DS/CSS Seminar Series: Fang-Yi Yu

Date: 
Thu, 01/16/2020 - 12:00pm

Ehrlicher Room, 3100 North Quad

Community and Deliberation

Abstract

We first consider how does community structures affect people to reach a consensus. We propose a family of majority-like dynamics which generalizes several previous opinion formation dynamics.  I show a dichotomy theorem of consensus time.  For any "majority-like" update function, there is a threshold of the level of intercommunity connectivity.  If the intercommunity connectivity of the graph is higher than the threshold, the system quickly converges to consensus.  Otherwise, the system can get "stuck" and take a long time to reach consensus. 

Finally, we consider when agents' opinions can induce community structure in social networks.  Besides agents changing their opinion, the network evolves due to disagreements between agents' opinions. We propose dynamics on graphs that consist of weak and strong ties, and only weak ties can be affected by the selection process. Through simulations, we discovered that with high selection, weak ties might inhibit consensus, and the graph forms communities of homogeneous opinions. 

Speaker Bio

Fang-Yi Yu is a Graduate Student Research Assistant and Research Fellow in the School of Information.