Google Translate is a must-have tool when you’re traveling. It’ll help you learn the local language. It’ll allow you to at least make an attempt at speaking in the native tongue. It’s not always accurate, though, especially if you’re trying to translate long items of text. Google says its researchers have found a fix using “Neural Machine Translation.”
“At a high level, the Neural system translates whole sentences at a time, rather than just piece by piece,” Google explained in a blog post on Wednesday. “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.”
A proper translation instead of a jumble of words
In other words, this should help you get an actual translation instead of a bunch of words that sort of mean the same thing in another language, but not quite. This is often a problem on websites, for example, where you can often translate the entire text of a page in one shot. Some of it makes sense, enough to give you a general idea of the translation, but it’s otherwise a mess.
“Since it’s easier to understand each sentence, translated paragraphs and articles are a lot smoother and easier to read,” Google added “And this is all possible because of end-to-end learning system built on Neural Machine Translation, which basically means that the system learns over time to create better, more natural translations.”
Google is deploying Neural Machine Translation to Google Search on the web, to its Google Translate application and the Google Translate website beginning today. It supports English, French, German, Spanish, Portuguese, Chinese, Japanese, Korean and Turkish to start, but Google says it hopes to one day expand to all of its 103 supported languages.