A block coordinate descent method for regularized multiconvex optimization with applications to nonnegative tensor factorization and completion

Yangyang Xu, Wotao Yin

Research output: Contribution to journalReview articlepeer-review

843 Scopus citations

Abstract

This paper considers regularized block multiconvex optimization, where the feasible set and objective function are generally nonconvex but convex in each block of variables. It also accepts nonconvex blocks and requires these blocks to be updated by proximal minimization. We review some interesting applications and propose a generalized block coordinate descent method. Under certain conditions, we show that any limit point satisfies the Nash equilibrium conditions. Furthermore, we establish global convergence and estimate the asymptotic convergence rate of the method by assuming a property based on the Kurdyka-Łojasiewicz inequality. The proposed algorithms are tested on nonnegative matrix and tensor factorization, as well as matrix and tensor recovery from incomplete observations. The tests include synthetic data and hyperspectral data, as well as image sets from the CBCL and ORL databases. Compared to the existing state-of-the-art algorithms, the proposed algorithms demonstrate superior performance in both speed and solution quality. The MATLAB code of nonnegative matrix/tensor decomposition and completion, along with a few demos, are accessible from the authors' homepages.

Original languageEnglish (US)
Pages (from-to)1758-1789
Number of pages32
JournalSIAM Journal on Imaging Sciences
Volume6
Issue number3
DOIs
StatePublished - 2013
Externally publishedYes

Keywords

  • Block coordinate descent
  • Block multiconvex
  • Kurdyka-Łojasiewicz inequality
  • Matrix completion
  • Nash equilibrium
  • Nonnegative matrix and tensor factorization
  • Proximal gradient method
  • Tensor completion

Fingerprint

Dive into the research topics of 'A block coordinate descent method for regularized multiconvex optimization with applications to nonnegative tensor factorization and completion'. Together they form a unique fingerprint.

Cite this