New paradigm for machine translation: How the neural machine translation works

DOI Web Site
  • NAKAZAWA Toshiaki
    Japan Science and Technology Agency, Department of Information Planning Kyoto University, Graduate School of Informatics

Bibliographic Information

Other Title
  • 機械翻訳の新しいパラダイム：ニューラル機械翻訳の原理
  • キカイ ホンヤク ノ アタラシイ パラダイム : ニューラル キカイ ホンヤク ノ ゲンリ

Search this article

Description

<p>A big paradigm shift has happened in the research field of machine translation: the main approach has switched from statistical machine translation (SMT) to neural machine translation (NMT), leading to dramatic improvements in translation quality. The types of translation errors produced in NMT are different from those in SMT, and some of them are unique to NMT. In this article, we briefly look back on the history of the machine translation research and review the mechanism of SMT. After that we explain the mechanism of NMT and compare it with SMT. NMT is basically composed of three parts: encoder, attention mechanism and decoder. We explain the details of each of the three components. The characteristics of each translation method are clarified and the pros and cons are mentioned. We also introduce an example application of NMT in Japan Science and Technology Agency (JST), and draw our conclusions.</p>

Journal

Keywords

Details 詳細情報について

Report a problem

Back to top