Abstract | ||
---|---|---|
The NTT Statistical Machine Translation System consists of two primary components: a statistical machine trans- lation decoder and a reranker. The decoder generates k- best translation canditates using a hierarchical phrase-based translation based on synchronous context-free grammar. The decoder employs a linear feature combination among sev- eral real-valued scores on translation and language models. The reranker reorders the k-best translation candidates using Ranking SVMs with a large number of sparse features. This paper describes the two components and presents the results for the evaluation campaign of IWSLT 2008. |
Year | Venue | DocType |
---|---|---|
2010 | IWSLT | Conference |
Citations | PageRank | References |
1 | 0.37 | 7 |
Authors | ||
5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Katsuhito Sudoh | 1 | 326 | 34.44 |
Taro Watanabe | 2 | 572 | 36.86 |
Junichi Suzuki | 3 | 1265 | 112.15 |
Hajime Tsukada | 4 | 449 | 29.46 |
Hideki Isozaki | 5 | 934 | 64.50 |