A text understanding system with learning capabilities is presented. New concepts are acquired by incorporating two kinds of evidence - knowledge about linguistic constructions in which unknown lexical items occur and knowledge about structural patterns in ontologies such that new concept descriptions can be compared with prior knowledge. On the basis of the quality of evidence gathered this way concept hypotheses are generated, ranked according to plausibility, and the most credible ones are selected for assimilation into the domain knowledge base.
Published Date: May 1998
Registration: ISBN 978-1-57735-051-4
Copyright: Published by The AAAI Press, Menlo Park, California