Umeå University's logo

umu.sePublications
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
A clearer view on fairness: visual and formal representations for comparative analysis
Umeå University, Faculty of Science and Technology, Department of Computing Science.ORCID iD: 0000-0002-7383-0529
Umeå University, Faculty of Science and Technology, Department of Computing Science.ORCID iD: 0000-0002-6458-2252
Umeå University, Faculty of Science and Technology, Department of Computing Science.ORCID iD: 0000-0002-8423-8029
Umeå University, Faculty of Science and Technology, Department of Computing Science.ORCID iD: 0000-0001-7409-5813
2024 (English)In: 14th Scandinavian Conference on Artificial Intelligence, SCAI 2024: June 10-11, 2024, Jönköping, Sweden / [ed] Florian Westphal; Einav Peretz-Andersson; Maria Riveiro; Kerstin Bach; Fredrik Heintz, Jönköping University , 2024, p. 112-120Conference paper, Published paper (Refereed)
Abstract [en]

The opaque nature of machine learning systems has raised concerns about whether these systems can guarantee fairness. Furthermore, ensuring fair decision making requires the consideration of multiple perspectives on fairness. 

At the moment, there is no agreement on the definitions of fairness, achieving shared interpretations is difficult, and there is no unified formal language to describe them. Current definitions are implicit in the operationalization of systems, making their comparison difficult.

In this paper, we propose a framework for specifying formal representations of fairness that allows instantiating, visualizing, and comparing different interpretations of fairness. Our framework provides a meta-model for comparative analysis. We present several examples that consider different definitions of fairness, as well as an open-source implementation that uses the object-oriented functional language Soda.

Place, publisher, year, edition, pages
Jönköping University , 2024. p. 112-120
Series
Linköping Electronic Conference Proceedings, ISSN 1650-3686, E-ISSN 1650-3740 ; 208
Keywords [en]
Responsible artificial intelligence, Ethics in artificial intelligence, Formal representation of fairness
National Category
Software Engineering Computer Sciences
Research subject
Computer Science; Ethics
Identifiers
URN: urn:nbn:se:umu:diva-232255DOI: 10.3384/ecp208013ISBN: 9789180757096 (electronic)OAI: oai:DiVA.org:umu-232255DiVA, id: diva2:1916584
Conference
14th Scandinavian Conference on Artificial Intelligence, Jönköping, Sweden, June 10-11, 2024
Funder
Wallenberg AI, Autonomous Systems and Software Program (WASP)Available from: 2024-11-27 Created: 2024-11-27 Last updated: 2024-12-02Bibliographically approved
In thesis
1. Making fairness actionable
Open this publication in new window or tab >>Making fairness actionable
2024 (English)Licentiate thesis, comprehensive summary (Other academic)
Alternative title[sv]
Att göra rättvisa handlingsbart
Abstract [en]

The opaque nature of machine learning systems has raised concerns about whether these systems can guarantee fairness. In addition, ensuring fair decision-making requires that multiple perspectives of fairness be considered.

Currently, there is no agreement on the definitions, the facilitation of shared interpretation is difficult, and there is a lack of a unified formal language to describe them. Current definitions are implicit in the operationalization of systems, making them difficult to compare.

In this thesis, we discuss how to make fairness actionable, providing concrete tools for that. We provide not only conceptual elements to model and abstract problems of fairness, but also a technical framework and a description language.

Abstract [sv]

Den opaka naturen hos maskininlärningssystem väcker oro kring om dessa system kan garantera rättvisa. Dessutom kräver rättvis beslutsfattande att flera perspektiv på rättvisa beaktas.

För närvarande finns det ingen enighet kring definitionerna, vilket försvårar delad tolkning, och det saknas ett enhetligt formellt språk för att beskriva dem. Nuvarande definitioner är inbyggda i hur systemen används, vilket gör dem svåra att jämföra.

I denna avhandling diskuterar vi hur rättvisa kan göras praktiskt tillämpbar och tillhandahåller konkreta verktyg för detta. Vi erbjuder både konceptuella element för att modellera och abstrahera rättviseproblem samt en teknisk ram och ett beskrivningsspråk.

Place, publisher, year, edition, pages
Umeå: Umeå University, 2024. p. 24
Series
Report / UMINF, ISSN 0348-0542 ; 24.12
Keywords
Algorithmic fairness, Ethics in artificial intelligence, Formal representation of fairness, Formal verification, Functional languages, Human-centered programming languages, Responsible artificial intelligence
National Category
Computer Sciences Software Engineering Ethics
Research subject
Computer Science; Computer Systems; Ethics
Identifiers
urn:nbn:se:umu:diva-232384 (URN)9789180705356 (ISBN)9789180705349 (ISBN)
Presentation
2024-12-13, MIT.A.121, MIT-huset, Campustorget 5, Umeå, Sweden, 13:00 (English)
Opponent
Supervisors
Funder
Wallenberg AI, Autonomous Systems and Software Program (WASP)
Available from: 2024-12-03 Created: 2024-11-28 Last updated: 2024-12-03Bibliographically approved

Open Access in DiVA

fulltext(290 kB)93 downloads
File information
File name FULLTEXT01.pdfFile size 290 kBChecksum SHA-512
fcf7e829ebabf22ce523733b670a2935e0258e9309a196728f411aeae87f6c907424bef3beba368cc52de3ca70724b6b6917c76ba09642c70f1529a4e37a777d
Type fulltextMimetype application/pdf

Other links

Publisher's full text

Authority records

Mendez, Julian AlfredoKampik, TimotheusAler Tubella, AndreaDignum, Virginia

Search in DiVA

By author/editor
Mendez, Julian AlfredoKampik, TimotheusAler Tubella, AndreaDignum, Virginia
By organisation
Department of Computing Science
Software EngineeringComputer Sciences

Search outside of DiVA

GoogleGoogle Scholar
Total: 94 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

doi
isbn
urn-nbn

Altmetric score

doi
isbn
urn-nbn
Total: 506 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf