Microsoft Translator adds AI-powered Offline Language Packs

Microsoft Translator adds AI-powered Offline Language Packs

Microsoft Translator is Microsoft’s competitor to Google Translate. The app has a broad feature set, and it features translation capabilities in many languages. Users can use it for text-to-speech translation in many languages.

Now, Microsoft has announced that Translator has added new capabilities that allow users and developers to get AI-powered translations whether they are connected to the Internet or not. The new capabilities allow users and third-party app developers to benefit from neural translation technology regardless of whether their device is connected to the cloud or offline.

End users can now download free AI-powered offline packs in Microsoft Translator. Microsoft states that this development comes after two years of work, and it’s said to complement its effort to make sure developers and users can access AI-powered tools for their data, regardless of whether it’s in the cloud or in a device. The ability is referred by experts as edge computing.

The company notes that Microsoft Translator released AI-powered online neural machine translation (NMT) in 2016. This capability was only available online because of “the computing power needed to run these high-quality translation models. In late 2017, the capability was launched for specific Android phones that had a dedicated AI chip–namely, the Huawei Mate 10 series as well as the Honor View 10, as they have a customized version of Translator. This was because the Kirin 970 SoC has a dedicated Neural Processing Unit (NPU). According to Microsoft, it allowed users of these devices to get offline translation quality on par with the quality of online neural translation.

The Microsoft Translator team was then able to further optimize the algorithms, which allowed them to run directly on any modern device’s CPU without needing a dedicated AI chip. The new Translator apps therefore now bring neural machine translation “to the edge of the cloud” for all Android, iOS, and Amazon Fire devices, with support for Windows devices coming soon.

According to Microsoft, the new NMT packs produce higher quality translations which are up to 23 percent better and about 50 percent smaller than the previous non-neural offline language packs. The NMT packs are available in Translator’s most popular languages, and the company states that new NMT languages will be added regularly. Users can check the complete up to date list here.

The second feature which Microsoft has announced today is a preview of the new local feature in the Translator Android app, which enables Android developers to “quickly and easily” add text translation to any Android app that benefits from translation capabilities. Additionally, Android developers can add offline NMT to their apps for the first time thanks to the new NMT offline packs, which allows their users to get access to NMT translated content without an Internet connection.

To integrate translation in their app, developers will need to add “some simple code” that will use Android’s bound service technology with an AIDL interface to silently call the Translator app, and the Translator app will then do the rest. If the user’s device is online, the Translator app will retrieve the information from the Microsoft Translator service on Azure. If there is no Internet connectivity, the app will use local NMT offline language packs to deliver the translation back to the developer’s app.

Microsoft states that this feature is expected to be generally available within 90 days of the preview release.  The company also notes that when the user’s device is online, translations can also leverage customized translation models that match the app and company’s unique terminology.

Developers can learn more about how the local feature preview works in the company’s GitHub documentation and sample app.


Source: Microsoft

About author

Idrees Patel
Idrees Patel

Idrees Patel is a smartphone enthusiast from India. He has been an Android user since the time he got the LG Optimus One in 2011. He has a bachelor's degree in Management Studies. The subjects in which he is interested are mobile processors, real-world UI performance, in-depth camera quality analysis, and many more. Contact him at [email protected]