Empirically-based control of natural language generation

Daniel Paiva; Roger Evans

Empirically-based control of natural language generation

Daniel Paiva, Roger Evans

University of Brighton

Research output: Chapter in Book/Conference proceeding with ISSN or ISBN › Conference contribution with ISSN or ISBN › peer-review

Abstract

In this paper we present a new approach to controlling the behaviour of a natural language generation system by correlating internal decisions taken during free generation of a wide range of texts with the surface stylistic characteristics of the resulting outputs, and using the correlation to control the generator. This contrasts with the generate-and-test architecture adopted by most previous empirically-based generation approaches, offering a more efficient, generic and holistic method of generator control. We illustrate the approach by describing a system in which stylistic variation (in the sense of Biber (1988)) can be effectively controlled during the generation of short medical information texts.

Original language	English
Title of host publication	Proceedings of the 43rd Annual Meeting of the ACL
Publisher	Association for Computational Linguistics
Pages	58-65
Number of pages	8
Publication status	Published - Jun 2005
Event	Proceedings of the 43rd Annual Meeting of the ACL - Ann Arbor, USA Duration: 1 Jun 2005 → …

Conference

Conference	Proceedings of the 43rd Annual Meeting of the ACL
Period	1/06/05 → …

Bibliographical note

Creative Commons Attribution-ShareAlike License

Keywords

Natural language generation

Access to Document

Empirically-basedNLG (2).pdfAccepted author manuscript, 180 KBLicence: CC BY-SA

http://acl.ldc.upenn.edu/P/P05/P05-1008.pdfLicence: Unspecified

Cite this

@inproceedings{ddccfa9c2a914b9ea8cbf3ace81cada5,

title = "Empirically-based control of natural language generation",

abstract = "In this paper we present a new approach to controlling the behaviour of a natural language generation system by correlating internal decisions taken during free generation of a wide range of texts with the surface stylistic characteristics of the resulting outputs, and using the correlation to control the generator. This contrasts with the generate-and-test architecture adopted by most previous empirically-based generation approaches, offering a more efficient, generic and holistic method of generator control. We illustrate the approach by describing a system in which stylistic variation (in the sense of Biber (1988)) can be effectively controlled during the generation of short medical information texts.",

keywords = "Natural language generation",

author = "Daniel Paiva and Roger Evans",

note = "Creative Commons Attribution-ShareAlike License; Proceedings of the 43rd Annual Meeting of the ACL ; Conference date: 01-06-2005",

year = "2005",

month = jun,

language = "English",

pages = "58--65",

booktitle = "Proceedings of the 43rd Annual Meeting of the ACL",

publisher = "Association for Computational Linguistics",

}

TY - GEN

T1 - Empirically-based control of natural language generation

AU - Paiva, Daniel

AU - Evans, Roger

N1 - Creative Commons Attribution-ShareAlike License

PY - 2005/6

Y1 - 2005/6

N2 - In this paper we present a new approach to controlling the behaviour of a natural language generation system by correlating internal decisions taken during free generation of a wide range of texts with the surface stylistic characteristics of the resulting outputs, and using the correlation to control the generator. This contrasts with the generate-and-test architecture adopted by most previous empirically-based generation approaches, offering a more efficient, generic and holistic method of generator control. We illustrate the approach by describing a system in which stylistic variation (in the sense of Biber (1988)) can be effectively controlled during the generation of short medical information texts.

AB - In this paper we present a new approach to controlling the behaviour of a natural language generation system by correlating internal decisions taken during free generation of a wide range of texts with the surface stylistic characteristics of the resulting outputs, and using the correlation to control the generator. This contrasts with the generate-and-test architecture adopted by most previous empirically-based generation approaches, offering a more efficient, generic and holistic method of generator control. We illustrate the approach by describing a system in which stylistic variation (in the sense of Biber (1988)) can be effectively controlled during the generation of short medical information texts.

KW - Natural language generation

M3 - Conference contribution with ISSN or ISBN

SP - 58

EP - 65

BT - Proceedings of the 43rd Annual Meeting of the ACL

PB - Association for Computational Linguistics

T2 - Proceedings of the 43rd Annual Meeting of the ACL

Y2 - 1 June 2005

ER -