https://arxiv.org/abs/1704.02798
Authors: Meire Fortunato, Charles Blundell, Oriol Vinyals
I thought the Bayesian RNN paper was cool so I gave it a crack. This was mostly an attempt to properly understand the paper and particularly, there are some bits of this which i'm not sure i've done correctly; particularly the approximate KL computation part.