Nonnegative tensor completion via low-rank Tucker decomposition: model and algorithm

Bilian Chen, Ting Sun, Zhehao Zhou, Yifeng Zeng, Langcai Cao

Research output: Contribution to journalArticle

86 Downloads (Pure)

Abstract

We consider the problem of low-rank tensor decomposition of incomplete tensors that has applications in many data analysis problems, such as recommender systems, signal processing, machine learning, and image inpainting. In this paper, we focus on nonnegative tensor completion via low-rank Tucker decomposition for dealing with it. The specialty of our model is that the ranks of nonnegative Tucker decomposition are no longer constants, while they all become a part of the decisions to be optimized. Our solving approach is based on the penalty method and blocks coordinate descent method with prox-linear updates for regularized multiconvex optimization. We demonstrate the convergence of our algorithm. The numerical results on the three image datasets show that the proposed algorithm offers competitive performance compared with other existing algorithms even though the data is highly sparse.

Original languageEnglish
Pages (from-to)95903-95914
JournalIEEE Access
Volume7
DOIs
Publication statusPublished - 16 Jul 2019

Fingerprint

Tensors
Decomposition
Naproxen
Recommender systems
Learning systems
Signal processing

Cite this

Chen, Bilian ; Sun, Ting ; Zhou, Zhehao ; Zeng, Yifeng ; Cao, Langcai. / Nonnegative tensor completion via low-rank Tucker decomposition: model and algorithm. In: IEEE Access. 2019 ; Vol. 7. pp. 95903-95914.
@article{aaa9ce55f80c483db3ea66ef0ad6f08d,
title = "Nonnegative tensor completion via low-rank Tucker decomposition: model and algorithm",
abstract = "We consider the problem of low-rank tensor decomposition of incomplete tensors that has applications in many data analysis problems, such as recommender systems, signal processing, machine learning, and image inpainting. In this paper, we focus on nonnegative tensor completion via low-rank Tucker decomposition for dealing with it. The specialty of our model is that the ranks of nonnegative Tucker decomposition are no longer constants, while they all become a part of the decisions to be optimized. Our solving approach is based on the penalty method and blocks coordinate descent method with prox-linear updates for regularized multiconvex optimization. We demonstrate the convergence of our algorithm. The numerical results on the three image datasets show that the proposed algorithm offers competitive performance compared with other existing algorithms even though the data is highly sparse.",
author = "Bilian Chen and Ting Sun and Zhehao Zhou and Yifeng Zeng and Langcai Cao",
year = "2019",
month = "7",
day = "16",
doi = "10.1109/ACCESS.2019.2929189",
language = "English",
volume = "7",
pages = "95903--95914",
journal = "IEEE Access",
issn = "2169-3536",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

Nonnegative tensor completion via low-rank Tucker decomposition: model and algorithm. / Chen, Bilian; Sun, Ting; Zhou, Zhehao; Zeng, Yifeng; Cao, Langcai.

In: IEEE Access, Vol. 7, 16.07.2019, p. 95903-95914.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Nonnegative tensor completion via low-rank Tucker decomposition: model and algorithm

AU - Chen, Bilian

AU - Sun, Ting

AU - Zhou, Zhehao

AU - Zeng, Yifeng

AU - Cao, Langcai

PY - 2019/7/16

Y1 - 2019/7/16

N2 - We consider the problem of low-rank tensor decomposition of incomplete tensors that has applications in many data analysis problems, such as recommender systems, signal processing, machine learning, and image inpainting. In this paper, we focus on nonnegative tensor completion via low-rank Tucker decomposition for dealing with it. The specialty of our model is that the ranks of nonnegative Tucker decomposition are no longer constants, while they all become a part of the decisions to be optimized. Our solving approach is based on the penalty method and blocks coordinate descent method with prox-linear updates for regularized multiconvex optimization. We demonstrate the convergence of our algorithm. The numerical results on the three image datasets show that the proposed algorithm offers competitive performance compared with other existing algorithms even though the data is highly sparse.

AB - We consider the problem of low-rank tensor decomposition of incomplete tensors that has applications in many data analysis problems, such as recommender systems, signal processing, machine learning, and image inpainting. In this paper, we focus on nonnegative tensor completion via low-rank Tucker decomposition for dealing with it. The specialty of our model is that the ranks of nonnegative Tucker decomposition are no longer constants, while they all become a part of the decisions to be optimized. Our solving approach is based on the penalty method and blocks coordinate descent method with prox-linear updates for regularized multiconvex optimization. We demonstrate the convergence of our algorithm. The numerical results on the three image datasets show that the proposed algorithm offers competitive performance compared with other existing algorithms even though the data is highly sparse.

UR - http://www.scopus.com/inward/record.url?scp=85070276622&partnerID=8YFLogxK

U2 - 10.1109/ACCESS.2019.2929189

DO - 10.1109/ACCESS.2019.2929189

M3 - Article

VL - 7

SP - 95903

EP - 95914

JO - IEEE Access

JF - IEEE Access

SN - 2169-3536

ER -