Yesterday afternoon I stopped by Intel Labs Seattle, the research lab run by the chip-making giant near the University of Washington campus, for the lab’s annual open house. It’s an extravaganza that always draws a big crowd from the local tech community. Besides the huge variety of lab demos, one of the most interesting things going on was a changing of the guard.
Dieter Fox, an associate professor of computer science and engineering at the UW, succeeded David Wetherall as director of the lab two weeks ago, when Wetherall’s three-year term officially finished (see photo below). Fox is the fourth director of the Seattle lab, formerly called Intel Research Seattle; all have been UW computer science professors. While Wetherall’s expertise is in wireless networks, mobile devices, and Internet protocols, Fox’s strengths are in robotics, artificial intelligence, and machine learning. (He is the co-author of the 2005 advanced textbook, Probabilistic Robotics, with Sebastian Thrun of Stanford University and Wolfram Burgard from the University of Freiburg.)
So, will Intel Labs Seattle now be doing all robotics, all the time? Will the first general-purpose household helper robot come out of Intel (NASDAQ: [[ticker:INTC]])? One can always hope—but Fox seems to have a broader and more practical outlook on the lab’s role in shaping the future of computing.
“Our role with respect to Intel is performing what they call disrupting research that is off-roadmap, but essentially our task is also to surprise Intel,” Fox says. “If we show what can be done with future computing systems, then we are serving our purpose. And beyond surprising Intel, we also want to surprise consumers by what can be done. It’s becoming more and more important that these computational systems are going to be observing the environment, using sensors. Today’s smartphones all have GPS, accelerometers, and all that. The key question is, how can we extract relevant information to make it more interesting for users?”
Seeing as robots are computing systems that sense and manipulate their environment, they will certainly figure prominently in the lab’s work—perhaps more than ever before. “For Intel, it’s clear the future of robotics is going to become extremely relevant. We need to see what are the key questions from a computational perspective, what kind of processing is needed for these systems,” Fox says. “Our key agenda is to inform Intel on what the future of computing looks like, especially computing connected to everyday scenarios.”
The idea is that if and when the market for intelligent household robots takes off, it’ll be Intel that provides their brains (in the form of microprocessor chips). But even beyond that, Fox says, “Intel could provide the processing that’s adapted to the specific needs of those systems, and along the way maybe also provide the computational toolset I need. So it’s not only the hardware, but it’s also a better understanding of how you extract information from these sensors. That’s also a theme for Intel—they want to go beyond just building the hardware, and show the whole user experience you can get if you have good computational power.”
Lastly, I got some closing thoughts on the lab’s evolution from its outgoing director. “The trajectory of the lab is, we’ve always done perception and sensing, starting with location, and we’re moving now to richer systems” like computer vision and robotic manipulation of objects, says Wetherall, who is going back to full-time teaching and research at UW this month (though he’ll stay involved with Intel Labs). “It’s quite a natural progression for the lab,” he notes. “That’s what leads to intelligent systems.”
