Microsoft has announced a breakthrough for speech recognition, producing a system that recognises words in a conversation as fast as a human.
The new speech recognition system, by Microsoft’s Artificial Intelligence and Research team, makes either the same or fewer errors than transcriptionists, having a word error rate (WER) of 5.9%.
No ad to show here.
“The 5.9% error rate is about equal to that of people who were asked to transcribe the same conversation, and it’s the lowest ever recorded against the industry standard Switchboard speech recognition task,” Microsoft explained on its blog.
The US tech firm says that the breakthrough marks the first time that a computer can recognise words in a conversation as accurately as a human.
The new speech recognition system means that Microsoft might get the upper hand in the assistant field
So then, what does this mean for the average person?
“The milestone will have broad implications for consumer and business products that can be significantly augmented by speech recognition. That includes consumer entertainment devices like the Xbox, accessibility tools such as instant speech-to-text transcription and personal digital assistants such as Cortana,” Microsoft explained.
The team stressed that the milestone doesn’t translate to perfection (as humans aren’t perfect either), merely meaning that the rate of errors is the same as that of a human hearing the conversation.
The system was made possible by the use of Microsoft’s Computational Network Toolkit, the researchers said, being a toolkit for deep learning.
