Multi-subspace power method for decomposing partially symmetric tensors

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wang, Kexin, Pereira, João M., Kileel, Joe, Seigal, Anna
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910241725087744
author Wang, Kexin
Pereira, João M.
Kileel, Joe
Seigal, Anna
author_facet Wang, Kexin
Pereira, João M.
Kileel, Joe
Seigal, Anna
contents We present an algorithm for low rank decomposition of tensors of any symmetry type, from fully asymmetric to fully symmetric. It recovers the decomposition one summand at a time via the higher-order power method. This approach is known to fail in general: there need not be a relationship between the summands of a decomposition and the (partially symmetric) singular vector tuples (pSVTs) of the tensor. Our approach overcomes this problem by transforming the input to a tensor with orthonormal slices, via orthogonalization of a flattening. The summands of the decomposition of the original tensor can be recovered from the pSVTs of this new transformed tensor. We introduce a shifted power method for computing pSVTs and prove its global convergence. Numerical experiments demonstrate that our algorithm achieves higher accuracy and faster runtime than existing methods.
format Preprint
id arxiv_https___arxiv_org_abs_2510_18627
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Multi-subspace power method for decomposing partially symmetric tensors
Wang, Kexin
Pereira, João M.
Kileel, Joe
Seigal, Anna
Numerical Analysis
We present an algorithm for low rank decomposition of tensors of any symmetry type, from fully asymmetric to fully symmetric. It recovers the decomposition one summand at a time via the higher-order power method. This approach is known to fail in general: there need not be a relationship between the summands of a decomposition and the (partially symmetric) singular vector tuples (pSVTs) of the tensor. Our approach overcomes this problem by transforming the input to a tensor with orthonormal slices, via orthogonalization of a flattening. The summands of the decomposition of the original tensor can be recovered from the pSVTs of this new transformed tensor. We introduce a shifted power method for computing pSVTs and prove its global convergence. Numerical experiments demonstrate that our algorithm achieves higher accuracy and faster runtime than existing methods.
title Multi-subspace power method for decomposing partially symmetric tensors
topic Numerical Analysis
url https://arxiv.org/abs/2510.18627