Billions of people worldwide exchange information through networks like Facebook and the Internet, while computer and real viruses use the network connections to spread havoc. Networks are modeled using random graphs, and the flow of information is modeled as a stochastic process living on the random graph. We investigate information diffusion processes on random networks, a key example being an epidemic on a random graph.
Of particular interest is the relation between network topology, such as the existence of hubs and communities, and the behavior of processes on the network. We propose to study the behavior of critical information diffusion on random intersection graphs, which are flexible random graph models having an inherent community structure and to study the critical behavior of multistage epidemics. Further, we aim to investigate how one can use the behavior of critical epidemics on random graphs to obtain algorithms for detection of communities. Currently, many of such algorithms are based on random walks, and we believe that information diffusion models lead to improved understanding of network community detection. Finally, we also consider to study the relations between critical information diffusion and the minimal spanning trr on such random graphs.
|Supervisors||Remco van der Hofstad (TU/e) and Johan van Leeuwaarden (TU/e)|
|PhD Student||Souvik Dhara|
|Location||Eindhoven University of Technology (TU/e)|