Information Dissipation in Complex Networks

Abstract: How much 'influence' does an individual have in a social network? How important is an individual website in the World Wide Web? The topologies of the interactions in these examples are complex, and these questions are difficult to answer for dynamic complex networks. We transform this question into the following: how fast does information about an individual node dissipate over time? If an individual node is influential then it will likely change the state of the network noticably, such that a future network state provides a good estimate of a previous node's state. Conversely, if an individual has no influence then one gains no such information about the node by knowing a future state. In order to make 'information dissipation' a generic quantitative property of the network itself we overlay the classic Ising-spin model over the network and study how uncertain a previous state of each spin becomes over time. In this talk I will introduce the Ising-spin model, complex networks, and our ongoing understanding of information dissipation in complex networks of spins.