Lars Hellström: Extreme eigenvalues and their eigenvectors

Abstract:

In my ongoing work on producing probability distributions using linear algebra, I've felt a need to produce examples to demonstrate that certain bounds are in fact tight. One approach for this leads to a study of extreme eigenvalues and eigenvectors for some very large matrices, which turn out to exhibit neat dynamics. This talk presents findings in this area obtained during the Christmas holidays. Most of the talk should be accessible also to students.