Colloquium Details
Principled Approaches for Trustworthy Algorithms, Statistics, and Machine Learning
Speaker: Gautam Kamath, University of Waterloo
Location: 60 Fifth Avenue Room 150
Date: March 25, 2024, 2 p.m.
Host: Mengye Ren
Synopsis:
Despite impressive recent advances, machine learning models exhibit a number of critical deficiencies. They are prone to leaking sensitive information about their training data. They remain alarmingly brittle to attacks by malicious parties. Troublingly, these issues stem from more fundamental statistical vulnerabilities, which remain unresolved even decades later, highlighting significant gaps in our understanding of how to deal with these important considerations. As long as these problems remain, our models will not be appropriate for use beyond deployment in toy settings.
In this talk, I will discuss recent advances on a number of these problems, which give key new algorithmic insights into how to address these considerations, and enable real-world deployments that were previously thought infeasible. In a first vignette, we will explore how to guarantee individual privacy in machine learning models, with a particular focus on large language models and the important role played by public data in the training pipeline. In a second vignette, we focus on how to robustly perform mean estimation, giving the first efficient and accurate algorithms for multivariate settings. We will go on to discuss connections to robustness against data poisoning attacks, robust exploratory data analysis, and surprising conceptual and technical connections with privacy.
Speaker Bio:
Gautam Kamath is an Assistant Professor at the University of Waterloo, and a Faculty Member and Canada CIFAR AI Chair at the Vector Institute for Artificial Intelligence. His research interests are in trustworthy algorithms, statistics, and machine learning, particularly focusing on considerations like data privacy and robustness. He has a B.S. from Cornell University and a Ph.D. from MIT. He is the recipient of the 2023 Golden Jubilee Research Excellence Award, recognizing him as the most outstanding junior researcher in the University of Waterloo’s Faculty of Math. Beyond research, he is celebrated for his teaching. His course on differential privacy is the most popular resource for learning the topic, with his online lecture videos having over 100,000 views. He has also given invited tutorials on the topic in multiple different countries. He is further well known for his passion and commitment to service and improving the community. Besides organizing and chairing several workshops and conferences, he is an Editor-in-Chief of Transactions on Machine Learning Research, and on the Executive Committee of the Learning Theory Alliance.
Notes:
In-person attendance only available to those with active NYU ID cards.