Note: low-rank tensor train completion with side information based on Riemannian optimization
Numerical Analysis
2020-06-24 v1 Numerical Analysis
Abstract
We consider the low-rank tensor train completion problem when additional side information is available in the form of subspaces that contain the mode- fiber spans. We propose an algorithm based on Riemannian optimization to solve the problem. Numerical experiments show that the proposed algorithm requires far fewer known entries to recover the tensor compared to standard tensor train completion methods.
Keywords
Cite
@article{arxiv.2006.12798,
title = {Note: low-rank tensor train completion with side information based on Riemannian optimization},
author = {Stanislav Budzinskiy and Nikolai Zamarashkin},
journal= {arXiv preprint arXiv:2006.12798},
year = {2020}
}