Deep Learning Approach of Sparse Autoencoders with Lp/L2 Regularization

Proceedings of the 4th International Conference on Computer Science and Application Engineering Pub Date : 2020-10-20 DOI:10.1145/3424978.3425011

Ziheng Wu, Cong Li, Baigen Pan

引用次数: 0

Abstract

In this paper, we put forward a novel deep learning approach with Lp / L2 regularization based on sparse autoencoders network, trained by a nonnegativity constraint algorithm. Since L2 norm regularization penalizes the negative weights with smaller magnitudes much weaker than those with bigger magnitudes, lots of the weights could take small negative values. In order to address this issue, non-Lipschitz nonconvex LP norm (0

查看原文本刊更多论文

Lp/L2正则化稀疏自编码器的深度学习方法

本文提出了一种基于稀疏自编码器网络的Lp / L2正则化深度学习方法，该方法采用非负性约束算法进行训练。由于L2范数正则化对较小的负权值的惩罚比较大的负权值要弱得多，因此许多权值可以取较小的负值。为了解决这一问题，引入了非lipschitz非凸LP范数(0

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊