David B. Newlin
The purpose for this discussion is to provide information on the human mirror neuron system (MNS) to aid in the design of artificial autonomous agents. Specifically, we emphasize the motivational and affective aspects of the MNS to allow the design of agents that themselves exhibit motivated behavior. Current evidence indicates that the MNS assigns motives to observed behavior. A tentative working model of the MNS is presented to organize this information and to highlight important issues that remain unresolved, such as the architecture of hemispheric asymmetry of function in the left and right MNSs and mesolimbic motivation systems. We conclude with recommendations, particularly in terms of dual, functionally asymmetric control systems, for designing motivated autonomous agents using information learned about the human MNS.
Subjects: 4. Cognitive Modeling; 17. Robotics
Submitted: Oct 17, 2007