Printable PDF
Department of Mathematics,
University of California San Diego


Math 278B - Mathematics of Information, Data, and Signals

Prof. Keaton Hamm

University of Texas at Arlington

Tensor decompositions by mode subsampling


We will overview variants of CUR decompositions for tensors. These are low-rank tensor approximations in which the constituent tensors or factor matrices are subtensors of the original data tensors. We will discuss variants of Tucker decompositions and those based on t-products in this framework. Characterizations of exact decompositions are given, and approximation bounds are shown for data tensors contaminated with Gaussian noise via perturbation arguments.  Experiments are shown for image compression and time permitting we will discuss applications to robust PCA.

Host: Alex Cloninger

April 18, 2024

11:30 AM

APM 2402