Some Recent Advances in Non- and Semiparametric Bayesian Modeling with Copulas, Mixtures, and Latent Variables