Finetuning an NLG system through experiments with human subjects: the case of vague descriptions.

Kees van Deemter

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Abstract. This discussion paper describes a sequence of experiments
with human subjects aimed at nding out how an nlg system should
choose between the dierent forms of a gradable adjective. This case
study highlights some general questions that one faces when trying to
base nlg systems on empirical evidence: one question is what task to set
a subject so as to obtain the most useful information about that subject,
another question has to do with dierences between subjects.
Original languageEnglish
Title of host publicationProc. of 3rd International Conference on Natural Language Generation
EditorsAnja Belz, Roger Evans, Paul Piwek
Place of PublicationBerlin and Heidelberg
PublisherSpringer Lecture Notes
Number of pages10
Publication statusPublished - 2004

    Fingerprint

Keywords

  • Vague adjectives.
  • Experiments with human subjects.

Cite this

van Deemter, K. (2004). Finetuning an NLG system through experiments with human subjects: the case of vague descriptions. In A. Belz, R. Evans, & P. Piwek (Eds.), Proc. of 3rd International Conference on Natural Language Generation Springer Lecture Notes. http://www.csd.abdn.ac.uk/~kvdeemte/vague/vDeemter-inlg04.pdf