Representing Knowledge As Arguments: Applying Expert System Technology to Judgemental Problem-Solving

Reference: P. Clark. Representing knowledge as arguments: Applying expert system technology to judgemental problem-solving. In T. R. Addis and R. M. Muir, editors, Research and Development in Expert Systems VII, pages 147-159. Cambridge Univ. Press, 1990.

Abstract: In many domains, it is not possible to easily gather a definitive body of expertise for problem-solving. A particularly well-known problem is that, when experts disagree, it is not easy or even possible to identify the `right' answer, a characteristic particularly true of problems involving human judgement. In such domains, the process of argumentation between experts plays a crucial role in pooling knowledge, locating inconsistencies and focusing attention on areas for further examination.

In this paper we recognise this process of argumentation - reasoning about why inferences are valid - as important in problem-solving, and present techniques for using AI technology to assist experts in this process. In the implementation we present, models of different opinions are represented separately rather than combined. The user and system interact to solve a given problem, the system arguing its case on the basis of consistency with previous decisions and the user supplying extra knowledge which the system is unaware of. Dialog focuses on the meta-level justifications for inferences which are made, not normally represented in the usual expert system methodology. By exploiting the computer's power of memory and search a powerful decision support tool can be constructed. We illustrate this with a full-scale system, named Optimist, for assisting geologists in oil exploration.