Adapting Static Fairness to Sequential Decision-Making: Bias Mitigation Strategies towards Equal Long-term Benefit Rate

TL;DR

This paper introduces ELBERT, a long-term fairness concept integrated into Markov Decision Processes, enabling bias mitigation in sequential decision-making while maintaining utility, and simplifies policy optimization.

Contribution

It proposes ELBERT, a novel long-term fairness measure for sequential decisions, and develops ELBERT-PO, a practical bias mitigation method with simplified policy gradients.

Findings

ELBERT-PO significantly reduces bias across various environments.

ELBERT effectively addresses temporal discrimination issues.

Policy gradients for ELBERT can be analytically simplified.

Abstract

Decisions made by machine learning models can have lasting impacts, making long-term fairness a critical consideration. It has been observed that ignoring the long-term effect and directly applying fairness criterion in static settings can actually worsen bias over time. To address biases in sequential decision-making, we introduce a long-term fairness concept named Equal Long-term Benefit Rate (ELBERT). This concept is seamlessly integrated into a Markov Decision Process (MDP) to consider the future effects of actions on long-term fairness, thus providing a unified framework for fair sequential decision-making problems. ELBERT effectively addresses the temporal discrimination issues found in previous long-term fairness notions. Additionally, we demonstrate that the policy gradient of Long-term Benefit Rate can be analytically simplified to standard policy gradients. This simplification makes conventional policy optimization methods viable for reducing bias, leading to our bias mitigation approach ELBERT-PO. Extensive experiments across various diverse sequential decision-making environments consistently reveal that ELBERT-PO significantly diminishes bias while maintaining high utility. Code is available at https://github.com/umd-huang-lab/ELBERT.