Some Properties of Most Relevant Explanation

Changhe Yuan

Abstract

This paper provides a study of the theoretical properties of Most Relevant Explanation (MRE). The study shows that MRE defines an implicit soft relevance measure that enables automatic pruning of less relevant or irrelevant variables when generating explanations. The measure also allows MRE to capture the intuitive phenomenon of explaining away encoded in Bayesian networks. Furthermore, we show that the solution space of MRE has a special lattice structure which yields interesting dominance relations among the candidate solutions.

PDF

Bibtex

@INPROCEEDINGS{Yuan09some,
author = {Changhe Yuan},
title = { Some Properties of Most Relevant Explanation },
booktitle = {Proceedings of IJCAI-09 Workshop on Explanation-aware Computing (ExaCt-09) },
address = {Pasadena, CA},
pages = {118--126},
year = {2009}
}