Gaussian elimination is among the oldest algorithms in mathematics and remains a standard method for solving linear systems. It can also be used to construct low-rank approximations of matrices and separable approximations of functions. In this talk, we discuss recent results on the quality of these approximations under greedy and randomized pivoting.
About the speaker:
Marc Aurèle Gilles is an Assistant Professor of Mathematics at Princeton University. His research develops theory, algorithms, and software for scientific computing, with a focus on numerical linear algebra and statistical inference from noisy cryo-EM data.
Date
Location
Amos Eaton 216
Speaker:
Marc Aurèle Gilles
from Princeton University
