A Two-Stage Optimization Algorithm for Tensor Decomposition

Hongchao Zhang, Zequn Zheng, Hongchao Zhang, Zequn Zheng · 2025

Read A Two-Stage Optimization Algorithm for Tensor Decomposition by Hongchao Zhang, Zequn Zheng, Hongchao Zhang, Zequn Zheng (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-00313.txt.

Description

Zequn Zheng, Hongchao Zhang, and Guangming Zhou decomposed hard tensors with generating math rules in a 2025 paper. Splitting a multi-way data array into a sum of simple rank-one pieces underlies squeezing and signal separation. The middle ranks, between the largest and second-largest dimensions, defeat standard iterative methods, and flattening tricks inherit the same hard case. Symbol-based direct methods exist, yet they demand uniqueness and grow expensive at high rank.

Their two-stage method first solves a generating math rule system, then refines by tuning. Handling even non-unique decompositions as long as rank stays within the largest dimension. The flattening-first workflow common in practice is exactly the regime targeted.

Catalog record

CREATORHongchao Zhang, Zequn Zheng, Hongchao Zhang, Zequn Zheng
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE64 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-00313

Browse the full library · About the Tower of Babel