Coordinate projected gradient descent minimization and its application to orthogonal nonnegative matrix factorization

Flavia Chorobura, Daniela Lupu, Ion Necoara · 2025

Read Coordinate projected gradient descent minimization and its application to orthogonal nonnegative matrix factorization by Flavia Chorobura, Daniela Lupu, Ion Necoara (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-00770.txt.

Description

Big maths problems often split into three parts: a smooth part, a curved part that will not split, and a wall that bars some values. Flavia Chorobura, Daniela Lupu, and Ion Necoara build a step by step method for such problems in a 2025 study. Their method moves one block of values at a time and picks its step size by a fresh rule.

The team proves the rule makes each step go down hill, as long as the curved part meets a few mild terms. They also bound how much work the method needs in the worst case. Tests on a matrix split task, where the parts must stay square and non minus, back the speed claims. The tool suits large scale fitting jobs.

Catalog record

CREATORFlavia Chorobura, Daniela Lupu, Ion Necoara
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE32 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-00770

Browse the full library · About the Tower of Babel