Norms and Learning in Probabilistic Logic-Based Agents

摘要

This paper proposes a new simulation approach for investigating phenomena such as norm emergence and internalization in large groups of learning agents. We define a probabilistic defeasible logic instantiating Dung's argumentation framework. Rules of this logic are attached to probabilities and describe the agents' minds and behaviour. We thus adopt the paradigm of reinforcement learning over this probability distribution to allow agents to adapt to their environment.