Winning the explanation game; providing user explanations for a model-based expert system

The authors develop an approach to organizing the presentation of large amounts of model-based data in an interactive format patterned after a model of human-human explanatory and argumentative discourse. Portions of this interface were implemented for Honeywell and the Air Force's model-based flight control maintenance and diagnostic system (FCMDS). It is concluded that sufficient information exists in a model-based system to provide a wide range of explanation types, and that, at least in FCMDS, this information is readily accessible though in a format difficult for the user to use. It is also concluded that the discourse approach is a convenient, powerful, and broadly applicable method of representing and controlling information exchange in human-machine, as well as human-human, interactions.<
>