Computational underpinnings of cooperative behaviour. From agent-based simulations to cognitive insights
In this talk I will discuss implementing cooperative behaviour in artificial agents based on BDI architectures. This will crucially involve insights into bounded agency and theory of mind from both cognitive and logical perspective. I will present a new benchmark task for cooperation. Within it, I will distinguish different types of implicit cooperation between many epistemic agents. Those results will be in the end compared to empirical findings based on studying human behaviour.
This is joint work with Mads Baatrup.