Google Translate has always been a life saver when it came to translating foreign languages in real time in the simplest way possible. Google Translate launched back in 2006, and in its ten year long journey, it began with support for a few languages to 103 now. Google has announced a new development for its Translate app, which is the adoption of Neural Machine Translation. Google Translate is primarily focused on using Phase-Based Machine Translation as the key algorithm. With Neural Machine Translation, Google Translate will now be able to translate whole sentences at a time and not just in parts.
Google had been working on Neural Machine Translation for quite some time now and in September this year, the company announced its own version of this technology, called the Google Neural Machine Translation system (GNMT). With the announcement, Google also released a technical report with full research results on the translation system. Google had announced the usage of Neural Machine Translation system on their official blog and also explained how this improves Google Translate – “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.” With this new development, Google also says that translated paragraphs and articles will now be smoother and easier to read. It doesn’t end here though. Neural Machine Learning system is an end-to-end learning system that “learns over time to create better, more natural translations.”
However, not all the 108 languages on Google Translate will be equipped with Neural Machine Translation. As of now, the system will work on eight major languages which include English, French, German, Spanish, Portuguese, Chinese, Japanese, Korean and Turkish. Google says that the languages chosen represent the native languages of around one-third of the world’s population, and hence it covers more than 35 percent of all Google Translate queries. In the blog post, Google has also given an example of how the new system plays better than the old system. A phrase written in German has been translated in English which shows the difference. In the old translation, the translated sentence reads, “No problem can be solved from the same consciousness that they have arisen.” The new translation with the usage of Neural Translation System reads, “Problems can never be solved with the same way of thinking that caused them.”
This difference in the old and new translation system clearly shows how Google Translate will now make translations not just efficient for basic use but also very helpful in bigger ways. Google also says that it will eventually roll out Neural Machine Translation to the remaining 103 languages and other platforms where Google Translate can be accessed.
Google has also made the system behind Neural Machine Translation available for all business through its Cloud Translation API which can be accessed through Google Cloud Platform. Google’s public cloud service offers Machine Learning APIs to make it easy for people to use their machine learning technology. ALSO READ: Google Translate for Android updated with Tap to Translate feature, here’s how to use it
Google Translate celebrated its ten year anniversary earlier this year. Google said that over 500 million users have been using Google Translate and over 100 billion words are translated every day. The most common translations that take place are between languages like English, Spanish, Arabic, Russian, Portuguese and Indonesian. Among the 108 languages supported on Google Translate, there are ten Indian languages. These are Hindi, Gujarati, Kannada, Malayalam, Marathi, Bengali, Tamil, Telugu, Punjabi and Urdu. ALSO READ: Google Translate turns 10; now supports 103 languages, translates 100 billion words every day
Google also mentioned how its growing Translate Community has helped and keeps helping in translations. Google’s Translate Community has over 3.5 million people who help in translating over 90 languages. Google had also written in its latest blog post about how it will continue to rely on the community for their contribution in sharing their language and reviewing translations.