Fairness and Bias in Robot Learning