The CASIA machine translation system for IWSLT 2013

Xingyuan Peng, Xiaoyin Fu, Wei Wei, Zhenbiao Chen, Wei Chen, Bo Xu


Abstract
In this paper, we describe the CASIA statistical machine translation (SMT) system for the IWSLT2013 Evaluation Campaign. We participated in the Chinese-English and English-Chinese translation tasks. For both of these tasks, we used a hierarchical phrase-based (HPB) decoder and made it as our baseline translation system. A number of techniques were proposed to deal with these translation tasks, including parallel sentence extraction, pre-processing, translation model (TM) optimization, language model (LM) interpolation, turning, and post-processing. With these techniques, the translation results were significantly improved compared with that of the baseline system.
Anthology ID:
2013.iwslt-evaluation.25
Volume:
Proceedings of the 10th International Workshop on Spoken Language Translation: Evaluation Campaign
Month:
December 5-6
Year:
2013
Address:
Heidelberg, Germany
Editor:
Joy Ying Zhang
Venue:
IWSLT
SIG:
SIGSLT
Publisher:
Note:
Pages:
Language:
URL:
https://aclanthology.org/2013.iwslt-evaluation.25
DOI:
Bibkey:
Cite (ACL):
Xingyuan Peng, Xiaoyin Fu, Wei Wei, Zhenbiao Chen, Wei Chen, and Bo Xu. 2013. The CASIA machine translation system for IWSLT 2013. In Proceedings of the 10th International Workshop on Spoken Language Translation: Evaluation Campaign, Heidelberg, Germany.
Cite (Informal):
The CASIA machine translation system for IWSLT 2013 (Peng et al., IWSLT 2013)
Copy Citation:
PDF:
https://aclanthology.org/2013.iwslt-evaluation.25.pdf