Fairness-Sensitive Policy-Gradient Reinforcement Learning for Reducing Bias in Robotic Assistance | IEEE Conference Publication | IEEE Xplore