Hierarchical Sub-sentential Alignment with IBM Models for Statistical Phrase-based Machine Translation
Hierarchical Sub-sentential Alignment with IBM Models for Statistical Phrase-based Machine Translation
复制标题
与基于统计短语的机器翻译的 IBM 模型进行分层子句子对齐
DOI：
10.5715/jnlp.24.619
复制
发表时间：
2017
期刊：
影响因子：
--
通讯作者：
Y. Lepage
中科院分区：
文献类型：
--
作者：
H. Wang;Y. Lepage
在本文中，我们描述了一种新的联合词对齐和对称方法。基于简单IBM模型的初始参数，我们在方括号转导语法约束的框架下对平行句对进行同步分析。我们的两阶段方法可以实现与FAST Align几乎相同的运行时间，同时在远距离相关的语言对(如英语-日语)上提供更好的比对。我们展示了如何将该方法集成到基于标准短语的SMT流水线中。尽管对齐结果好坏参半，但通过强制所有单词对齐(1对多/多对1)，我们的方法显著地减少了短语表的大小，在翻译质量上没有差别，在一些端到端的翻译实验中，我们的方法甚至优于fi。
In this paper, we describe a novel method for joint word alignment and symmetrization. Based on initial parameters from simple IBM models, we synchronously parse the parallel sentence pair under the framework of bracket transduction grammar constraints. Our 2-phase method can achieve nearly the same run-time as fast align while delivering better alignments on distantly-related language pairs such as English– Japanese. We show how to integrate this method into a standard phrase-based SMT pipeline. Although the alignment quality results are mixed, by forcing all words to be aligned ( 1-to-many/many-to-1 ), our method significantly reduces the phrase table size with no difference in translation quality and even outperforms fast align in some end-to-end translation experiments.
DOI：
10.5555/972705.972707
发表时间：
1997-09
期刊：
Comput. Linguistics
影响因子：
--
作者：
Dekai Wu
通讯作者：
Dekai Wu