Скачать или смотреть Recent Advances in Unsupervised Learning: Fundamental Limits and Efficient Algorithms (Part III)

This lecture explores optimal recovery, and the use of algorithms based on mean-field variational approximations. Finally, we will discuss the notion of computational/statistical gaps in these problems.

Abstract: Our understanding of unsupervised learning problems has progressed rapidly over the past decade. A synergy of diverse ideas from high-dimensional statistics, computer science and statistical physics has dramatically advanced the state-of-the-art, leading to powerful new algorithms and precise characterizations of the associated information theoretic limits. In addition, there is substantial emerging evidence regarding the existence of computational/statistical gaps—parameter regimes where signal recovery is information theoretically possible but is expected to be impossible using efficient algorithms. In this tutorial series, we will introduce some core techniques driving this recent progress. We will present these techniques in the context of a few concrete examples. We will end with a discussion of ongoing efforts in this domain and review several open problems.

Speaker biography: Subhabrata Sen is an Assistant Professor of Statistics at Harvard University. Prior to Harvard, he was a Schramm postdoctoral fellow at Microsoft Research New England and MIT. He obtained his PhD from Stanford Statistics in 2017. His research lies at the intersection of applied probability, statistics and machine learning. His research interests include high-dimensional and non-parametric statistics, random graphs and inference on networks. Subhabrata has received the Probability Dissertation Award for his thesis, an AMS Simons Travel grant, and an honorable mention at the Bernoulli Society New Researcher Award (2018). He has been a long-term visitor at the Simons Institute for the Theory of Computing in Fall 2021 and Fall 2022.

To continue receiving all AI4OPT seminar announcements, please sign up to the mailing list at https://lists.isye.gatech.edu/mailman....
For past seminars click the link provided: https://www.ai4opt.org/seminars/past-....