Content area
Full text
Abstract - One of the basic principles of knowledge representation is that it is a language by which people say things about the world. Visual depictions appear particularly useful for representation of knowledge, e.g., Peirce's Existential Graphs, and Sowa's Conceptual Graphs (CGs). Recently, a new flow-based model for representing knowledge, called the Flowthing Model (FM), has been proposed and used in several applications. This paper is an exploratory assessment of the capability of FM to express knowledge, in contrast to CGs. Initial examination suggests that FM contributes to expressing knowledge in a way not provided by CGs. In addition, FM seems to produce a new aspect that may complement the CG formalism. Such exploration can promote progress in knowledge representation and modeling paradigms and their utilization in various applications.
Keywords: Knowledge representation; conceptual graphs; diagrammatic representation; flow-based knowledge representation
(ProQuest: ... denotes formulae omitted.)
1 Introduction
During the past 40 years, visual depictions have been used in the area of knowledge representation, specifically using a semantic network, e.g., [1-2]. Many of these representations concentrate on the fields of linguistic knowledge (e.g., [3]), knowledge in large-scale development of applications, or logical aspects of semantic networks. A basic principle of knowledge representation, as a medium of human expression, is that it is "a language in which we say things about the world" [4]. This paper focuses on this aspect of knowledge representation, in contrast to such features as logical reasoning or computational efficiency.
To limit the scope of this paper, it examines Sowa's Conceptual Graphs (CGs) and contrasts them with a newly developed conceptual representation based on the notion of flow. The CG was developed as "a graph representation for logic based on the semantic networks of artificial intelligence and Peirce existential graphs" [5; italics added]. Initially, Sowa developed CGs as an intermediate language for mapping natural language assertions to a relational database. They have also been viewed as a diagrammatic system of logic to express meaning in a precise form, humanly readable, and computationally tractable [5].
CGs have been applied in a wide range of fields [5]. In artificial intelligence, CG formalism offers many benefits, including graph-based reasoning mechanisms, plug-in capabilities over data structures, and good visualization capabilities [6]. In addition, a conceptual graph...