Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Pre-training Multilingual Neural Machine Translation by Leveraging Alignment Information

About

We investigate the following question for machine translation (MT): can we develop a single universal MT model to serve as the common seed and obtain derivative and improved models on arbitrary language pairs? We propose mRASP, an approach to pre-train a universal multilingual neural machine translation model. Our key idea in mRASP is its novel technique of random aligned substitution, which brings words and phrases with similar meanings across multiple languages closer in the representation space. We pre-train a mRASP model on 32 language pairs jointly with only public datasets. The model is then fine-tuned on downstream language pairs to obtain specialized MT models. We carry out extensive experiments on 42 translation directions across a diverse settings, including low, medium, rich resource, and as well as transferring to exotic language pairs. Experimental results demonstrate that mRASP achieves significant performance improvement compared to directly training on those target pairs. It is the first time to verify that multiple low-resource language pairs can be utilized to improve rich resource MT. Surprisingly, mRASP is even able to improve the translation quality on exotic languages that never occur in the pre-training corpus. Code, data, and pre-trained models are available at https://github.com/linzehui/mRASP.

Zehui Lin, Xiao Pan, Mingxuan Wang, Xipeng Qiu, Jiangtao Feng, Hao Zhou, Lei Li• 2020

Related benchmarks

TaskDatasetResultRank
Machine TranslationWMT En-De 2014 (test)
BLEU30.3
379
Machine TranslationWMT En-Fr 2014 (test)
BLEU45.4
237
Machine TranslationIWSLT De-En 2014 (test)
BLEU30.3
146
Machine TranslationIWSLT En-De 2014 (test)
BLEU23.9
92
Machine TranslationWMT Ro-En 2016 (test)
BLEU36.9
82
Machine TranslationWMT 2016 (test)--
58
Machine TranslationWMT16 EN-RO (test)
BLEU38.9
56
Machine TranslationWMT English-French 2014 (test)
BLEU44.3
41
Machine TranslationWMT newstest 2014
Tokenized BLEU30.3
30
Machine TranslationWMT14 DE-EN (test)
BLEU29.8
28
Showing 10 of 46 rows

Other info

Code

Follow for update