How to improve the reliability of expert systems
Alan Bundy · Edinburgh Research Explorer (University of Edinburgh) · 1988
Reliability is likely to become an increasingly important issue for knowledge engineers. Without assurances of reliability they will be restricted in the scale and scope of the expert systems they can build and will not be able to capitalise on their current success. The key to greater reliability is a sound theoretical foundation for current and new knowledge engineering techniques making knowledge engineering a proper engineering science. We illustrate the kind of theoretical work that is required with a few case studies, in the areas of search control, fault diagnosis and learning.