Abstract:

Whom should we vaccinate in a large population contact network to mitigate an epidemic outbreak? Which places should one avoid visiting to save themself from potential infections? How to compute who gets infected in a large network as quickly as possible? These are important questions in epidemiology and public health. Epidemic outbreaks are often modeled as diffusion processes over underlying contact networks. The large size and the dynamic nature of real-world contact networks make the analysis of the diffusion process, and thus answering the questions above, a major challenge. In this talk, we explore provably near-optimal approaches which leverage the spectral radius of contact networks to facilitate diffusion analysis. The first half of the talk will focus on generating smaller representation of a large contact network to speed up existing techniques, and the second half will focus on directly answering questions related to vaccination and quarantining in people-location networks.