Digitala Vetenskapliga Arkivet

Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • 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
FairScore Transformer for fairness in supervised learning
Stockholm University, Faculty of Social Sciences, Department of Computer and Systems Sciences.
2024 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

Machine learning models are becoming more and more integrated into various decision-making processes, which calls for a rigorous analysis of their ethical implications, especially with regard to fairness. The use of the FairScore Transformer (FST), a post-processing method intended to reduce biases in supervised learning models, is examined in this thesis. First, a logistic regression model was trained with the Adult dataset in order to create a baseline for variables related to performance and fairness. The model's output probabilities were then modified using the FST in an effort to lessen demographic differences without materially affecting accuracy. The findings showed a minor loss in accuracy but a large reduction in fairness inequities, with demographic parity disparity falling from 15% to 5%. This work adds to the larger conversation on the development of ethical AI by providing empirical support for the FST's efficaciousness in improving fairness in machine learning models. According to the results, supervised learning models' fairness may be increased by post-processing methods like FST without significantly lowering the models' prediction accuracy.

Place, publisher, year, edition, pages
2024.
Keywords [en]
FairScore Transformer; fairness; supervised learning; logistic regression; post-processing; Adult dataset
National Category
Computer Sciences
Identifiers
URN: urn:nbn:se:su:diva-242852OAI: oai:DiVA.org:su-242852DiVA, id: diva2:1955835
Available from: 2025-05-01 Created: 2025-05-01

Open Access in DiVA

fulltext(1285 kB)13 downloads
File information
File name FULLTEXT01.pdfFile size 1285 kBChecksum SHA-512
b7a565af33bd338417bdbe17b173fa5268a06e6a056dfe5a2849dc24cfe8106c70e94f078ba85e7ddca12459f163a58d2ab9cf85c4eac465a2af4b609f8ba68c
Type fulltextMimetype application/pdf

Search in DiVA

By author/editor
Ntales, Athanasios
By organisation
Department of Computer and Systems Sciences
Computer Sciences

Search outside of DiVA

GoogleGoogle Scholar
Total: 13 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

urn-nbn

Altmetric score

urn-nbn
Total: 83 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • 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