Constrained sequence classification for lexical disambiguation

Tran The Truyen; Dinh Q. Phung; Svetha Venkatesh

doi:10.1007/978-3-540-89197-0_40

Constrained sequence classification for lexical disambiguation

Tran The Truyen, Dinh Q. Phung, Svetha Venkatesh

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

8 Citations (Scopus)

Abstract

This paper addresses lexical ambiguity with focus on a particular problem known as accent prediction, in that given an accentless sequence, we need to restore correct accents. This can be modelled as a sequence classification problem for which variants of Markov chains can be applied. Although the state space is large (about the vocabulary size), it is highly constrained when conditioned on the data observation. We investigate the application of several methods, including Powered Product-of-N-grams, Structured Perceptron and Conditional Random Fields (CRFs). We empirically show in the Vietnamese case that these methods are fairly robust and efficient. The second-order CRFs achieve best results with about 94% term accuracy.

Original language	English
Title of host publication	PRICAI 2008
Subtitle of host publication	Trends in Artificial Intelligence - 10th Pacific Rim International Conference on Artificial Intelligence, Proceedings
Pages	430-441
Number of pages	12
DOIs	https://doi.org/10.1007/978-3-540-89197-0_40
Publication status	Published - 1 Dec 2008
Externally published	Yes
Event	Pacific Rim International Conference on Artificial Intelligence 2008 - Hanoi, Vietnam Duration: 15 Dec 2008 → 19 Dec 2008 Conference number: 10th http://www.pricai.org/conferences/past-conferences/11-pricai-2008-conference.html https://link.springer.com/book/10.1007/978-3-540-89197-0 (Proceedings)

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	5351 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	Pacific Rim International Conference on Artificial Intelligence 2008
Abbreviated title	PRICAI 2008
Country/Territory	Vietnam
City	Hanoi
Period	15/12/08 → 19/12/08
Internet address	http://www.pricai.org/conferences/past-conferences/11-pricai-2008-conference.html https://link.springer.com/book/10.1007/978-3-540-89197-0 (Proceedings)

Keywords

Conditional random fields
Constrained sequence classification
Lexical disambiguation
Vietnamese accent restoration

Access to Document

10.1007/978-3-540-89197-0_40

Cite this

Truyen, T. T., Phung, D. Q., & Venkatesh, S. (2008). Constrained sequence classification for lexical disambiguation. In PRICAI 2008: Trends in Artificial Intelligence - 10th Pacific Rim International Conference on Artificial Intelligence, Proceedings (pp. 430-441). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5351 LNAI). https://doi.org/10.1007/978-3-540-89197-0_40

Truyen, Tran The ; Phung, Dinh Q. ; Venkatesh, Svetha. / Constrained sequence classification for lexical disambiguation. PRICAI 2008: Trends in Artificial Intelligence - 10th Pacific Rim International Conference on Artificial Intelligence, Proceedings. 2008. pp. 430-441 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{d4670e503eab4626859e0754c8e788cc,

title = "Constrained sequence classification for lexical disambiguation",

abstract = "This paper addresses lexical ambiguity with focus on a particular problem known as accent prediction, in that given an accentless sequence, we need to restore correct accents. This can be modelled as a sequence classification problem for which variants of Markov chains can be applied. Although the state space is large (about the vocabulary size), it is highly constrained when conditioned on the data observation. We investigate the application of several methods, including Powered Product-of-N-grams, Structured Perceptron and Conditional Random Fields (CRFs). We empirically show in the Vietnamese case that these methods are fairly robust and efficient. The second-order CRFs achieve best results with about 94% term accuracy.",

keywords = "Conditional random fields, Constrained sequence classification, Lexical disambiguation, Vietnamese accent restoration",

author = "Truyen, {Tran The} and Phung, {Dinh Q.} and Svetha Venkatesh",

year = "2008",

month = dec,

day = "1",

doi = "10.1007/978-3-540-89197-0_40",

language = "English",

isbn = "354089196X",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

pages = "430--441",

booktitle = "PRICAI 2008",

note = "Pacific Rim International Conference on Artificial Intelligence 2008, PRICAI 2008 ; Conference date: 15-12-2008 Through 19-12-2008",

url = "http://www.pricai.org/conferences/past-conferences/11-pricai-2008-conference.html, https://link.springer.com/book/10.1007/978-3-540-89197-0",

}

Truyen, TT, Phung, DQ & Venkatesh, S 2008, Constrained sequence classification for lexical disambiguation. in PRICAI 2008: Trends in Artificial Intelligence - 10th Pacific Rim International Conference on Artificial Intelligence, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 5351 LNAI, pp. 430-441, Pacific Rim International Conference on Artificial Intelligence 2008, Hanoi, Vietnam, 15/12/08. https://doi.org/10.1007/978-3-540-89197-0_40

Constrained sequence classification for lexical disambiguation. / Truyen, Tran The; Phung, Dinh Q.; Venkatesh, Svetha.
PRICAI 2008: Trends in Artificial Intelligence - 10th Pacific Rim International Conference on Artificial Intelligence, Proceedings. 2008. p. 430-441 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5351 LNAI).

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

TY - GEN

T1 - Constrained sequence classification for lexical disambiguation

AU - Truyen, Tran The

AU - Phung, Dinh Q.

AU - Venkatesh, Svetha

N1 - Conference code: 10th

PY - 2008/12/1

Y1 - 2008/12/1

N2 - This paper addresses lexical ambiguity with focus on a particular problem known as accent prediction, in that given an accentless sequence, we need to restore correct accents. This can be modelled as a sequence classification problem for which variants of Markov chains can be applied. Although the state space is large (about the vocabulary size), it is highly constrained when conditioned on the data observation. We investigate the application of several methods, including Powered Product-of-N-grams, Structured Perceptron and Conditional Random Fields (CRFs). We empirically show in the Vietnamese case that these methods are fairly robust and efficient. The second-order CRFs achieve best results with about 94% term accuracy.

AB - This paper addresses lexical ambiguity with focus on a particular problem known as accent prediction, in that given an accentless sequence, we need to restore correct accents. This can be modelled as a sequence classification problem for which variants of Markov chains can be applied. Although the state space is large (about the vocabulary size), it is highly constrained when conditioned on the data observation. We investigate the application of several methods, including Powered Product-of-N-grams, Structured Perceptron and Conditional Random Fields (CRFs). We empirically show in the Vietnamese case that these methods are fairly robust and efficient. The second-order CRFs achieve best results with about 94% term accuracy.

KW - Conditional random fields

KW - Constrained sequence classification

KW - Lexical disambiguation

KW - Vietnamese accent restoration

UR - http://www.scopus.com/inward/record.url?scp=58349091982&partnerID=8YFLogxK

U2 - 10.1007/978-3-540-89197-0_40

DO - 10.1007/978-3-540-89197-0_40

M3 - Conference Paper

AN - SCOPUS:58349091982

SN - 354089196X

SN - 9783540891963

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 430

EP - 441

BT - PRICAI 2008

T2 - Pacific Rim International Conference on Artificial Intelligence 2008

Y2 - 15 December 2008 through 19 December 2008

ER -

Truyen TT, Phung DQ, Venkatesh S. Constrained sequence classification for lexical disambiguation. In PRICAI 2008: Trends in Artificial Intelligence - 10th Pacific Rim International Conference on Artificial Intelligence, Proceedings. 2008. p. 430-441. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-540-89197-0_40

Constrained sequence classification for lexical disambiguation

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Cite this