Since 2007, Google Translate has been translating text using statistical machine translation — that is to say, generating translations based on patterns found on large amounts of text in books, documents, websites, and other sources. And it has so far done a great job, bar the occasional awkward translations of uncommon phrases and sentences. But Google apparently knows full well that it can do even better with the service, and so it has begun implementing a more advanced method called neural machine translation.
Neural machine translation is so named because it draws on a large artificial neural network to optimize translations. Unlike the phrase-based statistical machine translation, it’s designed to be able to translate entire sentences at a time, rather than translating one phrase or subcomponent after another. As a result, translations come out more accurate and natural-sounding.
At a high level, the Neural system translates whole sentences at a time, rather than just piece by piece. 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. Since it’s easier to understand each sentence, translated paragraphs and articles are a lot smoother and easier to read.
What’s more, neural machine translation entails machine learning, which means that with continuous use the system can improve itself on its own without being explicitly programmed.
Google intends to make neural machine translation work with all of the 103 languages that Google Translate supports. But initially, the system is made available for translations between English and French, German, Spanish, Portuguese, Chinese, Japanese, Korean, or Turkish. (According to Google, these represent the native languages of around one-third of the world’s population, covering more than 35 percent of all Google Translate queries.)