Bregman Proximal Gradient Algorithms for Deep Matrix Factorization

Mahesh Chandra Mukkamala, Felix Westerkamp, Emanuel Laude, Daniel Cremers, Peter Ochs

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

A typical assumption for the convergence of first order optimization methods is the Lipschitz continuity of the gradient of the objective function. However, for many practical applications this assumption is violated. To overcome this issue extensions based on generalized proximity measures, known as Bregman distances, were introduced. This initiated the development of the Bregman Proximal Gradient (BPG) algorithms, which, however, rely on problem dependent Bregman distances. In this paper, we develop Bregman distances for deep matrix factorization problems, which yields a BPG algorithm with theoretical convergence guarantees, while allowing for a constant step size strategy. Moreover, we demonstrate that the algorithms based on the developed Bregman distance outperform their Euclidean counterparts as well as alternating minimization based approaches.

Original languageEnglish
Title of host publicationScale Space and Variational Methods in Computer Vision - 8th International Conference, SSVM 2021, Proceedings
EditorsAbderrahim Elmoataz, Jalal Fadili, Yvain Quéau, Julien Rabin, Loïc Simon
PublisherSpringer Science and Business Media Deutschland GmbH
Pages204-215
Number of pages12
ISBN (Print)9783030755485
DOIs
StatePublished - 2021
Event8th International Conference on Scale Space and Variational Methods in Computer Vision, SSVM 2021 - Virtual, Online
Duration: 16 May 202120 May 2021

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12679 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference8th International Conference on Scale Space and Variational Methods in Computer Vision, SSVM 2021
CityVirtual, Online
Period16/05/2120/05/21

Fingerprint

Dive into the research topics of 'Bregman Proximal Gradient Algorithms for Deep Matrix Factorization'. Together they form a unique fingerprint.

Cite this