Data-driven Smooth-to-hard Spectral Regularization for General Linear Inverse Problems

DOI10.48550/ARXIV.2610.21664MaRDI QIDQ7400280arXivFDO

Authors Guanzhong Yang, Zi Wang, Guang Yang

Publication date 15 October 2026

Published in arXiv

Tldr This paper derives data-driven smooth-to-hard regularization methods for general linear inverse problems by using the logistic smooth-to-hard shrinker as a proximal map of explicit saturating spectral penalties to guarantee unique scaled updates and stationarity in proximal-gradient schemes.

Full work available at URL https://doi.org/10.48550/ARXIV.2610.21664

arXiv classification

eess.IVeess.SPmath.NAmath.OC

Cites work

• A Singular Value Thresholding Algorithm for Matrix Completion
• Generalized Singular Value Thresholding
• Guaranteed minimum-rank solutions of linear matrix equations via nuclear norm minimization
• Image Quality Assessment: From Error Visibility to Structural Similarity
• Nearly unbiased variable selection under minimax concave penalty
• Nonsmooth analysis of singular values. I: Theory

This page was built for publication: Data-driven Smooth-to-hard Spectral Regularization for General Linear Inverse Problems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7400280)