Improving neural networks by preventing co-adaptation of feature detectors

【文件属性】：

文件名称：Improving neural networks by preventing co-adaptation of feature detectors

文件大小：1.47MB

文件格式：ZIP

更新时间：2022-05-01 09:56:45

深度学习 模型训练

Hinton在2012年提出的Improving neural networks by preventing co-adaptation of feature detectors也就是dropout,常用于防止训练过拟合,主要是通过随机选择一部分神经元训练,而直接丢弃其他神经元.

【文件预览】：
2012-G. E. Hinton-Improving neural networks by preventing co-adaptation of feature detectors.pdf