Tech Report CS-91-12

Probabilistic Abduction for Plan Recognition

Eugene Charniak and Robert Goldman

February 1991


Plan recognition requires the construction of possible plans that could explain a set of observed actions, and the selection of one or more of them as providing the {\em best} explanation. In this paper we present a formal model of the latter process based upon probability theory. Our model consists of a knowledge-base of facts about the world expressed in a first-order language, and rules for using that knowledge base to construct a Bayesian network. The network is then evaluated, to find the plans with the highest probability.

(complete text in pdf)