Microsoft unveils new software that can translate spoken English into spoken Chinese almost instantly. The translator software was demonstrated by Microsoft at an event in China by Microsoft’s Research chief Rick Rashid.
The cadence (rhythm of the speech) and intonation (how the words are pronounced) are preserved by the software so that the translated speech still sounds like from the original speaker.
According to the BBC news,
Microsoft said research breakthroughs had reduced the number of errors made by the instant translation system.
It said it modelled the system on the way brains work to improve its accuracy.
The improved speech recognition system was used by Mr Rashid during his presentation. First, the audio of his speech was translated into English text. Next this was converted into Chinese and the words reordered so they made sense. Finally, the Chinese characters were piped through a text-to-speech system to emerge sounding like Mr Rashid.
The technique used by Microsoft is called the deep neural network which was developed by the Microsoft Research and the University of Toronto
The video regarding the demonstration is shown below. The beginning of the video explains the fundamental difficulty of computer translation and the last few minutes actually demonstrate the software’s English-to-Mandarin speech-to-speech translation engine. The software claims to have an improvement of accuracy of about 30% compare to the present English-Chinese translation models. The software is still a work in progress and may come out sometime developed as a universal language translator and the language barrier may become a part of history someday. Well, that still remains as a fiction as seen in the Star Trek now and the project will have to go a long phase to reach that point.
Video : Microsoft demonstration
- Microsoft demos speech translator (bbc.co.uk)
- Microsoft demonstrates “breakthrough” speech translation technology (sociable.co)
- Microsoft demos instant English-Chinese translation (updatednews.ca)
- Microsoft demos ‘human-brain-like’ instant language translation tool (indiavision.com)