Knowledge Representation: What’s Important About It?

Abstract

In this chapter the basic requirements for a knowledge representation scheme are examined. The nature of models and reasoning is discussed, and characteristics of various aspects of representation are outlined (including knowledge acquisition, perception, planning, and generalization). The knowledge representation language KL-One is introduced in order to clarify these notions with actual examples. Then, the expressive adequacy of a knowledge representation scheme (what it is capable of representing) is contrasted to the notational efficacy of a scheme (its actual shape and structure). Two aspects of notational efficacy receive special attention: computational efficiency (speed of inference) and conceptual efficiency (ease of representation). The chapter concludes with arguments about the role of predicate calculus in the representation of knowledge.