Dynamic fairness - Breaking vicious cycles in automatic decision making

Paaßen B, Bunge A, Hainke C, Sindelar L, Vogelsang M (2019)
In: Proceedings of the 27th European Symposium on Artificial Neural Networks (ESANN 2019). Verleysen M (Ed); Louvain-la-Neuve: i6doc: 477-482.

Konferenzbeitrag | Veröffentlicht | Englisch
 
Autor/in
; ; ; ;
Herausgeber*in
Abstract / Bemerkung
In recent years, machine learning techniques have been increasingly applied in sensitive decision making processes, raising fairness concerns. Past research has shown that machine learning may reproduce and even exacerbate human bias due to biased training data or flawed model assumptions, and thus may lead to discriminatory actions. To counteract such biased models, researchers have proposed multiple mathematical definitions of fairness according to which classifiers can be optimized. However, it has also been shown that the outcomes generated by some fairness notions may be unsatisfactory. In this contribution, we add to this research by considering decision making processes in time. We establish a theoretic model in which even perfectly accurate classifiers which adhere to almost all common fairness definitions lead to stable long-term inequalities due to vicious cycles. Only demographic parity, which enforces equal rates of positive decisions across groups, avoids these effects and establishes a virtuous cycle, which leads to perfectly accurate and fair classification in the long term.
Stichworte
fairness; automatic decision making; discrimination; classification
Erscheinungsjahr
2019
Titel des Konferenzbandes
Proceedings of the 27th European Symposium on Artificial Neural Networks (ESANN 2019)
Seite(n)
477-482
Konferenz
European Symposium on Artificial Neural Networks (ESANN 2019)
Konferenzort
Bruges
Konferenzdatum
2019-04-24 – 2019-04-26
ISBN
978-287-587-065-0
Page URI
https://pub.uni-bielefeld.de/record/2933502

Zitieren

Paaßen B, Bunge A, Hainke C, Sindelar L, Vogelsang M. Dynamic fairness - Breaking vicious cycles in automatic decision making. In: Verleysen M, ed. Proceedings of the 27th European Symposium on Artificial Neural Networks (ESANN 2019). Louvain-la-Neuve: i6doc; 2019: 477-482.
Paaßen, B., Bunge, A., Hainke, C., Sindelar, L., & Vogelsang, M. (2019). Dynamic fairness - Breaking vicious cycles in automatic decision making. In M. Verleysen (Ed.), Proceedings of the 27th European Symposium on Artificial Neural Networks (ESANN 2019) (pp. 477-482). Louvain-la-Neuve: i6doc.
Paaßen, B., Bunge, A., Hainke, C., Sindelar, L., and Vogelsang, M. (2019). “Dynamic fairness - Breaking vicious cycles in automatic decision making” in Proceedings of the 27th European Symposium on Artificial Neural Networks (ESANN 2019), Verleysen, M. ed. (Louvain-la-Neuve: i6doc), 477-482.
Paaßen, B., et al., 2019. Dynamic fairness - Breaking vicious cycles in automatic decision making. In M. Verleysen, ed. Proceedings of the 27th European Symposium on Artificial Neural Networks (ESANN 2019). Louvain-la-Neuve: i6doc, pp. 477-482.
B. Paaßen, et al., “Dynamic fairness - Breaking vicious cycles in automatic decision making”, Proceedings of the 27th European Symposium on Artificial Neural Networks (ESANN 2019), M. Verleysen, ed., Louvain-la-Neuve: i6doc, 2019, pp.477-482.
Paaßen, B., Bunge, A., Hainke, C., Sindelar, L., Vogelsang, M.: Dynamic fairness - Breaking vicious cycles in automatic decision making. In: Verleysen, M. (ed.) Proceedings of the 27th European Symposium on Artificial Neural Networks (ESANN 2019). p. 477-482. i6doc, Louvain-la-Neuve (2019).
Paaßen, Benjamin, Bunge, Astrid, Hainke, Carolin, Sindelar, Leon, and Vogelsang, Matthias. “Dynamic fairness - Breaking vicious cycles in automatic decision making”. Proceedings of the 27th European Symposium on Artificial Neural Networks (ESANN 2019). Ed. Michel Verleysen. Louvain-la-Neuve: i6doc, 2019. 477-482.
Alle Dateien verfügbar unter der/den folgenden Lizenz(en):
Copyright Statement:
This Item is protected by copyright and/or related rights. [...]
Link(s) zu Volltext(en)
Access Level
OA Open Access

Export

Markieren/ Markierung löschen
Markierte Publikationen

Open Data PUB

Quellen

arXiv: 1902.00375

Suchen in

Google Scholar
ISBN Suche