New Microsoft Lab in Cambridge to Combine Math and Social Science; Already Besieged By Potential Research Collaborators
There isn’t any shortage around here of potential collaborators and job seekers eager to work with Jennifer Tour Chayes, managing director of Microsoft’s newest research outpost, Microsoft Research New England. Chayes tells Xconomy that by 11:00 am Eastern time today, less than five hours after news of the lab’s creation hit the New York Times and the Boston Globe, she had already received more than 100 e-mails from the East Coast—about 80 percent of them from researchers at MIT, Harvard, Boston University, Brown, and Yale “telling us how excited they are and that they want to interact with us,” in her words.
Microsoft already has a major presence in Cambridge and greater Boston: in addition to the Boston-area employees of recently acquired companies such as Groove Networks, Softricity, and Fast Search & Transfer, the company is hiring staff for a new product-oriented “concept development center” adjacent to the MIT campus under the leadership of former Eons CTO Reed Sturtevant. But when Microsoft Research (MSR) comes to town, the academic world pays attention.
Chayes, 51, is the first woman appointed to lead a Microsoft Research lab and is a pioneer in areas of mathematics and theoretical computer science—such as network and graph theory, recommendation systems, and search filtering—that have increasing relevance in a world of Web-based communication and commerce. Chayes and her husband and close collaborator Christian Borgs, who will be the lab’s deputy managing director, have already begun to outline a vision for an interdisciplinary research center that will link experts in economics, psychology, and sociology with computer scientists who can translate their insights about human behavior into algorithms that will improve the growing range of products that Microsoft delivers over the Web.
“I think that putting the basic mathematics together with basic research in sociology, psychology, and economics will allow us to come up with the insights that we need to deliver a much better experience to our customers online,” says Chayes, who was formerly the research area manager for mathematics, theoretical computer science and cryptography for MSR Redmond and has been named as the Cambridge lab’s managing director.
The new group, to open in July in newly renovated space at One Memorial Drive, will be only a few stair-steps away from the product development groups that can translate its research findings into real software features—and from the new concept-development group headed by Sturtevant. “We’re really excited about that,” says Chayes. “We do long-term research and we come up with basic insights, but if there is some of that that can actually be turned into products, there is nobody better in the world to do it than Reed.”
Sturtevant, for his part, says the decision to open a Cambridge branch of Microsoft Research is “a real step forward for our emerging ‘vertical campus’ in Kendall Square…Having basic research, concept development and incubation, and full product development together at one site will bring great creative energy.”
To ensure communication between the groups, Chayes says the construction plans at One Memorial Drive include a space large enough for MSR staff and Sturtevant’s group to gather every afternoon for tea. “I’m a big believer in breaking bread,” says Chayes. “Probably a quarter of the projects in our theory group have started over our daily teas. People who wouldn’t ordinarily talk to each other start talking to each other, and before you know it they are writing on the walls and modeling.”
Chayes says she and Borgs haven’t yet identified the exact projects the new lab will pursue—that’ll have to wait until they can brainstorm with new staff members face-to-face. But it’s fair to predict that some of the work will reflect Chayes’ and Borgs’ own interests in using graphs to model the Internet and the Web, and, in particular, social interactions over these networks.
In one recent paper, for example, the couple focused on analyzing graph structure on the Web to defeat the “link spam” created by so-called search engine optimization services to artificially inflate various websites’ search-engine rankings. “We’re trying to use our understanding of graph structure to identify suspicious sites and come up with search algorithms that avoid ranking those sites higher than they should be ranked,” explains Chayes.
Related mathematical techniques could lead to better ways to guide Web users toward useful content, based on feedback from peers in their social networks. “We are just overwhelmed right now by everything that’s on the Web and what we really need are search engines that understand us and our social networks and are able to predict what we might like to see,” says Chayes. “If we can get sociologists and psychologists and economists interacting very closely with algorithms people and combinatorialists and people in graph theory, they’ll be able to build accurate models for things like recommendation system and filtering engines.”
Chayes says that she and Borgs have a unique relationship that should help them as co-managers. “We work about 16 hours a day, and we’re really enthusiastic about what we do—in fact, I think if our spouses were not as enthusiastic as we are about science and technology and we didn’t want to talk about it all day, we would drive each other crazy,” Chayes says. “But we do have different styles. I talk really quickly and I jump from one problem to the next trying to make connections from one field to the next. Christian goes deeply into problems and gets totally absorbed until he really understands the connections. And he has great intuition about people while I’m off making lots of connections between people. So as collaborators we are really complementary.”
But while Chayes and Borgs have similar academic backgrounds and research interests—both came from graduate programs in mathematical physics, Chayes at Princeton, Borgs at the University of Munich—they also share a belief in the value of collaboration across disciplinary lines. “People sometimes tend to think that only their subject is important, but the older we get, the more we realize that all of these different areas have so much to bring to problems, and these problems are not going to be solved unless we have deep understanding and deep respect for all these different disciplines,” says Chayes. “So that is a value we really want to see in everybody in the lab.”
It shouldn’t be a difficult value to inculcate, given the variety of researchers already contacting Chayes about potential collaborations and interactions. “I’ve already gotten several offers from faculty at MIT to have joint weekly seminars,” she says. “And here in Redmond we already have more student interns from MIT than from any other school; now we can have students from MIT and Harvard working with us year round. ”
A bigger Microsoft footprint in Cambridge could even become a magnet for researchers from other centers of science and technology. “I’ve already gotten a few notes from amazing people who are not in the Boston-Cambridge area but are hinting to me about how much they would love to be there,” says Chayes. “Which is just as we hoped. It’s really just a tribute to all the great universities there, but we think can leverage that to bring more great people into Microsoft Research.”