Bootstrapping statistical parsers from small datasets

Mark Steedman; Miles Osborne; Anoop Sarkar; Stephen Clark; Rebecca Hwa; Julia Hockenmaier; Paul Ruhlen; Steven Baker; Jeremiah Crim

Bootstrapping statistical parsers from small datasets

Mark Steedman, Miles Osborne, Anoop Sarkar, Stephen Clark, Rebecca Hwa, Julia Hockenmaier, Paul Ruhlen, Steven Baker, Jeremiah Crim

Research output: Contribution to conference › Paper › peer-review

Abstract

We present a practical co-training method for bootstrapping statistical parsers using a small amount of manually parsed training material and a much larger pool of raw sentences. Experimental results show that unlabelled sentences can be used to improve the performance of statistical parsers. In addition, we consider the problem of bootstrapping parsers when the manually parsed training material is in a different domain to either the raw sentences or the testing material. We show that bootstrapping continues to be useful, even though no manually produced parses from the target domain are used.

Original language	English (US)
Pages	331-338
Number of pages	8
State	Published - 2003
Externally published	Yes
Event	10th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2003 - Budapest, Hungary Duration: Apr 12 2003 → Apr 17 2003

Conference

Conference	10th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2003
Country/Territory	Hungary
City	Budapest
Period	4/12/03 → 4/17/03

ASJC Scopus subject areas

Language and Linguistics
Linguistics and Language

Library availability

Discover UIUC Full Text

Cite this

@conference{635c62510fdc4bff8076694c0896f7ed,

title = "Bootstrapping statistical parsers from small datasets",

abstract = "We present a practical co-training method for bootstrapping statistical parsers using a small amount of manually parsed training material and a much larger pool of raw sentences. Experimental results show that unlabelled sentences can be used to improve the performance of statistical parsers. In addition, we consider the problem of bootstrapping parsers when the manually parsed training material is in a different domain to either the raw sentences or the testing material. We show that bootstrapping continues to be useful, even though no manually produced parses from the target domain are used.",

author = "Mark Steedman and Miles Osborne and Anoop Sarkar and Stephen Clark and Rebecca Hwa and Julia Hockenmaier and Paul Ruhlen and Steven Baker and Jeremiah Crim",

note = "Publisher Copyright: {\textcopyright} 10th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2003. All rights reserved.; 10th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2003 ; Conference date: 12-04-2003 Through 17-04-2003",

year = "2003",

language = "English (US)",

pages = "331--338",

}

TY - CONF

T1 - Bootstrapping statistical parsers from small datasets

AU - Steedman, Mark

AU - Osborne, Miles

AU - Sarkar, Anoop

AU - Clark, Stephen

AU - Hwa, Rebecca

AU - Hockenmaier, Julia

AU - Ruhlen, Paul

AU - Baker, Steven

AU - Crim, Jeremiah

PY - 2003

Y1 - 2003

N2 - We present a practical co-training method for bootstrapping statistical parsers using a small amount of manually parsed training material and a much larger pool of raw sentences. Experimental results show that unlabelled sentences can be used to improve the performance of statistical parsers. In addition, we consider the problem of bootstrapping parsers when the manually parsed training material is in a different domain to either the raw sentences or the testing material. We show that bootstrapping continues to be useful, even though no manually produced parses from the target domain are used.

AB - We present a practical co-training method for bootstrapping statistical parsers using a small amount of manually parsed training material and a much larger pool of raw sentences. Experimental results show that unlabelled sentences can be used to improve the performance of statistical parsers. In addition, we consider the problem of bootstrapping parsers when the manually parsed training material is in a different domain to either the raw sentences or the testing material. We show that bootstrapping continues to be useful, even though no manually produced parses from the target domain are used.

UR - http://www.scopus.com/inward/record.url?scp=85118305024&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85118305024&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:85118305024

SP - 331

EP - 338

T2 - 10th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2003

Y2 - 12 April 2003 through 17 April 2003

ER -

Bootstrapping statistical parsers from small datasets

Abstract

Conference

ASJC Scopus subject areas

Library availability

Related links

Fingerprint

Cite this