In the past 10 years, Google Translate has gone up from offering support to just a few languages to support 103 languages now, apart from connecting strangers, to reach across different language barriers and also to help people find what and whom they love.
In the beginning, Google had pioneered a large-scale statistical machine translation that was using statistical models to translate the text. But now, Google has introduced the next big step towards making Google Translate much better with the Neural Machine Translation.
About the Neural Machine Translation
Neural Machine Translation has been producing exciting and amazing research results in the past few years. In September, the team of researchers had announced Google’s version of this same technique. The Neural system translates whole sentences at a time at an entirely different and high level of translation, rather than by doing it in bits and pieces. The Neural system uses a broader context that helps it to figure out the best and the most relevant translation, which will then be rearranged and adjusted to be presented more like what a human speaks with proper grammar. Since it is now easier to understand the sentences, the translated paragraphs and articles will be a lot smoother and easier to read and understand. And all this is possible because of the end-to-end learning system that has been built on the Neural Machine Translation, which essentially means that the system keeps learning over time to create better and more natural translations.
Into Action with 8 Languages
Google has now put the Neural Machine Translation into action with an out-and-out pairing of eight languages, to and from English with French, Spanish, German, Chinese, Portuguese, Turkish, Japanese, and Korean. These 8 languages represent the native dialects of almost one-third of the total world’s population, which covers more than 35% of all the queries on Google Translate.
With this latest update, Google Translate has improved more in just a single leap than what has happened in the last ten years combined. However, the company says in its blog post that this is just the beginning. While the company has started as of now with just eight language pairs within the Google Translate app, Google Search, and the website, their ultimate goal is to roll out the Neural Machine Translation eventually to all the 103 languages and surfaces where the Google Translate can be accessed.
Google Cloud Platform
That is not the end of the latest update. There is more coming to the Google Cloud Platform, Google’s cloud service for the public, which now offers Machine Learning APIs which will make it very easy for anyone to use Google’s machine learning technology. Apart from this, Google Cloud Platform will also be making the system that is used behind Neural Machine Translation available for all the businesses via the Google Cloud Translation API.
The step taken by Google towards the Neural Machine Translation is undoubtedly a significant milestone for Google Translate. However, there is still a lot more work to be done and Google promises to continue to keep learning over the time. Google has said that though the new Neural Machine Translation makes the Google Translate work better than ever, it will still continue to rely on the Translate Community, wherein the language-loving multilingual speakers help in sharing their language by way of contributing and reviewing the translations.
The enhanced service is available for the public at the official website of Google Translate, and also at its mobile apps for both the Android and iOS platforms.