Loading...
Please wait, while we are loading the content...
Similar Documents
Simplify Sentence Structure for Improving Human Post-editing Efficiency on Chinese-to-English Patent Machine Translation
| Content Provider | Semantic Scholar |
|---|---|
| Author | Ren, Xiaona Wei, Yongpeng Hu, Rile Dangdai, Qingyun |
| Copyright Year | 2015 |
| Abstract | In this paper, we propose a new approach to improve human post-editing efficiency by simplifying the sentence structure on Chinese-to-English patent machine translation (PMT). To simplify the structure of a patent sentence, we use a recognizer to recognize the max noun phrases (MNPs) in a Chinese sentence before translating the sentence. The MNPs are replaced with their head words in the sentence, which makes the sentence structure simpler to be translated. Therefore, the task of translating a complicated sentence is transformed into two subtasks: one is the translation task of MNPs and the other is the translation task of the simplified sentence. And then, the translation results of two subtasks are combined to get the final translation result of the input Chinese sentence. This method outperforms NTCIR-10 official baseline by approximate 2 BLEU points. Moreover, the translation results are beneficial for human post-editing, which can save human post-editing time and improve the quality of translation. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://www.mt-archive.info/15/MTS-2015-W2-Ren.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
