Harvard AND Deep latent variable models
Common descendants