BriefGPT.xyz
Aug, 2024
带有$\ell_1$正则化的稀疏深度学习模型
Sparse Deep Learning Models with the $\ell_1$ Regularization
HTML
PDF
Lixin Shen, Rui Wang, Yuesheng Xu, Mingsong Yan
TL;DR
本研究关注于稀疏神经网络的重要性,探讨正则化参数的选择如何影响学习到的神经网络的稀疏程度。通过从统计学角度导出带有$\ell_1$-范数的稀疏促进深度学习模型,本文发展了选择正则化参数的迭代算法,以实现预定的稀疏水平,并在数值实验中验证了方法的有效性。
Abstract
Sparse Neural Networks
are highly desirable in
Deep Learning
in reducing its complexity. The goal of this paper is to study how choices of
Regula
→