Department of Information Sciences and Technologies, ISCTE
Lisbon, Portugal
luis@iscte.pt
*Department of Computing Sciences, Faculty of Sciences, University of Lisbon,
Lisbon, Portugal
Helder.Coelho@di.fc.ul.pt
According to [Botelho and Coelho 1998a, 1998b] emotion plays an important role at the control level of emotion systems. Essentially, emotion may lead to reflexive behaviors, to the creation or intensification of motivators, and to set new criteria for plan selection. Emotion is a sequential, possibly iterative process that comprises appraisal stages, signal generation stages and emotion-responses. In this paper, we argue that adaptive behavior may be achieved through a process of emotion learning. We point out three ways in which an agent can improve its emotion processes and therefore become more adaptive: learning new affective appraisal rules, learning new emotion-signals, and learning new emotion-responses.
[Botelho and Coelho 1998a] Botelho, L.M. and Coelho, H. (1998) ``Artificial autonomous agents with artificial emotions'', Proc. of the Second International Conference on Autonomous Agents (Agents'98), p449450.