Apple working on in-house speech tech?

Is Apple Inc. developing its own homegrown speech technology team for its smart personal assistant Siri?

 

Apple has assembled a small but high-powered speech technology team and appears to be scouting for talent in Boston, business site Xconomy.com reported.

 

"Based on their online job profiles, we can say that members of the Apple speech team here are working on Siri, the company’s voice-activated virtual assistant. Details beyond that are hard to come by, however, even for others in the field," it said.

 

It said the group includes Gunnar Evermann, who stayed at Nuance for nearly four years before joining Apple in July 2011.

 

Evermann worked for Apple in California, but moved back to the Boston area recently. Xconomy.com said his job is listed as “manager, Siri Speech.”

 

Another member of the team is Larry Gillick, who was a vice president of research at Nuance and recently worked as a consultant and as chief scientist at EnglishCentral.

 

His job title at Apple is “chief speech scientist, Siri.”

 

Still another team member is Don McAllaster, who was a top research scientist with companies in the speech field, dating back to Dragon Systems. He worked with Gillick.

 

“They won’t tell us what they’re doing. We can only guess,” Xconomy quoted Jim Glass, who heads Massachussetts Institute of Technology’s Spoken Language Systems Group, as saying.

 

Notable presence

 

Xconomy said the Apple team, while small, is notable for its mere presence in the Boston area, as Apple rarely maintains engineering teams far from its corporate headquarters in Cupertino, California.

 

But more significantly, Apple does not have a track record of focusing on in-house research in this field, like Google or Microsoft.

 

Yet, Xconomy said such a move would make sense for Apple, which "doesn’t like giving other companies the keys to critical software services."

 

Talent scouting

 

It added that with its own speech team in Boston, Apple may be joining competitors Microsoft and Amazon in plumbing the area around MIT for talent.

 

Earlier, Xconomy noted Microsoft announced that former MIT Lincoln Laboratory staffer TJ Hazen would lead a new initiative at its NERD Center “to advance the state of the art in making all kinds of software easier, more natural and more intuitive to use, with a particular emphasis on speech and language processing capabilities.”

 

It added Amazon is looking to hire people with some of the same speech-technology skills to "push the envelope in automatic speech recognition (ASR), natural language understanding (NLU), and audio signal processing.” — TJD, GMA News