Incorporating Discrete Translation Lexicons into Neural Machine Translation
Citations Over TimeTop 1% of 2016 papers
Abstract
Neural machine translation (NMT) often makes mistakes in translating low-frequency content words that are essential to understanding the meaning of the sentence. We propose a method to alleviate this problem by augmenting NMT systems with discrete translation lexicons that efficiently encode translations of these low-frequency words. We describe a method to calculate the lexicon probability of the next word in the translation candidate by using the attention vector of the NMT model to select which source word lexical probabilities the model should focus on. We test two methods to combine this probability with the standard NMT probability: (1) using it as a bias, and (2) linear interpolation. Experiments on two corpora show an improvement of 2.0-2.3 BLEU and 0.13-0.44 NIST score, and faster convergence time.
Related Papers
- → ParaMT: A Paraphraser for Machine Translation(2008)13 cited
- A Low Cost Machine Translation Method for Cross-Lingual Information Retrieval(2008)
- → Machine translation using recursive chain‐link‐type learning based on translation examples(2004)3 cited
- → Automatic Post-Editing Method Using Translation Knowledge Based on Intuitive Common Parts Continuum for Statistical Machine Translation(2014)