Pruning Deep Neural Network Models via Minimax Concave Penalty Regression

In this study, we propose a filter pruning method based on MCP (Minimax Concave Penalty) regression. The convolutional process is conceptualized as a linear regression procedure, and the regression coefficients serve as indicators to assess the redundancy of channels. In the realm of feature selecti...

Full description

Bibliographic Details

Published in:	Applied Sciences
Main Authors:	Xinggu Liu, Lin Zhou, Youxi Luo
Format:	Article
Language:	English
Published:	MDPI AG 2024-04-01
Subjects:	model compression minimum and maximum concave penalties Lasso regression MCP regression
Online Access:	https://www.mdpi.com/2076-3417/14/9/3669