BERT-fused NMT

Reported on 4 benchmarks across 1 task · 1 paper · 2 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Natural Language Processing4 results

Machine TranslationonWMT2014 English-French
BLEU· 2020-02-17
38.27
SOTA
Incorporating BERT into Neural Machine Translation arXiv:2002.06823
Machine TranslationonWMT2016 English--Romanian
BLEU· 2020-02-17
36.02
SOTA
Incorporating BERT into Neural Machine Translation arXiv:2002.06823
Machine TranslationonWMT2014 English-German
BLEU score· 2020-02-17
30.75
best: 35.14 (Transformer Cycle (Rev))
Incorporating BERT into Neural Machine Translation arXiv:2002.06823
Machine TranslationonWMT2014 English-French
BLEU score· uses extra data· 2020-02-17
43.78
best: 46.4 (Transformer+BT (ADMIN init))
Incorporating BERT into Neural Machine Translation arXiv:2002.06823