A team of researchers at Microsoft says they have been able to make speech recognition smartphone, improving accuracy by approximately 15% as well as doubling the speed at which results are returned and speech is recognized. They also claim that users based in the U.S. will see results twice as fast when searching Bing or composing a text message through speech recognition. They say that all of this is possible due to a computation model that has been created which imitates the functioning of the human brain.
That said, the company didn’t say how well the speed and accuracy of speech recognition stacks up against its rivals, pointing out that different companies use different methods of measuring accuracy of these systems. They started updating U.S. data centers with this update in April, the entire update process is expected to be completed over the next couple of weeks. Microsoft hasn’t said when this improved speech recognition technology will be rolled out for users based outside the U.S. Microsoft has posted a video online that details these improvements, the video can be viewed here.