| Document Date: 2012-10-14 18:35:22
Open Document File Size: 1,24 MBShare Result on Facebook
City Perth / Manno-Lugano / AND SRN / Copenhagen / / Company Super Mario Bros / Multi-Dimensional Recurrent Neural Networks / Neural Networks / YouTube / Nintendo / Artificial Neural Networks / Markus Persson’s Infinite Mario Bros / Infinite Mario Bros. / / Country Switzerland / Australia / Denmark / / / Facility Charles University / IT University of Copenhagen / / IndustryTerm feedforward network / simulated car racing / evolutionary algorithms / car racing skills / simulated car racing competition / large input networks / indirect encoding algorithm / reinforcement learning algorithms / large network / recurrent networks / competition web page / generated large network / hybrid neuroevolution/genetic programming algorithm / simulated car racing competitions / search space / highdimensional search spaces / learning algorithm / learning algorithms / / MarketIndex set 4000 / / NaturalFeature Mario falls / / Organization Charles University in Prague / Computational Intelligence / Congress / Federal Government / University of Copenhagen / / Person Mario Big / Satisfaction / Jan Koutn´ık / Julian Togelius / Sergey Karakovskiy / Tom Schaul / / Position large SRN controller / large controller / neural network location evaluator / player / human player / cyclic controller / HyperGP controller / controller / / ProgrammingLanguage Java / Python / / Technology YPER GP ALGORITHM / RL algorithms / operating system / RL algorithm / intelligent agent / learning algorithm / reinforcement learning algorithms / neural network / indirect encoding algorithm / Java / learning algorithms / hybrid neuroevolution/genetic programming algorithm / / URL http /
SocialTag |
Super Mario Evolution Julian Togelius, Sergey Karakovskiy, Jan Koutn´ık and J¨urgen Schmidhuber Abstract— We introduce a new reinforcement learning benchmark based on the classic platform game Super Mario Bros. The