Read the original document by opening this link in a new tab.
Table of Contents
1 Introduction
1.1 Background
1.1.1 Dimension reduction via (safe) screening
1.2 Our contributions
2 Safe screening for sparse regularization
2.1 Safe screening
2.1.1 Gap-safe screening
2.2 Gap-safe screening for Prox-SGD
3 Screening for online algorithms
4 Paper organization
Summary
The document discusses the use of sparsity-promoting regularizers in online regression problems. It introduces the concept of safe screening rules to identify and eliminate useless features in the optimization process. The contributions include adapting gap-safe screening rules for online algorithms to enforce sparsity identification. The paper explores the advantages of screening techniques in improving computational efficiency and dimension reduction. Numerical experiments on LASSO and sparse logistic regression are presented in the document.