Multi-Source Neural Translation
This work addresses the challenge of enhancing translation accuracy for multilingual applications, though it appears incremental as it builds on existing neural encoder-decoder frameworks.
The paper tackled the problem of improving machine translation by using multiple source languages (French and German) to predict a target English string, resulting in up to a +4.8 Bleu score increase over a strong baseline model.
We build a multi-source machine translation model and train it to maximize the probability of a target English string given French and German sources. Using the neural encoder-decoder framework, we explore several combination methods and report up to +4.8 Bleu increases on top of a very strong attention-based neural translation model.