Advances in Memetic Automaton: Toward Human-Like Autonomous Agents in Complex Multi-Agent Learning Problems

The meme-centric memetic automaton (MA) was recently proposed as an adaptive entity or a software agent wherein memes are defined as the building blocks of knowledge. The conceptualization of MA has led to the development of a large number of potentially rich meme-inspired designs that form a cornerstone of memetic computation as tools for problem-solving. In this study, we investigate the use of memetic multi-agent systems to develop more intelligent and human-like autonomous agents by taking MA as the essential backbone of the agent. Taking inspiration from a psychological Broadbent-Treisman Attenuation Model, we propose an attention intensity control method in meme expression for enhancing agents' perception of the value of all kinds of information captured from the environment, hence leading to a greater capability of meme knowledge generalization. Our particular focus is placed on the design of meme selection for more effective knowledge transmission across the population. To this end, we introduce a bidirectional imitation strategy based on agents' estimation of the importance and/or uncertainty of decision making in a dynamic environment. Experiments on a minefield navigation simulation as well as a commercial video game demonstrate the superior performance of our proposed method compared to state-of-the-art methods.