Thesis title: Exploiting the partially-separable structure in quasi-Newton methods for unconstrained optimization and deep learning
Presented by: Paul Raynaud
Program: Mathematics
Department: Dép. de mathématiques et génie industriel
Jury
President: Sébastien Le Digabel, Ph. D.
Research Director: Dominique Orban, Doctorat
Research Co-Director: Jean Bigeon, Doctorat
Member: Pierre Lemaire, Doctorat
External members: Margherita Porcelli, Ph. D. & Sébastien Bourguignon, Doctorat
Representative of the Director of Graduate Studies: Daniel Aloise, Ph. D.
Free entry
Welcome to all!