Quantifying explainable discrimination and removing illegal discrimination in automated decision making

Handle URI:
http://hdl.handle.net/10754/594195
Title:
Quantifying explainable discrimination and removing illegal discrimination in automated decision making
Authors:
Kamiran, Faisal; Žliobaite, Indre; Calders, Toon
Abstract:
Recently, the following discrimination-aware classification problem was introduced. Historical data used for supervised learning may contain discrimination, for instance, with respect to gender. The question addressed by discrimination-aware techniques is, given sensitive attribute, how to train discrimination-free classifiers on such historical data that are discriminative, with respect to the given sensitive attribute. Existing techniques that deal with this problem aim at removing all discrimination and do not take into account that part of the discrimination may be explainable by other attributes. For example, in a job application, the education level of a job candidate could be such an explainable attribute. If the data contain many highly educated male candidates and only few highly educated women, a difference in acceptance rates between woman and man does not necessarily reflect gender discrimination, as it could be explained by the different levels of education. Even though selecting on education level would result in more males being accepted, a difference with respect to such a criterion would not be considered to be undesirable, nor illegal. Current state-of-the-art techniques, however, do not take such gender-neutral explanations into account and tend to overreact and actually start reverse discriminating, as we will show in this paper. Therefore, we introduce and analyze the refined notion of conditional non-discrimination in classifier design. We show that some of the differences in decisions across the sensitive groups can be explainable and are hence tolerable. Therefore, we develop methodology for quantifying the explainable discrimination and algorithmic techniques for removing the illegal discrimination when one or more attributes are considered as explanatory. Experimental evaluation on synthetic and real-world classification datasets demonstrates that the new techniques are superior to the old ones in this new context, as they succeed in removing almost exclusively the undesirable discrimination, while leaving the explainable differences unchanged, allowing for differences in decisions as long as they are explainable. © 2012 Springer-Verlag London.
KAUST Department:
Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division
Citation:
Kamiran F, Žliobaitė I, Calders T (2012) Quantifying explainable discrimination and removing illegal discrimination in automated decision making. Knowl Inf Syst 35: 613–644. Available: http://dx.doi.org/10.1007/s10115-012-0584-8.
Publisher:
Springer Nature
Journal:
Knowledge and Information Systems
Issue Date:
18-Nov-2012
DOI:
10.1007/s10115-012-0584-8
Type:
Article
ISSN:
0219-1377; 0219-3116
Appears in Collections:
Articles; Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division

Full metadata record

DC FieldValue Language
dc.contributor.authorKamiran, Faisalen
dc.contributor.authorŽliobaite, Indreen
dc.contributor.authorCalders, Toonen
dc.date.accessioned2016-01-19T13:23:36Zen
dc.date.available2016-01-19T13:23:36Zen
dc.date.issued2012-11-18en
dc.identifier.citationKamiran F, Žliobaitė I, Calders T (2012) Quantifying explainable discrimination and removing illegal discrimination in automated decision making. Knowl Inf Syst 35: 613–644. Available: http://dx.doi.org/10.1007/s10115-012-0584-8.en
dc.identifier.issn0219-1377en
dc.identifier.issn0219-3116en
dc.identifier.doi10.1007/s10115-012-0584-8en
dc.identifier.urihttp://hdl.handle.net/10754/594195en
dc.description.abstractRecently, the following discrimination-aware classification problem was introduced. Historical data used for supervised learning may contain discrimination, for instance, with respect to gender. The question addressed by discrimination-aware techniques is, given sensitive attribute, how to train discrimination-free classifiers on such historical data that are discriminative, with respect to the given sensitive attribute. Existing techniques that deal with this problem aim at removing all discrimination and do not take into account that part of the discrimination may be explainable by other attributes. For example, in a job application, the education level of a job candidate could be such an explainable attribute. If the data contain many highly educated male candidates and only few highly educated women, a difference in acceptance rates between woman and man does not necessarily reflect gender discrimination, as it could be explained by the different levels of education. Even though selecting on education level would result in more males being accepted, a difference with respect to such a criterion would not be considered to be undesirable, nor illegal. Current state-of-the-art techniques, however, do not take such gender-neutral explanations into account and tend to overreact and actually start reverse discriminating, as we will show in this paper. Therefore, we introduce and analyze the refined notion of conditional non-discrimination in classifier design. We show that some of the differences in decisions across the sensitive groups can be explainable and are hence tolerable. Therefore, we develop methodology for quantifying the explainable discrimination and algorithmic techniques for removing the illegal discrimination when one or more attributes are considered as explanatory. Experimental evaluation on synthetic and real-world classification datasets demonstrates that the new techniques are superior to the old ones in this new context, as they succeed in removing almost exclusively the undesirable discrimination, while leaving the explainable differences unchanged, allowing for differences in decisions as long as they are explainable. © 2012 Springer-Verlag London.en
dc.publisherSpringer Natureen
dc.subjectClassificationen
dc.subjectDiscrimination-aware data miningen
dc.subjectIndependenceen
dc.titleQuantifying explainable discrimination and removing illegal discrimination in automated decision makingen
dc.typeArticleen
dc.contributor.departmentComputer, Electrical and Mathematical Sciences and Engineering (CEMSE) Divisionen
dc.identifier.journalKnowledge and Information Systemsen
dc.contributor.institutionBournemouth University, Poole, United Kingdomen
dc.contributor.institutionEindhoven University of Technology, Eindhoven, Netherlandsen
kaust.authorKamiran, Faisalen
All Items in KAUST are protected by copyright, with all rights reserved, unless otherwise indicated.