Massively Multilingual Neural Machine Translation in the Wild: Findings and Challenges
This work addresses the challenge of scalable and practical machine translation for a wide range of languages, including low-resource ones, though it is incremental in advancing towards a universal system.
The authors tackled the problem of building a universal neural machine translation system by developing a single model that handles 103 languages trained on over 25 billion examples, resulting in significant improvements for low-resource languages while maintaining competitive quality for high-resource languages.
We introduce our efforts towards building a universal neural machine translation (NMT) system capable of translating between any language pair. We set a milestone towards this goal by building a single massively multilingual NMT model handling 103 languages trained on over 25 billion examples. Our system demonstrates effective transfer learning ability, significantly improving translation quality of low-resource languages, while keeping high-resource language translation quality on-par with competitive bilingual baselines. We provide in-depth analysis of various aspects of model building that are crucial to achieving quality and practicality in universal NMT. While we prototype a high-quality universal translation system, our extensive empirical analysis exposes issues that need to be further addressed, and we suggest directions for future research.