Towards the Next 1000 Languages in Multilingual Machine Translation: Exploring the Synergy Between Supervised and Self-Supervised Learning

01/09/2022
by   Aditya Siddhant, et al.
8

Achieving universal translation between all human language pairs is the holy-grail of machine translation (MT) research. While recent progress in massively multilingual MT is one step closer to reaching this goal, it is becoming evident that extending a multilingual MT system simply by training on more parallel data is unscalable, since the availability of labeled data for low-resource and non-English-centric language pairs is forbiddingly limited. To this end, we present a pragmatic approach towards building a multilingual MT model that covers hundreds of languages, using a mixture of supervised and self-supervised objectives, depending on the data availability for different language pairs. We demonstrate that the synergy between these two training paradigms enables the model to produce high-quality translations in the zero-resource setting, even surpassing supervised translation quality for low- and mid-resource languages. We conduct a wide array of experiments to understand the effect of the degree of multilingual supervision, domain mismatches and amounts of parallel and monolingual data on the quality of our self-supervised multilingual models. To demonstrate the scalability of the approach, we train models with over 200 languages and demonstrate high performance on zero-resource translation on several previously under-studied languages. We hope our findings will serve as a stepping stone towards enabling translation for the next thousand languages.

READ FULL TEXT

Please sign up or login with your details

Forgot password? Click here to reset

Sign in with Google

×

Use your Google Account to sign in to DeepAI

×

Consider DeepAI Pro