20. Dependence maps for many variables
Represent large probability models with diagrams that show which variables directly depend on which others. Bayesian networks, Markov random fields, and factor graphs make complex dependence easier to calculate and check.