As machine learning algorithms are increasingly deployed for high-impact
automated decision making, ethical and increasingly also legal standards demand
that they treat all individuals fairly, without discrimination based on their
age, gender, race or other sensitive traits. In recent