Decision Theory for Discrimination-Aware Classification

By Faisal Kamiran, Asim Karim, Xiangliang Zhang. 2012

381 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Safety Research, AI Ethics, Fairness & Society, Game & Decision Theory

How researchers see this paper

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

In Artificial Intelligence

In Safety Research

In AI Ethics, Fairness & Society

In Game & Decision Theory

Citations in the corpus

The Price of Fair PCA: One Extra Dimension (2018, background): “…sensitive attribute directly [Pedreshi et al., 2008; Kamiran et al., 2010; Calders and Verwer, 2010; Kamiran and Calders, 2011; Luong et al., 2011; Kamiran et al., 2012; Kamishima et al., 2012; Hajian and Domingo-Ferrer, 2013; Feldman et al., 2015; Zafar et al., 2015; Fish et al., 2016; Adler et…”
AI Fairness 360: An Extensible Toolkit for Detecting, Understanding, and Mitigating Unwanted Algorithmic Bias (2018, background): “Re-ject option classiﬁcation (Kamiran et al., 2012) gives favorable outcomes to unprivileged groups and unfavorable outcomes to privileged groups in a conﬁdence band around the decision boundary with the highest uncertainty.”
Fairness in Criminal Justice Risk Assessments: The State of the Art (2017, background): “Instructive deﬁnitions can be found in computer science (Pedreschi, 2008; Kamishima et al., 2011; Dwork et al., 2012; Kamiran et al., 2012; Chouldechova, 2016; Friedler et al., 2016; Hardt et al., 2016; Joseph et al., 2016; Kleinberg et al., 2016; Calmon et al., 2017; Corbitt-Davies et al., 2017),…”
Learning Fair Classifiers (2015, background): “The second strategy consists of modifying existing classifiers to limit discrimination (Cadlers & Verwer, 2010; Calders et al., 2013; Kamiran et al., 2012; Kamishima et al., 2011; 2013; Pedreschi et al., 2009).”
Fairness in Reinforcement Learning (2016, background): “[11] Faisal Kamiran, Asim Karim, and Xiangliang Zhang.”
Bias Mitigation Post-processing for Individual and Group Fairness (2018, methodology): “Kamiran et al. [14] elect to change values of ˆ y i to ˇ y i in a reject option band to conform to one type of human sensibility, but the same effect on disparate impact can be achieved using the same numbers of samples from elsewhere in X .”
A multidisciplinary survey on discrimination analysis (2013, background): “Kamiran et al. (2012) propose correcting predictions of probabilistic classifiers that are close to the decision boundary, given that (statistical) discrimination may occur when there is no clear feature supporting a positive or a negative decision.”
Controlling Attribute Effect in Linear Regression (2013, background): “More recently, another type of bias arising from dependence on a socially sensitive attribute was identified in the data mining community [1]–[3].”
Building Better Open-Source Tools to Support Fairness in Automated Scoring (2017, background): “, 2013; Calders and Verwer, 2010), and the output decisions can be changed after-the-fact (Kamiran et al., 2012).”
Discrimination- and privacy-aware patterns (2015, methodology): “Lastly, Kamiran et al. (2010) propose correcting predictions of probabilistic classifiers that are close to the decision boundary, given that (statistical) discrimination may occur when there is no clear feature supporting a positive or a negative decision.”
Anti-discrimination Analysis Using Privacy Attack Strategies (2014, methodology): “We exploit the observation proposed in [10] that discrimination affects the tuples close to the decision boundary of a classifier.”
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,”
Fairness Constraints: Mechanisms for Fair Classification (2015, background): “The second strategy consists of modifying existing learning algorithms to limit discrimination (Kamiran et al., 2012; Calders et al., 2013; Cadlers & Verwer, 2010; Kamishima et al., 2011; 2013; Pedreschi et al., 2009).”
FairBatch: Batch Selection for Model Fairness (2020, methodology): “…2018; Zhang et al., 2018; Cotter et al., 2019; Roh et al., 2020), which tailor the model training for fairness, and (3) postprocessing techniques (Kamiran et al., 2012; Hardt et al., 2016; Pleiss et al., 2017; Chzhen et al., 2019), which perturb only the model output without touching upon the…”
Your fairness may vary: Pretrained language model fairness in toxic text classification (2021, methodology): “…other post-processing methods for bias mitigation include assigning favorable labels to unprivileged groups in regions of high classiﬁer uncertainty (Kamiran et al., 2012), minimizing error disparity while maintaining clas-siﬁer calibration (Pleiss et al., 2017), a relaxed nearly-optimal procedure…”