Microsoft has developed itself to become more than just a familiar computer technology brand name. It continues to innovate, experiment and launch new projects that we are all familiar with. Amongst many of their successful launches lies the creation of speech recognition by Microsoft. Microsoft launched it for the first time in Windows Vista.
Of course, speech based command prompting is nothing new. Siri, Google Assistant and ALEXA are popular services by iOS, Google and Amazon. Microsoft’s Cortana, its speech recognition software, is available on Windows based phones, and most of the new versions of Windows (7, 8, 8.1, 10).
Human level accuracy by Microsoft’s speech recognition software
Microsoft issued a press release recently that it has achieved human level accuracy in its speech recognition. Its error rate has reduced to 5.1% error rate vs. 5.9% last year. It states that this error rate is at par with professional transcribers which is quite a feat.
Now, this may take a while to sink in. The fact that computers and phones in today’s world have the ability to understand us exactly the way other humans do is almost overwhelming. Think of it this way, AI would be able to understand your speech and provide you with information or data that you request, without having to enunciate or repeat yourself. Ever encountered bots on Facebook messenger, or on Skype? The most basic AI algorithms work on simple response patterns such as basic Facebook bots. However, more complex versions can work with complex sentence structures and questions as well – perhaps like Replika as our best friend, and Ava from Zoom managing our schedules for us. We may even have Holly from Red Dwarf or VICI from Small Wonder in our lives one day.
Microsoft has also been working on speech recognition for full conversations. It can now understand the context of the conversation and predict ahead of what you are going to say – similar to how humans interact with each other. Perhaps, this is exactly what inspired Microsoft to make sure their speech recognition truly understands the conversational context:
This is a huge achievement as it puts us one step ahead in terms of having technology that is not only more user-friendly but also more intuitive. Human-accurate speech recognition gives us great leverage in enhancing the way we use technology today.
What do you think about this amazing feat by Microsoft? Do you think it can help improve the way we look at AI today? Cortana might just become your new best friend.