Microsoft is celebrating the official opening of its newest research lab in Cambridge, MA, today with a day-long symposium at MIT on the intersection of computer science and the social sciences. That’s also where the lab itself —the company’s fifth research outpost outside its Redmond, WA, home base—will focus its efforts, at least initially, says Rick Rashid, senior vice president of Microsoft Research.
We covered the news back in February when Microsoft first announced the creation of Microsoft Research New England and the appointment of Jennifer Chayes—a former area manager for mathematics, theoretical computer science and cryptography at MSR Redmond—to direct it. Since then, Chayes and her husband Christian Borgs, the new lab’s deputy managing director, have been working to assemble a group of researchers who will bring a combination of economic, sociological, and psychological perspectives to bear on the ways people use computers and the Internet.
That’s bound to be a fruitful area for Microsoft, though in the long run, the lab’s direction will be determined by the people who join it, Rashid said in an interview this morning. “Whenever you start a research organization there tends to be a bit of a focus on the things that the people who started it want to do,” he says. “Ultimately, this lab will be no different from the other ones, in that we’ll hire great people and you’ll see the breadth of the organization develop. But certainly this whole area of interdisciplinary work, social computing, the social sciences area as it intersects with computer science, is a very interesting way to get the lab started.”
Already the lab includes an initial team of 33 full-time researchers, visiting researchers, post-docs, and interns, according to Microsoft. At its core, aside from Chayes and Borgs, are four other researchers with backgrounds in mathemematics and computer-science theory, including mathematician Henry Cohn, former head of the cryptography group at MSR Redmond; the husband-and-wife team of Adam Kalai, an expert on game theory and machine learning, and Yael Kalai, a cryptography expert, both briefly part of a theory group at Georgia Tech; and Butler Lampson, a pioneer of areas such as time-sharing, personal computing, WYSIWYG editing, and tablet computing. Also on board, at least temporarily, are visiting researchers Susan Athey, a rising star in microeconomics and econometrics at Harvard, and prominent MIT economists Glenn Ellison and Daren Acemoglu. (We published a full list of the lab’s founding staff back in July.)
More hiring announcements are expected today. The lab is also busy building connections with local computer science departments—it’s already running joint seminars with the Laboratory of Information and Decision Science and the Computer Science and Artificial Intelligence Laboratory (CSAIL) at MIT. (Today’s symposium is at at MIT’s Stata Center, home to CSAIL.) Chayes and Borgs have been “doing a great job of making connections and building strong relationships with MIT, Harvard, and other universities in the Northeast,” says Rashid.
I asked Rashid how he expects Microsoft Research New England to contribute to the company. “I have no idea,” he answered. “If I knew, then we wouldn’t really need it. The difference between product development and research activity is that in product groups you may or may not do what you thought you would do, but at least you had a specific outcome in mind when you started, and you have a schedule and a plan. In basic research you don’t have those things—you have smart people and good intentions. The goal is to give them an opportunity to do what they want to do, and then step back.”
Microsoft is one of the last high-tech behemoths that can still afford to operate this way; with the exceptions of Hewlett-Packard, IBM, Intel, and Xerox, most computing companies have long since folded their research operations into product development groups (witness last year’s saga at Mitsubishi’s Cambridge lab). And Microsoft rival Google famously has no formal research wing at all—its software engineers are encouraged, instead, to spend 20 percent of their time on fringe projects that don’t necessarily relate to the company’s existing business lines.
When I asked Rashid whether Microsoft had considered similar arrangements—perhaps giving product developers more time to do speculative work, or encouraging researchers to work alongside product developers more closely—he said that product developers inside Microsoft have no trouble innovating, even without a “20 percent time” tradition like Google’s. “I think a lot of people don’t quite realize how many of the exciting things that happen inside product groups come from individuals who have great ideas,” Rashid says. “That’s what Microsoft product groups are known for. We don’t call it something special; we just expect people to do intelligent work.”
And while there is plenty of collaboration between researchers and product developers inside Microsoft, Rashid says, there’s still enormous value to freeing the best people to do long-term basic research. “If you want to really contribute to the science of your field, that is a full-time job,” he says. “It’s not something you do in your spare time. The reality is that you want people to be able to take lots of risk, to go where their ideas take them independent of what might seem relevant at the time. And you want them to participate in the academic community, to be critiqued by their peers within that community. Historically, organizations that don’t subject themselves to peer review aren’t able to maintain an edge.”
In fact, one of the main ways Microsoft evaluates its researchers—and maintains accountability despite the open-ended nature of its brief—is to monitor their contributions to academic publications, Rashid says. “What is their impact on the field? That’s really measured through peer-reviewed publication, not so much numbers but quality—where are you publishing.”
But Rashid uses other criteria as well. “What is your impact on the researchers around you? Some researchers may not publish as much but are people who have this ability to synergize activity—causing things to happen that wouldn’t otherwise. You want to reward that too. And a third thing to look at, obviously, is technology transfer, whether it’s into Microsoft or into the world more broadly.”
It will be at least two years, Rashid says, before Microsoft can begin to tell how its research investment in New England is paying off. “This is just really early days, but the trajectory looks good,” he says. “I think people are excited.”
