We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Machine learning fairness notions: Bridging the gap with real-world applications

Abstract: Fairness emerged as an important requirement to guarantee that Machine Learning (ML) predictive systems do not discriminate against specific individuals or entire sub-populations, in particular, minorities. Given the inherent subjectivity of viewing the concept of fairness, several notions of fairness have been introduced in the literature. This paper is a survey that illustrates the subtleties between fairness notions through a large number of examples and scenarios. In addition, unlike other surveys in the literature, it addresses the question of: which notion of fairness is most suited to a given real-world scenario and why? Our attempt to answer this question consists in (1) identifying the set of fairness-related characteristics of the real-world scenario at hand, (2) analyzing the behavior of each fairness notion, and then (3) fitting these two elements to recommend the most suitable fairness notion in every specific setup. The results are summarized in a decision diagram that can be used by practitioners and policymakers to navigate the relatively large catalog of ML.
Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computers and Society (cs.CY); Machine Learning (stat.ML)
Journal reference: Information Processing and Management, 58(5). pp. 107-132 (2021)
DOI: 10.1016/j.ipm.2021.102642
Cite as: arXiv:2006.16745 [cs.LG]
  (or arXiv:2006.16745v5 [cs.LG] for this version)

Submission history

From: Karima Makhlouf [view email]
[v1] Tue, 30 Jun 2020 13:01:06 GMT (532kb,D)
[v2] Mon, 19 Oct 2020 07:50:19 GMT (543kb,D)
[v3] Thu, 24 Mar 2022 14:39:51 GMT (209kb)
[v4] Thu, 31 Mar 2022 07:48:14 GMT (209kb)
[v5] Tue, 7 Jun 2022 17:59:57 GMT (209kb)

Link back to: arXiv, form interface, contact.