The Way We Think About What It Means to Know ThingsDecember 22, 2006
Computer scientist Jon Kleinberg of Cornell University.
For Nevanlinna Prize recipient Jon Kleinberg, the social and information networks created by the Internet provide not only an important new research theme but also "data on human interactions at unprecedented levels of scale and resolution."
Established in 1981 by the International Mathematical Union, the Rolf Nevanlinna Prize is awarded every four years, at the International Congress of Mathematicians, for "outstanding contributions in Mathematical Aspects of Information Sciences." Refining the description, the IMU lists numerous areas in which eligible researchers might work, including mathematical aspects of computer science, such as complexity theory, analysis of algorithms, and modeling of intelligence, as well as scientific computing and numerical analysis.
It's not difficult to see why the 2006 prize was awarded to Jon Kleinberg.
A professor of computer science at Cornell University, Kleinberg has been a major contributor to the study and evolution of search engine technology, to small-world network theory (most popularly associated with the "six degrees of separation" idea), and to the analysis of large data streams for the purpose of identifying bursts, which in turn identify topics receiving significant attention at a certain period in time (an analysis of the news in August and September of 2005, for example, would have identified a burst of data related to Hurricane Katrina).
Kleinberg's interests do not stop there, however. At the moment, among the host of areas he continues to explore is one that he finds particularly compelling: online community modeling.
"Technological networks and social networks are currently in a process of convergence," Kleinberg explains. "They are becoming intertwined with one another. We need to study the social and information networks that have been created by the rise of the Internet and the Web, and the interactions that take place in these networks."
Given the vastly interdisciplinary nature of such research, what Kleinberg describes is no trivial task. Indeed, he is already working with sociologists as well as computer scientists and mathematicians to research the ways in which online communities develop and in which better tools could be built for online communication.
"These developments do not just provide us with an important new research theme," Kleinberg says. "They are also providing us with data on human interactions at unprecedented levels of scale and resolution. This produces exciting opportunities for interaction with the social sciences--they have the most experience, of course, in framing complex, nuanced questions about human interaction and global phenomena in human populations."
Such collaborations, he says, could eventually lead to an understanding of the ways in which the design of online systems influences the types of discussions that unfold there and, more generally, of the ways the technology affects society. He points to the history and evolution of online search as an example.
"If 15 years ago someone had tried to convince you how much value you would get out of this piece of technology [online search] that didn't exist at the time, it would have been very hard to imagine," he says. "Now, if we have some question, we simply assume that we're going to type a query and within half a minute have the answer, or be on our way to learning more about something. It's really the sense that we are now hyper-informed in a way we wouldn't have thought possible."
The history of online search, from its initial development to its eventual role as a highly integrated part of everyday life, is a perfect example of the kind of effect that new technology can have, says Kleinberg.
"[Web search] is an example of a series of technical achievements over the past 15 years, many of them quite challenging and involving the efforts of a huge number of people," he says, "and they've changed our lives and the way we think about what it means to know things. That's a trajectory, a useful analogy, as we think about building other tools."
Turning the eventual results of such research into tools that can improve online communication is a long way off, Kleinberg admits. But mathematicians and computer scientists are drawing on previous networking research to develop models that could lead to better understanding of community development.
"At a general level, the models that we tend to use are based on networks and based on tools from algorithm design and probability," he explains. "Graph theory, probability . . . in some sense, that was sort of inevitable when you start looking at these problems. Many of the great technological programs in the past century were fundamentally engineered systems, but the current generation of large-scale network systems such as the Web are fundamentally not engineered---they're sort of just happening.
"So we have to build models to explain how they're evolving, how they're growing. In that way, we hope to understand them well enough to understand how our attempts at design can be more informed by theory," he adds.
On its face, Kleinberg admits, finding ways to improve communication and informed discussion in online communities might not seem like a problem for computer scientists.
"You might say that helping people understand each other is not really the realm of computer science," he says. "This is about how people relate to each other, and aren't these things related to psychology? At a superficial level that's true, but you could have said the same thing about Web search.
"I think what search has shown us is that now when people want to understand something better, their first step is to go to a computer," he says, "and computer science brought that about."
Michelle Sipics is a contributing editor at SIAM News.