Inverse circumscription
Hubie Chen · BIROn (Birkbeck, University of London) · 2003
Inverse (or identification) problems involve deciding whether or not an explicitly given set of data points have an implicit description, for instance, in the form of a constraint network. Such problems provide insight into the relationships among various representations of knowledge, which may have differing computational properties. This paper formalizes and studies the inverse circumscription problem, which (roughly speaking) is to decide, given a set of models, if there exists a formula whose circumscription describes the input set.