Explanations in knowledge systems: the roles of the task structure and domain functional models

An investigation of explanation (the process by which an expert system explains how it arrived at a particular solution) in the domains of logistics planning and diagnosis is discussed. The research was performed as part of the Strategic Computing Initiative of the US Dept. of Defense´s Defence Advanced Research Projects Agency (DARPA). Types and aspects of explanation are examined. The generic task approach to generating explanations of problem-solving strategies and steps and its implementation in the Mission Planning Assistant (MPA), which plans a particular kind of US Air Force mission are described. MPA uses a generic task tool called the design specialists and plans language. Explanation based on logical task structure is then considered. The use of causal models to confirm the plausibility of diagnostic hypothesis is discussed.<>