Parametric Policy Gradients for Robotics Frank Sehnke, Thomas R¨uckstieß, Martin Felder and J¨urgen Schmidhuber Abstract— Slow convergence is a major problem for policy gradient methods. It is a consequence of the f - R. J. Thomas - Document - PDFSEARCH.IO - Document Search Engine

Back to Results

First Page	Meta Content
	Parametric Policy Gradients for Robotics Frank Sehnke, Thomas R¨uckstieß, Martin Felder and J¨urgen Schmidhuber Abstract— Slow convergence is a major problem for policy gradient methods. It is a consequence of the f Add to Reading List Document Date: 2013-05-15 08:23:49 Open Document File Size: 2,05 MB Share Result on Facebook City Aberdeen / San Francisco / Manno-Lugano / Cambridge / Beijing / New York / / Company Neural Information Processing Systems / MIT Press / ICANN / / Country Germany / Switzerland / Jordan / China / / / Facility Institute of Applied Mechanics / Courtesy Institute of Automatic Control Engineering / Courtesy Institute of Applied Mechanics / Australian National University / Institute of Automatic Control Engineering / / IndustryTerm statistical gradient-following algorithms / stochastic optimization algorithms / typical solution / policy gradient algorithms / online policy gradient learning / simultaneous perturbation algorithm / / Organization vol. / Cognitive Science Society / Institute of Automatic Control Engineering / MIT / Institute of Applied Mechanics / Faculty of Robotics and Embedded Systems / Courtesy Institute of Automatic Control Engineering / Courtesy Institute of Applied Mechanics / Australian National University / Research Foundation / / Person Martin Felder / Frank Sehnke / Morgan Kaufmann / / Position Natural actor-critic / linear controller / rt / scalar reward rt / differentiable controller / natural actor critic / head / final controller / same controller / controller / / Product Franklin / / ProgrammingLanguage J / APL / / ProvinceOrState California / / PublishedMedium Machine Learning / / Technology policy gradient algorithms / stochastic optimization algorithms / PGPE algorithm / machine learning / simulation / statistical gradient-following algorithms / Policy-gradient algorithms / simultaneous perturbation algorithm / / URL http / SocialTag Estimation theory Dimensional analysis CMA-ES Applied mathematics Science Measurement Statistics Reinforcement learning