Bias-corrected Q-learning to control max-operator bias in Q-learning | IEEE Conference Publication | IEEE Xplore