“Dynamic control of spreading processes on networks”

Speaker: Soheil Eshghi, Ph.D.
Postdoc, Crawford Lab

Abstract: Epidemics, ideas, and behaviors spread in networks. Ultimately, we seek to characterize these spreading processes, to understand how we can intervene in them, and to finally devise interventions that lead to favorable outcomes at low cost, e.g., curbing the spread of an epidemic through targeted vaccination, maintaining ethical behavior in an organization through network-based incentive design.  

The choice of models for these phenomena has typically been constrained by the dearth of data, leading to simplifications that merge heterogeneous populations into “homogeneous” groups in the models. With the rapid increase in the scale of available data, some of the bottle-necks necessitating these simplifications have changed; while complicated networked models may now be identified, the sheer number of possible time-varying interventions exploiting these heterogeneities leads to a “curse of dimensionality”, requiring the use of principled approaches to resource allocation in the design of policies and interventions.  

I will present some of our work on epidemics, advertising, and organizational behavior that highlight the benefits of using such principled dynamic network-based policies.  

Bio: Soheil Eshghi is a postdoctoral associate in the Crawford Lab at the Yale School of Public Health (YSPH), working on the control of epidemics. He received his PhD in Electrical and Systems Engineering from the University of Pennsylvania, and conducted postdoctoral research at Cornell University and Yale Institute for Network Science (YINS). His research interests are in constrained dynamic decision-making for biological, technological, and social systems.

.