A single encoder-decoder trained with both decoding directions beats a unidirectional Transformer by 0.8 to 1.3 BLEU and saves about half the parameters of a two-model ensemble.
In WMT 2017, Copenhagen, Denmark, September 7-8
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Efficient Bidirectional Neural Machine Translation
A single encoder-decoder trained with both decoding directions beats a unidirectional Transformer by 0.8 to 1.3 BLEU and saves about half the parameters of a two-model ensemble.