Incorporating Word and Subword Units in Unsupervised Machine Translation Using Language Model Rescoring
This work addresses the challenge of unsupervised translation for morphologically rich languages, though it is incremental as it builds on existing methods like BPE and MUSE.
The paper tackled unsupervised machine translation from German to Czech by combining word-level and subword-level neural models with a phrase-based statistical model and language model rescoring, achieving results in the WMT'19 shared task without parallel data.
This paper describes CAiRE's submission to the unsupervised machine translation track of the WMT'19 news shared task from German to Czech. We leverage a phrase-based statistical machine translation (PBSMT) model and a pre-trained language model to combine word-level neural machine translation (NMT) and subword-level NMT models without using any parallel data. We propose to solve the morphological richness problem of languages by training byte-pair encoding (BPE) embeddings for German and Czech separately, and they are aligned using MUSE (Conneau et al., 2018). To ensure the fluency and consistency of translations, a rescoring mechanism is proposed that reuses the pre-trained language model to select the translation candidates generated through beam search. Moreover, a series of pre-processing and post-processing approaches are applied to improve the quality of final translations.