Scalable Deep Unsupervised Clustering with Concrete GMVAEs

Mark Collier, Hector Urdiales

Discrete random variables are natural components of probabilistic clustering models. A number of VAE variants with discrete latent variables have been developed. Training such meth…