Non-smooth Higher-order Optimization on Manifolds
W. Diepeveen (TU Delft - Electrical Engineering, Mathematics and Computer Science)
Jan Lellmann – Mentor (University of Lübeck)
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
This thesis introduces a higher-order optimization method for solving non-smooth variational problems on Riemannian manifolds. In this work, we apply the Riemannian Semismooth Newton (RSSN) method to a non-smooth non-linear optimality system derived in recent advances in manifold duality theory. In particular we will show a novel local convergence result for an inexact version of the Riemannian Semismooth Newton method and show state-of-the-art performance in numerical experiments by solving several `2-TV-like problems on manifolds with positive and negative curvature.