Fairness in Reinforcement Learning with Bisimulation Metrics
It addresses fairness for automated decision-making systems in dynamic environments, offering a novel approach to mitigate group disparities.
The paper tackles fairness in reinforcement learning by connecting bisimulation metrics to group fairness, proposing a method that learns reward functions and observation dynamics to ensure fair treatment of groups while maintaining problem fidelity. It demonstrates effectiveness on lending and college admission benchmarks.
Ensuring long-term fairness is crucial when developing automated decision making systems, specifically in dynamic and sequential environments. By maximizing their reward without consideration of fairness, AI agents can introduce disparities in their treatment of groups or individuals. In this paper, we establish the connection between bisimulation metrics and group fairness in reinforcement learning. We propose a novel approach that leverages bisimulation metrics to learn reward functions and observation dynamics, ensuring that learners treat groups fairly while reflecting the original problem. We demonstrate the effectiveness of our method in addressing disparities in sequential decision making problems through empirical evaluation on a standard fairness benchmark consisting of lending and college admission scenarios.