Using Context Vectors in Improving a Machine Translation System with Bridge Language
Samira Tofighi Zahabi, Somayeh Bakhshaei and Shahram Khadivi
The 51st Annual Meeting of the Association for Computational Linguistics - Short Papers (ACL Short Papers 2013)
Sofia, Bulgaria, August 4-9, 2013
Mapping phrases between languages as translation of each other by using an intermediate language (pivot language) when the pivot phrases are equal, may lead to translations that are not similar in context, since a word or phrase has different meaning in different situations. In other words, ploysemous can make wrong translations. Therefore a pruning method based on the context vectors can be applied to determine the concept of each phrase in the source and target languages. This pruning method will delete irrelevant translations. In the following, we will show that by using this pruning method the BLEU of the system will increase about 2.86% BLEU.
Conference Manager (V2.61.0 - Rev. 2792M)