Quantifying explainable discrimination and removing illegal discrimination in automated decision making

By Faisal Kamiran, Indrė Žliobaitė, Toon Calders. 2013

180 citations recorded by the source corpus.

Subfields: Artificial Intelligence, AI Ethics, Fairness & Society

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 AI Ethics, Fairness & Society

Citations in the corpus

Conscientious Classification: A Data Scientist's Guide to Discrimination-Aware Classification (2017, positive): “Kamiran et al. incorporate this notion into a metric called unexplained difference , 14 which loops through the entire data population and attempts to decompose total mean difference into explainable and unexplained portions.”
Learning Optimal Fair Policies (2018, background): “…interest in the issues of fairness in machine learning (Pedreshi, Ruggieri, and Turini 2008; Feldman et al. 2015; Hardt, Price, and Srebro 2016; Kamiran, Zliobaite, and Calders 2013; Corbe-Davies et al. 2017; Jabbari et al. 2017; Kusner et al. 2017; Zhang and Bareinboim 2018; Mitchell and…”
Avoiding Discrimination through Causal Reasoning (2017, background): “A related notion of “explanatory features” in a non-causal setting was introduced in [21].”
Algorithmic Decision Making and the Cost of Fairness (2017, methodology): “al’s local massaging algorithm [22] for achieving conditional statistical parity yields a non-optimal multiplethreshold rule, even if one starts with the optimal single threshold rule.”
Discrimination- and privacy-aware patterns (2015, methodology): “This requirement refers to discrimination that may be partly explained by attributes not in DIb (Zliobaite et al. 2011; Kamiran et al. 2013), e.”
Measuring discrimination in algorithmic decision making (2017, methodology): “Unexplained difference (Kamiran et al. 2013) is measured, as the name suggests, as the overall mean difference minus the differences that can be explained by another legitimate variable.”
The Measure and Mismeasure of Fairness: A Critical Review of Fair Machine Learning (2018, methodology): “…received considerable attention in the machine learning community (Agarwal et al., 2018; Calders and Verwer, 2010; Chouldechova, 2017; Edwards and Storkey, 2015; Feldman et al., 2015; Hardt et al., 2016; Kamiran et al., 2013; Pedreshi et al., 2008; Zafar et al., 2015, 2017; Zemel et al., 2013).”
Using sensitive personal data may be necessary for avoiding discrimination in data-driven decision models (2016, methodology): “This is a generalization over previously used conditional discrimination measure (Kamiran et al. 2013), the latter does not require explicitly designating explanatory variables.”
Better decision support through exploratory discrimination-aware data mining: foundations and empirical evidence (2014, background): “…of discrimination-indexed features with legitimate grounds for differentiation—for example, ‘‘no known savings’’ (Luong 2011, p. 59) as a legitimate ground for rejecting a loan application, or women on average missing specific requirements for a job (Kamiran et al. 2013; Kamishima et al. 2012).”
Impartial Predictive Modeling: Ensuring Fairness in Arbitrary Models (2016, methodology): “Multiple authors have raised doubts that the legal requirement of removing race prior to fitting a model is sufficient to achieve fairness (Kamishima et al., 2012; Kamiran et al., 2013).”