Cleveland and Philadelphia, the Cambridge Centre for Mathematical Sciences, and the Royal Society. His work has appeared in the Guardian, Forbes, and the Harvard Business Review.
Awards
- Leverhulme Trust Senior Research Fellowship, The Royal Society, (2024-2025)
- Research Fellowship, The Alan Turing Institute (2016-2018)
- Turner Dissertation Award, Washington University in St. Louis (2016)
Relevant Publications
- Richter, L., He, X., Minervini, P., & Kusner, M. An Auditing Test to Detect Behavioral Shift in Language Models. (2025) The International Conference on Learning Representations.
- Dovonon, G. J.-S., Bronstein, M., & Kusner, M. (2025) Setting the Record Straight on Transformer Oversmoothing. Transactions on Machine Learning Research.
- Kaddour, J., Key, O., Nawrot, P., Minervini, P., & Kusner, M. J. (2023). No train no gain: Revisiting efficient training algorithms for transformer-based language models. Advances in Neural Information Processing Systems, 36, 25793-25818.
- Kaddour, J., Liu, L., Silva, R., & Kusner, M. J. (2022). When do flat minima optimizers work?. Advances in Neural Information Processing Systems, 35, 16577-16595.
- Kusner, M. J., Loftus, J., Russell, C., & Silva, R. (2017). Counterfactual fairness. Advances in neural information processing systems, 30.