Microsoft wants Skype to be the world’s interpreter. The company announced this week that Skype Translator, the real-time speech translation service included with the popular communications software, hit a significant milestone.
Microsoft has added Japanese to the mix, bringing the total number of supported languages to 10. This enables automatic translation services for Japanese speakers who conduct video and voice calls with other Skype users speaking in any of the nine other languages. The other languages are English, Arabic, Chinese (Mandarin), French, German, Italian, Portuguese (Brazilian), Russian and Spanish.
Japanese language supports is not only a milestone for Skype Translator, but also for Microsoft’s artificial intelligence (AI) and machine learning technologies.
“Speech translation is a hard problem to solve, as is always the case when machines are trying to mimic a deeply human capability,” stated a Microsoft in an April 6 blog post. Adding to the challenge is a vocabulary and grammatical structure that is completely different from English, making it difficult for English-speakers to learn, according to the software giant.
Skype Translator uses two neural-network based Artificial Intelligence (AI) technologies: Automatic Speech Recognition and Machine Translation, according to Microsoft. It also uses Microsoft’s TrueText natural language processing technology and a text to speech synthesizer which enables users to hear, and not only read, the translation, according to Microsoft.
The company’s machine translation AI translates each word in context of the full sentence, producing output sentences that sound more fluent than word-for-word translations. Meanwhile, Microsoft’s TrueText technology polishes up redundant words, filler sounds and other quirks to improve the overall quality of the translated speech.
Microsoft isn’t the only American technology company trying to break down language barriers between English and Japanese speakers.
In January, Google added real-time Japanese translation capabilities to its World Lens app, a favorite among travelers. Although the Google Translate app already allows users to snap a picture of Japanese text and view an English translation, “it’s a whole lot more convenient if you can just point your camera and instantly translate text on the go,” said Masakazu Seno, a Google Translate software engineer, in a Jan. 26 announcement.
“With Word Lens, you just need to fire up the Translate app, point your camera at the Japanese text, and the English translations will appear overlaid on your screen—even if you don’t have an Internet or data connection,” Seno added.
And last month, the company teased that Google Translate users should soon start getting higher-quality translations from the service, beginning with the Hindi, Russian and Vietnamese languages. In a March 6 blog post, Google Translate lead Barak Turovsky announced that his group would begin rolling out a new experience powered by Google’s deep neural network technologies.
Similar to Microsoft’s AI-based approach, Google’s system translates whole sentences at a time. This results in more accurate, natural sounding translations that better simulates how people speak to one another, according to Google.