Representing Bayesian Networks within Probabilistic Horn Abduction
This work provides a new logical-probabilistic integration for evidential reasoning, potentially extending Bayesian networks beyond propositional languages, but it appears incremental as it builds on existing abduction and network concepts.
The paper tackles the problem of representing Bayesian belief networks within a probabilistic Horn abduction framework, showing that any probabilistic knowledge from Bayesian networks can be represented and enabling approximations to posterior probabilities.
This paper presents a simple framework for Horn clause abduction, with probabilities associated with hypotheses. It is shown how this representation can represent any probabilistic knowledge representable in a Bayesian belief network. The main contributions are in finding a relationship between logical and probabilistic notions of evidential reasoning. This can be used as a basis for a new way to implement Bayesian Networks that allows for approximations to the value of the posterior probabilities, and also points to a way that Bayesian networks can be extended beyond a propositional language.