Markov Chain Monte Carlo and Variational Inference: Bridging the Gap

Recent advances in stochastic gradient variational inference have made it possible to perform variational Bayesian inference with posterior approximations containing auxiliary random variables. This e