Tensor train completion: local recovery guarantees via Riemannian optimization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Budzinskiy, Stanislav, Zamarashkin, Nikolai
Format: Preprint
Published: 2021
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912314431635456
author Budzinskiy, Stanislav
Zamarashkin, Nikolai
author_facet Budzinskiy, Stanislav
Zamarashkin, Nikolai
contents In this work, we estimate the number of randomly selected elements of a tensor that with high probability guarantees local convergence of Riemannian gradient descent for tensor train completion. We derive a new bound for the orthogonal projections onto the tangent spaces based on the harmonic mean of the unfoldings' singular values and introduce a notion of core coherence for tensor trains. We also extend the results to tensor train completion with auxiliary subspace information and obtain the corresponding local convergence guarantees.
format Preprint
id arxiv_https___arxiv_org_abs_2110_03975
institution arXiv
publishDate 2021
record_format arxiv
spellingShingle Tensor train completion: local recovery guarantees via Riemannian optimization
Budzinskiy, Stanislav
Zamarashkin, Nikolai
Numerical Analysis
Machine Learning
In this work, we estimate the number of randomly selected elements of a tensor that with high probability guarantees local convergence of Riemannian gradient descent for tensor train completion. We derive a new bound for the orthogonal projections onto the tangent spaces based on the harmonic mean of the unfoldings' singular values and introduce a notion of core coherence for tensor trains. We also extend the results to tensor train completion with auxiliary subspace information and obtain the corresponding local convergence guarantees.
title Tensor train completion: local recovery guarantees via Riemannian optimization
topic Numerical Analysis
Machine Learning
url https://arxiv.org/abs/2110.03975