Interpreting clinical latent representations using autoencoders and probabilistic models | Publicación