k-NN as an implementation of situation testing for discrimination discovery and prevention

By Binh Luong Thanh, Salvatore Ruggieri, Franco Turini. 2011

325 citations recorded by the source corpus.

Subfields: Artificial Intelligence, General Social Sciences, Safety Research, 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 General Social Sciences

In Safety Research

In AI Ethics, Fairness & Society

Citations in the corpus

Deep k-Nearest Neighbors: Towards Confident, Interpretable and Robust Deep Learning (2018, background): “their lack of fairness towards minorities [94], [95], [96], [97],”
On formalizing fairness in prediction with machine learning (2017, background): “Alternatively, in the cases where disproportionality in the respective outcomes can be justified by using non-protected attributes (which don’t merely serve as a proxy for protected attributes), imposing statistical parity leads to incorrect outcomes and may amount to discrimination against qualifie”
Conscientious Classification: A Data Scientist's Guide to Discrimination-Aware Classification (2017, neutral): “15 Calders et al. similarly use ‘‘propensity score stratiﬁcation’’ to group observations into cohorts, where the propensity score is based on the likelihood to be a member of a protected class.”
Data preprocessing techniques for classification without discrimination (2011, background): “There is, however, one big difference: whereas in massaging only the minimal number of objects is changed to remove all discrimination from the dataset, the authors of [19] propose to continue relabeling until all labels are consistent.”
Causal Modeling-Based Discrimination Discovery and Removal: Criteria, Bounds, and Algorithms (2019, background): “The authors in [16] exploited the idea of situation testing to discover individual discrimination.”
Learning Optimal and Fair Decision Trees for Non-Discriminative Decision-Making (2019, methodology): “First, pre-processing approaches, which rely on modifying the data to eliminate or neutralize any preexisting bias and subsequently apply standard ML techniques (Kamiran and Calders 2012; 2009; Luong, Ruggieri, and Turini 2011).”
Stable and Fair Classification (2019, methodology): “Finally, a diﬀerent line of research pre-processes the training data with the goal of removing the bias for learning, e.g., [39, 51, 40, 76, 26, 44].”
The Price of Fair PCA: One Extra Dimension (2018, background): “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; ”
A Causal Framework for Discovering and Removing Direct and Indirect Discrimination (2016, background): “(Luong, Ruggieri, and Turini 2011; Zhang, Wu, and Wu 2016b) dealt with the individual discrimination by ﬁnding a group of similar individuals.”
Fairness Beyond Disparate Treatment & Disparate Impact: Learning Classification without Disparate Mistreatment (2016, methodology): “Finally, we experiment with synthetic as well as real world datasets and show that our methodology can be eﬀectively used to avoid disparate mistreatment.”
A survey on measuring indirect discrimination in machine learning (2015, background): “for each individual in the dataset to identify whether s/he is discriminated, and then analyze how many individuals in the dataset are affected. Currently 4.4.1. Situation testing. Situation testing [Luong et al. 2011] measures which fraction of individuals in the protected group are considered diP ”
Fairness in Learning: Classic and Contextual Bandits (2016, background): “…otherwise known as equality of outcomes , statistical parity – or algorithms that avoid disparate impact (see e.g. Calders and Verwer [2010], Luong et al. [2011], Kamishima et al. [2011], Feldman et al. [2015], Fish et al. [2016] and Adler et al. [2016] for a study of auditing existing…”
From Parity to Preference-based Notions of Fairness in Classification (2017, background): “A number of learning mechanisms have been proposed to achieve parity in treatment [24], An open-source code implementation of our scheme is available at: http://fate-computing.”
Fairness in Reinforcement Learning (2016, background): “Early work in data mining [8, 11, 12, 17, 21, 29] considered the question from a primarily empirical standpoint, often using statistical parity as a fairness goal.”
Exposing the probabilistic causal structure of discrimination (2015, methodology): “For an individual record r in original data table D, individual discrimination quantifies as follows [25]:”