Open access
2026
Deployment-Aware Comparison of Lightweight Bias Mitigation Techniques in Machine Learning Pipelines
A deployment-aware empirical comparison of three lightweight bias mitigation methods (Reweighing, Disparate Impact Remover, and Equalized Odds post-processing) against an unmitigated baseline is presented.
S. Parthasarathy
· IEEE Access · 0 citations