2008 | OriginalPaper | Buchkapitel
Modelling Behaviour Cycles for Life-Long Learning in Motivated Agents
verfasst von : Kathryn Merrick
Erschienen in: Simulated Evolution and Learning
Verlag: Springer Berlin Heidelberg
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
Natural systems such as plants, animals and humans exhibit behaviour that forms distinct, rhythmic cycles. These cycles permit individuals and societies to learn, adapt and evolve in complex, dynamic environments. This paper introduces a model of behaviour cycles for artificial systems. This model provides a new way to conceptualise and evaluate life-long learning in artificial agents. The model is demonstrated for evaluating the sensitivity of motivated reinforcement learning agents. Results show that motivated reinforcement learning agents can learn behaviour cycles that are relatively robust to changes in motivation parameters.