| A model-based problem solving approach uses `deep knowledge' to solve problems. We will show how a model-based approach is applied to the diagnosis of a digital circuit system. A model which specifies how the system is supposed to work normally is required to deduce all possible fault hypotheses from the given model and the observations obtained so far. An association-based problem solving approach achieves the same competence level of human experts by using `shallow knowledge,' i.e., direct associations between conditions and conclusions. We will show how an association-based approach derives conclusions and how shallow knowledge can be obtained. |