Microsoft's latest machine-learning poster child: Translator Hub

Microsoft's Translator Hub , based on the Research unit's Azure-based Translator service, is now commercially available.
Written by Mary Jo Foley, Senior Contributing Editor

In April, I blogged about Microsoft execs' latest favorite buzzword: Machine learning.
This week, the Softies -- specifically, Chief Operating Officer Kevin Turner -- mentioned machine learning again. This time it was in the context of the top eight tech trends he and other Microsoft execs see driving the industry and Microsoft. (The other seven: Cloud, explosion of data, social computing, natural user interfaces, mobility/connectivity, and the consumerization of IT.)

The specific technology highlighted this week at the Microsoft Worldwide Partner Conference as an example of machine learning technology is the Microsoft Translator Hub. The Translator Hub is commercially available, as of this week.


The Hub is one of several different Microsoft products and services based on the Microsoft Translator technology from Microsoft Research. There's a Translator application programming interface for developers that is available via the Windows Azure Marketplace. There's a Translator app for Windows Phone. There's Bing Translator. There's a Translator widget. And there's the Translator Hub.
The Hub allows for customized automatic translation between any two languages. It is built around a self-service model, meaning users can train and customize it for their particular needs. The Hub is actually a Windows Azure cloud service that allows users to upload language data and build and dploy custom translation models. These are then accessible using the Microsoft Translator APIs or widgets.
Inside Microsoft, this translation technology is used by the Internet Explorer, Lync and Windows Phone teams. It also is used by businesses including Facebook, Trip Advisor and Twitter, according to Microsoft.
What's the partner angle? "Developers and partners are incorporating Microsoft Translator’s application programming interface (API) to build even more translation tools and services," according to the company.

Editorial standards