Why we need to keep the E[Z^4] from dropping a3?

Course Q&A Deep Learning Specialization Improving Deep Neural Networks: Hyperparameter tun
1

In the video, " Dropout Regularization, inverted dropout", teacher said " So, in order to not reduce the expected value of z^4, what you do is you need to take this, and divide it by 0.8 because this will correct or just a bump that back up by roughly 20% that you need."

So. my question is why we need to keep the E[Z^4] ? If E[Z^4] is reduced from the a3, what will happen??

2

This is a common question. Here’s another recent thread about it. Which also links to this one. On both those links, please read from the given post to the end of the thread to get the full discussion.

3

thanks, thats a loooong one!