Hypothesis Management in Situation-Specific Network Construction
This addresses the challenge of balancing tractability and accuracy in uncertain model construction for domain-specific applications like military intelligence.
The paper tackles the problem of constructing knowledge-based models under uncertainty in entity-variable associations, using Multi-entity Bayesian Networks (MEBNs) to manage hypotheses, and applies it to infer military vehicle organization from intelligence reports, comparing it to existing tracking and fusion methods.
This paper considers the problem of knowledge-based model construction in the presence of uncertainty about the association of domain entities to random variables. Multi-entity Bayesian networks (MEBNs) are defined as a representation for knowledge in domains characterized by uncertainty in the number of relevant entities, their interrelationships, and their association with observables. An MEBN implicitly specifies a probability distribution in terms of a hierarchically structured collection of Bayesian network fragments that together encode a joint probability distribution over arbitrarily many interrelated hypotheses. Although a finite query-complete model can always be constructed, association uncertainty typically makes exact model construction and evaluation intractable. The objective of hypothesis management is to balance tractability against accuracy. We describe an application to the problem of using intelligence reports to infer the organization and activities of groups of military vehicles. Our approach is compared to related work in the tracking and fusion literature.