Talk by Marco Donatelli, University of Insubria, Italy.
Regularization techniques for image reconstruction problems typically involve a smooth term and a potentially non-smooth convex term. A common approach to solving these problems is the use of proximal gradient methods. To accelerate the convergence of these first-order iterative algorithms, extrapolation strategies have been introduced in the literature. We propose a variable metric strategy that can be reinterpreted as a right-preconditioning method. Consequently, we also investigate a left-preconditioned version of the same proximal gradient method, together with possible approximation techniques. We prove the convergence of the preconditioned iterations under a suitable choice of the sequence of preconditioners. Numerical results show that preconditioning accelerates convergence while reducing the overall CPU time.
Sign-up below to receive the Zoom link.
Professor Marco Donatelli is a mathematician whose research focuses on numerical analysis and inverse problems. His main interests include iterative methods, multigrid techniques, preconditioning, and the regularization of ill-posed problems. He develops efficient numerical methods for large-scale computational problems arising from differential equations and image processing. He has authored more than 120 scientific publications and is actively involved in international research collaborations.
He is currently Full Professor of Numerical Analysis at the University of Insubria, Italy. He obtained his degree in Computer Science from the University of Florence in 2002 and earned his PhD in Mathematics and Statistics for Computational Sciences from the University of Milan in 2006. He joined the University of Insubria in 2005 as an Assistant Professor and became Associate Professor in 2014 before being appointed Full Professor in 2023.
This event is organized by Prof. Dr Matthias Voigt from the Faculty of Mathematics and Computer Science.
