- Abstract: Most state-of-the-art neural machine translation systems, despite being different in architectural skeletons (e.g., recurrence, convolutional), share an indispensable feature: the Attention. However, most existing attention methods are token-based and ignore the importance of phrasal alignments, the key ingredient for the success of phrase-based statistical machine translation. In this paper, we propose novel phrase-based attention methods to model n-grams of tokens as attention entities. We incorporate our phrase-based attentions into the recently proposed Transformer network, and demonstrate that our approach yields improvements of 1.3 BLEU for English-to-German and 0.5 BLEU for German-to-English translation tasks, and 1.75 and 1.35 BLEU points in English-to-Russian and Russian-to-English translation tasks on WMT newstest2014 using WMT’16 training data.
- Keywords: neural machine translation, natural language processing, attention, transformer, seq2seq, phrase-based, phrase, n-gram
- TL;DR: Phrase-based attention mechanisms to assign attention on phrases, achieving token-to-phrase, phrase-to-token, phrase-to-phrase attention alignments, in addition to existing token-to-token attentions.