What is understandable in Bayesian network explanations?

TL;DR

This paper investigates how well humans understand explanations of Bayesian network predictions by comparing four different explanation methods through a survey with human participants.

Contribution

It introduces an empirical study comparing multiple explanation approaches for Bayesian networks based on human interpretability.

Findings

Different explanation methods vary in human interpretability

Survey results highlight which explanations are more understandable

Initial insights into effective explanations for non-expert users

Abstract

Explaining predictions from Bayesian networks, for example to physicians, is non-trivial. Various explanation methods for Bayesian network inference have appeared in literature, focusing on different aspects of the underlying reasoning. While there has been a lot of technical research, there is very little known about how well humans actually understand these explanations. In this paper, we present ongoing research in which four different explanation approaches were compared through a survey by asking a group of human participants to interpret the explanations.