Artificial Emotion Model Based on Neuromodulators and Q-learning

摘要

An emotion behavior learning model is presented for intelligent system in this paper. The proposed model was constructed on the basis of those corresponding relationships between emotions and neurotransmitters which was indicated by the brain science, neuromodulators and meta-parameters of Q-learning which made the learning process more adaptive to environment. The emotional effect on the learning capability of intelligent system was observed in two stimulation experiments. The results confirm that artificial emotion model could improve the actions of intelligent system.