Using Pareto Simulated Annealing to Address Algorithmic Bias in Machine Learning

TL;DR

This paper introduces a multi-objective optimization approach using Pareto Simulated Annealing to explicitly incorporate bias minimization alongside accuracy in machine learning models, aiming to reduce algorithmic bias.

Contribution

It presents a novel multi-objective optimization method that explicitly considers bias as a criterion during model training, improving fairness without sacrificing accuracy.

Findings

Effective bias reduction demonstrated on real-world datasets

Multi-objective optimization balances accuracy and fairness

Method outperforms traditional single-objective approaches

Abstract

Algorithmic Bias can be due to bias in the training data or issues with the algorithm itself. These algorithmic issues typically relate to problems with model capacity and regularisation. This underestimation bias may arise because the model has been optimised for good generalisation accuracy without any explicit consideration of bias or fairness. In a sense, we should not be surprised that a model might be biased when it hasn't been "asked" not to be. In this paper, we consider including bias (underestimation) as an additional criterion in model training. We present a multi-objective optimisation strategy using Pareto Simulated Annealing that optimise for both balanced accuracy and underestimation. We demonstrate the effectiveness of this strategy on one synthetic and two real-world datasets.