Gibbs-Sampling

approximate inference in graphical models — forward and rejection sampling, likelihood weighting, gibbs sampling and mcmc, and loopy belief propagation, with the trade-offs that decide which to reach for.