Learning from Sparse and Delayed Rewards with a Multilayer Spiking Neural Network | IEEE Conference Publication | IEEE Xplore