Generating Non-Projective Word Order in Statistical Linearization

Bernd Bohnet; Jonas Kuhn; Wolfgang Seeker; Anders Björkelund

Generating Non-Projective Word Order in Statistical Linearization

Bernd Bohnet, Jonas Kuhn, Wolfgang Seeker, Anders Björkelund

Computer Science

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

5 Citations (Scopus)

Abstract

We propose a technique to generate nonprojective word orders in an efﬁcient statistical linearization system. Our approach predicts liftings of edges in an unordered syntactic tree by means of a classiﬁer, and uses a
projective algorithm for tree linearization. We
obtain statistically signiﬁcant improvements
on six typologically different languages: English, German, Dutch, Danish, Hungarian, and
Czech.

Original language	English
Title of host publication	Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning - Proceedings
Publisher	Association for Computational Linguistics, ACL
Pages	928-939
Number of pages	11
ISBN (Print)	978-1-937284-43-5
Publication status	Published - 2012

Access to Document

http://www.aclweb.org/anthology-new/D/D12/D12-1085.pdf?CFID=222971427&CFTOKEN=12587301

Cite this

Bohnet, B., Kuhn, J., Seeker, W., & Björkelund, A. (2012). Generating Non-Projective Word Order in Statistical Linearization. In Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning - Proceedings (pp. 928-939). Association for Computational Linguistics, ACL. http://www.aclweb.org/anthology-new/D/D12/D12-1085.pdf?CFID=222971427&CFTOKEN=12587301

@inproceedings{0e2dc94664c348c8840829eeb12e5a3a,

title = "Generating Non-Projective Word Order in Statistical Linearization",

abstract = "We propose a technique to generate nonprojective word orders in an efﬁcient statistical linearization system. Our approach predicts liftings of edges in an unordered syntactic tree by means of a classiﬁer, and uses aprojective algorithm for tree linearization. Weobtain statistically signiﬁcant improvementson six typologically different languages: English, German, Dutch, Danish, Hungarian, andCzech.",

author = "Bernd Bohnet and Jonas Kuhn and Wolfgang Seeker and Anders Bj{\"o}rkelund",

year = "2012",

language = "English",

isbn = "978-1-937284-43-5",

pages = "928--939",

booktitle = "Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning - Proceedings",

publisher = "Association for Computational Linguistics, ACL",

}

Bohnet, B, Kuhn, J, Seeker, W & Björkelund, A 2012, Generating Non-Projective Word Order in Statistical Linearization. in Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning - Proceedings. Association for Computational Linguistics, ACL, pp. 928-939. <http://www.aclweb.org/anthology-new/D/D12/D12-1085.pdf?CFID=222971427&CFTOKEN=12587301>

Generating Non-Projective Word Order in Statistical Linearization. / Bohnet, Bernd; Kuhn, Jonas; Seeker, Wolfgang et al.
Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning - Proceedings. Association for Computational Linguistics, ACL, 2012. p. 928-939.

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

TY - GEN

T1 - Generating Non-Projective Word Order in Statistical Linearization

AU - Bohnet, Bernd

AU - Kuhn, Jonas

AU - Seeker, Wolfgang

AU - Björkelund, Anders

PY - 2012

Y1 - 2012

N2 - We propose a technique to generate nonprojective word orders in an efﬁcient statistical linearization system. Our approach predicts liftings of edges in an unordered syntactic tree by means of a classiﬁer, and uses aprojective algorithm for tree linearization. Weobtain statistically signiﬁcant improvementson six typologically different languages: English, German, Dutch, Danish, Hungarian, andCzech.

AB - We propose a technique to generate nonprojective word orders in an efﬁcient statistical linearization system. Our approach predicts liftings of edges in an unordered syntactic tree by means of a classiﬁer, and uses aprojective algorithm for tree linearization. Weobtain statistically signiﬁcant improvementson six typologically different languages: English, German, Dutch, Danish, Hungarian, andCzech.

M3 - Conference contribution

SN - 978-1-937284-43-5

SP - 928

EP - 939

BT - Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning - Proceedings

PB - Association for Computational Linguistics, ACL

ER -

Generating Non-Projective Word Order in Statistical Linearization

Abstract

Access to Document

Fingerprint

Cite this