.png){kind=icon}
Return
Randomized algorithms for computing tensor ring rounding
.png){kind=icon}
PRE
AI
DOI：10.1007/s40314-026-03911-1.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Tensor ring (TR) decomposition is an effective method for representation of higher-order tensors, with broad applications in domains ranging from machine learning to quantum physics. This work introduces some novel randomized extensions to the conventional TR rounding technique, generalizing established matrix randomization principles to the tensor setting. We first present an Orthogonalize-then-Randomize strategy that injects randomness directly into the rounding procedure. Furthermore, by adopting randomized tensor chains, we reduce the reliance on costly orthogonalization and propose a more computationally economical Randomize-then-Orthogonalize variant. Experimental results validate that both randomized algorithms outperform the standard deterministic TR rounding approach in terms of efficiency.
Keywords:
Tensor ring decomposition
Tensor ring rounding
Randomized algorithm
Journal
C
IF:
2.5
Papers:
89
Citations:
0
Organization
No organization information available
Cited Papers
.png){kind=icon}
ShareSave
ShareSaveShareSaveShareSave
View more