The company on Thursday said it had selected 10 startups out of 720 applicants to participate in a four-month accelerator in and around Seattle.
Tech accelerators are designed to help new companies get off to a running start, offering guidance, networking and funding. Microsoft has gathered clusters of companies in Seattle since 2012, first in partnership with the Techstars accelerator organization, and later on its own.
Hanan Lavy, the director of Microsoft’s Seattle accelerator, said each member of the current class of companies specializes in some component of machine learning, or the computer science behind algorithms that lets software spot patterns and react accordingly. The startups range from an artificial-intelligence-powered doctor bot to a university recruiting tool.
“We see machine learning popping up all over the place,” Lavy said, “all sorts of (industries), from DNA sequencing to financial applications.”
During the four-month program, companies will have access to Microsoft’s own machine-learning engineers, among other experts. Also included are grants that give companies, if they choose, access to Microsoft’s cloud-computing platform.
Lavy, who helped found Microsoft’s first startup accelerator, in Tel Aviv, Israel, in 2012, said lending a hand to startups makes business sense for Microsoft. It gives the company a better view of the emerging trends that entrepreneurs are chasing, he said. It also introduces startups in a fiercely competitive technology market to Microsoft’s programs.
“When we started, what we had in mind is, ‘What can we do (to help) great startups that will eventually put Microsoft in the center of the ecosystem?’,” Lavy said.
A perk of Microsoft’s program, Lavy says, is its cost. Some accelerators take an equity stake in the company at the end of the program. Microsoft doesn’t.
That helped attract Daniela Braga and Amy Du, co-founders of DefinedCrowd, a Seattle-area startup. Braga and Du got together late last year to found the startup, which is aiming to build better algorithms that help computers understand spoken commands and the way people use language.
Before you’re able to get Apple’s Siri or Microsoft’s Cortana to reliably reply to a request for a list of nearby coffee shops, for example, companies need to collect and analyze a mountain of data, Braga said.
She and Du hope that by bringing together crowdsourcing and machine-learning algorithms, they can simplify that process.
“There’s contextual information, accents, background noise, then you need to detect intent,” said Braga, who previously worked on speech-recognition technologies with Microsoft and VoiceBox Technologies. “There’s many different ways of asking, ‘Find me the nearest Starbucks.’”
———
©2016 The Seattle Times
Visit The Seattle Times at www.seattletimes.com
Distributed by Tribune Content Agency, LLC.
