Topic maps offer flexible and powerful techniques for knowledge representation (KR). They define the general concepts and provide - with Intention - just the necessary minimum of semantics. But KR requires more semantic to model ontologies, class hierarchies, association properties, inference rules, and constraint-based validation.
This chapter explains why these semantics are needed, gives some examples of applications, and presents an approach to a technical solution. The solution itself will make use of the topic maps paradigm.
KR is already well understood within the field of artificial intelligence (AI) research. Concepts like semantic networks and conceptual graphs were developed to model knowledge. The general approach of the topic maps paradigm defines the basic constructs for KR with topic maps, but supporting this particular application domain was not a design goal of the topic map standardization effort. Therefore, the required semantics must be defined as a kind of application profile.