Exogenous control of belief learning in social networks

Abstract

Emerging growth of social media and convergence of mobile computing technologies offer unprecedented opportunities for targeting personalized purposeful communications into social networks. Leveraging reach and influence of such networks requires deeper understanding of how persuasive bias can be infused at individual and group level through exogenous control of the belief learning process in a network. Embedding linear control lends itself to a direct way to relate persuasive bias to the notion of communication exposure. We show in this paper that by applying linear exogenous control over DeGroot update cycles of beliefs, the agents can be persuaded to shift their beliefs in desired ways. We present the theoretical framework for such control and illustrate the theory with examples to show how consensus can be amplified or even destroyed.