Computing Low-Rank Approximations of Large-Scale Matrices with the Tensor Network Randomized SVD
Name
17m1140480.pdf
Description
Published version
Size
613.34 KB
Format
Adobe PDF
Checksum (MD5)
d8f7a09323393cf18143875c288ef4ff
Author(s)
Batselier, Kim
Yu, Wenjian
Daniel, Luca
Wong, Ngai
Date Issued
January 2018
Journal
SIAM Journal on Matrix Analysis and Applications
Publisher
Society for Industrial & Applied Mathematics (SIAM)
Citation
Batselier, Kim, et al. “Computing Low-Rank Approximations of Large-Scale Matrices with the Tensor Network Randomized SVD.” SIAM Journal on Matrix Analysis and Applications 39, no. 3 (January 2018): 1221–44. © 2018 Society for Industrial and Applied Mathematics.
Version
Final published version
Abstract
We propose a new algorithm for the computation of a singular value decomposition (SVD) low-rank approximation of a matrix in the matrix product operator (MPO) format, also called the tensor train matrix format. Our tensor network randomized SVD (TNrSVD) algorithm is an MPO implementation of the randomized SVD algorithm that is able to compute dominant singular values and their corresponding singular vectors. In contrast to the state-of-the-art tensor-based alternating least squares SVD (ALS-SVD) and modified alternating least squares SVD (MALS-SVD) matrix approximation methods, TNrSVD can be up to 13 times faster while achieving better accuracy. In addition, our TNrSVD algorithm also produces accurate approximations in particular cases where both ALS-SVD and MALS-SVD fail to converge. We also propose a new algorithm for the fast conversion of a sparse matrix into its corresponding MPO form, which is up to 509 times faster than the standard tensor train SVD method while achieving machine precision accuracy. The efficiency and accuracy of both algorithms are demonstrated in numerical experiments. Key words: curse of dimensionality, low-rank tensor approximation, matrix factorization, matrix product operator, singular value decompositon (SVD), tensor network, tensor train (TT) decomposition, randomized algorithm
MIT Department
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
Terms of Use
Article is made available in accordance with the publisher's policy and may be subject to US copyright law. Please refer to the publisher's site for terms of use.
Persistent DSpace Link
DOI of Published Version
https://doi.org/10.1137/17m1140480
