Are latent variables normal distributed in VAE? Actually, there are other including uniformly distributed etc… The normal distribution is quite popular. Mathematically, a normal distribution is easy to manipulate and sample.

Removing information from your input forces you not to be overconfidence in your model which can be a sign of overfitting by training data. This is very similar to the idea of dropout in DL.

Written by

Deep Learning

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store