Identifying Computer-Translated Paragraphs using Coherence Features

Hoang-Quoc  Nguyen-Son; Ngoc-Dung  T. Tieu; Huy H. Nguyen; Junichi Yamagishi; Isao Echizen

Identifying Computer-Translated Paragraphs using Coherence Features

Hoang-Quoc Nguyen-Son, Ngoc-Dung T. Tieu, Huy H. Nguyen, Junichi Yamagishi, Isao Echizen

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

Abstract / Description of output

We have developed a method for extracting the coherence features from a paragraph by matching similar words in its sentences. We conducted an experiment with a parallel German corpus containing 2000 human-created and 2000 machine-translated paragraphs. The result showed that our method achieved the best performance (accuracy = 72.3%, equal error rate = 29.8%) when it is compared with previous methods on various computer-generated text including translation and paper generation (best accuracy = 67.9%, equal error rate = 32.0%). Experiments on Dutch, another rich resource language, and a low resource one (Japanese) attained similar performances. It demonstrated the efficiency of the coherence features at distinguishing computer-translated from human-created paragraphs on diverse languages.

Original language	English
Title of host publication	Proceedings of the 32nd Pacific Asia Conference on Language, Information and Computation (PACLIC 32)
Place of Publication	Hung Hom, Kowloon Hong Kong
Publisher	Association for Computational Linguistics (ACL)
Number of pages	9
Publication status	E-pub ahead of print - 3 Dec 2018
Event	32nd Pacific Asia Conference on Language, Information and Computation - Kowloon, Hong Kong Duration: 1 Dec 2018 → 3 Dec 2018 http://www.cbs.polyu.edu.hk/2018paclic/

Conference

Conference	32nd Pacific Asia Conference on Language, Information and Computation
Abbreviated title	PACLIC 32
Country/Territory	Hong Kong
City	Kowloon
Period	1/12/18 → 3/12/18
Internet address	http://www.cbs.polyu.edu.hk/2018paclic/

Access to Document

Identifying computer-translated paragraphs using coherence featuresAccepted author manuscript, 490 KB

https://www.aclweb.org/anthology/papers/Y/Y18/Y18-1056/

Cite this

Nguyen-Son, H.-Q., T. Tieu, N.-D., H. Nguyen, H., Yamagishi, J., & Echizen, I. (2018). Identifying Computer-Translated Paragraphs using Coherence Features. In Proceedings of the 32nd Pacific Asia Conference on Language, Information and Computation (PACLIC 32) Association for Computational Linguistics (ACL). Advance online publication. https://www.aclweb.org/anthology/papers/Y/Y18/Y18-1056/

@inproceedings{4f22e8f7115f4115856086d94dd2b092,

title = "Identifying Computer-Translated Paragraphs using Coherence Features",

abstract = "We have developed a method for extracting the coherence features from a paragraph by matching similar words in its sentences. We conducted an experiment with a parallel German corpus containing 2000 human-created and 2000 machine-translated paragraphs. The result showed that our method achieved the best performance (accuracy = 72.3%, equal error rate = 29.8%) when it is compared with previous methods on various computer-generated text including translation and paper generation (best accuracy = 67.9%, equal error rate = 32.0%). Experiments on Dutch, another rich resource language, and a low resource one (Japanese) attained similar performances. It demonstrated the efficiency of the coherence features at distinguishing computer-translated from human-created paragraphs on diverse languages.",

author = "Hoang-Quoc Nguyen-Son and {T. Tieu}, Ngoc-Dung and {H. Nguyen}, Huy and Junichi Yamagishi and Isao Echizen",

year = "2018",

month = dec,

day = "3",

language = "English",

booktitle = "Proceedings of the 32nd Pacific Asia Conference on Language, Information and Computation (PACLIC 32)",

publisher = "Association for Computational Linguistics (ACL)",

note = "32nd Pacific Asia Conference on Language, Information and Computation, PACLIC 32 ; Conference date: 01-12-2018 Through 03-12-2018",

url = "http://www.cbs.polyu.edu.hk/2018paclic/",

}

Nguyen-Son, H-Q, T. Tieu, N-D, H. Nguyen, H, Yamagishi, J & Echizen, I 2018, Identifying Computer-Translated Paragraphs using Coherence Features. in Proceedings of the 32nd Pacific Asia Conference on Language, Information and Computation (PACLIC 32). Association for Computational Linguistics (ACL), Hung Hom, Kowloon Hong Kong, 32nd Pacific Asia Conference on Language, Information and Computation, Kowloon, Hong Kong, 1/12/18. <https://www.aclweb.org/anthology/papers/Y/Y18/Y18-1056/>

Identifying Computer-Translated Paragraphs using Coherence Features. / Nguyen-Son, Hoang-Quoc ; T. Tieu, Ngoc-Dung ; H. Nguyen, Huy et al.
Proceedings of the 32nd Pacific Asia Conference on Language, Information and Computation (PACLIC 32). Hung Hom, Kowloon Hong Kong: Association for Computational Linguistics (ACL), 2018.

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

TY - GEN

T1 - Identifying Computer-Translated Paragraphs using Coherence Features

AU - Nguyen-Son, Hoang-Quoc

AU - T. Tieu, Ngoc-Dung

AU - H. Nguyen, Huy

AU - Yamagishi, Junichi

AU - Echizen, Isao

PY - 2018/12/3

Y1 - 2018/12/3

N2 - We have developed a method for extracting the coherence features from a paragraph by matching similar words in its sentences. We conducted an experiment with a parallel German corpus containing 2000 human-created and 2000 machine-translated paragraphs. The result showed that our method achieved the best performance (accuracy = 72.3%, equal error rate = 29.8%) when it is compared with previous methods on various computer-generated text including translation and paper generation (best accuracy = 67.9%, equal error rate = 32.0%). Experiments on Dutch, another rich resource language, and a low resource one (Japanese) attained similar performances. It demonstrated the efficiency of the coherence features at distinguishing computer-translated from human-created paragraphs on diverse languages.

AB - We have developed a method for extracting the coherence features from a paragraph by matching similar words in its sentences. We conducted an experiment with a parallel German corpus containing 2000 human-created and 2000 machine-translated paragraphs. The result showed that our method achieved the best performance (accuracy = 72.3%, equal error rate = 29.8%) when it is compared with previous methods on various computer-generated text including translation and paper generation (best accuracy = 67.9%, equal error rate = 32.0%). Experiments on Dutch, another rich resource language, and a low resource one (Japanese) attained similar performances. It demonstrated the efficiency of the coherence features at distinguishing computer-translated from human-created paragraphs on diverse languages.

M3 - Conference contribution

BT - Proceedings of the 32nd Pacific Asia Conference on Language, Information and Computation (PACLIC 32)

PB - Association for Computational Linguistics (ACL)

CY - Hung Hom, Kowloon Hong Kong

T2 - 32nd Pacific Asia Conference on Language, Information and Computation

Y2 - 1 December 2018 through 3 December 2018

ER -

Identifying Computer-Translated Paragraphs using Coherence Features

Abstract / Description of output

Conference

Access to Document

Fingerprint

Cite this