深度学习论文随记(四)ResNet 残差网络
DeepResidual Learning for Image Recognition
Author:Kaiming He, XiangyuZhang, Shaoqing Ren, Jian Sun,
Microsoft Research
Year:2015
1、导引
之前文章谈到GoogLeNet和VGG,人们开始认为增加网络的层数,即让网络变深似乎能进一步提高分类任务的准确性。于是,微软研究院的何凯明团队提出了这样一个问题:
Is learning better networks as easy as stacking more layers?
然后他们通过增加层数就发现一个奇怪的现象: