Resampled Datasets Are Not Enough: Mitigating Societal Bias Beyond Single Attributes

Yusuke Hirota, Jerone Andrews, Dora Zhao, Orestis Papakyriakopoulos, Apostolos Modas, Yuta Nakashima, Alice Xiang


Abstract
We tackle societal bias in image-text datasets by removing spurious correlations between protected groups and image attributes. Traditional methods only target labeled attributes, ignoring biases from unlabeled ones. Using text-guided inpainting models, our approach ensures protected group independence from all attributes and mitigates inpainting biases through data filtering. Evaluations on multi-label image classification and image captioning tasks show our method effectively reduces bias without compromising performance across various models. Specifically, we achieve an average societal bias reduction of 46.1% in leakage-based bias metrics for multi-label classification and 74.8% for image captioning.
Anthology ID:
2024.emnlp-main.471
Volume:
Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing
Month:
November
Year:
2024
Address:
Miami, Florida, USA
Editors:
Yaser Al-Onaizan, Mohit Bansal, Yun-Nung Chen
Venue:
EMNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
8249–8267
Language:
URL:
https://aclanthology.org/2024.emnlp-main.471
DOI:
Bibkey:
Cite (ACL):
Yusuke Hirota, Jerone Andrews, Dora Zhao, Orestis Papakyriakopoulos, Apostolos Modas, Yuta Nakashima, and Alice Xiang. 2024. Resampled Datasets Are Not Enough: Mitigating Societal Bias Beyond Single Attributes. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 8249–8267, Miami, Florida, USA. Association for Computational Linguistics.
Cite (Informal):
Resampled Datasets Are Not Enough: Mitigating Societal Bias Beyond Single Attributes (Hirota et al., EMNLP 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.emnlp-main.471.pdf