Google’s smarter, A.I.-powered translation system expands to more languages
In Google's brilliant intention to make language translation sound more natural, It began rolling from phrase-based machine translation and brought about an AI-powered system called Google Neural Machine Translation (GNMT) in the previous years which takes advantage of broad neural networks to translate entire sentences not just phrases, for significantly improved translations. The company put the system to work in Google Translate for eight language pairs in November and is today expanding support to three more: Russian, Hindi, and Vietnamese.
Also, the company has announced the extension of its neural network integration with Google Translate. The GNMT enables translation to be generated more accurately for keywords and phrases. This new improvement is now being programmed for various new languages.
The newly announced machine-assisted language translations, Google Neural Machine Translation system (GNMT) went into action late with support for translating to and from English and French, German, Spanish, Portuguese, Chinese, Japanese, Korean and Turkish. These represent the native languages of around one-third of the world’s population, covering more than 35% of all Google Translate queries, the company said at the time.
In fact, the current news about the improvement is also reasonably essential regarding scale. For instance, in the U.S. alone, 1,292,448 people converse in Vietnamese; another 836,171 speak Russian, and 586,173 speak Hindi, Google says, citing U.S. census data. And several languages will be added in weeks ahead, and Thai inclusively, which didn’t quite make today’s release.
Google Translate itself today serves over 500 million monthly users in need of 140 billion words in a day. As curled from, the New York Times reported in December. Based on the report, Google’s switch to this A.I.-powered machine translation system is to be completed this year.
Neural translation is a significant leap over unique translation systems, as it’s able to take advantage of the progress made in the machine learning field to make translations more natural and correct, and sound more like the way people speak the language. What makes the difference is that the system doesn’t translate each part of a sentence piece by piece, but looks at the sentence as a whole. This assists the system figure out the broader context and the most related translation. It then rearranges and adjusts the sentence using the right grammar.
Also, the Neural Machine Translation system learns over time and improves, resulting in better and more original and natural translations the longer it works.
The new translations powered by this system will go live on the Google Translate platform, starting today. This includes online at translate.google.com, through Google search and the Google Search app, and in the Google Translate apps for iOS and Android. The translations will soon be made available for automatic page translations on Google Chrome.
On the general look of it, the use of machine learning translates on more useful language support. When working with common phrases and words, the neural network should make light work of most translations. It'll begin to fall with more unclear vocabulary.