Abstract:

Many opinion dynamics models have been studied under a variety of conditions. In particular much attention goes to the influence of network characteristics on the outcomes of the models (be they polarisation, consensus or fragmentation). Seemingly innocuous assumptions are however, often overlooked. In this talk I present results studying models (Axelrod model of culture formation and Social Reinforcement Learning) while breaking these kinds of assumptions: whether the interaction between agents is symmetric, and whether the opinions live on a space with extremes or not. We see that these assumptions can be important to the emergent dynamics, and deserve more attention.