A kernel-based reinforcement learning approach to stochastic pole balancing control systems | IEEE Conference Publication | IEEE Xplore