X hits on this document





3 / 43

physics, as well as in scaling phenomena in physics.

During postdoctoral work at IBM, Barabási wrote a paper on network topology which, he explained to THE BIOLOGICAL PHYSICIST, was “rejected everywhere”. When he left IBM to begin a faculty position at Notre Dame in 1995, he initially focused his research on more traditional problems, in materials science. But networks continued to fascinate him. His interest was caught in particular by a network unique in that we are watching it grow under our eyes, and in that we ourselves are the microscopic forces that drive its growth – the internet. In collaboration with his “best graduate student”, Réka Albert, and postdoctoral researcher Hawoong Jeong, map the topology of the www. They considered it, at first, as “exploratory research”. But when they noticed the prevalence of hublike behavior in the web (for example, the New York Times web site has an immense number of links, whereas an obscure blogger may have none), and recognized that the classical network theory of Erdős and Rényi was totally unequipped to deal with hubs, they began to map the topology of other networks as well.

A breakthrough came in the summer of 1999, while Barabási was attending a workshop on nonequilibrium dynamics in Porto, Portugal. His group had, by that point, recognized another critical feature of the web’s hublike structure – it exhibited power law scaling. In other words, the plot of the number of nodes with a given number of connections vs. the number of connections followed could be fit with a power law. An immediate consequence of this scaling was that there existed only a few nodes with a large number of links (nytimes.com, bbcnews.co.uk), and very many hubs with only a few links (obscuresolipsisticblogger.com). Another essential aspect of the internet’s power law

structure is that there is no such thing as a “typical” website. There is no characteristic “size” of a website (measured by the number of links), as there would be if the network followed a Gaussian distribution. (This is why, as many of our readers know, such systems are described as scale-free.)

Albert-László Barabási, speaking about his research at a recent conference.

“At that time,” recounts Barabási in Linked, “the web was the only network mathematically proven to have hubs. Struggling to understand it, we were searching for its distinguishing features. At the same time, we wanted to learn more about the structure of other real networks. Therefore, just before leaving for Porto, I had contacted Duncan Watts, who kindly provided us the data describing the power grid of the western United States and the C. elegans topology. Brett Tjaden, the former graduate student behind The Oracle of Bacon Website, now assistant professor of computer science in Athens, Ohio, sent us the Hollywood actor database. Jay Brockman, a computer science professor at Notre Dame, gave us data on a man-made network, the

wiring diagram of a computer manufactured by IBM.” (pp. 79-80)


Before leaving for Porto, Barabási asked Réka Albert to analyze these other networks. She wrote to him that she had “looked at the degree


Document info
Document views106
Page views106
Page last viewedFri Oct 28 18:25:42 UTC 2016