Neural Machine Translation of Rare Words with Subword Units

ACL 2016  ·  Rico Sennrich, Barry Haddow, Alexandra Birch ·

Neural machine translation (NMT) models typically operate with a fixed vocabulary, but translation is an open-vocabulary problem. Previous work addresses the translation of out-of-vocabulary words by backing off to a dictionary. In this paper, we introduce a simpler and more effective approach, making the NMT model capable of open-vocabulary translation by encoding rare and unknown words as sequences of subword units. This is based on the intuition that various word classes are translatable via smaller units than words, for instance names (via character copying or transliteration), compounds (via compositional translation), and cognates and loanwords (via phonological and morphological transformations). We discuss the suitability of different word segmentation techniques, including simple character n-gram models and a segmentation based on the byte pair encoding compression algorithm, and empirically show that subword models improve over a back-off dictionary baseline for the WMT 15 translation tasks English-German and English-Russian by 1.1 and 1.3 BLEU, respectively.

PDF Abstract ACL 2016 PDF ACL 2016 Abstract

Datasets


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Machine Translation WMT2015 English-German BPE word segmentation BLEU score 22.8 # 4
Machine Translation WMT2015 English-Russian C2-50k Segmentation BLEU score 20.9 # 1

Methods