EiFFFeL: Enforcing Fairness in Forests by Flipping Leaves

TL;DR

EiFFFeL is a post-processing method that enhances fairness in decision forests by relabeling leaves, achieving user-defined fairness levels with minimal accuracy loss.

Contribution

The paper introduces EiFFFeL, a novel leaf-flipping approach for enforcing fairness in tree-based models through post-processing relabeling strategies.

Findings

Achieves user-defined group fairness levels

Maintains high accuracy with fairness enforcement

Effective across different decision forest configurations

Abstract

Nowadays Machine Learning (ML) techniques are extensively adopted in many socially sensitive systems, thus requiring to carefully study the fairness of the decisions taken by such systems. Many approaches have been proposed to address and to make sure there is no bias against individuals or specific groups which might originally come from biased training datasets or algorithm design. In this regard, we propose a fairness enforcing approach called EiFFFeL:Enforcing Fairness in Forests by Flipping Leaves which exploits tree-based or leaf-based post-processing strategies to relabel leaves of selected decision trees of a given forest. Experimental results show that our approach achieves a user defined group fairness degree without losing a significant amount of accuracy.