Skip to Main content Skip to Navigation
Book sections

A Causal Bayesian Networks Viewpoint on Fairness

Abstract : We offer a graphical interpretation of unfairness in a dataset as the presence of an unfair causal effect of the sensitive attribute in the causal Bayesian network representing the data-generation mechanism. We use this viewpoint to revisit the recent debate surrounding the COMPAS pretrial risk assessment tool and, more generally, to point out that fairness evaluation on a model requires careful considerations on the patterns of unfairness underlying the training data. We show that causal Bayesian networks provide us with a powerful tool to measure unfairness in a dataset and to design fair models in complex unfairness scenarios.
Document type :
Book sections
Complete list of metadata

Cited literature [46 references]  Display  Hide  Download

https://hal.inria.fr/hal-02271668
Contributor : Hal Ifip <>
Submitted on : Tuesday, August 27, 2019 - 10:23:36 AM
Last modification on : Tuesday, August 27, 2019 - 10:34:50 AM
Long-term archiving on: : Friday, January 10, 2020 - 8:41:25 AM

File

 Restricted access
To satisfy the distribution rights of the publisher, the document is embargoed until : 2022-01-01

Please log in to resquest access to the document

Licence


Distributed under a Creative Commons Attribution 4.0 International License

Identifiers

Citation

Silvia Chiappa, William Isaac. A Causal Bayesian Networks Viewpoint on Fairness. Eleni Kosta; Jo Pierson; Daniel Slamanig; Simone Fischer-Hübner; Stephan Krenn. Privacy and Identity Management. Fairness, Accountability, and Transparency in the Age of Big Data : 13th IFIP WG 9.2, 9.6/11.7, 11.6/SIG 9.2.2 International Summer School, Vienna, Austria, August 20-24, 2018, Revised Selected Papers, AICT-547, Springer International Publishing, pp.3-20, 2019, IFIP Advances in Information and Communication Technology, 978-3-030-16743-1. ⟨10.1007/978-3-030-16744-8_1⟩. ⟨hal-02271668⟩

Share

Metrics

Record views

61