Project:
Host Institution: UniGRAZ
Supervisor: Kristian Bredies
Workpackages to which it is related: WP1, WP4
Objectives: Establishment of convergence properties for non-stationary preconditioners for specific first-order splitting methods. Application to CG iteration and Douglas-Rachford-type algorithms. Development of multiscale optimization formulations within a preconditioned splitting framework. Application to large-scale variational image reconstruction problems in MRI.
Biography
I was born in Genoa in 1994 and attended scientific high school Liceo G. D. Cassini, where I joined the mathematics school team and took part in maths competitions. I have always liked scientific subjects; this passion led me to enrol in the Mathematics faculty in Genoa, where I completed my bachelor's. I then decided to have a one-year period abroad, since the Erasmus program offered me the opportunity to study at Marie Curie university in Paris. After that, I completed my master's thesis about a theoretical result on the learning process of neural networks, obtained using techniques from optimal transport theory.
Motivation: I probably started enjoying mathematics as a kid because it was the first thing I could understand clearly. It certainly gets harder later on, but more rewarding all at once. During my studies, I first turned to pure mathematics, but then I moved towards more applied subjects. My position allows me to fruitfully join the research in a thriving field that I believe is going to be one of the trademarks of our time.
Hobbies: I have always practiced sports, especially football and horse-riding - but I am always up for new ones. My position will also allow me to dive into a different culture, which is a thing I've always enjoyed. I try to be open and meet people, and this is one of the reasons I was particularly happy to apply and be accepted in the TraDE-OPT network.