On the use of backpropagation in associative reinforcement learning | IEEE Conference Publication | IEEE Xplore