Bitte benutzen Sie diese Kennung, um auf die Ressource zu verweisen:
https://doi.org/10.21256/zhaw-21602
Publikationstyp: | Working Paper – Gutachten – Studie |
Titel: | Learning to ignore : fair and task independent representations |
Autor/-in: | Bödi, Linda Helen Grabner, Helmut |
et. al: | No |
DOI: | 10.21256/zhaw-21602 |
Umfang: | 14 |
Erscheinungsdatum: | 2020 |
Verlag / Hrsg. Institution: | ZHAW Zürcher Hochschule für Angewandte Wissenschaften |
Sprache: | Englisch |
Fachgebiet (DDC): | 006: Spezielle Computerverfahren |
Zusammenfassung: | Training fair machine learning models, aiming for their interpretability and solving the problem of domain shift has gained a lot of interest in the last years. There is a vast amount of work addressing these topics, mostly in separation. In this work we show that they can be seen as a common framework of learning invariant representations. The representations should allow to predict the target while at the same time being invariant to sensitive attributes which split the dataset into subgroups. Our approach is based on the simple observation that it is impossible for any learning algorithm to differentiate samples if they have the same feature representation. This is formulated as an additional loss (regularizer) enforcing a common feature representation across subgroups. We apply it to learn fair models and interpret the influence of the sensitive attribute. Furthermore it can be used for domain adaptation, transferring knowledge and learning effectively from very few examples. In all applications it is essential not only to learn to predict the target, but also to learn what to ignore. |
URI: | https://arxiv.org/abs/2101.04047 https://digitalcollection.zhaw.ch/handle/11475/21602 |
Lizenz (gemäss Verlagsvertrag): | CC BY 4.0: Namensnennung 4.0 International |
Departement: | School of Engineering |
Organisationseinheit: | Institut für Datenanalyse und Prozessdesign (IDP) |
Enthalten in den Sammlungen: | Publikationen School of Engineering |
Dateien zu dieser Ressource:
Datei | Beschreibung | Größe | Format | |
---|---|---|---|---|
2020_Boedi-Grabner_Learning-to-ignore.pdf | 3.81 MB | Adobe PDF | ![]() Öffnen/Anzeigen |
Zur Langanzeige
Bödi, L. H., & Grabner, H. (2020). Learning to ignore : fair and task independent representations. ZHAW Zürcher Hochschule für Angewandte Wissenschaften. https://doi.org/10.21256/zhaw-21602
Bödi, L.H. and Grabner, H. (2020) Learning to ignore : fair and task independent representations. ZHAW Zürcher Hochschule für Angewandte Wissenschaften. Available at: https://doi.org/10.21256/zhaw-21602.
L. H. Bödi and H. Grabner, “Learning to ignore : fair and task independent representations,” ZHAW Zürcher Hochschule für Angewandte Wissenschaften, 2020. doi: 10.21256/zhaw-21602.
BÖDI, Linda Helen und Helmut GRABNER, 2020. Learning to ignore : fair and task independent representations [online]. ZHAW Zürcher Hochschule für Angewandte Wissenschaften. Verfügbar unter: https://arxiv.org/abs/2101.04047
Bödi, Linda Helen, and Helmut Grabner. 2020. “Learning to Ignore : Fair and Task Independent Representations.” ZHAW Zürcher Hochschule für Angewandte Wissenschaften. https://doi.org/10.21256/zhaw-21602.
Bödi, Linda Helen, and Helmut Grabner. Learning to Ignore : Fair and Task Independent Representations. ZHAW Zürcher Hochschule für Angewandte Wissenschaften, 2020, https://doi.org/10.21256/zhaw-21602.
Alle Ressourcen in diesem Repository sind urheberrechtlich geschützt, soweit nicht anderweitig angezeigt.