Convex Coupled Matrix and tensor completion

Tutkimustuotos: Lehtiartikkeli

Standard

Convex Coupled Matrix and tensor completion. / Wimalawarne, Kishan; Yamada, Makoto; Mamitsuka, Hiroshi.

julkaisussa: Neural Computation, Vuosikerta 30, Nro 11, 01.11.2018, s. 3095-3127.

Tutkimustuotos: Lehtiartikkeli

Harvard

Wimalawarne, K, Yamada, M & Mamitsuka, H 2018, 'Convex Coupled Matrix and tensor completion', Neural Computation, Vuosikerta. 30, Nro 11, Sivut 3095-3127. https://doi.org/10.1162/neco_a_01123

APA

Vancouver

Author

Wimalawarne, Kishan ; Yamada, Makoto ; Mamitsuka, Hiroshi. / Convex Coupled Matrix and tensor completion. Julkaisussa: Neural Computation. 2018 ; Vuosikerta 30, Nro 11. Sivut 3095-3127.

Bibtex - Lataa

@article{bca3ccfad65446e58d9d1a1df1a843ee,
title = "Convex Coupled Matrix and tensor completion",
abstract = "We propose a set of convex low-rank inducing norms for coupled matrices and tensors (hereafter referred to as coupled tensors), in which information is shared between thematrices and tensors through commonmodes. More specifically,we first propose a mixture of the overlapped trace norm and the latent normswith thematrix trace norm, and then, propose a completion model regularized using these norms to impute coupled tensors. A key advantage of the proposed norms is that they are convex and can be used to find a globally optimal solution, whereas existingmethods for coupled learning are nonconvex.We also analyze the excess risk bounds of the completionmodel regularized using our proposed norms and show that they can exploit the low-rankness of coupled tensors, leading to better bounds compared to those obtained using uncoupled norms. Through synthetic and real-data experiments, we show that the proposed completion model compares favorably with existing ones.",
author = "Kishan Wimalawarne and Makoto Yamada and Hiroshi Mamitsuka",
year = "2018",
month = "11",
day = "1",
doi = "10.1162/neco_a_01123",
language = "English",
volume = "30",
pages = "3095--3127",
journal = "Neural Computation",
issn = "0899-7667",
number = "11",

}

RIS - Lataa

TY - JOUR

T1 - Convex Coupled Matrix and tensor completion

AU - Wimalawarne, Kishan

AU - Yamada, Makoto

AU - Mamitsuka, Hiroshi

PY - 2018/11/1

Y1 - 2018/11/1

N2 - We propose a set of convex low-rank inducing norms for coupled matrices and tensors (hereafter referred to as coupled tensors), in which information is shared between thematrices and tensors through commonmodes. More specifically,we first propose a mixture of the overlapped trace norm and the latent normswith thematrix trace norm, and then, propose a completion model regularized using these norms to impute coupled tensors. A key advantage of the proposed norms is that they are convex and can be used to find a globally optimal solution, whereas existingmethods for coupled learning are nonconvex.We also analyze the excess risk bounds of the completionmodel regularized using our proposed norms and show that they can exploit the low-rankness of coupled tensors, leading to better bounds compared to those obtained using uncoupled norms. Through synthetic and real-data experiments, we show that the proposed completion model compares favorably with existing ones.

AB - We propose a set of convex low-rank inducing norms for coupled matrices and tensors (hereafter referred to as coupled tensors), in which information is shared between thematrices and tensors through commonmodes. More specifically,we first propose a mixture of the overlapped trace norm and the latent normswith thematrix trace norm, and then, propose a completion model regularized using these norms to impute coupled tensors. A key advantage of the proposed norms is that they are convex and can be used to find a globally optimal solution, whereas existingmethods for coupled learning are nonconvex.We also analyze the excess risk bounds of the completionmodel regularized using our proposed norms and show that they can exploit the low-rankness of coupled tensors, leading to better bounds compared to those obtained using uncoupled norms. Through synthetic and real-data experiments, we show that the proposed completion model compares favorably with existing ones.

UR - http://www.scopus.com/inward/record.url?scp=85055662568&partnerID=8YFLogxK

U2 - 10.1162/neco_a_01123

DO - 10.1162/neco_a_01123

M3 - Letter

VL - 30

SP - 3095

EP - 3127

JO - Neural Computation

JF - Neural Computation

SN - 0899-7667

IS - 11

ER -

ID: 29456660