Accelerating Non-Negative and Bounded-Variable Linear Regression Algorithms with Safe Screening - Signaux et Images Accéder directement au contenu
Pré-Publication, Document De Travail Année : 2022

Accelerating Non-Negative and Bounded-Variable Linear Regression Algorithms with Safe Screening

Résumé

Non-negative and bounded-variable linear regression problems arise in a variety of applications in machine learning and signal processing. In this paper, we propose a technique to accelerate existing solvers for these problems by identifying saturated coordinates in the course of iterations. This is akin to safe screening techniques previously proposed for sparsity-regularized regression problems. The proposed strategy is provably safe as it provides theoretical guarantees that the identified coordinates are indeed saturated in the optimal solution. Experimental results on synthetic and real data show compelling accelerations for both non-negative and bounded-variable problems.
Fichier principal
Vignette du fichier
NN-BV_Screening_v1.pdf (637.68 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03564336 , version 1 (11-02-2022)
hal-03564336 , version 2 (19-06-2023)

Identifiants

Citer

Cassio F. Dantas, Emmanuel Soubies, Cédric Févotte. Accelerating Non-Negative and Bounded-Variable Linear Regression Algorithms with Safe Screening. 2022. ⟨hal-03564336v1⟩
234 Consultations
103 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More