Fair-LS: A Group-Specific Clamping Factor for Fair Label Spreading in Graph-Based Semi-Supervised Learning

• Willian Dihanster Gomes de Oliveira UNIFESP
• Lilian Berton UNIFESP

DOI: https://doi.org/10.5753/eniac.2025.14368

Resumo

Graph-based semi-supervised learning methods, such as Label Spreading, can propagate bias when the graph structure reflects group disparities. We propose Fair-LS, an extension of Label Spreading that introduces group-specific clamping factors to control label diffusion differently for privileged and unprivileged groups. Experiments on benchmark datasets, Adult Census Income, Compas Recidivism, and German Credit, show that Fair-LS could improve fairness metrics (Disparate Impact and Average Absolute Odds Difference) with minimal reduction in AUC-ROC. The results suggest that adapting propagation dynamics by group is an effective strategy to reduce bias in semi-supervised learning.

Referências

Angwin, J., Larson, J., Mattu, S., and Kirchner, L. (2016). How we analyzed the compas recidivism algorithm.

Barocas, S., Hardt, M., and Narayanan, A. (2023). Fairness and machine learning: Limitations and opportunities. MIT press.

Becker, B. and Kohavi, R. (1996). Adult. UCI Machine Learning Repository. DOI: 10.24432/C5XW20.

Bellamy, R. K. E., Dey, K., Hind, M., Hoffman, S. C., Houde, S., Kannan, K., Lohia, P., Martino, J., Mehta, S., Mojsilovic, A., Nagar, S., Ramamurthy, K. N., Richards, J., Saha, D., Sattigeri, P., Singh, M., Varshney, K. R., and Zhang, Y. (2018). AI Fairness 360: An extensible toolkit for detecting, understanding, and mitigating unwanted algorithmic bias.

Chakraborty, J., Tu, H., Majumder, S., and Menzies, T. (2021). Can we achieve fairness using semi-supervised learning? arXiv preprint arXiv:2111.02038.

Chong, Y., Ding, Y., Yan, Q., and Pan, S. (2020). Graph-based semi-supervised learning: A review. Neurocomputing, 408:216–230.

de Oliveira, W. D. G. and Berton, L. (2023). A systematic review for class-imbalance in semi-supervised learning. Artificial Intelligence Review, 56(Suppl 2):2349–2382.

de Oliveira, W. D. G. and Berton, L. (2024). A comparative study of pre-processing algorithms for fair classification in few labeled data context. AI and Ethics, pages 1–21.

Hofmann, H. (1994). Statlog (German Credit Data). UCI Machine Learning Repository. DOI: 10.24432/C5NC77.

Miyazaki, T. and Sumikawa, Y. (2018). Label propagation using amendable clamping. In IUI Workshops.

Nadkarni, P. (2016). Chapter 10 - core technologies: Data mining and “big data”. In Nadkarni, P., editor, Clinical Research Computing, pages 187–204. Academic Press.

Rabonato, R. T. and Berton, L. (2025). A systematic review of fairness in machine learning. AI and Ethics, 5(3):1943–1954.

Shi, Y., Ke, G., Soukhavong, D., Lamb, J., Meng, Q., Finley, T., Wang, T., Chen, W., Ma, W., Ye, Q., Liu, T.-Y., Titov, N., and Cortes, D. (2025). lightgbm: Light Gradient Boosting Machine. R package version 4.6.0.99.

Song, Z., Yang, X., Xu, Z., and King, I. (2022). Graph-based semi-supervised learning: A comprehensive review. IEEE Transactions on Neural Networks and Learning Systems, 34(11):8174–8194.

Sulaiman, M. and Roy, K. (2025). Gflc: Graph-based fairness-aware label correction for fair classification. arXiv preprint arXiv:2506.15620.

Tsang, A., Wilder, B., Rice, E., Tambe, M., and Zick, Y. (2019). Group-fairness in influence maximization. arXiv preprint arXiv:1903.00967.

Yang, S., Cerrato, M., Ienco, D., Pensa, R. G., and Esposito, R. (2023). Fair swirl: fair semi-supervised classification with representation learning. Machine Learning, 112(9):3051–3076.

Zhang, T., Zhu, T., Han, M., Chen, F., Li, J., Zhou, W., and Yu, P. S. (2023). Fairness in graph-based semi-supervised learning. Knowledge and Information Systems, 65(2):543–570.

Zhang, T., Zhu, T., Han, M., Li, J., Zhou, W., and Yu, P. S. (2020). Fairness constraints in semi-supervised learning. arXiv preprint arXiv:2009.06190.

Zhou, D., Bousquet, O., Lal, T., Weston, J., and Schölkopf, B. (2003). Learning with local and global consistency. Advances in neural information processing systems, 16.

Zhu, X., Ghahramani, Z., and Lafferty, J. (2002). Learning from labeled and unlabeled data with label propagation. In CMU CALD Tech Report.