Bayesian Interpolants as Explanations for Neural Inferences

04/08/2020

∙

by   Kenneth L. McMillan, et al.

∙

0

∙

share

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 made by neural networks. The method produces explanations that are at the same time concise, understandable and precise.