Microsoft’s AI-driven Translator supports more than 100 languages

Get a free Techzine subscription!

Microsoft added support for 12 languages and dialects to its Translator service in the Azure public cloud. The addition uses artificial intelligence to automate text translation.

The languages and dialects added include Georgian, Mongolian (Traditional), Mongolian (Cyrillic), Turkmen, Tibetan, Tatar, Kyrgyz, Uyghur, Dhivehi, Bashkir, Macedonian, and Uzbek (Latin). The update constitutes a notable milestone for Translator.

The total number of languages and dialects supported by the service now number more than 100, an improvement from about 80 when 2021 began.

How to access the Translator

Microsoft enables users to access Translator’s automatic translation features via several routes. The service is available to enterprises via an API (application programming interface) in the company’s Azure public cloud.

Consumers can get it via the Bing search engine’s built-in tool, or the standalone translation apps Microsoft built for iOS and Android.

Alternatively, users can integrate the Translation features into the Office suite of apps too. The Azure version of Translator has features you will not find in consumer iterations. Companies can customize the AI models powering the translation by adding support for industry-specific terms (like product names and such).

The way Microsoft does it

Microsoft says one of the companies using the translator is Volkswagen AG, which translated about 1 billion words every year across 60+ languages.

Announcing the update, Microsoft shared details on how it is working to enhance the neural networks powering the Translator. The researcher working on it developed a multilingual AI model dubbed Z-Code that allows neural networks optimized for different languages to learn from each other and become more accurate.

The ability to reuse some information reduces the amount of training data Microsoft has to collect to make new AI models.