Higher-order logics for handling uncertainty in expert systems

✍ Scribed by E.H. Mamdani; H.J. Efstathiou

• Publisher: Elsevier Science
• Year: 1985
• Weight: 665 KB
• Volume: 22
• Category: Article
• ISSN: 0020-7373
• DOI: 10.1016/s0020-7373(85)80004-3

✦ Synopsis

We argue that the current performance of expert systems is being limited by their capacity to cope with uncertainty. Probabilistic logics alone are not enough to cope with the many kinds of uncertainty that can occur. We show how modal and quantified logics have been devised to express different types of knowledge and are each a partial solution to the problem. Logic, however, can only express a limited amount of knowledge and this shortcoming is crucially affecting the knowledge engineering of expert systems. We argue that fuzzy logic and the associated Test-Score Semantics have a role in the development of expert systems, since they enable the representation of uncertainty and a topic-dependent semantics.