Probabilistic reasoning in intelligent systems: networks of plausible inference
Probabilistic reasoning in intelligent systems: networks of plausible inference
Probabilistic Reasoning in Multi-Agent Systems: A Graphical Models Approach
Probabilistic Reasoning in Multi-Agent Systems: A Graphical Models Approach
An Algebraic Characterization of Equivalent Bayesian Networks
Proceedings of the IFIP 17th World Computer Congress - TC12 Stream on Intelligent Information Processing
Justifying Multiply Sectioned Bayesian Networks
ICMAS '00 Proceedings of the Fourth International Conference on MultiAgent Systems (ICMAS-2000)
Hi-index | 0.00 |
Multiply Sectioned Bayesian Networks (MSBNs) provide a highly modular and efficient framework for uncertain reasoning in multiagent distributed system. The formal structure of MSBN was developed based on a set of five assumptions. A critical requirement in building a MSBN is its structure verifications which include directed acyclic graph (DAG) verification and d-sepset verification. Previous works on structure verification in MSBNs have been based on graphical methods which exploit mainly the graphical information carried by the subnets of MSBNs. In this paper, we approach the structure verification process in MSBNs from the symbolic perspetive. Our methods utilize mainly the probability information carried by the MSBN instead of the graphical information. We introduce several symbolic operations to support structure verification based on the algebraic description of the Joint Probability Distribution (JPD) in each local subnet. We also discuss the issue of dealing verification failure in order to direct attentions towards fixing the offending parts of the subnets.