A Bayesian Account of Measures of Interpretability in Human-AI Interaction