Second-order optimization based adaptive PARAFAC decomposition of three-way tensors

A fast adaptive parallel factor (PARAFAC) decomposition algorithm is proposed for a class of third order tensors that have one dimension growing linearly with time. It is based on an alternating least squares approach in conjunction with a Newton-type optimization technique. By preserving the Khatri-Rao product and exploiting the reduced-rank update structure of the estimated subspace at each time instant, the algorithm achieves linear complexity and superior convergence performance. A modified version of the algorithm is also proposed to deal with the non-negative constraint. In addition, parallel implementation issues are investigated. Finally, the performance of the algorithm is numerically studied and compared to several state-of-the-art algorithms.

Title:	Second-order optimization based adaptive PARAFAC decomposition of three-way tensors
Authors:	Nguyen Viet-Dung Abed-Meraim, Karim Nguyen Linh-Trung
Keywords:	Fast adaptive PARAFAC;Big data;Parallel computing;Non-negative constraint
Issue Date:	2017
Publisher:	ACADEMIC PRESS INC ELSEVIER SCIENCE, 525 B ST, STE 1900, SAN DIEGO, CA 92101-4495 USA
Citation:	ISIKNOWLEDGE
Abstract:	A fast adaptive parallel factor (PARAFAC) decomposition algorithm is proposed for a class of third order tensors that have one dimension growing linearly with time. It is based on an alternating least squares approach in conjunction with a Newton-type optimization technique. By preserving the Khatri-Rao product and exploiting the reduced-rank update structure of the estimated subspace at each time instant, the algorithm achieves linear complexity and superior convergence performance. A modified version of the algorithm is also proposed to deal with the non-negative constraint. In addition, parallel implementation issues are investigated. Finally, the performance of the algorithm is numerically studied and compared to several state-of-the-art algorithms.
Description:	TNS07080 ; DIGITAL SIGNAL PROCESSING Volume: 63 Pages: 100-111 Published: APR 2017
URI:	http://repository.vnu.edu.vn/handle/VNU_123/29799
ISSN:	1051-2004
Appears in Collections:	Bài báo của ĐHQGHN trong Web of Science