A Theoretical Consideration on Artificial Imitation Based on Mental Image Directed Semantic Theory

In artificial or robotic imitation, it is extremely difficult for an artifact to understand what in human demonstration is significant or not because there are too many things to attend to as it is. The Mental Image Directed Semantic Theory (MIDST) has proposed an omnisensory mental image model and its description language L_md intended to facilitate intuitive human- artifact interaction such that happens between non-expert people and home robots. This paper presents a brief sketch of L_md and a theoretical consideration on artificial imitation of human action driven by human suggestion interpreted in L_md, controlling the artificial attention mechanism efficiently.