Read the Original Article at http://www.informationweek.com/news/showArticle.jhtml?articleID=229100305
Intel is planning to invest $100 million over five years to establish university research centers in the U.S. that will develop computing technology with clear commercial applications.
The chipmaker announced Wednesday that the first center would be led by Stanford University, which would act as the hub for coordinating research between Stanford and seven other participating universities: the Universities of California at Davis, Berkeley and Irvine; the University of Washington; Cornell; Princeton; and Harvard.
Four Intel researchers will be assigned to each center to push developed technology out of the universities and into the commercial computing market. The plan is to eventually have a total of six centers, each receiving $2.5 million in funding. Only U.S. universities will be involved. "This is a U.S.-only program, and for the foreseeable future, we expect this to be only a U.S. program," Justin Rattner, Intel CTO, said during a teleconference with reporters.
Each center will be focused on a particular research area, which will be broad in scope in order to provide researchers with lots of flexibility in direction. "We want to encourage new thinking and out-of-the-box thinking," Rattner said.
An advisory board comprised of university and Intel representatives will guide the research and the operations of a center. Other companies, including Intel competitors, could apply to participate in the research, but their acceptance would be up to the board. "It's not our goal to have as many possible companies or (government) agencies participating, but we certainly recognize that in the right situation, it's the best thing for the center," Rattner said.
Other organizations joining a center would have to contribute to the funding and agree to an "open" intellectual property model, which means that technology advancements would be publicly available. The centers will focus on the use of open-source software in their research. "We're treating the research as pre-competitive, and we want to make sure we're not limiting our academic partners," Rattner said.
Intel plans to monitor the progress of the centers before committing to funding beyond five years. "We didn't want this to be an open-ended thing, and then one day wake up and say this doesn't make sense," Rattner said. Intel plans to review the progress of each of the centers at the three-year mark.
The center led by Stanford University will focus on visual computing in personal computers, Internet-enabled televisions, and emerging mobile devices, such as tablets and smartphones. Researchers will focus on several technology areas, such as the development of consumer-oriented software for building 3D models of places and buildings, including interiors and exteriors. Other areas include real-world simulation, so a person, for example, could build a virtual 3D character with human behaviors and natural motion. Such a character could be used for trying on clothes at a retail site.
The center also will work on technology that would enable computers to understand the movements and facial expressions of people for a more natural interaction with users. In addition, researchers will work on new types of cameras that can recognize objects and send the information for use in applications. The simplest example would be pointing a camera at a restaurant menu in a foreign language to get an immediate translation.
Visual computing and the ability to weave computers into people's lives more naturally are not new to Intel. The company launched last summer the Interaction and Experience Research division to develop technology for more seamless interaction between people and computers.
Stanford and the other participating universities are committing a total of 30 faculty and 50 graduate students to the center. Stanford professor Pat Hanrahan, who will lead the university's work, has high hopes for the outcome. "The return on this investment is going to be enormous," he said during the teleconference with Rattner.
Hanrahan is a big supporter of industry-funded research. Government-funded research tends to be more restrictive, while companies providing money tend to encourage "wild, way-out" research to push the envelope on what's possible, he said. "I think this is actually the best model than any other I have seen," Hanrahan said of the Intel-funded centers.