Multi-dimensional Fair Federated Learning (2312.05551v1)
Abstract: Federated learning (FL) has emerged as a promising collaborative and secure paradigm for training a model from decentralized data without compromising privacy. Group fairness and client fairness are two dimensions of fairness that are important for FL. Standard FL can result in disproportionate disadvantages for certain clients, and it still faces the challenge of treating different groups equitably in a population. The problem of privately training fair FL models without compromising the generalization capability of disadvantaged clients remains open. In this paper, we propose a method, called mFairFL, to address this problem and achieve group fairness and client fairness simultaneously. mFairFL leverages differential multipliers to construct an optimization objective for empirical risk minimization with fairness constraints. Before aggregating locally trained models, it first detects conflicts among their gradients, and then iteratively curates the direction and magnitude of gradients to mitigate these conflicts. Theoretical analysis proves mFairFL facilitates the fairness in model development. The experimental evaluations based on three benchmark datasets show significant advantages of mFairFL compared to seven state-of-the-art baselines.
- Mitigating bias in federated learning. arXiv preprint arXiv:2012.02447.
- Fairness-aware agnostic federated learning. In SDM, 181–189.
- UCI machine learning repository. http://archive.ics.uci.edu/ml.
- EU, E.; et al. 2012. Charter of fundamental rights of the European Union. The Review of International Affairs, 63(1147): 109–123.
- Fairfed: Enabling group fairness in federated learning. In AAAI, 7494–7502.
- Certifying and removing disparate impact. In SIGKDD, 259–268.
- Predicting ac optimal power flows: Combining deep learning and lagrangian dual methods. In AAAI, 630–637.
- Lagrangian duality for constrained deep learning. In ECML PKDD, 118–135.
- Enforcing fairness in private federated learning via the modified method of differential multipliers. In NeurIPS 2021 Workshop Privacy in Machine Learning.
- Counterfactual fairness in text classification through robustness. In AAAI/ACM Conference on AI, Ethics, and Society, 219–226.
- Equality of opportunity in supervised learning. In NeurIPS, 3315–3323.
- Federated learning meets multi-objective optimization. TNSE, 9(4): 2039–2051.
- Ditto: Fair and robust federated learning through personalization. In ICML, 6357–6368.
- Fair Resource Allocation in Federated Learning. In ICLR.
- On the Convergence of FedAvg on Non-IID Data. In International Conference on Learning Representations.
- Communication-efficient learning of deep networks from decentralized data. In AISTAT, 1273–1282.
- Fairness in risk assessment instruments: Post-processing to achieve counterfactual equalized odds. In ACM Conference on Fairness, Accountability, and Transparency, 386–400.
- Agnostic federated learning. In ICML, 4615–4625.
- A data-driven approach to predict the success of bank telemarketing. Decision Support Systems, 62: 22–31.
- A Review on Fairness in Machine Learning. ACM Computing Surveys, 55(3): 1–44.
- ProPublica. 2021. Compas recidivism risk score data and analysis. https://www.propublica.org/datastore/dataset/compas-recidivism-risk-score-data-and-analysis.
- Adaptive federated optimization. arXiv preprint arXiv:2003.00295.
- FairBatch: batch selection for model fairness. In ICLR.
- Fairness without harm: Decoupled classifiers with preference guarantees. In ICML, 6373–6382.
- A field guide to federated optimization. arXiv preprint arXiv:2107.06917.
- Federated learning with fair averaging. In IJCAI, 1615–1623.
- Gifair-fl: A framework for group and individual fairness in federated learning. INFORMS Journal on Data Science.
- Fairness constraints: mechanisms for fair classification. In AISTAT, 962–970.
- Improving fairness via federated learning. In ICLR.
- A dynamic reweighting strategy for fair federated learning. In ICASSP, 8772–8776.