Advancements in voice recognition technology have come a long way since Siri, OK Google ,and Cortana were first introduced. Still, it seems like there’s way more room for improvement. Fortunately, Microsoft’s Chief Scientist of Speech, Xuedong Huang says that the technology improves at a 20% rate every year for the past 20 years.
Ultimately, this means that “In the next four to five years, computers will be as good as humans” when it comes to understanding human speech.
Earlier this year, Google announced that they got their speech recognition error rate down to 8%. When Microsoft made its first-ever speech recognition technology available with Windows 95, the error rate was almost 100%. And, if the experts are correct, that 8% error rate is about to get whole lot smaller.
Beyond the convenience that precision in voice recognition technology can bring, Huang, along with Microsoft and their tech community, assert that it paves the way for what they believe to be the next phase: Real artificial intelligence.
The existence of continuously improving speech interface apps for consumers allows humans to have more complex conversations with gadgets. And as speech becomes a more accepted way to interact with technology, it creates a foundation for improved artificial intelligence, which could see better, more holistic applications. Potentially, advancing speech recognition technology can allow users to control, not just their phones, but every single device in their home.
“The language barrier is going to be essentially nonexistent in four years, for the major languages and the major scenarios,” said Arul Menezes, who heads the Machine Translation team at Microsoft Research, says in a press release.
It’s an amazing idea,