Mapping Knowledge to Language with LOOM

Manfred Stede and Dietmar Rosner

In language generation, one of the central tasks is to map a "deep" representation of some kind to lexical units, from which a grammatical utterance can be produced. Unless a strict one-to-one relationship between concepts and words is assumed, which, we argue, is too simplistic for the general case, some sort of structure-mapping process is required for accomplishing this step. We investigate the mechanisms that one particular knowledge representation language of the KL-ONE family has to offer for this purpose, and compare their relative merits.


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.