H Geffner - Intelligenza Artificiale, 2011 - content.iospress.com
… from experience: the agent’s own experience, in reinforcement learning [48], or the experience
… Finally, in the model-based approach, the controller is not learned from experience but is …