Penalizing Unfairness in Binary Classification

By Yahav Bechavod, Katrina Ligett. 2017

126 citations recorded by the source corpus.

How researchers see this paper

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

Citations in the corpus

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…”
Algorithmic Fairness (2020, methodology): “This dataset was used, for example, by [13] and [141], where gender was studied as the sensitive feature.”
Adversarial Removal of Demographic Attributes Revisited (2019, background): “Several approaches have been proposed to learn classiﬁers that are invariant (unbiased with respect) to protected attributes: cost-sensitive (Agar-wal et al., 2018), regularization-based (Bechavod and Ligett, 2017), and adversarial (Ganin and Lempitsky, 2015).”
Rényi Fair Inference (2019, background): “To do that, various independence proxies such as mutual information Kamishima et al. (2011), false positive/negative rates Bechavod & Ligett (2017), equalized odds Donini et al. (2018), Pearson correlation coefﬁcient Zafar et al. (2015; 2017), Hilbert Schmidt independence criterion (HSIC) P ´…”
Putting Fairness Principles into Practice: Challenges, Metrics, and Improvements (2019, background): “Others have focused on constrained optimization [1, 20], or using a variety of regularization techniques [3, 24, 36].”
A Review on Fairness in Machine Learning (2022, methodology): “As a result, the mechanisms also differ in the techniques used for solving them, depending on linearity and computational complexity of the optimization problem (e.g., see the works of Bechavod and Ligett [16, 17], which use a regularization penalty term that aims to minimize a proxy for equalized o”
Why Is My Classifier Discriminatory? (2018, methodology): “Fairness in prediction has been encouraged by adjusting models through regularization (Bechavod & Ligett, 2017; Kamishima et al., 2011), constraints (Kamiran et al., 2010; Zafar et al., 2017), and representation learning (Zemel et al., 2013).”
Fairness through Causal Awareness: Learning Causal Latent-Variable Models for Biased Data (2018, background): “Much work has been done recently on the problem of fair classiﬁcation [1, 4, 15, 50], yielding an abundant supply of deﬁnitions, models, and algorithms for the purposes of learning classiﬁers whose outputs satisfy distributional constraints.”
Welfare and Distributional Impacts of Fair Classification (2018, background): “Recent efforts have incorporated fairness constraints into the learning process itself by way of regularization (Bechavod & Ligett, 2017) or convex proxies for constrained optimization (Zafar et al., 2017) to generate a new classifier h′(x).”
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]”
Fairness in Recommendation Ranking through Pairwise Comparisons (2019, background): “Rather, numerous approaches have been developed for optimizing fairness metrics during classifier training, such as constraint-based optimization [2, 22], adversarial learning [8, 20, 35, 37, 53, 54], and regularization over model predictions [5, 7, 30, 51].”
Explainable Deep Learning: A Field Guide for the Uninitiated (2020, unknown): “In-process methods (Calders et al., 2009; Kamishima et al., 2011; Zafar et al., 2017a; Woodworth et al., 2017; Zafar et al., 2017b; Bechavod & Ligett, 2017; Kearns et al., 2018; Pérez-Suay et al., 2017; Berk et al., 2017; Olfat & Aswani, 2018; Agarwal et al., 2018; Menon & Williamson, 2018; Donini…”
Fairness in Machine Learning (2020, methodology): “Some methods transform the constrained optimization problem via the method of Lagrange multipliers [3, 15, 34, 37, 38, 59, 97, 135, 183, 185] or add penalties to the objective [5, 14, 22, 44, 46, 56, 58, 62, 73, 74, 75, 79, 82, 87, 88, 89, 93, 94, 95, 98, 104, 115, 119, 123, 133, 134, 135, 138, 139,”