UCAM Biomedical translation at WMT19: Transfer learning multi-domain ensembles

06/13/2019

∙

by   Danielle Saunders, et al.

∙

0

∙

share

The 2019 WMT Biomedical translation task involved translating Medline abstracts. We approached this using transfer learning to obtain a series of strong neural models on distinct domains, and combining them into multi-domain ensembles. We further experiment with an adaptive language-model ensemble weighting scheme. Our submission achieved the best submitted results on both directions of English-Spanish.