A Multilevel Approach For Nonnegative Matrix Factorization

TL;DR

This paper introduces a multilevel framework to accelerate Nonnegative Matrix Factorization algorithms, significantly improving convergence speed on image datasets by leveraging lower-dimensional representations.

Contribution

It presents a novel multilevel approach that embeds NMF algorithms into a framework to enhance their efficiency and convergence speed.

Findings

Multilevel strategy significantly speeds up NMF algorithms.

Applicable to image datasets with good low-dimensional representations.

Enhances popular NMF algorithms like ANLS, multiplicative updates, HALS.

Abstract

Nonnegative Matrix Factorization (NMF) is the problem of approximating a nonnegative matrix with the product of two low-rank nonnegative matrices and has been shown to be particularly useful in many applications, e.g., in text mining, image processing, computational biology, etc. In this paper, we explain how algorithms for NMF can be embedded into the framework of multilevel methods in order to accelerate their convergence. This technique can be applied in situations where data admit a good approximate representation in a lower dimensional space through linear transformations preserving nonnegativity. A simple multilevel strategy is described and is experimentally shown to speed up significantly three popular NMF algorithms (alternating nonnegative least squares, multiplicative updates and hierarchical alternating least squares) on several standard image datasets.