Representing Knowledge with Rules. Overview of the Semantic Knowledge Engineering Approach

Knowledge engineering belongs to the core of the classic symbolic Artificial Intelligence. Rules are one of the most successful methods to represent knowledge and provide effective reasoning. However, building knowledge bases with rules is not always straightforward. Some of the problems that need to be addressed include logical description of rules, practical interpretation of the rule base, scalability, intelligibility of the design methods, logical correctness. Practical design of rule bases can be simplified with the use of visual representations such as decision tables and decision trees. The Semantic Knowledge Engineering Approach (SKE) aims at solving the above mentioned problems. It introduces a formal description of the rule base with the use of the ALSV(FD) logic and the XTT2 (eXtended Tabular Trees) representation method. With SKE practical engineering issues also include the integration with the Semantic Web approach, as well as Software Engineering methods, and Business Process Modeling. Moreover, the approach is supported by a set of software tools. SKE has been developed in the GEIST research group at AGH and is being used in number of projects.