UC San Diego

The study of human interaction has taken an unprecedented quantitative and technological turn, driven by the advent of online social networks, the consequent availability of massive tracks of human behavior, and advancements in computing and analytical tools. To provide a rigorous understanding of social networks, this dissertation proposes theories of human networked interaction, able to explain and predict individual and global outcomes observable in a population. Populations of heterogeneous, complex individuals are modeled using homogeneous, simple agents who act according to simple incentives and local rules. The resulting models are prone to analysis using algorithmic complexity, game theory and statistics, enabling the test of predictions about the original population. Rigorous models and methods are proposed to study several aspects of human interaction: scenarios of social computation, in which interconnected individuals cooperate to solve a problem in a distributed fashion; the analysis of online social networks, to study influence spreading over networks and to test hypothesis of behavioral interaction; the theoretical analysis of human populations, to understand fundamental capabilities and limits of social coordination in complex scenarios