Name
Title | ||
|---|---|---|
Extracting Pre-ordering Rules from Chunk-based Dependency Trees for Japanese-to-English Translation. |
Abstract | ||
|---|---|---|
Word ordering remains as an essential problem for translating between languages with substantial structural differences, such as SOV and SVO languages. In this paper, we propose to use source chunk-based dependency trees to extract pre-ordering rules, which are customized for translating from Japanese (SOV) to English (SVO). In order to obtain a finegrained classification of the reordering phenomena, Japanese function words and punctuation marks are included in the pre-ordering rules as additional lexical clues. We further prune the pre-ordering rule set by referring the predicate-argument structures of the target language side. Experimental results are reported for large-scale Japanese-to-English translation, showing a significant improvement of 1.48 BLEU points compared with the baseline SMT system. |
Year | Venue | DocType |
|---|---|---|
2011 | MTSummit | Conference |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
5 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Xianchao Wu | 1 | 1 | 3.73 |
| Katsuhito Sudoh | 2 | 326 | 34.44 |
| Kevin Duh | 3 | 819 | 72.94 |
| Hajime Tsukada | 4 | 449 | 29.46 |
| Masaaki Nagata | 5 | 573 | 77.86 |