Highlights of Microsoft Translator in 2021

Throughout 2021, Translator delivered new innovations to help people communicate regardless of the language they speak. Here are some important points.

Overview of Document Translation

Document translation is a new feature of the Translator service that translates entire documents, or batches of documents, into a variety of file formats preserving their original structure and format.

Document translation has been specially designed to translate large files with rich content. You can also apply custom glossaries and custom translation templates created with Custom Translator to ensure your documents translate exactly the way you want.

  • Translate large files: Translate entire documents asynchronously.
  • Translate multiple files: Translate multiple files into all supported languages ​​and dialects while preserving document structure and data format.
  • Preserve source file layout: Translate files while preserving the original layout and format.
  • Apply custom translation: Translate documents using general and custom translation templates.
  • Apply custom glossaries: Translate documents using custom glossaries.
  • Automatically detect document language: Let the document translation service determine the document language.
  • Translate documents with content in multiple languages: Use the auto-detect feature to translate documents with content in multiple languages ​​into your target language.

Learn more:

Containers available in secure preview

The Translator service on Azure is now available in closed preview containers. Containers allow you to run certain functionality of the Translator service in your own environment. They are ideal for organizations with limited internet connectivity or for organizations with specific security and data governance requirements that prevent them from sending information to the cloud.

Some of the benefits of containers for some organizations include:

  • Keep data in-house: Some industries may have regulations that prohibit sending data to the cloud. Containers keep data in place to comply with these regulations.
  • Portable architecture: Containers enable the creation of a portable application architecture that can be deployed across Azure, on-premises, and at the edge.
  • Flexibility: With containers, you can run the latest models in the Translator service and update those models when you want to update them.

Learn more:
The translation service is now available in containers

Translator now supports over 100 languages ​​and dialects!

In 2021, Translator added 23 new languages ​​and dialects for a grand total of over 100! The addition of these languages ​​was made possible by the advent of the transformer architecture which opened up new avenues for creating machine translation models, allowing training with smaller amounts of material than before. Thanks to the Multilingual Transformer Architecture, we can now supplement learning data with material from other languages, often in the same language family or a related language family, to produce models for languages containing small amounts of data, commonly referred to as low-resource languages.

Learn more:

  • Inuktitut: Inuktitut is now available in Microsoft Translator!
  • Albanian, Amharic, Armenian, Azerbaijani, Khmer, Lao, Burmese, Nepalese and Tigrinya: Microsoft Translator releases nine new languages ​​for International Mother Language Day 2021
  • Chinese (literary): Microsoft Translator publishes literary translation in Chinese
  • Bashkir, Dhivehi, Georgian, Kyrgyz, Macedonian, Mongolian (Cyrillic), Mongolian (Traditional), Tatar, Tibetan, Turkmen, Uyghur and Uzbek (Latin): the translator now translates more than 100 languages

Award-Winning Research – Large Scale Multilingual Translation

This year, at the annual machine translation conference, WMT 2021, the Microsoft Translator ZCode team, together with the Microsoft Project Turing team and Microsoft Research Asia, competed in the “Large-Scale Multilingual Translation” track and won. The track consisted of a full translation task between all 10,000 directions in 101 languages ​​and two small tasks: one focused on 5 Central and Southern European languages ​​and the other on 5 Southeast Asian languages. The Microsoft ZCode-DeltaLM model won all three tasks by huge margins, including an incredible gain of more than 10 points over the M2M100 model in the large task rated on 10,000 language pairs.

Learn more
Large-scale multilingual translation: 10,000 language pairs and beyond

Group Transcribe, a Microsoft Garage project

Group Transcribe, a new app from Microsoft Garage, launched in March. Group Transcribe provides real-time transcription and translation for in-person meetings and conversations. Powered by cutting-edge artificial intelligence speech and language technology, Group Transcribe allows each participant to use their own device, providing highly accurate speaker attribution so conversation participants can see who said what in their language favourite.

Group Transcribe provides high-quality, real-time transcription and translation:

  • Start a conversation from your phone and easily invite others to join you
  • Stay focused without taking notes or pushing to speak
  • Follow the conversation in real time in the language of your choice
  • Automatically save transcription after each session
  • Browse and view previous transcripts from your home screen
  • Easily share transcriptions with others
  • Supports languages ​​in over 80 locales

Group Transcribe is now available on iOS.

Learn more:
New Garage Group Transcribe project helps you transcribe and translate while advancing inclusive voice AI

Persistent conversation codes in the Web Translator app

In June, we updated the cross-device conversations feature of Web Translator (translate.it). With the new version of web chats, you can now create a predefined chat code that can be reused for your future chats. You can then invite anyone to join the conversation from their own phone, tablet or laptop.

Canned conversations are designed for educators, presenters, or for recurring meetings where you need live transcription or translation to keep everyone in the conversation, no matter what language they speak.

Learn more:
Access your conversations with a permanent conversation code using Microsoft Translator

New voice regions and updated language switcher in the Translator app

This summer, we added regional accents to the Microsoft Translator app on iOS and Android. Known as Speech Regions, you can now select the accent of text-to-speech audio output you want to hear while using the app. For example, if you wanted to hear a translation in English, you could select American, British, Australian or other accents.

Along with this update, we’ve also updated the app’s language switcher to make it easier to see which translation features are available for each language.

Learn more:
New voice regions and updated language switcher in the Translator app