2CS24 - TOPICS IN INFORMATION PROCESSING

EXPERT SYSTEMS


1. Overview

Expert systems (also sometimes referred to as a Knowledge Based Systems - KBS) are decision support system designed to behave like an expert in some given application domain. They comprise a Knowledge Base (KB) and an inference engine. The KB contains domain specific information obtained from expert sources, including such things as simple facts about the domain and rules that describe relations or phenomena within the domain. The inference engine then contains methods of manipulating the knowledge in the KB. As such the development of expert systems entails additional processes such as knowledge elicitation, knowledge analysis, domain conceptualisation and knowledge representation, which are not usually found in the development of more traditional software systems.

Expert systems became popular in the early 1980s when it was extravagantly claimed by supporters that this "new" technology would solve the worlds computing problems. However, in their initial enthusiasm, the case was somewhat overstated; thus when the technology inevitable failed to meet expectations (late 1980s) research in the field suffered a significant set back. However, the field is now undergoing something of a revival in acknowledgement that not all of the original claims were "bogus".

Keywords. Expert Systems, Intelligent Systems, Knowledge based systems, Artificial Intelligence (AI).


2. Reading

There are many works in our library that cover the field of Expert Systems and KBS. Worth a look are:

  1. Bench-Capon, T.J.M. Knowledge Representation: An Approach to Artificial Intelligence. Academic Press.
  2. Buchanan, B.G. (1984) Rule-based expert systems: the MYCIN experiments of the Stanford Heuristic Programming Project Addison-Wesley.

Web searching is not so easy given the number of hits that you are likely to encounter, however "http://www2.csc.liv.ac.uk/~tbc/kbs.html" is worth a look.


3. Lecture notes

Click here to obtain a printed copy of the lecture slides.


4. Seminar title

Suggested seminar titles:

  1. ES Applications.
  2. Techniques to develop ES.
  3. Knowledge elicitation and analysis.
  4. nowledge representation.



Created and maintained by Frans Coenen. Last updated 11 October 1999