Learning Action Sequences Through Imitation in Behavior Based Architectures

摘要: In this paper a new architecture for learning action sequences through imitation is proposed. Imitation occurs by means of observing and applying basic behaviors. When an agent has observed another applied the sequence later on, imitated can be seen as meme. Agents that behave similarly therefore grouped their typical behavioral patterns. This thus explores from view memetic proliferation. Combining with meme theory we show simulating societies significant performance improvements achieved. The quantified using entropy measure to qualitatively evaluating emerging clusters. Our approach demonstrated example society emotion driven agents imitate each other reach pleasant emotional state.