Projective simulation for classical learning agents: a comprehensive investigation

We study the model of projective simulation (PS), a novel approach to artificial intelligence based on the stochastic processing of episodic memory, which was first introduced in [H.J. Briegel and G. De las Cuevas. Sci. Rep. 2, 400, (2012)]. Throughout the paper we provide a detailed analysis of the PS features and examine its performances, where we mainly focus on its achievable efficiency and learning times. We situate the PS agent in different learning scenarios, and study its learning abilities. A variety of scenarios are being considered, thereby demonstrating the model's flexibility. In addition, to put the PS scheme in context, we compare its performance with those of Q-learning and learning classifier systems, two popular representatives of artificial intelligence models. It is shown that PS is a competitive artificial intelligence model of unique properties and strengths.