Convention emergence through spreading mechanisms

In this paper we identify the spreading components required to design convention emergence mechanisms. We have empirically tested the effectiveness of our approach to emerge conventions in scenarios where the space of conventions is large. Moreover, we identify their shortcomings when used for the emergence of stable, global conventions despite unreliable communications (be them because of noise, maliciousness, or errors). Therefore, in order to guarantee the robust emergence of conventions we propose to endow agents with a self-protection component.