Generation of Dutch referring expressions using the D-TUNA corpus

Marissa Hoek, Mariet Theune

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

13 Downloads (Pure)

Abstract

This paper describes our research into generating Dutch noun phrases as descriptions of furniture objects or people. This is usually done in two steps: attribute selection and realisation. This research focuses only on the realisation step: generating a noun phrase from given attributes. The research is done on the Dutch version of the TUNA-corpus, which contains annotated human-produced descriptions. Three algorithms were developed for this task, each an improvement over the last. We extracted the lexical choice from the D-TUNA corpus, and used templates generated from the corpus which specified the order of attributes. The algorithms were then evaluated for string similarity with the original human descriptions from the corpus, and a human evaluation was carried out which tested clarity and fluency. A steady improvement of scores in both the automatic and human evaluation was observed for each new version of the algorithm.
Original languageUndefined
Title of host publicationProceedings of PRE-CogSci 2013 – Bridging the Gap between Cognitive and Computational Approaches to Reference
EditorsA. Gatt, R. van Gompel, E. Gurman-Bard, E. Krahmer, K. van Deemter
Place of PublicationTilburg, The Netherlands
PublisherTilburg University
Pages-
Number of pages6
ISBN (Print)not assigned
Publication statusPublished - Jul 2013
EventPRE-CogSci 2013 – Bridging the Gap between Cognitive and Computational Approaches to Reference - Berlin, Germany
Duration: 31 Jul 201331 Jul 2013
https://pre2013.uvt.nl/

Publication series

Name
PublisherTilburg University

Workshop

WorkshopPRE-CogSci 2013 – Bridging the Gap between Cognitive and Computational Approaches to Reference
CountryGermany
CityBerlin
Period31/07/1331/07/13
Internet address

Keywords

  • EWI-23554
  • IR-87092
  • METIS-297758

Cite this

Hoek, M., & Theune, M. (2013). Generation of Dutch referring expressions using the D-TUNA corpus. In A. Gatt, R. van Gompel, E. Gurman-Bard, E. Krahmer, & K. van Deemter (Eds.), Proceedings of PRE-CogSci 2013 – Bridging the Gap between Cognitive and Computational Approaches to Reference (pp. -). Tilburg, The Netherlands: Tilburg University.
Hoek, Marissa ; Theune, Mariet. / Generation of Dutch referring expressions using the D-TUNA corpus. Proceedings of PRE-CogSci 2013 – Bridging the Gap between Cognitive and Computational Approaches to Reference. editor / A. Gatt ; R. van Gompel ; E. Gurman-Bard ; E. Krahmer ; K. van Deemter. Tilburg, The Netherlands : Tilburg University, 2013. pp. -
@inproceedings{7f3131ed00124717ac22ac090cc600af,
title = "Generation of Dutch referring expressions using the D-TUNA corpus",
abstract = "This paper describes our research into generating Dutch noun phrases as descriptions of furniture objects or people. This is usually done in two steps: attribute selection and realisation. This research focuses only on the realisation step: generating a noun phrase from given attributes. The research is done on the Dutch version of the TUNA-corpus, which contains annotated human-produced descriptions. Three algorithms were developed for this task, each an improvement over the last. We extracted the lexical choice from the D-TUNA corpus, and used templates generated from the corpus which specified the order of attributes. The algorithms were then evaluated for string similarity with the original human descriptions from the corpus, and a human evaluation was carried out which tested clarity and fluency. A steady improvement of scores in both the automatic and human evaluation was observed for each new version of the algorithm.",
keywords = "EWI-23554, IR-87092, METIS-297758",
author = "Marissa Hoek and Mariet Theune",
year = "2013",
month = "7",
language = "Undefined",
isbn = "not assigned",
publisher = "Tilburg University",
pages = "--",
editor = "A. Gatt and {van Gompel}, R. and E. Gurman-Bard and E. Krahmer and {van Deemter}, K.",
booktitle = "Proceedings of PRE-CogSci 2013 – Bridging the Gap between Cognitive and Computational Approaches to Reference",

}

Hoek, M & Theune, M 2013, Generation of Dutch referring expressions using the D-TUNA corpus. in A Gatt, R van Gompel, E Gurman-Bard, E Krahmer & K van Deemter (eds), Proceedings of PRE-CogSci 2013 – Bridging the Gap between Cognitive and Computational Approaches to Reference. Tilburg University, Tilburg, The Netherlands, pp. -, PRE-CogSci 2013 – Bridging the Gap between Cognitive and Computational Approaches to Reference, Berlin, Germany, 31/07/13.

Generation of Dutch referring expressions using the D-TUNA corpus. / Hoek, Marissa; Theune, Mariet.

Proceedings of PRE-CogSci 2013 – Bridging the Gap between Cognitive and Computational Approaches to Reference. ed. / A. Gatt; R. van Gompel; E. Gurman-Bard; E. Krahmer; K. van Deemter. Tilburg, The Netherlands : Tilburg University, 2013. p. -.

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

TY - GEN

T1 - Generation of Dutch referring expressions using the D-TUNA corpus

AU - Hoek, Marissa

AU - Theune, Mariet

PY - 2013/7

Y1 - 2013/7

N2 - This paper describes our research into generating Dutch noun phrases as descriptions of furniture objects or people. This is usually done in two steps: attribute selection and realisation. This research focuses only on the realisation step: generating a noun phrase from given attributes. The research is done on the Dutch version of the TUNA-corpus, which contains annotated human-produced descriptions. Three algorithms were developed for this task, each an improvement over the last. We extracted the lexical choice from the D-TUNA corpus, and used templates generated from the corpus which specified the order of attributes. The algorithms were then evaluated for string similarity with the original human descriptions from the corpus, and a human evaluation was carried out which tested clarity and fluency. A steady improvement of scores in both the automatic and human evaluation was observed for each new version of the algorithm.

AB - This paper describes our research into generating Dutch noun phrases as descriptions of furniture objects or people. This is usually done in two steps: attribute selection and realisation. This research focuses only on the realisation step: generating a noun phrase from given attributes. The research is done on the Dutch version of the TUNA-corpus, which contains annotated human-produced descriptions. Three algorithms were developed for this task, each an improvement over the last. We extracted the lexical choice from the D-TUNA corpus, and used templates generated from the corpus which specified the order of attributes. The algorithms were then evaluated for string similarity with the original human descriptions from the corpus, and a human evaluation was carried out which tested clarity and fluency. A steady improvement of scores in both the automatic and human evaluation was observed for each new version of the algorithm.

KW - EWI-23554

KW - IR-87092

KW - METIS-297758

M3 - Conference contribution

SN - not assigned

SP - -

BT - Proceedings of PRE-CogSci 2013 – Bridging the Gap between Cognitive and Computational Approaches to Reference

A2 - Gatt, A.

A2 - van Gompel, R.

A2 - Gurman-Bard, E.

A2 - Krahmer, E.

A2 - van Deemter, K.

PB - Tilburg University

CY - Tilburg, The Netherlands

ER -

Hoek M, Theune M. Generation of Dutch referring expressions using the D-TUNA corpus. In Gatt A, van Gompel R, Gurman-Bard E, Krahmer E, van Deemter K, editors, Proceedings of PRE-CogSci 2013 – Bridging the Gap between Cognitive and Computational Approaches to Reference. Tilburg, The Netherlands: Tilburg University. 2013. p. -