A Dynamic Lexical Knowledge Base for Natural Language Processing

Caroline Barriere and Fred Popowich

In this paper we present the design of a Lexical Knowledge Base (LKB) that uses Conceptual Graphs (CGs) to incorporate the information extracted from a children’s dictionary. We show how CGs are used to build the concept hierarchy and to express the definitions of words. As an addition to the standard CG formalism, we introduce a notion of certainty on the facts extracted from the dictionary and present in the LKB. This allows us to assign weight to the different types of information (description, usage, examples) given by the dictionary definitions. A representation formalism like CGs, which are advocated as being closely related to natural language, facilitates the building of a dynamic LKB that can be augmented and updated during the processing of natm'al language text.

This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.