Does Size Matter – How Much Data is Required to Train a REG Algorithm?

Mariet Theune, Ruud Koolen, Emiel Krahmer, Sander Wubben

    Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

    6 Citations (Scopus)
    20 Downloads (Pure)

    Abstract

    In this paper we investigate how much data is required to train an algorithm for attribute selection, a subtask of Referring Expressions Generation (REG). To enable comparison between different-sized training sets, a systematic training method was developed. The results show that depending on the complexity of the domain, training on 10 to 20 items may already lead to a good performance.
    Original languageUndefined
    Title of host publicationProceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies, HLT 2011
    Place of PublicationStroudsburg, PA, USA
    PublisherAssociation for Computational Linguistics (ACL)
    Pages660-664
    Number of pages5
    ISBN (Print)978-1-932432-88-6
    Publication statusPublished - Jun 2011
    Event49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies 2011 - Portland, United States
    Duration: 19 Jun 201124 Jun 2011
    Conference number: 49

    Publication series

    Name
    PublisherAssociation for Computational Linguistics

    Conference

    Conference49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies 2011
    Abbreviated titleHLT 2011
    CountryUnited States
    CityPortland
    Period19/06/1124/06/11

    Keywords

    • IR-78517
    • EWI-20799
    • METIS-281562

    Cite this