Natural Language Output in Information-Seeking Dialogues Using RhetoricalRelations

摘要

The paper presents an approach to unify a speech-act oriented model forinformation-seeking dialogues (COR) with a model to describe the structure of monological text units (RST). The proposed approach, which also leads to essential extensions of both models, is verified by interfacing the dialogue component of an intelligent multi-media retrieval system with a component for natural language generation. The resulting retrieval system which uses natural language as an output modality facilitates the user interaction in problematic situations. (Copyright (c) GMD 1992.)