Pretreatment for speech machine translation

Xiaofei Zhang; Chong Feng; Heyan Huang

doi:10.1007/978-3-642-16732-4_13

Pretreatment for speech machine translation

Xiaofei Zhang^*, Chong Feng, Heyan Huang

^*Corresponding author for this work

School of Computer Science and Technology

Chinese Academy of Sciences

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

Abstract

There are many meaningless modal particles and dittographes in natural spoken language, furthermore ASR (automatic speech recognition) often has some recognition errors and the ASR results have no punctuations. And thus the translation would be rather poor if the ASR results are directly translated by MT (machine translation). Therefore, it is necessary to transform the abnormal ASR results into normative texts to fit machine translation. In this paper, a pretreatment approach which based on conditional random field model was introduced to delete the meaningless modal particles and dittographes, correct the recognition errors, and punctuated the ASR results before machine translation. Experiments show that the MT BLEU of 0.2497 is obtained, that improved by 18.4% over the MT baseline without pretreatment.

Original language	English
Title of host publication	Computational Collective Intelligence
Subtitle of host publication	Technologies and Applications - Second International Conference, ICCCI 2010, Proceedings
Pages	113-121
Number of pages	9
Edition	PART 2
DOIs	https://doi.org/10.1007/978-3-642-16732-4_13
Publication status	Published - 2010
Event	2nd International Conference on Computational Collective Intelligence - Technologies and Applications, ICCCI 2010 - Kaohsiung, Taiwan, Province of China Duration: 10 Nov 2010 → 12 Nov 2010

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Number	PART 2
Volume	6422 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	2nd International Conference on Computational Collective Intelligence - Technologies and Applications, ICCCI 2010
Country/Territory	Taiwan, Province of China
City	Kaohsiung
Period	10/11/10 → 12/11/10

Keywords

Automatic speech recognition
Conditional random field model
Pretreatment
Speech machine translation

Access to Document

10.1007/978-3-642-16732-4_13

Cite this

Zhang, X., Feng, C., & Huang, H. (2010). Pretreatment for speech machine translation. In Computational Collective Intelligence: Technologies and Applications - Second International Conference, ICCCI 2010, Proceedings (PART 2 ed., pp. 113-121). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6422 LNAI, No. PART 2). https://doi.org/10.1007/978-3-642-16732-4_13

Zhang, Xiaofei ; Feng, Chong ; Huang, Heyan. / Pretreatment for speech machine translation. Computational Collective Intelligence: Technologies and Applications - Second International Conference, ICCCI 2010, Proceedings. PART 2. ed. 2010. pp. 113-121 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2).

@inproceedings{ef6d08e5d60c499fb0a685a6108e33d8,

title = "Pretreatment for speech machine translation",

abstract = "There are many meaningless modal particles and dittographes in natural spoken language, furthermore ASR (automatic speech recognition) often has some recognition errors and the ASR results have no punctuations. And thus the translation would be rather poor if the ASR results are directly translated by MT (machine translation). Therefore, it is necessary to transform the abnormal ASR results into normative texts to fit machine translation. In this paper, a pretreatment approach which based on conditional random field model was introduced to delete the meaningless modal particles and dittographes, correct the recognition errors, and punctuated the ASR results before machine translation. Experiments show that the MT BLEU of 0.2497 is obtained, that improved by 18.4\% over the MT baseline without pretreatment.",

keywords = "Automatic speech recognition, Conditional random field model, Pretreatment, Speech machine translation",

author = "Xiaofei Zhang and Chong Feng and Heyan Huang",

year = "2010",

doi = "10.1007/978-3-642-16732-4\_13",

language = "English",

isbn = "3642167314",

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

number = "PART 2",

pages = "113--121",

booktitle = "Computational Collective Intelligence",

edition = "PART 2",

note = "2nd International Conference on Computational Collective Intelligence - Technologies and Applications, ICCCI 2010 ; Conference date: 10-11-2010 Through 12-11-2010",

}

Zhang, X, Feng, C & Huang, H 2010, Pretreatment for speech machine translation. in Computational Collective Intelligence: Technologies and Applications - Second International Conference, ICCCI 2010, Proceedings. PART 2 edn, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 2, vol. 6422 LNAI, pp. 113-121, 2nd International Conference on Computational Collective Intelligence - Technologies and Applications, ICCCI 2010, Kaohsiung, Taiwan, Province of China, 10/11/10. https://doi.org/10.1007/978-3-642-16732-4_13

Pretreatment for speech machine translation. / Zhang, Xiaofei; Feng, Chong ; Huang, Heyan.
Computational Collective Intelligence: Technologies and Applications - Second International Conference, ICCCI 2010, Proceedings. PART 2. ed. 2010. p. 113-121 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6422 LNAI, No. PART 2).

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

TY - GEN

T1 - Pretreatment for speech machine translation

AU - Zhang, Xiaofei

AU - Feng, Chong

AU - Huang, Heyan

PY - 2010

Y1 - 2010

N2 - There are many meaningless modal particles and dittographes in natural spoken language, furthermore ASR (automatic speech recognition) often has some recognition errors and the ASR results have no punctuations. And thus the translation would be rather poor if the ASR results are directly translated by MT (machine translation). Therefore, it is necessary to transform the abnormal ASR results into normative texts to fit machine translation. In this paper, a pretreatment approach which based on conditional random field model was introduced to delete the meaningless modal particles and dittographes, correct the recognition errors, and punctuated the ASR results before machine translation. Experiments show that the MT BLEU of 0.2497 is obtained, that improved by 18.4% over the MT baseline without pretreatment.

AB - There are many meaningless modal particles and dittographes in natural spoken language, furthermore ASR (automatic speech recognition) often has some recognition errors and the ASR results have no punctuations. And thus the translation would be rather poor if the ASR results are directly translated by MT (machine translation). Therefore, it is necessary to transform the abnormal ASR results into normative texts to fit machine translation. In this paper, a pretreatment approach which based on conditional random field model was introduced to delete the meaningless modal particles and dittographes, correct the recognition errors, and punctuated the ASR results before machine translation. Experiments show that the MT BLEU of 0.2497 is obtained, that improved by 18.4% over the MT baseline without pretreatment.

KW - Automatic speech recognition

KW - Conditional random field model

KW - Pretreatment

KW - Speech machine translation

UR - http://www.scopus.com/inward/record.url?scp=78649621878&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-16732-4_13

DO - 10.1007/978-3-642-16732-4_13

M3 - Conference contribution

AN - SCOPUS:78649621878

SN - 3642167314

SN - 9783642167317

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

SP - 113

EP - 121

BT - Computational Collective Intelligence

T2 - 2nd International Conference on Computational Collective Intelligence - Technologies and Applications, ICCCI 2010

Y2 - 10 November 2010 through 12 November 2010

ER -

Zhang X, Feng C , Huang H. Pretreatment for speech machine translation. In Computational Collective Intelligence: Technologies and Applications - Second International Conference, ICCCI 2010, Proceedings. PART 2 ed. 2010. p. 113-121. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2). doi: 10.1007/978-3-642-16732-4_13

Pretreatment for speech machine translation

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this