Improving reinforcement learning control via online bilinear action interpolation | IEEE Conference Publication | IEEE Xplore