Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Yfantidou, Sofia, Spathis, Dimitris, Constantinides, Marios, Vakali, Athena, Quercia, Daniele, Kawsar, Fahim
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2406.02361
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866910472036417536
author Yfantidou, Sofia
Spathis, Dimitris
Constantinides, Marios
Vakali, Athena
Quercia, Daniele
Kawsar, Fahim
author_facet Yfantidou, Sofia
Spathis, Dimitris
Constantinides, Marios
Vakali, Athena
Quercia, Daniele
Kawsar, Fahim
contents Self-supervised learning (SSL) has become the de facto training paradigm of large models, where pre-training is followed by supervised fine-tuning using domain-specific data and labels. Despite demonstrating comparable performance with supervised methods, comprehensive efforts to assess SSL's impact on machine learning fairness (i.e., performing equally on different demographic breakdowns) are lacking. Hypothesizing that SSL models would learn more generic, hence less biased representations, this study explores the impact of pre-training and fine-tuning strategies on fairness. We introduce a fairness assessment framework for SSL, comprising five stages: defining dataset requirements, pre-training, fine-tuning with gradual unfreezing, assessing representation similarity conditioned on demographics, and establishing domain-specific evaluation processes. We evaluate our method's generalizability on three real-world human-centric datasets (i.e., MIMIC, MESA, and GLOBEM) by systematically comparing hundreds of SSL and fine-tuned models on various dimensions spanning from the intermediate representations to appropriate evaluation metrics. Our findings demonstrate that SSL can significantly improve model fairness, while maintaining performance on par with supervised methods-exhibiting up to a 30% increase in fairness with minimal loss in performance through self-supervision. We posit that such differences can be attributed to representation dissimilarities found between the best- and the worst-performing demographics across models-up to x13 greater for protected attributes with larger performance discrepancies between segments.
format Preprint
id arxiv_https___arxiv_org_abs_2406_02361
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Using Self-supervised Learning Can Improve Model Fairness
Yfantidou, Sofia
Spathis, Dimitris
Constantinides, Marios
Vakali, Athena
Quercia, Daniele
Kawsar, Fahim
Machine Learning
Self-supervised learning (SSL) has become the de facto training paradigm of large models, where pre-training is followed by supervised fine-tuning using domain-specific data and labels. Despite demonstrating comparable performance with supervised methods, comprehensive efforts to assess SSL's impact on machine learning fairness (i.e., performing equally on different demographic breakdowns) are lacking. Hypothesizing that SSL models would learn more generic, hence less biased representations, this study explores the impact of pre-training and fine-tuning strategies on fairness. We introduce a fairness assessment framework for SSL, comprising five stages: defining dataset requirements, pre-training, fine-tuning with gradual unfreezing, assessing representation similarity conditioned on demographics, and establishing domain-specific evaluation processes. We evaluate our method's generalizability on three real-world human-centric datasets (i.e., MIMIC, MESA, and GLOBEM) by systematically comparing hundreds of SSL and fine-tuned models on various dimensions spanning from the intermediate representations to appropriate evaluation metrics. Our findings demonstrate that SSL can significantly improve model fairness, while maintaining performance on par with supervised methods-exhibiting up to a 30% increase in fairness with minimal loss in performance through self-supervision. We posit that such differences can be attributed to representation dissimilarities found between the best- and the worst-performing demographics across models-up to x13 greater for protected attributes with larger performance discrepancies between segments.
title Using Self-supervised Learning Can Improve Model Fairness
topic Machine Learning
url https://arxiv.org/abs/2406.02361