Learning Deep Latent Gaussian Models with Markov Chain Monte Carlo

  Abstract