Bayesian Interpolants as Explanations for Neural Inferences

Kenneth L. McMillan

The notion of Craig interpolant, used as a form of explanation in automated reasoning, is adapted from logical inference to statistical inference and used to explain inferences mad…