Generation of Referring Expressions in Large Domains

Roman Kutlak, Kees van Deemter, Christopher Stuart Mellish

Research output: Contribution to conferencePaper

Abstract

In this article, we investigate generation of referring expressions from large knowledge bases. We discuss some of the issues that arise when using existing referring expression generation algorithms and introduce a corpus-based algorithm that aims to overcome these issues. The new algorithm is based on the idea of communal common ground and uses a search engine to estimate what properties of a referent are likely to be known by hearers. The algorithm was evaluated against the Incremental Algorithm and human-created descriptions in a hearer oriented experiment where hearers attempted to identify described people and provide judgements about the used descriptions.

The algorithm outperformed the Incremental Algorithm in terms of the number of correctly identified referents.
Original languageEnglish
Publication statusPublished - 31 Jul 2013
EventPRE-CogSci - Berlin, Germany
Duration: 31 Jul 201331 Jul 2013

Workshop

WorkshopPRE-CogSci
CountryGermany
CityBerlin
Period31/07/1331/07/13

Fingerprint

Search engines
Experiments

Keywords

  • Referring Expression Generation
  • Common Ground
  • Heuristic
  • Search Engine

ASJC Scopus subject areas

  • Human-Computer Interaction

Cite this

Kutlak, R., van Deemter, K., & Mellish, C. S. (2013). Generation of Referring Expressions in Large Domains. Paper presented at PRE-CogSci, Berlin, Germany.

Generation of Referring Expressions in Large Domains. / Kutlak, Roman; van Deemter, Kees; Mellish, Christopher Stuart.

2013. Paper presented at PRE-CogSci, Berlin, Germany.

Research output: Contribution to conferencePaper

Kutlak, R, van Deemter, K & Mellish, CS 2013, 'Generation of Referring Expressions in Large Domains' Paper presented at PRE-CogSci, Berlin, Germany, 31/07/13 - 31/07/13, .
Kutlak R, van Deemter K, Mellish CS. Generation of Referring Expressions in Large Domains. 2013. Paper presented at PRE-CogSci, Berlin, Germany.
Kutlak, Roman ; van Deemter, Kees ; Mellish, Christopher Stuart. / Generation of Referring Expressions in Large Domains. Paper presented at PRE-CogSci, Berlin, Germany.
@conference{603b9c138c57401ebd1152bde45662fb,
title = "Generation of Referring Expressions in Large Domains",
abstract = "In this article, we investigate generation of referring expressions from large knowledge bases. We discuss some of the issues that arise when using existing referring expression generation algorithms and introduce a corpus-based algorithm that aims to overcome these issues. The new algorithm is based on the idea of communal common ground and uses a search engine to estimate what properties of a referent are likely to be known by hearers. The algorithm was evaluated against the Incremental Algorithm and human-created descriptions in a hearer oriented experiment where hearers attempted to identify described people and provide judgements about the used descriptions.The algorithm outperformed the Incremental Algorithm in terms of the number of correctly identified referents.",
keywords = "Referring Expression Generation, Common Ground, Heuristic, Search Engine",
author = "Roman Kutlak and {van Deemter}, Kees and Mellish, {Christopher Stuart}",
year = "2013",
month = "7",
day = "31",
language = "English",
note = "PRE-CogSci ; Conference date: 31-07-2013 Through 31-07-2013",

}

TY - CONF

T1 - Generation of Referring Expressions in Large Domains

AU - Kutlak, Roman

AU - van Deemter, Kees

AU - Mellish, Christopher Stuart

PY - 2013/7/31

Y1 - 2013/7/31

N2 - In this article, we investigate generation of referring expressions from large knowledge bases. We discuss some of the issues that arise when using existing referring expression generation algorithms and introduce a corpus-based algorithm that aims to overcome these issues. The new algorithm is based on the idea of communal common ground and uses a search engine to estimate what properties of a referent are likely to be known by hearers. The algorithm was evaluated against the Incremental Algorithm and human-created descriptions in a hearer oriented experiment where hearers attempted to identify described people and provide judgements about the used descriptions.The algorithm outperformed the Incremental Algorithm in terms of the number of correctly identified referents.

AB - In this article, we investigate generation of referring expressions from large knowledge bases. We discuss some of the issues that arise when using existing referring expression generation algorithms and introduce a corpus-based algorithm that aims to overcome these issues. The new algorithm is based on the idea of communal common ground and uses a search engine to estimate what properties of a referent are likely to be known by hearers. The algorithm was evaluated against the Incremental Algorithm and human-created descriptions in a hearer oriented experiment where hearers attempted to identify described people and provide judgements about the used descriptions.The algorithm outperformed the Incremental Algorithm in terms of the number of correctly identified referents.

KW - Referring Expression Generation

KW - Common Ground

KW - Heuristic

KW - Search Engine

M3 - Paper

ER -