Integrating Rules in Term Subsumption Knowledge Representation Servers

Abstract

This paper addresses the integration of services for rule-based reasoning in knowledge representation servers based on term subsumption languages. As an alternative to previous constructions of rules as concept→concept links, a mechanism is proposed based on intensional roles implementing the axiom of comprehension in set theory. This has the benefit of providing both rules as previously defined, and set aggregation, using a simple mechanism that is of identical computational complexity to that for rules alone. The extensions proposed have been implemented as part of KRS, a knowledge representation server written as a class library in C++. The paper gives an example of their application to the ripple-down rule technique for large-scale knowledge base operation, acquisition and maintenance.