[工程科技]Inference meta models A new perspective on belief propagation with bayesian networks.pdfVIP

Universiteit van Amsterdam IAS technical report IAS-UVA-06-01 Inference Meta Models: A New Perspective On Belief Propagation With Bayesian Net- works Gregor Pavlin, Jan Nunnink, and Frans Groen Intelligent Systems Laboratory Amsterdam, University of Amsterdam The Netherlands We investigate properties of Bayesian networks (BNs) in the context of robust state estimation. We focus on problems where state estimation can be viewed as a classification of the possible states. We introduce a coarse perspective of the in- ference processes and show that classification with BNs can be very robust, even if we use models and evidence associated with significant uncertainties. By making coarse and realistic assumptions we can formulate asymptotic properties of the classification performance. In addition, we identify situations in which Bayesian fusion supports robust inference and we introduce techniques that support de- tection of potentially misleading inference results. The presented coarse grained analysis of inference processes from the runtime perspective is relevant for a sig- nificant class of real world domains, where it is difficult to obtain domain models that precisely describe the true probability distributions over the combinations of states of interest. Keywords: Bayesian networks, Robust information fusion, Heterogeneous infor- mation. IAS intellige