Randomized ULV decomposition for approximating low-rank matrices

Maboud F. Kaloorazi, Jie Chen

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

2 Scopus citations

Abstract

In this paper, we present a low-rank matrix approximation algorithm called Randomized Rank-k ULV (RR-ULV) decomposition. Fundamental in our work is the exploitation of the randomized sampling paradigm, which provides an efficient strategy in order to construct an approximation of a large input matrix. Our proposed RR-ULV is computationally efficient, robust, highly accurate, and can also harness modern computational platforms. We apply RR-ULV to randomly generated data and real world data: We consider reconstructing a low-rank image, and further solving the robust principal component analysis task to validate its efficacy and efficiency. Our experimental results demonstrate that RR-ULV outperforms the existing methods.

Original languageEnglish
Title of host publication2019 IEEE International Conference on Signal Processing, Communications and Computing, ICSPCC 2019
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781728117072
DOIs
StatePublished - Sep 2019
Event2019 IEEE International Conference on Signal Processing, Communications and Computing, ICSPCC 2019 - Dalian, Liaoning, China
Duration: 20 Sep 201922 Sep 2019

Publication series

Name2019 IEEE International Conference on Signal Processing, Communications and Computing, ICSPCC 2019

Conference

Conference2019 IEEE International Conference on Signal Processing, Communications and Computing, ICSPCC 2019
Country/TerritoryChina
CityDalian, Liaoning
Period20/09/1922/09/19

Keywords

  • Low-rank image recovery
  • Low-rank-plus-sparse matrix decomposition
  • Matrix factorization
  • Randomized methods

Fingerprint

Dive into the research topics of 'Randomized ULV decomposition for approximating low-rank matrices'. Together they form a unique fingerprint.

Cite this