Removing biased data to improve fairness and accuracy
arXiv:2102.03054
Abstract
Machine learning systems are often trained using data collected from historical decisions. If past decisions were biased, then automated systems that learn from historical data will also be biased. We propose a black-box approach to identify and remove biased training data. Machine learning models trained on such debiased data (a subset of the original training data) have low individual discrimination, often 0%. These models also have greater accuracy and lower statistical disparity than models trained on the full historical data. We evaluated our methodology in experiments using 6 real-world datasets. Our approach outperformed seven previous approaches in terms of individual discrimination and accuracy.
16 pages, 5 Figures, 8 Tables
References in corpus (18)
- Equality of Opportunity in Supervised Learning
- Understanding Black-box Predictions via Influence Functions
- Counterfactual Fairness
- Fairness Testing: Testing Software for Discrimination
- Data Decisions and Theoretical Implications when Adversarially Learning Fair Representations
- On Fairness and Calibration
- AI Fairness 360: An Extensible Toolkit for Detecting, Understanding, and Mitigating Unwanted Algorithmic Bias
- Automated Directed Fairness Testing
- A Convex Framework for Fair Regression
- Learning Non-Discriminatory Predictors
- Achieving Fairness through Adversarial Learning: an Application to Recidivism Prediction
- Operationalizing Individual Fairness with Pairwise Fair Representations
- Does mitigating ML's impact disparity require treatment disparity?
- What's Sex Got To Do With Fair Machine Learning?
- A statistical framework for fair predictive algorithms
- Deep Weighted Averaging Classifiers
- Provably Fair Representations
- Learning unbiased features