Show simple item record

dc.contributor.author
Plečko, Drago
dc.contributor.author
Meinshausen, Nicolai
dc.date.accessioned
2021-02-24T07:08:31Z
dc.date.available
2021-02-24T07:04:34Z
dc.date.available
2021-02-24T07:08:31Z
dc.date.issued
2020-11
dc.identifier.issn
1532-4435
dc.identifier.issn
1533-7928
dc.identifier.uri
http://hdl.handle.net/20.500.11850/471220
dc.identifier.doi
10.3929/ethz-b-000459191
dc.description.abstract
Fairness of classification and regression has received much attention recently and various, partially non-compatible, criteria have been proposed. The fairness criteria can be enforced for a given classifier or, alternatively, the data can be adapted to ensure that every classifier trained on the data will adhere to desired fairness criteria. We present a practical data adaption method based on quantile preservation in causal structural equation models. The data adaptation is based on a presumed counterfactual model for the data. While the counterfactual model itself cannot be verified experimentally, we show that certain population notions of fairness are still guaranteed even if the counterfactual model is misspecified. The nature of the fulfilled observational non-causal fairness notion (such as demographic parity, separation or sufficiency) depends on the structure of the underlying causal model and the choice of resolving variables. We describe an implementation of the proposed data adaptation procedure based on Random Forests and demonstrate its practical use on simulated and real-world data.
en_US
dc.format
application/pdf
en_US
dc.language.iso
en
en_US
dc.publisher
Journal of Machine Learning Research
en_US
dc.rights.uri
http://creativecommons.org/licenses/by/4.0/
dc.subject
Supervised learning
en_US
dc.subject
Fairness in machine learning
en_US
dc.subject
Causality
en_US
dc.subject
Graphical models
en_US
dc.subject
Counterfactual fairness
en_US
dc.title
Fair data adaptation with quantile preservation
en_US
dc.type
Journal Article
dc.rights.license
Creative Commons Attribution 4.0 International
ethz.journal.title
Journal of Machine Learning Research
ethz.journal.volume
21
en_US
ethz.journal.abbreviated
J. mach. learn. res.
ethz.pages.start
242
en_US
ethz.size
44 p.
en_US
ethz.version.deposit
publishedVersion
en_US
ethz.identifier.wos
ethz.identifier.scopus
ethz.publication.place
s.l.
en_US
ethz.publication.status
published
en_US
ethz.identifier.url
https://www.jmlr.org/papers/v21/19-966.html
ethz.date.deposited
2021-01-05T03:57:16Z
ethz.source
SCOPUS
ethz.source
WOS
ethz.eth
yes
en_US
ethz.availability
Open access
en_US
ethz.rosetta.installDate
2021-02-24T07:04:51Z
ethz.rosetta.lastUpdated
2021-02-24T07:04:51Z
ethz.rosetta.exportRequired
true
ethz.rosetta.versionExported
true
dc.identifier.olduri
http://hdl.handle.net/20.500.11850/459191
dc.identifier.olduri
http://hdl.handle.net/20.500.11850/467607
ethz.COinS
ctx_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.atitle=Fair%20data%20adaptation%20with%20quantile%20preservation&rft.jtitle=Journal%20of%20Machine%20Learning%20Research&rft.date=2020-11&rft.volume=21&rft.spage=242&rft.issn=1532-4435&1533-7928&rft.au=Ple%C4%8Dko,%20Drago&Meinshausen,%20Nicolai&rft.genre=article&
 Search print copy at ETH Library

Files in this item

Thumbnail

Publication type

Show simple item record