Split federated learning gains ability to selectively forget data contributions
Split Federated Unlearning
Machine LearningCryptography and Security
Summary
When multiple devices work together to train a model without sharing raw data, it's hard to erase the impact of specific data once learned. The authors identify the problem of ‘unlearning’ in split federated learning, where model parts are shared and coupled. They propose a method called SURE that can selectively remove client or class data effects while keeping the rest of the model useful. Their approach balances forgetting unwanted information and retaining overall model accuracy better than previous approaches adapted for this setup.
What this means in practice
- •For mobile device teams: Remove individual user data influence from split federated learning models without full retraining, improving privacy compliance.
- •For edge computing providers: Maintain and update split models across distributed edge nodes by selectively unlearning outdated or incorrect local contributions.
Authors
Jia Chen, Yecheng Lv, Xinrong Hu, Mingdi Huang, Jun Bai
Abstract
Split federated learning (SFL) combines partitioned model execution with federated aggregation to support collaborative learning across distributed Web and edge environments. However, removing learned contributions in SFL is challenging due to cross-partition coupling and historical influence propagation through dual-side aggregation. To the best of our knowledge, we provide the first systematic formulation of Split Federated Unlearning (SFU), encompassing client-wise and class-wise removal, counterfactual retraining, and partition-specific update constraints. We theoretically characterize historical influence propagation and mode-dependent unlearning feasibility, revealing structural limitations of restricted model updates. As an initial solution, we propose SURE, integrating dual-importance estimation, selective parameter perturbation, and importance-guided recovery. Extensive experiments with existing unlearning methods adapted to SFL reveal substantial forgetting-utility trade-offs across model architectures, data distributions, and update modes. SURE achieves competitive agreement with mode-matched retraining at modest additional cost, while further analyses highlight the impact of split position and update constraints on unlearning behavior. These findings establish SFU as a distinct research problem and provide a foundation for unlearning in partitioned federated systems.