Microsoft Translator is an artificial intelligence (AI) based text translation service on Azure and part of the cognitive services. Recently, the company announced that 12 new languages and dialects have been added to the service, meaning it can now translate between more than 100 languages and dialects.
More than 20 years ago, Microsoft started the machine translation project. In 2002 the Machine translation group, part of Microsoft AI and Research’s larger Speech and Language group, was created to break down language barriers and enable global communication for both written and spoken languages. They work on the Microsoft Translator APIused by various Microsoft products since 2006, including the Translate AI service in Azure.
Translator covers most spoken languages including English, Chinese, Hindi, Arabic and Spanish right from the start leveraging statistical machine translation (SMT) models. When AI started to evolve further, the company adopted neural machine translation (NMT) and further expanded its language library. And now with a new version of Bashkir, Dhivehi, Georgian, Kyrgyz, Macedonian, Mongolian (Cyrillic), Mongolian (Traditional), Tatar, Tibetan, Turkmen, Uyghur and Uzbek (Latin) translation languages spoken by 84.6 million people have been added. Moreover, the Optical character recognition (OCR) function is powered by the same AI models; thus, any written text can be absorbed and translated into any supported language.
The technology behind Translate’s upgrades is Microsoft’s multilingual AI model known as Z-Codepart of Microsoft XYZ Code Initiative to combine AI models. It combines multiple languages based on language families so models can learn from each other. Additionally, sharing linguistic elements between similar languages and transfer learning greatly reduces the amount of data needed for a good translation, making the process faster and more accurate.
Microsoft isn’t the only public cloud provider to offer a translation service that supports over 100 languages. Google Translate first achieved this milestone in February 2016. AWS, however, with Amazon Translationonly supports 71.
One hundred languages is a good step towards achieving our ambition that everyone can communicate regardless of the language they speak.
Not only are we celebrating what we’ve done in translation – reaching 100 languages - but also in speech and OCR. We want to remove language barriers.
The company says the service can now translate text and documents to and from languages natively spoken by 5.66 billion people worldwide (72% of the world’s population).