notes-cog-ai-reasoning-bayesNets

"The universality of this intuition ((the Sure-Thing principle)) suggests that a structure like causal Bayesian networks is used by people to represent and process causal knowledge. Aside from sanctioning the sure-thing principle, Bayesian networks also enjoy transparency and parsimony, two features that are not shared by other theories of actions, (e.g., possible-world counterfactuals (Lewis, 1973) or potential outcome theory (Neyman, 1923; Rubin, 1974). "

-- Pearl, The Sure-Thing Principle