แนวคิดหลัก
This work investigates the use of subjective logic to model opinions and belief change in social networks, proposing a subjective logic belief/opinion update function to represent belief change as communication occurs in social networks.
บทคัดย่อ
The authors propose a model for social networks using elements of subjective logic, such as multinomial opinions, trust opinions, and belief fusion operators. They introduce a belief/opinion update function that uses subjective logic's trust discount and belief fusion to represent belief change in social networks.
The authors find that an update function with belief fusion from subjective logic does not have ideal properties, such as weak convergence and non-increasing uncertainty, to model a rational update in a social network. However, they show that an update function with cumulative belief fusion can represent different scenarios not described in previous work, such as consensus, balanced opposite, and unbalanced opposite opinions among agents.
The key insights are:
Consensus case: Agents with agreeing or disagreeing opinions will accumulate evidence and converge to complete certainty about the proposition.
Balanced opposite case: Agents with exactly opposite opinions will converge to complete indecision about the proposition.
Unbalanced opposite case: Agents with conflicting but unbalanced opinions will either radicalize or converge to a non-radical point, depending on the initial distance between their opinions.
The authors suggest that the update function with cumulative belief fusion has the potential to model belief dynamics in social networks that were not captured by previous models.