Re:Survey of Dropout Methods for Deep Neural Networks
by bumby on Tuesday May 07, 2019 @09:23AM
(#19849072)
My understanding is that it avoids overfitting when data points are highly correlated.
For example, if you use image augmentation to generate additional data, your augmented images are going to be highly correlated to their parent image leading to overfitting of the data. By using random dropout, this overfitting can be somewhat mitigated.
For example, if you use image augmentation to generate additional data, your augmented images are going to be highly correlated to their parent image leading to overfitting of the data. By using random dropout, this overfitting can be somewhat mitigated.