Extending Idioms for Bayesian Network Construction with Qualitative Constraints
Annet Onnes (Universiteit Utrecht)
Mehdi Dastani (Universiteit Utrecht)
R.I.J. Dobbe (TU Delft - Information and Communication Technology)
Silja Renooij (Universiteit Utrecht)
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
Bayesian networks (BNs) are compact representations of probability distributions that allow for supporting reasoning and decision making under uncertainty. Their interpretable structure and probability parameters allow for integrating human knowledge in their construction and explanation. For BN construction, reusable building blocks, or idioms, exist that describe the dependencies and reasoning patterns among small sets of variables. In this paper we formalise the concept of an idiom, explicitly including qualitative constraints that capture the reasoning patterns among variables as stated in the informal descriptions that accompany the idioms in literature. Our proposed formalisation ensures that idioms can be applied more consistently and reliably, improving the BN’s accountability.