Robust Low-Tubal-Rank Tensor Sensing via Preconditioned Subgradient Descent
IEEE Signal Processing Letters · 2026 · 33 · 2725-2729
DOI: 10.1109/lsp.2026.3698753 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Signal Processing Letters, 2026. Verfasst von Ran Chen und 1 weiteren Autor:innen. Die Arbeit nennt 40 Literaturangaben.
Abstract (Originalsprache)
We study robust low-tubal-rank tensor sensing from corrupted linear measurements. We propose a preconditioned subgradient descent (PSGD) method for a factorized $\ell _{1}$ -loss formulation, which avoids repeated t-SVD computations during the iterative updates. Under the $\ell _{1}$ -RIP assumption, PSGD enjoys a linear convergence rate independent of the tensor condition number. Experiments on both synthetic and real datasets demonstrate significant acceleration and robustness over the standard subgradient-based method.
Zitieren
APA
Chen, R., & Huang, L. (2026). Robust Low-Tubal-Rank Tensor Sensing via Preconditioned Subgradient Descent. IEEE Signal Processing Letters, 33, 2725-2729. https://doi.org/10.1109/lsp.2026.3698753
Vancouver
Chen R, Huang L. Robust Low-Tubal-Rank Tensor Sensing via Preconditioned Subgradient Descent. IEEE Signal Processing Letters. 2026;33:2725-2729. doi:10.1109/lsp.2026.3698753
BibTeX
@article{Chen2026,
author = {Chen, Ran and Huang, Linyu},
title = {Robust Low-Tubal-Rank Tensor Sensing via Preconditioned Subgradient Descent},
journal = {IEEE Signal Processing Letters},
year = {2026},
volume = {33},
pages = {2725--2729},
doi = {10.1109/lsp.2026.3698753}
}
EUVIMED