Logic-Based Networks: Concept Graphs and Conceptual Structures

Logic-based networks are semantic networks that support reasoning capabilities. In this paper, knowledge processing within logic-based networks is viewed as three stages. The first stage involves the formation of concepts and relations: the basic primitives with which we wish to formulate knowledge. The second stage involves the formation of well-formed formulas that express knowledge about the primitive concepts and relations once isolated. The final stage involves efficiently processing the wffs to the desired end. Our research involves each of these steps as they relate to Sowa’s conceptual structures and Wille’s concept lattices. Formal Concept Analysis gives us a capability to perform concept formation via symbolic machine learning. Concept(ual) Graphs provide a means to describe relational properties between primitive concept and relation types. Finally, techniques from other areas of computer science are required to compute logic-based networks efficiently. This paper illustrates the three stages of knowledge processing in practical terms using examples from our research.