Model-agnostic bias mitigation methods with regressor distribution control for Wasserstein-based fairness metrics.
Alexey MiroshnikovKonstandinos KotsiopoulosRyan FranksArjun Ravi KannanPublished in: CoRR (2021)
Keyphrases
- probability distribution
- mathematical model
- computational cost
- qualitative and quantitative
- probabilistic model
- statistical models
- computational model
- feature selection
- linear models
- similarity measure
- high level
- support vector machine
- benchmark datasets
- significant improvement
- preprocessing
- machine learning methods
- probability density function
- classification models
- control method
- neural network
- similarity metrics
- learned models