Exploring Localization in Bayesian Networks for Large Expert Systems
This work addresses computational inefficiency for human reasoners in large domains by enabling localized queries, though it is incremental as it builds on existing Bayesian network methods.
The paper tackles the inefficiency of propagating evidence through homogeneous Bayesian networks in large expert systems by introducing multiply sectioned Bayesian networks, which enable localization-preserving representation and reduce computational requirements to the size of a single junction tree while maintaining identical probabilities.
Current Bayesian net representations do not consider structure in the domain and include all variables in a homogeneous network. At any time, a human reasoner in a large domain may direct his attention to only one of a number of natural subdomains, i.e., there is ?localization' of queries and evidence. In such a case, propagating evidence through a homogeneous network is inefficient since the entire network has to be updated each time. This paper presents multiply sectioned Bayesian networks that enable a (localization preserving) representation of natural subdomains by separate Bayesian subnets. The subnets are transformed into a set of permanent junction trees such that evidential reasoning takes place at only one of them at a time. Probabilities obtained are identical to those that would be obtained from the homogeneous network. We discuss attention shift to a different junction tree and propagation of previously acquired evidence. Although the overall system can be large, computational requirements are governed by the size of only one junction tree.