Apple has improved the performance of its Siri personal digital assistant so much that it now outperforms alternatives like Google Now and Microsoft’s Cortana.
The speech recognition capability in Siri now has a 5 percent word error rate, thanks to a 40 percent reduction on the part of Apple, Craig Federighi, Apple’s senior vice president of software engineering, said today at Apple’s 2015 Worldwide Developers Conference in San Francisco.
The word error rate is “an industry-leading number,” Federighi told the crowd.
That’s a major challenge to Google, which has made great strides for speech recognition in Android. At the Google I/O conference a couple of weeks ago, Google executive Sundar Pichai said Google’s speech recognition now has an 8 percent word error rate.
Siri now serves more than 1 billion requests a week, Federighi said.
The Apple personal assistant comes with a new user interface in the new iOS 9 mobile operating system, he said.
Apple has made Siri better and better in the past few years, following Apple’s 2010 acquisition of the startup named Siri, which had built a personal assistant app for the iPhone.
Last year Apple reportedly acquired speech recognition company Novauris Technologies, a spinoff of Dragon Systems. Apple has also hired several speech recognition experts trained in a type of artificial intelligence called deep learning. These moves may well have helped Apple do more with Siri speech queries from iOS device owners.