Yes, but take it way further. i.e. How can you use a Deep Network to direct NEAT? Can you model the topology generation as a latent space, and what would some o

Yes, but take it way further. i.e. How can you use a Deep Network to direct NEAT? Can you model the topology generation as a latent space, and what would some of the properties of that space be? Can we integrate the best parts of transfer learning to speed up the generate -> evaluate cycle. That sort of thing