Grammar Variational Autoencoder

Matt J. Kusner, Brooks Paige, José Miguel Hernández-Lobato

Deep generative models have been wildly successful at learning coherent latent representations for continuous data such as video and audio. However, generative modeling of discrete…