home | about

Contemporary robots travel on the Web and create maps like this.

The Human Genome Project

Think Networks:
Albert-László Barabási:

By Roy Christopher

We all know our world is held together through a vast network of connections, and we're all coming to realize that it's becoming more connected and interdependent with every passing day. The question is how? In what ways are we altering our lives with this network, and how do we deal with the negative aspects of the overwhelming connectivity?

Enter Albert-László Barabási and his new book, Linked: The New Science of Networks. Underneath our online world of seemingly random connections, the cells of our bodies and our social ties lies a network of hubs and ever-growing links with surprisingly not-random patterns. Dr. Barabási digs deep into the world of links on the Web, social networks, cellular connections and other fields. He returns with a clear picture of how these connections operate and how they're re-shaping our lives.

Barabási is the Emil T. Hofman Professor of Physics at the University of Notre Dame, and teaches and directs research on complex networks. His varied contributions have been featured and acclaimed in the media, including magazines such as Nature, New Scientist and National Geographic, to name only a few. If he has one message to deliver, it is, 'Think networks.'

Roy Christopher: Can you give our readers a brief overview of the ideas in Linked?

Albert-László Barabási: For many decades we believed that networks are random. Whenever we had to face a very complex system, such as people are connected by social links (society), chemicals in the cell connected by chemical reactions, webpages connected by URLs, we assumed that the links are thrown randomly around. In the last few years we learned that this is not the case. Instead, networks hide wonderful order and are described by rather rigid evolutionary laws. These laws lead to the emergence of hubs, nodes with an extraordinary large number of links, that partly dominate real networks but they also keep them together. For example, on the Web most people point to a few webpages, such as Google or Yahoo!; the web within the cell is held together by a few very active chemicals; in the economy a few companies, such as IBM or Microsoft, have economic links to an extraordinary number of other companies. The number and the size of these hubs is fundamentally determined by the network's size, and they are responsible for a large number of rather unexpected phenomena. For example, sexual hubs are responsible for turning AIDS from a local epidemic into a pandemic and these hubs make the Internet and the cell extremely resilient against failures. In short, the emergence of the hubs have completely reshaped our understanding of complexity in general. Probably the most interesting aspect of the network laws, is that just about all networks out there follow them. In addition to the sexual network, Internet and society, we and other have studies the structure of the language, the connections between comic book characters of the network of scientists connected by collaborations, and many other systems, and the same laws apply to them; the same structures emerge. This universality of network structure and evolution tells us that if we understand one network, we can apply this knowledge to understand all other complex webs out there.

A scale-free network of 130 nodes generated by the scale-free model. The five biggest nodes are shown in red, and they are in contact with 60% of nodes (green).
(from http://www.nd.edu/~networks)

Your book seems to explain away the role of chaos in the development of complex networks. Do you think the new discoveries in the behavior of scale-free networks will replace chaos theory as a way to understand large, seemingly chaotic systems?

Chaos is talking about the interactions between a few agents, showing that even simple systems can behave in a rather complicated manner. Network theory does the opposite -- it tells us that rather complex systems follow simple rules and laws. Network theory has allowed us to understand the architecture of complexity, a completely neglected area until recently. As networks are pervasive, we simply cannot ignore this message.

Mark Granovetter's idea of the 'strength of weak ties' has intrigued me since reading Malcolm Gladwell's The Tipping Point. It seems such a simple idea that explains so much and to which I hadn't really given much thought before. Do you see this idea as one of the unsung traits of the behavior of networks?

Granovetter's wonderful and rather influential insight, that weak ties play a key role in our social network, is still valid. Network theory has shown us, however, that there is a brand new world out there beyond the weak ties, that without data Granovetter could not even imagine addressing. The role of the hubs could not be appreciated until a few years ago, when we finally got real data on complex networks. Yet, following up on Granovetter's insight is still a dream for us-- while we have detailed topologies describing everything from the www to the collaboration networks, the strength of the ties is most of the cases is not available. This is one of the future directions for network research.

Tell me about your interest in 'parasitic computing.' Where do you see this area going in the future?

Parasitic computing was a proof of a concept for us -- to show that one could turn the whole Internet into a computer, distribute a large computation on thousands of unsuspecting computers. For this we designed a way to turn anybody's computer into a slave for us, making them to unwillingly perform computation on our behalf. Do not worry -- we did not used your computers. We just demonstrated on several computers that this is possible. The future of parasitic computing could lead us in many directions. It need to be made efficient, and in this case it could turn into a mainstream distributed computation tool. For that to happen, we need to address the technical, ethical, and legal aspects of the process, which I believe that it could be worked out. It could also turn into an underground method for computation, where people will quietly use it. We published our results so that people are aware of the possibility, and we will have safeguards against it before too late. Finally, parasitic computing could simply allow us to think outside of the box. My collaborator on the project, Vince Freeh, is doing just that, and he has some wonderful ideas on the future of parasitic computing, that will surprise many, I am sure, once he is ready to talk about them.
p53, a tumor suppressor molecule, one of the hubs of the cell's regulatory network.

Is there anything else on which you're working that you'd like to mention here?

My interest has lately turned towards biology. I see cell biology the area where network thinking would have a biggest impact in the near future. With the genome project finished, we have all the pieces, and we need to understand how the cell works as a whole. It is a wonderful challenge, with sure impact on everything from cancer research to drug development.

© 2002 21C Magazine