Achieving Long-Term Fairness in Sequential Decision Making
This work addresses fairness in sequential decision-making systems, which is crucial for mitigating biases in applications like hiring or lending, but it appears incremental as it builds on existing causal and optimization methods.
The paper tackles the problem of ensuring long-term fairness in sequential decision making by proposing a framework that uses path-specific effects on time-lagged causal graphs to measure fairness, formulating it as a constrained optimization problem converted to performative risk optimization, and applying repeated risk minimization with theoretical convergence analysis. Empirical results on synthetic and semi-synthetic datasets demonstrate the algorithm's effectiveness.
In this paper, we propose a framework for achieving long-term fair sequential decision making. By conducting both the hard and soft interventions, we propose to take path-specific effects on the time-lagged causal graph as a quantitative tool for measuring long-term fairness. The problem of fair sequential decision making is then formulated as a constrained optimization problem with the utility as the objective and the long-term and short-term fairness as constraints. We show that such an optimization problem can be converted to a performative risk optimization. Finally, repeated risk minimization (RRM) is used for model training, and the convergence of RRM is theoretically analyzed. The empirical evaluation shows the effectiveness of the proposed algorithm on synthetic and semi-synthetic temporal datasets.