You are here
Microsoft’s Latest Speech Recognition Software Is As Accurate As Humans Innovation Trending 

Microsoft’s Latest Speech Recognition Software Is As Accurate As Humans

 Artificial intelligence is really the “thing” of our close future, and it appears that Microsoft is making some ground breaking steps in this field.

Recently, Microsoft researchers have been successful in creating a speech recognition software which, as previously announced, is competent to hear languages as precisely as people. Microsoft in fact broke the ceiling with their sophisticated speech recognition software.

The software has a word error rate of 5.9 percent, which is somewhat equal to that of humans.

Microsoft described this in an article:

The research milestone doesn’t mean the computer recognized every word perfectly. In fact, humans don’t do that, either. Instead, it means that the error rate – or the rate at which the computer misheard a word like “have” for “is” or “a” for “the” – is the same as you’d expect from a person hearing the same conversation.

Just recently before reaching this milestone, Microsoft managed to reach an error rate of 6.3 percent after relying on deep neural networks. This technology operates similar to how the human brain works and interprets info using specialised graphics processing units (GPUs) enabling an unprecedented applications learning speeds.

The landmark will have far reaching consequences on our daily life. For starters, Microsoft’s products like Xbox and personal assistant Cortana would be a great deal better at comprehending voice commands. Availability applications, like immediate transcription services would be more exact.

Additionally, it may result in a worth addition to Microsoft’s productivity applications like Office, or MS Word’s dictation attribute – getting to near-human levels is an enticing prospect really.

But in the bigger scheme of things, this indicates a major landmark for AI research. Geoffrey Zweig, from Microsoft’s Language and Dialog research group shed light on this prospect and points out that the next stage is to construct a speech recognition software that not only attempts to merely transcribe human languages, but also make sense of it. At this point, with the latest development in mind, a tremendous step towards reaching that supreme target has been made.

What do you think about the latest Microsoft’s speech recognition software?

Related posts