This talk describes the development of reinforcement learning within the Sigma graphical cognitive architecture. Reinforcement learning has been deconstructed in terms of the interactions among more basic mechanisms and knowledge in Sigma, making it a derived capability rather than a de novo mechanism. Basic reinforcement learning -- both model-based and model-free -- are demonstrated, along with the intertwining of model learning. (Winter Intelligence Oxford - AGI12 - http://agi-conference.org/2012)




This talk describes the development of reinforcement learning within the Sigma graphical cognitive architecture. Reinforcement learning has been deconstructed in terms of the interactions among more basic mechanisms and knowledge in Sigma, making it a derived capability rather than a de novo mechanism. Basic reinforcement learning -- both model-based and model-free -- are demonstrated, along with the intertwining of model learning. (Winter Intelligence Oxford - AGI12 - http://agi-conference.org/2012)




http://www.youtube.com/watch?v=yLo_-yjBwpA