Does the GaussianDropout Layer in Keras retain probability like the Dropout Layer?

Question

I was wondering, if the GaussianDropout Layer in Keras retains probability like the Dropout Layer. The Dropout Layer is implemented as an Inverted Dropout which retains probability.

If you aren't aware of the problem you may have a look at the discussion and specifically at the linxihui's answer. The crucial point which makes the Dropout Layer retaining the probability is the call of K.dropout, which isn't called by a GaussianDropout Layer.

Is there any reason why GaussianDropout Layer does not retain probability? Or is it retaining but in another way being unseen?

Similar - referring to the Dropout Layer: Is the Keras implementation of dropout correct?

here the differences between GausDropout Dropout and GausNoise https://stackoverflow.com/a/65504785/10375049 — Marco Cerliani, Dec 30 '20 at 10:10

score 1 · Accepted Answer · edited Mar 07 '17 at 09:57

1

So Gaussian dropout doesn't need to retain probability as its origin comes from applying Central Limit Theorem to inverted dropout. The details might be found here in 2nd and 3rd paragraph of Multiplicative Gaussian Noise chapter (chapter 10, p. 1951).

edited Mar 07 '17 at 09:57

D.Laupheimer

1,074
1
9
21

answered Mar 03 '17 at 17:25

Marcin Możejko

39,542
10
109
120

Is there anything more you want to know? You haven't upvoted my answer. – Marcin Możejko Mar 07 '17 at 07:32
Atm I am satisfied. But I accepted your answer. Better? ;D – D.Laupheimer Mar 07 '17 at 07:34
Cool :) I thought thay you weren't fully satisfied. – Marcin Możejko Mar 07 '17 at 07:40
I don't understand the theory entirely, but I guess it doesn't matter. The most important information (to me) is: I can use GaussianDropout layer in Keras with a clear conscience! – D.Laupheimer Mar 07 '17 at 08:21

Does the GaussianDropout Layer in Keras retain probability like the Dropout Layer?

1 Answers1