A Knowledge Representation Language

Abstract

In this chapter we present a knowledge representation language which was developed in accordance with the criteria and conclusions described in the previous chapter. The skeleton of the language is formed by a semantic network scheme. The definition of the language covers its syntactical properties, the semantic of all the slots and facets as well as the inference processes. One of the main goals is the explicit representation of all the different types and aspects of knowledge as deeply discussed in the last chapter. The semantic network language together with the control algorithms which will be presented later define the knowledge representation system ERNEST (Erlangen semantic NEtwork System and Tools). But ERNEST also includes two further problem independent modules for supervised learning and for the explanation of system resources and results. Both will be described in further chapters. The complete ERNEST system has been implemented in the programming language C and the UNIX operating system. Different knowledge based systems for image, speech, and other signal analysis tasks are based on the ERNEST system.