Dynamics of Uncertain Opinions in Social Networks

We study the impact of uncertainty or lack of confidence on the propagation of opinions and the formation of consensus or divergence in social networks. Many social and political issues often lead to extreme divergence of opinions, however, very often people hold opinions without a full sense of confidence. In this work, we are interested in how uncertain opinions evolve through interactions in a social network. Our study considers two types of agents, informed agents with high confidence (low uncertainty) and uninformed agents with low confidence (high uncertainty), and examines the impact of the numbers of informed vs. Uninformed agents on opinion dynamics. In particular, we use the well-known Subjective Logic approach to model uncertain opinions (i.e., An opinion consisting of belief, disbelief, and uncertainty). We devise two different trust attitudes, certainty-based trust and similarity-based trust, and study how the opinions are affected by different trust attitudes and how this impacts consensus.