Discrimination Aware Decision Tree Learning

By Faisal Kamiran, Toon Calders, Mykola Pechenizkiy. 2010

394 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Safety Research, AI Ethics, Fairness & Society, Algorithms & Complexity

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 Algorithms & Complexity

Citations in the corpus

Prediction-Based Decisions and Fairness: A Catalogue of Choices, Assumptions, and Definitions (2018, background): “Therefore, solutions such as disparate learning processes (DLPs) , which allow the use of sensitive variables during training but not prediction, are either sub-or equi-optimal [104, 67, 68, 69, 128].”
Conscientious Classification: A Data Scientist's Guide to Discrimination-Aware Classification (2017, neutral): “10,11 Post-processing is the ﬁnal class of methods and can be performed post-training.”
Causal Modeling-Based Discrimination Discovery and Removal: Criteria, Bounds, and Algorithms (2019, background): “sion tree [18], naive Bayes classifier [19], and logistic”
Efficient candidate screening under multiple tests and implications for fairness (2019, background): “In the more recent literature on fairness in machine learning, researchers often focus on binary classification, with employees characterized by a protected characteristic (group membership), and other (non-protected) covariates [15, 11, 12].”
Learning Optimal and Fair Decision Trees for Non-Discriminative Decision-Making (2019, methodology): “Our approach most closely relates to the work in (Kamiran, Calders, and Pechenizkiy 2010), where the authors propose a heuristic algorithm for learning fair decision-trees for classiﬁcation .”
The Price of Fair PCA: One Extra Dimension (2018, neutral): “Other work has developed techniques to obfuscate a 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;…”
FairGAN: Fairness-aware Generative Adversarial Networks (2018, background): “remove discriminatory effect before conducting predictive analytics [4], [7], [11]–[13], the in-process approaches that enforce fairness to classifiers by introducing constraints or regularization terms to the objective functions [14], [15], and the post-process approaches that directly change the p”
Adaptive Sensitive Reweighting to Mitigate Bias in Fairness-aware Classification (2018, background): “Disparate impact elimination reflects the ability of a classifier to achieve statistical parity [25, 27, 28], i.”
Achieving Non-Discrimination in Data Release (2016, methodology): “Proposed methods for discrimination prevention using algorithm tweaking require some tweak of predictive models, including the decision tree [17], the naive bayes classifier [5], the logistic regression [18], and the log-linear model [33].”
A survey on measuring indirect discrimination in machine learning (2015, methodology): “13a], or modify input data [Feldman et al. 2015; Zemel et al. 2013]. Model postprocessing produces a regular model and then modiﬁes it (e.g. by changing the labels of some leaves in a decision tree) [Kamiran et al. 2010; Calders and Verwer 2010]. Model regularisation adds optimization constraints in”
MurTree: Optimal Decision Trees via Dynamic Programming and Search (2020, background): “In this instance, optimal decision trees provide the best tree that balanced accuracy and fairness, albeit with a high computational time when compared to specialised heuristic methods (Kamiran et al. (2010)).”
Learning Fair Rule Lists (2019, methodology): “particular subgroup. Hereafter, we focus on the related work with respect to training fair and interpretable classiﬁers, which is the objective of our method. Kamiran, Calders and Pechenizkiy (2010) (Kamiran et al., 2010) have proposed a learning algorithm incorporating the discrimination and accura”
On the relation between accuracy and fairness in binary classification (2015, background): “It has been observed (Kamiran et al., 2010) that, assuming the labels in data are correct, discrimination removal comes at a cost – it reduces prediction accuracy.”
A comparative study of fairness-enhancing interventions in machine learning (2018, background): “We find surprising variability in fairness measures arising from variations in training-test splits; this appears to not have been previously mentioned in the literature.”
A Confidence-Based Approach for Balancing Fairness and Accuracy (2016, methodology): “While DADT appears to achieve lower label error and comparable bias, we note that the standard deviation of the bias reported in [7] is 0.015 while for SDB (on the Census Income dataset) the standard deviations are at least one order of magnitude smaller.”
Algorithmic Fairness (2020, background): “[77] suggest adjusting a decision tree split criterion to maximize information gain between the split attribute and the class label while minimizing information gain with respect to the sensitive attribute.”