Sky of Unlearning (SoUL): Rewiring Federated Machine Unlearning via Selective Pruning
Md Mahabub Uz Zaman, Xiang Sun, Jingjing Yao
TL;DR
The paper addresses federated unlearning in privacy-preserving IoD-FL under limited resources. It introduces Sky of Unlearning (SoUL), a framework that combines a server-side unlearning update with a selective pruning mechanism to remove the influence of unlearned data while preserving overall accuracy. A key contribution is the SP pruning strategy, which uses L1-norm based importance to prune neurons primarily associated with unlearning, alongside an explicit aggregation rule $\hat{\theta} = \alpha \theta + (1-\alpha) \theta_k^u$. Empirical results show SoUL achieves near retraining accuracy while reducing computation and communication time, demonstrating scalability for large IoD deployments facing data-removal requests.
Abstract
The Internet of Drones (IoD), where drones collaborate in data collection and analysis, has become essential for applications such as surveillance and environmental monitoring. Federated learning (FL) enables drones to train machine learning models in a decentralized manner while preserving data privacy. However, FL in IoD networks is susceptible to attacks like data poisoning and model inversion. Federated unlearning (FU) mitigates these risks by eliminating adversarial data contributions, preventing their influence on the model. This paper proposes sky of unlearning (SoUL), a federated unlearning framework that efficiently removes the influence of unlearned data while maintaining model performance. A selective pruning algorithm is designed to identify and remove neurons influential in unlearning but minimally impact the overall performance of the model. Simulations demonstrate that SoUL outperforms existing unlearning methods, achieves accuracy comparable to full retraining, and reduces computation and communication overhead, making it a scalable and efficient solution for resource-constrained IoD networks.