Back
Much existing work in reinforcement learning involves environments that are either intentionally neutral, lacking a role for cooperation and competition, or intentionally simple, when agents need imagine nothing more than that they are playing versions of themselves. Richer game theoretic notions become important as these constraints are relaxed. For humans, this encompasses issues that concern utility, such as envy and guilt, and that concern inference, such as recursive modeling of other players, I will discuss studies treating a paradigmatic game of trust as an interactive partially-observable Markov decision process, and will illustrate the solution concepts with evidence from interactions between various groups of subjects, including those diagnosed with borderline and anti-social personality disorders.
Peter Dayan (Max Planck Institute for Biological Cybernetics: Computational Neuroscience)
Director
Peter Dayan read Mathematics at Cambridge, studied for his PhD with David Willshaw in Edinburgh, and did postdocs with Terry Sejnowski at the Salk Institute and Geoff Hinton in Toronto. He was an assistant professor in the Department of Brain and Cognitive Sciences at MIT, and was a founding faculty member of the Gatsby Computational Neuroscience Unit at UCL. He is currently a Director at the Max Planck Institute for Biological Cybernetics. His interests include affective decision making and neural reinforcement learning.