A model for justification production by expert planning systems

Abstract

Although explanation capability is one of the distinguishing characteristics of expert systems, the explanation facilities of most existing systems are quite primitive. This paper describes an architecture for production of explanations that justify reasoning that is based on the premises that (1) the task of justifying expert decisions is an intelligence-requiring activity and (2) the appropriate model for machine-produced justifications should be explanations written by people. The architecture consists of (1) a processing strategy for an Augmented Phrase Structured Grammar (APSG) for the purpose of structuring text, (2) a special language designed for writing text structuring grammars, and (3) a Justification Grammar (JG) representing the structure of justification texts. A prototype justification production system is presented to demonstrate how the model can be used to synthesize knowledge from a variety of sources and produce coherent, multisentential text similar to that produced by a domain expert.