Local Non-Bayesian Social Learning With Stubborn Agents

In this article, we study a social learning model in which agents iteratively update their beliefs about the true state of world using private signals and other non-Bayesian manner. Some are stubborn, meaning they attempt to convince others an erroneous (modeling fake news). We show that while learn on short timescales, “forget” it believe be longer timescales. Using these results, devise strategies for seeding stubborn so as disrupt learning, outperforms intuitive heuristics gives novel insights regarding vulnerabilities learning.