Policy Gradient Methods for Reinforcement Learning with Function Approximation

@misc{citeulike:749436, abstract = {Function approximation is essential to reinforcement learning, but the standard approach of approximating a value function and determining a policy from it has so far proven theoretically intractable.}, author = {Sutton, R. and Mcallester, D. and Singh, S. and Mansour, Y. }, citeulike-article-id = {749436}, keywords = {learning, reinforcement}, posted-at = {2006-07-10 15:52:34}, priority = {2}, title = {Policy Gradient Methods for Reinforcement Learning with Function Approximation}, url = {http://citeseer.ist.psu.edu/sutton99policy.html}, year = {1999} }

TY - GEN ID - citeulike:749436 L3 - citeulike-article-id:749436 TI - Policy Gradient Methods for Reinforcement Learning with Function Approximation N2 - Function approximation is essential to reinforcement learning, but the standard approach of approximating a value function and determining a policy from it has so far proven theoretically intractable. KW - learning KW - reinforcement AU - Sutton, R AU - Mcallester, D AU - Singh, S AU - Mansour, Y PY - 1999/// UR - http://citeseer.ist.psu.edu/sutton99policy.html ER -