Wissensrepräsentation für Fortgeschrittene Computer-Anwendungen

Summary

In recent years concepts and techniques have been developed in Computer Science and, in particular, in Artificial Intelligence which open up possibilities of a special kind. It is now possible to apply sophisticated programs to tasks which would require intelligence if carried out by humans, for example intelligent decision making, medical consultation or automatic image interpretation. One of the key reasons for this progress is the development of knowledge representation techniques and knowledge-based system architectures. It is shown by means of introductory examples that a knowledge-based approach offers several advantages, including transparency, adaptability, and an improved user interface. The main part of the presentation will deal with different ways of representing knwoledge. It is shown how IF-THEN rules may be used to represent knowledge in so-called expert systems which are designed to outperform human experts in certain domains, e.g. in medical diagnosis. While the resulting system architecture is particularly simple, such rules are certainly not adequate for representing highly structured knowledge. Several other techniques are described, including semantic nets which expose the interrelationships between pieces of knowledge by named links and lend themselves to an illustrative graphical representation. As the need arises to represent more aspects of this world in more detail and more depth, knowledge representation mechanisms have to fullfill sophisticated and mainfold requirements. The presentation concludes with some advanced methods and forthcoming applications.