Models for Sentence Compression: A Comparison across Domains, Training Requirements and Evaluation Measures

James Clarke; Mirella Lapata

Models for Sentence Compression: A Comparison across Domains, Training Requirements and Evaluation Measures

School of Informatics

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

Abstract

Sentence compression is the task of producing a summary at the sentence
level. This paper focuses on three aspects of this task which have not received detailed treatment in the literature: training requirements, scalability, and automatic evaluation. We provide a novel comparison between a supervised constituent-based and an weakly supervised word-based compression algorithm and examine how these models port to different domains (written vs. spoken text). To achieve this, a human-authored compression corpus has been created and our study high-lights potential problems with the automatically gathered compression corpora currently used. Finally, we assess whether automatic evaluation measures can be used to determine compression quality.

Original language	English
Title of host publication	ACL 2006, 21st International Conference on Computational Linguistics and 44th Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference, Sydney, Australia, 17-21 July 2006
Publisher	Association for Computational Linguistics
Pages	377-384
Number of pages	8
Publication status	Published - 2006

Access to Document

2006_Clarke_Lapata_Models for Sentence Compression A Comparision across Domains, Training Requirements and Evaluation MeasuresFinal published version, 146 KB

http://aclweb.org/anthology/P06-1048

Statistical model for text-to-text generation
Lapata, M.
EPSRC
1/02/05 → 28/02/11
Project: Research

Cite this

Clarke, J., & Lapata, M. (2006). Models for Sentence Compression: A Comparison across Domains, Training Requirements and Evaluation Measures. In ACL 2006, 21st International Conference on Computational Linguistics and 44th Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference, Sydney, Australia, 17-21 July 2006 (pp. 377-384). Association for Computational Linguistics. http://aclweb.org/anthology/P06-1048

Clarke, James ; Lapata, Mirella. / Models for Sentence Compression: A Comparison across Domains, Training Requirements and Evaluation Measures. ACL 2006, 21st International Conference on Computational Linguistics and 44th Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference, Sydney, Australia, 17-21 July 2006. Association for Computational Linguistics, 2006. pp. 377-384

@inproceedings{4f4e62031e1f4946a5de7deaa7260296,

title = "Models for Sentence Compression: A Comparison across Domains, Training Requirements and Evaluation Measures",

abstract = "Sentence compression is the task of producing a summary at the sentencelevel. This paper focuses on three aspects of this task which have not received detailed treatment in the literature: training requirements, scalability, and automatic evaluation. We provide a novel comparison between a supervised constituent-based and an weakly supervised word-based compression algorithm and examine how these models port to different domains (written vs. spoken text). To achieve this, a human-authored compression corpus has been created and our study high-lights potential problems with the automatically gathered compression corpora currently used. Finally, we assess whether automatic evaluation measures can be used to determine compression quality.",

author = "James Clarke and Mirella Lapata",

year = "2006",

language = "English",

pages = "377--384",

booktitle = "ACL 2006, 21st International Conference on Computational Linguistics and 44th Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference, Sydney, Australia, 17-21 July 2006",

publisher = "Association for Computational Linguistics",

}

Clarke, J & Lapata, M 2006, Models for Sentence Compression: A Comparison across Domains, Training Requirements and Evaluation Measures. in ACL 2006, 21st International Conference on Computational Linguistics and 44th Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference, Sydney, Australia, 17-21 July 2006. Association for Computational Linguistics, pp. 377-384. <http://aclweb.org/anthology/P06-1048>

Models for Sentence Compression: A Comparison across Domains, Training Requirements and Evaluation Measures. / Clarke, James; Lapata, Mirella.
ACL 2006, 21st International Conference on Computational Linguistics and 44th Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference, Sydney, Australia, 17-21 July 2006. Association for Computational Linguistics, 2006. p. 377-384.

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

TY - GEN

T1 - Models for Sentence Compression: A Comparison across Domains, Training Requirements and Evaluation Measures

AU - Clarke, James

AU - Lapata, Mirella

PY - 2006

Y1 - 2006

N2 - Sentence compression is the task of producing a summary at the sentencelevel. This paper focuses on three aspects of this task which have not received detailed treatment in the literature: training requirements, scalability, and automatic evaluation. We provide a novel comparison between a supervised constituent-based and an weakly supervised word-based compression algorithm and examine how these models port to different domains (written vs. spoken text). To achieve this, a human-authored compression corpus has been created and our study high-lights potential problems with the automatically gathered compression corpora currently used. Finally, we assess whether automatic evaluation measures can be used to determine compression quality.

AB - Sentence compression is the task of producing a summary at the sentencelevel. This paper focuses on three aspects of this task which have not received detailed treatment in the literature: training requirements, scalability, and automatic evaluation. We provide a novel comparison between a supervised constituent-based and an weakly supervised word-based compression algorithm and examine how these models port to different domains (written vs. spoken text). To achieve this, a human-authored compression corpus has been created and our study high-lights potential problems with the automatically gathered compression corpora currently used. Finally, we assess whether automatic evaluation measures can be used to determine compression quality.

M3 - Conference contribution

SP - 377

EP - 384

BT - ACL 2006, 21st International Conference on Computational Linguistics and 44th Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference, Sydney, Australia, 17-21 July 2006

PB - Association for Computational Linguistics

ER -

Clarke J, Lapata M. Models for Sentence Compression: A Comparison across Domains, Training Requirements and Evaluation Measures. In ACL 2006, 21st International Conference on Computational Linguistics and 44th Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference, Sydney, Australia, 17-21 July 2006. Association for Computational Linguistics. 2006. p. 377-384

Models for Sentence Compression: A Comparison across Domains, Training Requirements and Evaluation Measures

Abstract

Access to Document

Fingerprint

Projects

Statistical model for text-to-text generation

Cite this