Checking the Sufficiently Scattered Condition using a Global Non-Convex Optimization Software

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Gillis, Nicolas, Luce, Robert
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866929666697199616
author Gillis, Nicolas
Luce, Robert
author_facet Gillis, Nicolas
Luce, Robert
contents The sufficiently scattered condition (SSC) is a key condition in the study of identifiability of various matrix factorization problems, including nonnegative, minimum-volume, symmetric, simplex-structured, and polytopic matrix factorizations. The SSC allows one to guarantee that the computed matrix factorization is unique/identifiable, up to trivial ambiguities. However, this condition is NP-hard to check in general. In this paper, we show that it can however be checked in a reasonable amount of time in realistic scenarios, when the factorization rank is not too large. This is achieved by formulating the problem as a non-convex quadratic optimization problem over a bounded set. We use the global non-convex optimization software Gurobi, and showcase the usefulness of this code on synthetic data sets and on real-world hyperspectral images.
format Preprint
id arxiv_https___arxiv_org_abs_2402_06019
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Checking the Sufficiently Scattered Condition using a Global Non-Convex Optimization Software
Gillis, Nicolas
Luce, Robert
Machine Learning
Signal Processing
Optimization and Control
The sufficiently scattered condition (SSC) is a key condition in the study of identifiability of various matrix factorization problems, including nonnegative, minimum-volume, symmetric, simplex-structured, and polytopic matrix factorizations. The SSC allows one to guarantee that the computed matrix factorization is unique/identifiable, up to trivial ambiguities. However, this condition is NP-hard to check in general. In this paper, we show that it can however be checked in a reasonable amount of time in realistic scenarios, when the factorization rank is not too large. This is achieved by formulating the problem as a non-convex quadratic optimization problem over a bounded set. We use the global non-convex optimization software Gurobi, and showcase the usefulness of this code on synthetic data sets and on real-world hyperspectral images.
title Checking the Sufficiently Scattered Condition using a Global Non-Convex Optimization Software
topic Machine Learning
Signal Processing
Optimization and Control
url https://arxiv.org/abs/2402.06019