Bayesian Interpolants as Explanations for Neural Inferences
Artificial Intelligence
2020-04-10 v1 Machine Learning
Machine Learning
Abstract
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.
Keywords
Cite
@article{arxiv.2004.04198,
title = {Bayesian Interpolants as Explanations for Neural Inferences},
author = {Kenneth L. McMillan},
journal= {arXiv preprint arXiv:2004.04198},
year = {2020}
}
Comments
10 pages, 4 figures