Efficient Algorithms for Low-Rank Matrix Optimization with Applications to Matrix Factorization

Abstract

Matrix factorization is a fundamental computational problem. The first goal of this proposal is to develop a new theory of structured matrix and operator factorization. This theory leads to largescale matrix optimization problems that cannot be solved with contemporary algorithms. The second goal of this proposal is to develop a new class of storage-optimal algorithms for largescale matrix optimization. These algorithms will support the computation of new types of matrix factorizations.

Document Details

Document Type
DoD Grant Award
Publication Date
Jan 04, 2017
Source ID
N000141712146

Entities

People

  • Joel Tropp

Organizations

  • California Institute of Technology
  • Office of Naval Research
  • United States Navy

Tags

Fields of Study

  • Computer science

Readers

  • Linear Algebra
  • Operations Research