Neural Machine Translation with Word Predictions

Abstract

In the encoder-decoder architecture for neural machine translation (NMT), the hidden states of the recurrent structures in the encoder and decoder carry the crucial information about the sentence.These vectors are generated by parameters which are updated by back-propagation of translation errors through time. We argue that propagating errors through the end-to-end recurrent structures are not a direct way of control the hidden vectors. In this paper, we propose to use word predictions as a mechanism for direct supervision. More specifically, we require these vectors to be able to predict the vocabulary in target sentence. Our simple mechanism ensures better representations in the encoder and decoder without using any extra data or annotation. It is also helpful in reducing the target side vocabulary and improving the decoding efficiency. Experiments on Chinese-English and German-English machine translation tasks show BLEU improvements by 4.53 and 1.3, respectively

arXiv commentAccepted at EMNLP2017

References in corpus (4)

• Neural Machine Translation by Jointly Learning to Align and Translate
• Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling
• ADADELTA: An Adaptive Learning Rate Method
• Vocabulary Selection Strategies for Neural Machine Translation

Cited by in corpus (7)

• Source-side Prediction for Neural Headline Generation
• Why Do Neural Response Generation Models Prefer Universal Replies?
• Learning to Discriminate Noises for Incorporating External Information in Neural Machine Translation
• Modeling Future Cost for Neural Machine Translation
• Modeling Past and Future for Neural Machine Translation
• Accelerated Reinforcement Learning for Sentence Generation by Vocabulary Prediction
• GRET: Global Representation Enhanced Transformer