Abductive Reasoning with Filtered Circumscription∗

For logical artificial intelligence to be truly useful, its methods must scale to problems of realistic size. An interruptible algorithm enables a logical agent to act in a timely manner to the best of its knowledge, given its reasoning so far. This seems necessary to avoid analysis paralysis, trying to think of every potentiality, however unlikely, beforehand. These considerations prompt us to look for alternative reasoning mechanisms for filtered circumscription, a nonmonotonic reasoning formalism used e.g. by Temporal Action Logic and Event Calculus. We generalize Ginsberg’s circumscriptive theorem prover and describe an interruptible theorem prover based on abduction that has been used to unify planning and reasoning in a logical agent architecture.