CIF: Small: Contagion Processes in Multi-layer and Multiplex Networks

Carnegie Mellon University, Pittsburgh PA

Investigators

Abstract

Online social networks are increasingly serving as a primary source of real time news for billions of people. In addition, online social networks often act as a medium to initiate and facilitate collective action. This project advances our understanding of contagion processes such as the spread of information and influence in online social networks. Unlike much of the current literature focusing on single and isolated networks, the project considers contagions taking place over multiple networks. By a combination of approaches involving mathematical modeling and analysis as well as real-life data sets, this project seeks a better understanding of i) how the participation of multiple networks affects the speed and extent that information propagates; ii) why different topics have different spreading characteristics over the same population; and iii) the ways misinformation spreads in social networks and how this might be countered by efficiently injecting correct information. The project will engage students from underrepresented groups and will include outreach activities to high-schools to encourage them to pursue a STEM education. This project seeks to advance the state-of-the art in modeling and analysis of two major classes of contagion processes over multi-layer and multiplex networks: i) Information propagation (i.e., simple contagion), and ii) Influence propagation (i.e., complex contagion). Advanced tools from random graph theory and percolation theory are leveraged to derive fundamental relations between network parameters (e.g., degree distribution, size, clustering, degree correlations, coupling strength, and transmissibility) and the threshold, probability, and expected size of information cascades. The analysis of contagion is extended to a wide range of network models including real-world topologies. The modeling of complex contagion is extended from binary-state dynamics to multi-stage dynamics and this will be utilized in studying the spread of misinformation. On a broader level, this project contributes to a largely unexplored field of analyzing random graph models formed by the union of two or more random graphs. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.