Word translation based on machine learning models using translation memory and corpora

摘要

The second contest on word sense disambiguation, SENSEVAL-2, was held in Spring, 2001. It consists of several tasks in various languages. In this paper, we describe our system that is used for one of these tasks: the Japanese translation task. In this task, senses of a word are defined in terms of the word's translations. Given an input sentence and a target word in the sentence, our system first estimates the similarity between the input sentence and parallel example sets called translation memory. It then selects an appropriate translation of the target word by using the example set with the highest similarity. The similarity is calculated using dynamic programming and a machine learning model which uses the following features: similarity of a string, words to the left and to the right of the target word in the input sentence, content words in the input sentence and their translations, and co-occurrence of content words in bilingual and monolingual corpora in English and Japanese. Our system achieves an accuracy of 63.4%, finishing the contest in third place among nine systems developed by seven groups.