Microsoft Translator, the company’s speech translation technology, used by the cross-platform Microsoft Translator app and Skype Translator, is now powered by neural network technology, the company announced in a blog post. This new technology is available to all developers through the Microsoft Translator speech API, which currently supports ten languages including Arabic, Chinese Mandarin, English, French, German, Italian, Brazilian Portuguese, Russian, Spanish and Japanese (the company says that more languages will be supported over time).
While Microsoft is only getting started with the use of neural networks for speech and text translation, the technology giant believes that this new technology can deliver much better results than preexisting tools:
Neural networks are making in-roads into the machine translation industry, providing major advances in translation quality over the existing industry-standard Statistical Machine Translation (SMT) technology. Because of how the technology functions, neural networks better capture the context of full sentences before translating them, providing much higher quality and more human-sounding output.
For now, the ten languages mentioned by Microsoft are supported in the Microsoft Translator app for iOS and Android as well as through Skype Translator on Skype for Windows desktop and the Skype Preview UWP app on Windows 10 devices. Microsoft has also set a dedicated website to let you see how this new neural network technology compares to the previous SMT technology.
Lastly, the neural network based Microsoft Translator API is now accessible by Azure customers on the Azure portal. For developers looking to move their Microsoft Translator subscription to Azure, you can learn how to do so on this support page.