Fairness Oriented Multiobjective Optimization#
Fomo is a Fairness Oriented Multiobjective Optimization toolbox for training regression and classification models. Fomo focuses on optimizing the trade-offs between various metrics in ML fairness that are often in direct conflict. The goal is to assist decision makers in weighing multiple criteria by generating good solution sets, rather than singular models.
Cite#
If you use Fomo please cite the following:
W.G. La Cava (2023). Optimizing fairness tradeoffs in machine learning with multiobjective meta-models. GECCO ‘23. arXiv:2304.12190