Degree of abstraction in referring expression generation and its relation with the construction of the contrast set

Authors:
Raquel Hervás;Pablo Gervás
Affiliations:
Universidad Complutense, de Madrid Madrid, Spain;Universidad Complutense, de Madrid Madrid, Spain
Venue:
INLG '08 Proceedings of the Fifth International Natural Language Generation Conference
Year:
2008

Citing 2
Cited 1

A fast algorithm for the generation of referring expressions

COLING '92 Proceedings of the 14th conference on Computational linguistics - Volume 1
Building Natural Language Generation Systems (Studies in Natural Language Processing)

Building Natural Language Generation Systems (Studies in Natural Language Processing)

Assessing the influence of personal preferences on the choice of vocabulary for natural language generation

Information Processing and Management: an International Journal

Quantified Score

Hi-index	0.00

Visualization

Abstract

Referring Expression Generation (REG) is the task that deals with references to entities appearing in a spoken or written discourse. If these referents are organized in terms of a taxonomy, there are two problems when establishing a reference that would distinguish an intended referent from its possible distractors. The first one is the choice of the set of possible distractrors or contrast set in the given situation. The second is to identify at what level of the taxonomy to phrase the reference so that it unambiguously picks out only the intended referent, leaving all possible distractors in different branches of the taxonomy. We discuss the use of ontologies to deal with the REG task, paying special attention to the choice of the the contrast set and to the use of the information of the ontology to select the most appropriate type to be used for the referent.