Reinforcement Learning with Stepwise Fairness Constraints
This work addresses fairness in dynamic AI applications like credit and employment, introducing a novel framework for sequential settings.
The paper tackles the problem of ensuring group fairness at each time step in sequential decision-making settings using reinforcement learning, and presents learning algorithms with strong theoretical guarantees for policy optimality and fairness violation.
AI methods are used in societally important settings, ranging from credit to employment to housing, and it is crucial to provide fairness in regard to algorithmic decision making. Moreover, many settings are dynamic, with populations responding to sequential decision policies. We introduce the study of reinforcement learning (RL) with stepwise fairness constraints, requiring group fairness at each time step. Our focus is on tabular episodic RL, and we provide learning algorithms with strong theoretical guarantees in regard to policy optimality and fairness violation. Our framework provides useful tools to study the impact of fairness constraints in sequential settings and brings up new challenges in RL.