A Convex Framework for Fair Regression

By Richard A. Berk, Hoda Heidari, Shahin Jabbari, Matthew Joseph, Michael Kearns, Jamie H. Morgenstern, Seth Neel, Aaron Roth. 2017

379 citations recorded by the source corpus.

We introduce a flexible family of fairness regularizers for (linear and logistic) regression problems. These regularizers all enjoy convexity, permitting fast optimization, and they span the rang from notions of group fairness to strong individual fairness. By varying the weight on the fairness regularizer, we can compute the efficient frontier of the accuracy-fairness trade-off on any given dataset, and we measure the severity of this trade-off via a numerical quantity we call the Price of Fairness (PoF). The centerpiece of our results is an extensive comparative study of the PoF across six different datasets in which fairness is a primary consideration.

Subfields: Safety Research, AI Ethics, Fairness & Society, Mathematical Optimization

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Safety Research

In AI Ethics, Fairness & Society

In Mathematical Optimization

Citations in the corpus

Does Gender Matter? Towards Fairness in Dialogue Systems (2019, neutral): “, 2015), regression (Berk et al., 2017), graph embedding (Bose”
A Survey on Bias and Fairness in Machine Learning (2019, methodology): “[13] proposes a fair regression method along with evaluating it with a measure introduced as the "price of fairness" (POF) to measure accuracy-fairness trade-offs.”
Calibration for the (Computationally-Identifiable) Masses (2017, positive): “This result adds to the growing literature on fairness-accuracy trade-oﬀs [FKL16,BHJ + 17,CG17].”
Empirical Risk Minimization under Fairness Constraints (2018, background): “Methods in the second family enforce fairness directly during the training step: [18, 19, 20, 3, 21, 22, 23, 24, 25, 26] are examples of this method which provide non-convex approaches to the solution of the problem or they derive consistency results just for the non-convex formulation resorting lat”
Algorithmic Fairness (2020, methodology): “This dataset was used, for example, by [14], where gender was studied as the sensitive feature, and [13], where race was studied as the sensitive feature.”
Nonconvex Optimization for Regression with Fairness Constraints (2018, methodology): “Berk et al. (2017) considered a convex framework where fairness is imposed by the regularizer, and Pérez-Suay et al. (2017) introduced a reguralizer inspired by the Hilbert-Schmidt Independence Criteria (Gretton et al., 2005).”
Trustworthy Artificial Intelligence: A Review (2022, background): “In-the-loop 3 Data/Model Bias Classification [13, 91, 99, 223] Regression [16] Classification, Regression [225] Decision Tree [98]”
Avoiding Resentment Via Monotonic Fairness (2019, background): “An alternative approach is to learn a single classifier on X to predict Y , and to encourage fairness by regularization using a fairness-promoting penalty [27, 26, 5] or constraints [43, 44, 1].”
A Geometric Solution to Fair Representations (2020, methodology): “, penalizing unfair inferences [2, 11], or creating representations that do not strongly depend on protected features [14, 20, 23].”
Debiasing Pre-Trained Language Models via Efficient Fine-Tuning (2022, methodology): “To mitigate biased behaviors in machine learning models, researchers have proposed meth-ods targeting different tasks and domains, such as classification (Menon and Williamson, 2018; Roh et al., 2021), regression (Agarwal et al., 2019; Berk et al., 2017), and adversarial learning (Xu et al., 2018).”
Multicalibration: Calibration for the (Computationally-Identifiable) Masses (2018, result): “This result adds to the growing literature on fairness-accuracy trade-offs (Fish et al., 2016; Berk et al., 2017; Chouldechova & G’Sell, 2017).”