So I am trying to do pretraining on MIDI (music input) of humans using recurrent neural networks. I read the papers https://arxiv.org/pdf/1412.6581v6.pdf and http://www.hexahedria.com/2015/08/03/composing-music-with-recurrent-neural-networks/ And this video https://www.youtube.com/watch?v=cu1_uJ9qkHA
but I am not sure I am understand the structure of the nets (it is not well defined in the papers, especially in a case when I am merging two of them).
- How should I combine a convolutional neural network and a recurrent neural network? What should be the proper intuition for do so?
- Can I use a guassian mixture model as a decoder alongside variational recurrent autoencoder?
- Can any one point me to a url/paper which has detailed the architecture of such a variational recurrent auto encoder to do pre training on MIDI?