The SCD semismooth* Newton method for the efficient minimization of Tikhonov functionals

Simon Hubmer (JKU, Linz)

Jun 11. 2026, 11:20 — 11:50

We consider the efficient numerical minimization of Tikhonov functionals with nonlinear operators and non-smooth and non-convex penalty terms, which appear e.g. in variational regularization. For this, we consider a new class of SCD semismooth* Newton methods, which are based on a novel concept of graphical derivatives, and exhibit locally superlinear convergence. We present a detailed description of these methods, and provide explicit algorithms in the case of sparsity ($\ell_p$, $0\leq p < \infty$) and TV penalty terms. The numerical performance of these methods is then illustrated on a number of tomographic imaging problems.

Further Information
Venue:
ESI Boltzmann Lecture Hall
Associated Event:
Applications of Tomographic Methods (Workshop)
Organizer(s):
Wolfgang Drexler (Med U Vienna)
Peter Elbau (U of Vienna)
Ronny Ramlau (RICAM, Linz)
Monika Ritsch-Marte (Med Uni Innsbruck)
Otmar Scherzer (U of Vienna)
Gerhard Schütz (TU Wien)
Gabriele Steidl (TU Berlin)
Glenn van de Ven (U of Vienna)