Knowledge Acquisition using Documents, Conceptual Graphs and a Semantically Structured Dictionary

In this paper, we first show how in CGKAT, our knowledge acquisition tool, any document element and its semantics may be represented using the Conceptual Graphs formalism (Sowa, 1984) and a structured document editor. Then, we study the kinds of hypertext links that may be set between documents elements and concepts or relations of the knowledge base (such links enables the use of search techniques on the KB for finding information within the documents). In a second part, we detail the top-level ontologies (for concepts and relations) proposed by CGKAT and its exploitation of a semantically structured dictionary for guiding knowledge representation and easing its later reuse.