Learning user´s preference in controlling facial expressions of an embodied agent by the user´s face

In order to transmit facial expressions of the user with the face of an embodied agent in agent-mediated distance communication, it is necessary to realize "facial expression mapping", which gives synthetic facial expressions of the agent appropriate for representing the facial expressions of the user. However, it depends on the user\´s preference which synthetic facial expression should appear on the agent\´s face to describe each facial expression of the user. We discuss how to learn this user\´s preference in facial expression mapping through the interaction with the user during agent-mediated communication with facial expression mapping.