计算机应用与软件2024，Vol.41Issue(6)：243-249,7.DOI:10.3969/j.issn.1000-386x.2024.06.036
尺度因子正则化BN算法
BN ALGORITHM OF SCALE FACTOR REGULARIZATION
摘要
Abstract
In order to further improve the convergence speed of deep neural network training,using the characteristics of batch normalization(BN)algorithm as reference,a BN algorithm of scale factor regularization is proposed.By applying L2 regularization to the learnable scale factor y in the BN layer,y was attenuated,the gradient upper bound of the parameter was reduced,and the optimization space was smoother.Based on VGG16 Net and AlexNet,the image classification comparison experiments between this algorithm and the BN algorithm were carried out on the cifar10,cifar100 and crack image datasets.The results show that the proposed algorithm not only improves the convergence speed of network training,but also improves the accuracy rate at the same training times.关键词
批规范化/尺度因子/L2正则化/图像分类Key words
Batch normalization/Scale factor/L2 regularization/Image classification分类
信息技术与安全科学引用本文复制引用
刘向阳,汪琦..尺度因子正则化BN算法[J].计算机应用与软件,2024,41(6):243-249,7.基金项目
国家自然科学基金项目(61001139) （61001139）
云南省重大科技专项计划项目(202002AE090010). （202002AE090010）
