An Inertial Newton Algorithm for Deep Learning

Bolte, Jérôme, Castera, Camille, Pauwels, Edouard and Févotte, Cédric (2019) An Inertial Newton Algorithm for Deep Learning. TSE Working Paper, n. 19-1043, Toulouse

[img]
Preview
Text
Download (710kB) | Preview
Official URL: https://www.tse-fr.eu/sites/default/files/TSE/docu...

Abstract

We devise a learning algorithm for possibly nonsmooth Deep Neural Networks
featuring inertia and Newtonian directional intelligence only by means of a backpropagation
oracle. Our algorithm has an appealing mechanical interpretation making the role of its 2 hyper-parameters transparent. An elementary phase space lifting allows both for its implementation and its theoretical study under very general assumptions. Our algorithm shows high performances and appears to be faster than the state of art on [some genuine problems].

Item Type: Monograph (Working Paper)
Language: English
Date: October 2019
Place of Publication: Toulouse
Subjects: B- ECONOMIE ET FINANCE
Divisions: TSE-R (Toulouse)
Institution: Université Toulouse 1 Capitole
Site: UT1
Date Deposited: 17 Oct 2019 14:05
Last Modified: 21 Oct 2019 12:52
["eprint_fieldname_oai_identifier" not defined]: oai:tse-fr.eu:123630
URI: http://publications.ut-capitole.fr/id/eprint/32854

Actions (login required)

View Item View Item

Downloads

Downloads per month over past year