The LMU Munich Unsupervised Machine Translation System for WMT19

Dario Stojanovski, Viktor Hangya, Matthias Huck, Alexander Fraser

Fourth Conference on Machine Translation 2019

We describe LMU Munich's machine translation system for German to
Czech translation which was used to participate in the WMT19 shared
task on unsupervised news translation. We train our model using
monolingual data only from both languages. The final model is an
unsupervised neural model using established techniques for
unsupervised translation such as denoising autoencoding and online
back-translation. We bootstrap the model with masked language model
pretraining and enhance it with back-translations from an unsupervised
phrase-based system which is itself bootstrapped using unsupervised
bilingual word embeddings.