Issue No. 01 - Spring (1989 vol. 4)
pp: 9-15, 19-24
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/64.21896
<p>Explaining how knowledge-based systems reason involves presentation user modeling, dialogue structure, and the way systems understand their own problem-solving knowledge and strategies. The authors concentrate on the last of these, noting that such understanding provides any explanations's content. The authors also note that most current approaches to knowledge-based system construction require expressing knowledge and control at such low levels that it's hard to give high-level explanations. Providing an explanation example from a prototypical system (MYCIN) built using generic-task methods, they propose generic-task methodology as one way to build knowledge-based systems that contain basic explanation constructs at appropriate abstraction levels. The central concept of generic tasks is what input-output behavior (i.e. that task function), knowledge needed to perform the task, and inferences appropriate for the task are all specified together.</p>
B. Chandrasekaran, M. C. Tanner and J. R. Josephson, "Explaining Control Strategies in Problem Solving," in IEEE Intelligent Systems, vol. 4, no. , pp. 9-15, 19-24, 1989.