Oh me of little faith; It turns out that Google hasn’t been dragging its feet when it comes to bringing its latest translation advances to Bengali.
The language is one of nine Indian tongues, alongside Hindi, Marathi, Tamil, Telugu, Gujarati, Punjabi, Malayalam and Kannada, to benefit from the search giant’s ‘neural machine translation’ approach.
Barak Turovsky, group product manager at Google Translate, says the new technique improves translation quality more in a single jump than has seen in the last ten years. He explains:
Neural translation is a lot better than our old phrase-based system, translating full sentences at a time, instead of pieces of a sentence. It uses this broader context to help it figure out the most relevant translation, which it then rearranges and adjusts to be more like a human speaking with proper grammar.
Just like it’s easier to learn a language when you already know a related language, we’ve discovered that our neural technology speaks each language better when it learns several at a time.
For example, we have a whole lot more sample data for Hindi than its relatives Marathi and Bengali, but when we train them all together, the translations for all improve more than if we’d trained each individually.
Alongside that step Google is also adding more Indian languages to its Gboard keyboard app and, in collaboration with Oxford University Press, a new online Hindi dictionary.