Partitioned Alternating Least Squares Technique for Canonical Polyadic Tensor Decomposition

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F67985556%3A_____%2F16%3A00460710" target="_blank" >RIV/67985556:_____/16:00460710 - isvavai.cz</a>

Výsledek na webu

<a href="http://dx.doi.org/10.1109/LSP.2016.2577383" target="_blank" >http://dx.doi.org/10.1109/LSP.2016.2577383</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1109/LSP.2016.2577383" target="_blank" >10.1109/LSP.2016.2577383</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Partitioned Alternating Least Squares Technique for Canonical Polyadic Tensor Decomposition

Popis výsledku v původním jazyce

Canonical polyadic decomposition (CPD), also known as parallel factor analysis, is a representation of a given tensor as a sum of rank-one components. Traditional method for accomplishing CPD is the alternating least squares (ALS) algorithm. Convergence of ALS is known to be slow, especially when some factor matrices of the tensor contain nearly collinear columns. We propose a novel variant of this technique, in which the factor matrices are partitioned into blocks, and each iteration jointly updates blocks of different factor matrices. Each partial optimization is quadratic and can be done in closed form. The algorithm alternates between different random partitionings of the matrices. As a result, a faster convergence is achieved. Another improvement can be obtained when the method is combined with the enhanced line search of Rajih et al. Complexity per iteration is between those of the ALS and the Levenberg–Marquardt (damped Gauss–Newton) method.

Název v anglickém jazyce

Partitioned Alternating Least Squares Technique for Canonical Polyadic Tensor Decomposition

Popis výsledku anglicky

Canonical polyadic decomposition (CPD), also known as parallel factor analysis, is a representation of a given tensor as a sum of rank-one components. Traditional method for accomplishing CPD is the alternating least squares (ALS) algorithm. Convergence of ALS is known to be slow, especially when some factor matrices of the tensor contain nearly collinear columns. We propose a novel variant of this technique, in which the factor matrices are partitioned into blocks, and each iteration jointly updates blocks of different factor matrices. Each partial optimization is quadratic and can be done in closed form. The algorithm alternates between different random partitionings of the matrices. As a result, a faster convergence is achieved. Another improvement can be obtained when the method is combined with the enhanced line search of Rajih et al. Complexity per iteration is between those of the ALS and the Levenberg–Marquardt (damped Gauss–Newton) method.

Druh

J_x - Nezařazeno - Článek v odborném periodiku (Jimp, Jsc a Jost)

CEP obor

BB - Aplikovaná statistika, operační výzkum

OECD FORD obor

—

Projekt

<a href="/cs/project/GA14-13713S" target="_blank" >GA14-13713S: Metody dekompozice tenzorů a jejich aplikace</a><br>

Návaznosti

I - Institucionalni podpora na dlouhodoby koncepcni rozvoj vyzkumne organizace

Rok uplatnění

2016

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název periodika

IEEE Signal Processing Letters

ISSN

1070-9908

e-ISSN

—

Svazek periodika

23

Číslo periodika v rámci svazku

7

Stát vydavatele periodika

US - Spojené státy americké

Počet stran výsledku

5

Strana od-do

993-997

Kód UT WoS článku

000379694800005

EID výsledku v databázi Scopus

2-s2.0-84978100769