Paul Rosenbloom - Deconstructing Reinforcement Learning in Sigma
494 views · Published 30 May 2013 · 26:50 · Indexed 31 October 2026
Channel: Science, Technology & the Future · 2013 · Science & Technology
Winter Intelligence Oxford - AGI12 - http://agi-conference.org/2012 ==Deconstructing Reinforcement Learning in Sigma== Abstract. This article 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. Keywords: Reinforcement learning, cognitive architecture, graphical models. http://agi-conference.org/2012/wp-content/uploads/2012/12/paper_44.pdf Paul Rosenbloom, University of Southern California, USA
More from this channel
-
5:26
Matrioshka Brain - Novart - Scifi inspired music with Medley of 90s Bryce Experiments
-
14:03
Aussiecon4 Hand Waving, Rule Bending and Other Dirty Tricks of SciFi p1
-
5:17
Alex Holmes - The ANU League of Extraordinary Atheists
-
5:13
Reidar Lystad - Macquarie University Atheists League
-
10:52
PZ Myers at Freethought University Alliance 3 of 5
-
2:09
PZ Myers at Freethought University Alliance 5 of 5
-
5:42
Short talk at Bayesian Artificial Intelligence Book Launch - Ann E. Nicholson
-
8:35
An Interview with Russell Blackford part 1 of 6