Estimating word-level quality of statistical machine translation output using monolingual information alone
- Author
- Arda Tezcan (UGent) , Veronique Hoste (UGent) and Lieve Macken (UGent)
- Organization
- Abstract
- Various studies show that statistical machine translation (SMT) systems suffer from fluency errors, especially in the form of grammatical errors and errors related to idiomatic word choices. In this study, we investigate the effectiveness of using monolingual information contained in the machine-translated text to estimate word-level quality of SMT output.We propose a recurrent neural network architecture which uses morpho-syntactic features and word embeddings as word representations within surface and syntactic n-grams.We test the proposed method on two language pairs and for two tasks, namely detecting fluency errors and predicting overall post-editing effort. Our results show that this method is effective for capturing all types of fluency errors at once. Moreover, on the task of predicting post-editing effort, while solely relying on monolingual information, it achieves on-par results with the state-of-the-art quality estimation systems which use both bilingual and monolingual information.
- Keywords
- Machine translation, Quality estimation, Post-editing, Neural Networks, lt3
Downloads
-
(...).pdf
- full text (Published version)
- |
- UGent only
- |
- |
- 732.95 KB
Citation
Please use this url to cite or link to this publication: http://hdl.handle.net/1854/LU-8606927
- MLA
- Tezcan, Arda, et al. “Estimating Word-Level Quality of Statistical Machine Translation Output Using Monolingual Information Alone.” NATURAL LANGUAGE ENGINEERING, vol. 26, no. 1, 2020, pp. 73–94, doi:10.1017/S1351324919000111.
- APA
- Tezcan, A., Hoste, V., & Macken, L. (2020). Estimating word-level quality of statistical machine translation output using monolingual information alone. NATURAL LANGUAGE ENGINEERING, 26(1), 73–94. https://doi.org/10.1017/S1351324919000111
- Chicago author-date
- Tezcan, Arda, Veronique Hoste, and Lieve Macken. 2020. “Estimating Word-Level Quality of Statistical Machine Translation Output Using Monolingual Information Alone.” NATURAL LANGUAGE ENGINEERING 26 (1): 73–94. https://doi.org/10.1017/S1351324919000111.
- Chicago author-date (all authors)
- Tezcan, Arda, Veronique Hoste, and Lieve Macken. 2020. “Estimating Word-Level Quality of Statistical Machine Translation Output Using Monolingual Information Alone.” NATURAL LANGUAGE ENGINEERING 26 (1): 73–94. doi:10.1017/S1351324919000111.
- Vancouver
- 1.Tezcan A, Hoste V, Macken L. Estimating word-level quality of statistical machine translation output using monolingual information alone. NATURAL LANGUAGE ENGINEERING. 2020;26(1):73–94.
- IEEE
- [1]A. Tezcan, V. Hoste, and L. Macken, “Estimating word-level quality of statistical machine translation output using monolingual information alone,” NATURAL LANGUAGE ENGINEERING, vol. 26, no. 1, pp. 73–94, 2020.
@article{8606927,
abstract = {{Various studies show that statistical machine translation (SMT) systems suffer from fluency errors, especially in the form of grammatical errors and errors related to idiomatic word choices. In this study, we
investigate the effectiveness of using monolingual information contained in the machine-translated text
to estimate word-level quality of SMT output.We propose a recurrent neural network architecture which
uses morpho-syntactic features and word embeddings as word representations within surface and syntactic n-grams.We test the proposed method on two language pairs and for two tasks, namely detecting fluency errors and predicting overall post-editing effort. Our results show that this method is effective for capturing all types of fluency errors at once. Moreover, on the task of predicting post-editing effort, while solely relying on monolingual information, it achieves on-par results with the state-of-the-art quality estimation systems which use both bilingual and monolingual information.}},
author = {{Tezcan, Arda and Hoste, Veronique and Macken, Lieve}},
issn = {{1351-3249}},
journal = {{NATURAL LANGUAGE ENGINEERING}},
keywords = {{Machine translation,Quality estimation,Post-editing,Neural Networks,lt3}},
language = {{eng}},
number = {{1}},
pages = {{73--94}},
title = {{Estimating word-level quality of statistical machine translation output using monolingual information alone}},
url = {{http://doi.org/10.1017/S1351324919000111}},
volume = {{26}},
year = {{2020}},
}
- Altmetric
- View in Altmetric
- Web of Science
- Times cited: