GRE3D7: A Corpus of distinguishing descriptions for objects in visual scenes

Jette Viethen; Robert Dale

GRE3D7: A Corpus of distinguishing descriptions for objects in visual scenes

Jette Viethen
, Robert Dale

Research output: Chapter in Book/Report/Conference proceeding › Conference proceeding contribution › peer-review

109 Downloads (Pure)

Abstract

Recent years have seen a trend towards empirically motivated and more data-driven approaches in the field of referring expression generation (REG). Much of this work has focussed on initial reference to objects in visual scenes. While this scenario of use is one of the strongest contenders for real-world applications of referring expression generation, existing data sets still only embody very simple stimulus scenes. To move this research forward, we require data sets built around increasingly complex scenes, and we need much larger data sets to accommodate their higher dimensionality. To control the complexity, we also need to adopt a hypothesis-driven approach to scene design. In this paper, we describe GRE3D7, the largest corpus of humanproduced distinguishing descriptions available to date, discuss the hypotheses that underlie its design, and offer a number of analyses of the 4480 descriptions it contains.

Original language	English
Title of host publication	Proceedings of the UCNLG+Eval
Subtitle of host publication	Language Generation and Evaluation Workshop
Publisher	Association for Computational Linguistics (ACL)
Pages	12-22
Number of pages	11
ISBN (Print)	9781937284183
Publication status	Published - 2011
Event	Workshop on Language Generation and Evaluation - Edinburgh Duration: 31 Jul 2011 → 31 Jul 2011

Workshop

Workshop	Workshop on Language Generation and Evaluation
City	Edinburgh
Period	31/07/11 → 31/07/11

Bibliographical note

Copyright the Publisher 2011. Version archived for private and non-commercial use with the permission of the author/s and according to publisher conditions. For further rights please contact the publisher.

Access to Document

Mq-18199-Publisher version (open access)Final published version, 1.57 MBLicence: Unspecified

Cite this

@inproceedings{6ca0e2d576274e8d9bde8c30c68c1e0d,

title = "GRE3D7: A Corpus of distinguishing descriptions for objects in visual scenes",

abstract = "Recent years have seen a trend towards empirically motivated and more data-driven approaches in the field of referring expression generation (REG). Much of this work has focussed on initial reference to objects in visual scenes. While this scenario of use is one of the strongest contenders for real-world applications of referring expression generation, existing data sets still only embody very simple stimulus scenes. To move this research forward, we require data sets built around increasingly complex scenes, and we need much larger data sets to accommodate their higher dimensionality. To control the complexity, we also need to adopt a hypothesis-driven approach to scene design. In this paper, we describe GRE3D7, the largest corpus of humanproduced distinguishing descriptions available to date, discuss the hypotheses that underlie its design, and offer a number of analyses of the 4480 descriptions it contains.",

author = "Jette Viethen and Robert Dale",

note = "Copyright the Publisher 2011. Version archived for private and non-commercial use with the permission of the author/s and according to publisher conditions. For further rights please contact the publisher.; Workshop on Language Generation and Evaluation ; Conference date: 31-07-2011 Through 31-07-2011",

year = "2011",

language = "English",

isbn = "9781937284183",

pages = "12--22",

booktitle = "Proceedings of the UCNLG+Eval",

publisher = "Association for Computational Linguistics (ACL)",

}

TY - GEN

T1 - GRE3D7

T2 - Workshop on Language Generation and Evaluation

AU - Viethen, Jette

AU - Dale, Robert

N1 - Copyright the Publisher 2011. Version archived for private and non-commercial use with the permission of the author/s and according to publisher conditions. For further rights please contact the publisher.

PY - 2011

Y1 - 2011

N2 - Recent years have seen a trend towards empirically motivated and more data-driven approaches in the field of referring expression generation (REG). Much of this work has focussed on initial reference to objects in visual scenes. While this scenario of use is one of the strongest contenders for real-world applications of referring expression generation, existing data sets still only embody very simple stimulus scenes. To move this research forward, we require data sets built around increasingly complex scenes, and we need much larger data sets to accommodate their higher dimensionality. To control the complexity, we also need to adopt a hypothesis-driven approach to scene design. In this paper, we describe GRE3D7, the largest corpus of humanproduced distinguishing descriptions available to date, discuss the hypotheses that underlie its design, and offer a number of analyses of the 4480 descriptions it contains.

AB - Recent years have seen a trend towards empirically motivated and more data-driven approaches in the field of referring expression generation (REG). Much of this work has focussed on initial reference to objects in visual scenes. While this scenario of use is one of the strongest contenders for real-world applications of referring expression generation, existing data sets still only embody very simple stimulus scenes. To move this research forward, we require data sets built around increasingly complex scenes, and we need much larger data sets to accommodate their higher dimensionality. To control the complexity, we also need to adopt a hypothesis-driven approach to scene design. In this paper, we describe GRE3D7, the largest corpus of humanproduced distinguishing descriptions available to date, discuss the hypotheses that underlie its design, and offer a number of analyses of the 4480 descriptions it contains.

M3 - Conference proceeding contribution

SN - 9781937284183

SP - 12

EP - 22

BT - Proceedings of the UCNLG+Eval

PB - Association for Computational Linguistics (ACL)

Y2 - 31 July 2011 through 31 July 2011

ER -

GRE3D7: A Corpus of distinguishing descriptions for objects in visual scenes

Abstract

Workshop

Bibliographical note

Access to Document

Fingerprint

Cite this