Randomized sketching is an established technique to speed up the solution of large-scale overdetermined least squares problems. Only recently, a rather different approach has enabled the application of this powerful technique to most other core problems in numerical computing, including linear systems of equations, linear and nonlinear eigenvalue problems, and matrix functions. I will discuss these advances and highlight key remaining challenges for future research.
Recent advances in randomized sketching
16.04.2024 10:05 - 10:50
Organiser:
Fakultät für Mathematik
Location:
SR 06