Researchers at Sandia National Laboratories have been studying the ways that information, ideas and behaviors propagate through social networks in order to gain advance warning of cyber attacks or other threatening behavior.
The initial problem is how to explain the disparate consequences of seemingly similar triggering events. Thus, in 2005, the Danish newspaper Jyllands-Posten published cartoons featuring the Muslim Prophet Muhammad, prompting widespread protests. In 2006, by contrast, the Pope gave a lecture in which he made comments about Islam that were considered derogatory by some, but the ensuing controversy quickly faded away.
“While each event appeared at the outset to have the potential to trigger significant protests, the ‘Danish cartoons’ incident ultimately led to substantial Muslim mobilization, including massive protests and considerable violence, while outrage triggered by the pope lecture quickly subsided with essentially no violence,” wrote Sandia authors Richard Colbaugh and Kristin Glass. “It would obviously be very useful to have the capability to distinguish these two types of reaction as early in the event lifecycle as possible.”
What accounts for the difference in these outcomes? The intrinsic qualities of the events are not sufficient to explain why one had disruptive consequences and the other did not. Rather, the authors say, one must factor in the mechanisms of influence by which individual responses are shaped and spread.
By way of analogy, it has been shown that “it is likely to be impossible to predict movie revenues, even very roughly, based on the intrinsic information available concerning the movie” such as cast or genre, but that “it *is* possible to identify early indicators of movie success, such as temporal patterns in pre-release ‘buzz’, and to use these indicators to accurately predict ultimate box office revenues.”
The Sandia authors developed a methodology that reflects the “topological properties” of social and information networks — including the density and hierarchy of connections among network members — and modeled the dynamics of “social diffusion events” in which individuals exercise influence on one another.
They report that their model lends itself, among other things, to “distinguishing successful mobilization and protest events, that is, mobilizations that become large and self-sustaining, from unsuccessful ones early in their lifecycle.”
They tested the model to predict the spread of textual memes, to distinguish between events that generated significant protest (a May 2005 Quran desecration) and those that did not (the knighting of Salman Rushdie in 2007), and to provide early warning of cyber attacks.
The authors’ research was sponsored by the Department of Defense and the Department of Homeland Security, among others. See Early warning analysis for social diffusion events by Richard Colbaugh and Kristin Glass, originally published in Security Informatics, Vol. 1, 2012, SAND 2010-5334C.