Individually Fair Gradient Boosting
arXiv:2103.16785
Abstract
We consider the task of enforcing individual fairness in gradient boosting. Gradient boosting is a popular method for machine learning from tabular data, which arise often in applications where algorithmic fairness is a concern. At a high level, our approach is a functional gradient descent on a (distributionally) robust loss function that encodes our intuition of algorithmic fairness for the ML task at hand. Unlike prior approaches to individual fairness that only work with smooth ML models, our approach also works with non-smooth models such as decision trees. We show that our algorithm converges globally and generalizes. We also demonstrate the efficacy of our algorithm on three ML problems susceptible to algorithmic bias.
ICLR Camera-Ready Version
References in corpus (19)
- Scikit-learn: Machine Learning in Python
- Towards Deep Learning Models Resistant to Adversarial Attacks
- Preventing Fairness Gerrymandering: Auditing and Learning for Subgroup Fairness
- Fast learning rates for plug-in classifiers
- Certifying Some Distributional Robustness with Principled Adversarial Training
- Robust Wasserstein Profile Inference and Applications to Machine Learning
- AI Fairness 360: An Extensible Toolkit for Detecting, Understanding, and Mitigating Unwanted Algorithmic Bias
- The Frontiers of Fairness in Machine Learning
- Fairness Without Demographics in Repeated Loss Minimization
- Flexibly Fair Representation Learning by Disentanglement
- Fairness Through Computationally-Bounded Awareness
- Operationalizing Individual Fairness with Pairwise Fair Representations
- Online Learning with an Unknown Fairness Metric
- Minimax Statistical Learning with Wasserstein Distances
- Training individually fair ML models with Sensitive Subspace Robustness
- Calibration for the (Computationally-Identifiable) Masses
- An Empirical Study on Learning Fairness Metrics for COMPAS Data with Human Supervision
- SenSeI: Sensitive Set Invariance for Enforcing Individual Fairness
- Debiasing representations by removing unwanted variation due to protected attributes