Federated Unlearning: a Perspective of Stability and Fairness
This work addresses the challenge of managing trade-offs in federated unlearning for systems with heterogeneous data, which is incremental to existing research.
The paper tackles the problem of federated unlearning with data heterogeneity by introducing metrics and an optimization framework, and it proposes mechanisms that empirically balance trade-offs in verification, stability, and fairness.
This paper explores the multifaceted consequences of federated unlearning (FU) with data heterogeneity. We introduce key metrics for FU assessment, concentrating on verification, global stability, and local fairness, and investigate the inherent trade-offs. Furthermore, we formulate the unlearning process with data heterogeneity through an optimization framework. Our key contribution lies in a comprehensive theoretical analysis of the trade-offs in FU and provides insights into data heterogeneity's impacts on FU. Leveraging these insights, we propose FU mechanisms to manage the trade-offs, guiding further development for FU mechanisms. We empirically validate that our FU mechanisms effectively balance trade-offs, confirming insights derived from our theoretical analysis.