no code implementations • 5 Jun 2018 • Erik Holmer, Andreas Marfurt
Our model builds on the variational autoencoder framework to define a generative document model based on next-word prediction.
