pFedUL: Layer-Aware Federated Unlearning for Personalized Federated Learning

2026-06-15

AI Summary

This work addresses the challenge in personalized federated learning (pFL) where existing unlearning methods struggle to effectively remove the influence of a target client on shared layers without degrading the personalized performance of remaining clients. We formally define the problem of federated unlearning under pFL and propose a layer-aware selective unlearning framework. Our approach quantifies each client’s contribution to both shared and personalized layers via gradient-based attribution, enabling adaptive, differential unlearning strategies. A lightweight recalibration protocol is further introduced to restore personalization capabilities. We also introduce two new evaluation metrics, PPS and CFI. Experiments on CIFAR-10/100 and FEMNIST demonstrate that our method achieves unlearning efficacy comparable to full retraining while preserving 97.3% of the average personalized accuracy for retained clients, significantly outperforming six state-of-the-art baselines.

Abstract

Federated unlearning (FU) enables the removal of specific data contributions from federated learning (FL) models to comply with regulations such as the General Data Protection Regulation (GDPR). However, most existing FU methods are designed for the FedAvg paradigm, where all clients share a single global model. In practice, personalized federated learning (pFL) methods such as FedPer, FedRep, Ditto, and FedBN have become widely adopted due to their superior handling of non-IID data. These methods decompose the model into shared global layers and client-specific personalized layers, fundamentally altering the semantics of unlearning, yet this setting has received little attention. We formalize FU under the pFL paradigm, identifying a tension between unlearning completeness on shared layers and personalization preservation for remaining clients. We then propose pFedUL, a layer-aware selective unlearning framework comprising three components: (1) gradient-based layer-wise contribution attribution that separately quantifies the target client's influence on shared and personalized parameters, (2) adaptive selective unlearning that applies differentiated forgetting strategies across layer types, and (3) a lightweight recalibration protocol enabling remaining clients to restore personalization with minimal overhead. We further introduce two new metrics, Personalization Preservation Score (PPS) and Cross-client Fairness Index (CFI), to evaluate pFL-specific unlearning quality. Experiments on CIFAR-10, CIFAR-100, and FEMNIST under varying non-IID settings indicate that pFedUL achieves unlearning effectiveness comparable to full retraining while maintaining an average of 97.3\% personalized accuracy for remaining clients. Compared with six state-of-the-art FU methods adapted to the pFL setting, pFedUL consistently achieves superior personalization preservation.

Problem

• Federated Unlearning
• Personalized Federated Learning
• Non-IID Data
• Model Personalization
• GDPR Compliance

Innovation

• Federated Unlearning
• Personalized Federated Learning
• Layer-aware Unlearning
• Contribution Attribution
• Selective Forgetting