Towards A Computational Model for Learning Affective Empathy Responses in Robotic System
Mark Allison · 2024
Appropriate behavioral responses are critical for social robots to achieve trust and to build and maintain long term relationships. In particular, empathetic robots can facilitate complex decision making in collaborative environments as human perspectives and emotional states are considered. Furthermore, empathy may broaden the application of robotic systems into domains such as healthcare and even crisis response, thereby mitigating social isolation, facilitating therapy or saving lives. In this paper, we build upon psychological and neuroscience models towards a computational model of prosocial empathy that actively tracks and considers the human emotional state. By leveraging a reinforcement learning model we provide for a tailored experience with more precise responses over time.