Do you know that you can factorize a matrix into a product of its submatrices? How can we use this to build algorithms for data analysis? Can we build algorithms to denoise or complete matrices by only viewing small parts of them? Of course! This will be demonstrated by Dr Keaton Hamm from the Univeristy in Texas at Arlington, whom NASK SCIENCE will host in Warsaw as part of its Do (N)ASK series.
Keaton will overview some matrix factorizations that allow one to observe only small randomly chosen submatrices of a data matrix, and how these factorizations can be applied in the design of fast algorithms for certain tasks such as Robust PCA or matrix completion. He will show how to obtain state-of-the-art runtime for these tasks and apply the algorithms to some image and video processing tasks and discuss some natural generalizations of this approach to tensor data.
The lecture may be of interest to those involved in computational linear and multilinear algebra; those involved in computer vision - clustering, motion segmentation, foreground/background separation or face modelling and imaging. Manifold learning is one facet of unsupervised learning in which one tries to find a low-dimensional representation of a data manifold. The manifold hypothesis is heavily utilized in ML applications (often implicitly).
Join the lecture: 1 March, Wednesday, at 10.30am CET on the NASK YouTube channel:
Keaton Hamm has been an Assistant Professor of Mathematics at the @UTArlington since 2020. He previously held postdoc positions at Vanderbilt University and the University of Arizona, and obtained his Ph.D. from Texas A&M University in 2015. His current research interests are in computational and randomized linear (and multilinear) algebra with applications to computer vision and image processing, and nonlinear dimensionality reduction and optimal transport with applications to imaging. His research has been funded by the National Science Foundation, the Army Research Office, and the U.S. Department of Agriculture.
The NASK SCIENCE series of meetings is an educational initiative of the Institute, which aims to openly share interesting knowledge with a wider audience. The meetings will be hosted by experts from Poland and abroad with extensive scientific portfolios and inspiring research results and a readiness to critically evaluate and discuss them. These events are held in Polish or English, as lectures or Do (N)ASK seminars.