Private and Robust Contribution Evaluation in Federated Learning
This addresses a critical issue for organizations in cross-silo federated learning by enabling fair, private, and robust contribution evaluation, though it builds incrementally on existing marginal-contribution methods.
The paper tackled the problem of evaluating client contributions in federated learning with secure aggregation, which protects privacy but complicates fairness and robustness. It introduced two new contribution scores that outperform baselines, better approximate Shapley rankings, and improve model performance and misbehavior detection on medical and CIFAR10 datasets.
Cross-silo federated learning allows multiple organizations to collaboratively train machine learning models without sharing raw data, but client updates can still leak sensitive information through inference attacks. Secure aggregation protects privacy by hiding individual updates, yet it complicates contribution evaluation, which is critical for fair rewards and detecting low-quality or malicious participants. Existing marginal-contribution methods, such as the Shapley value, are incompatible with secure aggregation, and practical alternatives, such as Leave-One-Out, are crude and rely on self-evaluation. We introduce two marginal-difference contribution scores compatible with secure aggregation. Fair-Private satisfies standard fairness axioms, while Everybody-Else eliminates self-evaluation and provides resistance to manipulation, addressing a largely overlooked vulnerability. We provide theoretical guarantees for fairness, privacy, robustness, and computational efficiency, and evaluate our methods on multiple medical image datasets and CIFAR10 in cross-silo settings. Our scores consistently outperform existing baselines, better approximate Shapley-induced client rankings, and improve downstream model performance as well as misbehavior detection. These results demonstrate that fairness, privacy, robustness, and practical utility can be achieved jointly in federated contribution evaluation, offering a principled solution for real-world cross-silo deployments.