analiza aiqbench
This paper introduces f-FUM, a federated unlearning framework that leverages teacher-student optimization and $f$-divergence to efficiently remove specific data contributions without retraining or moving raw data. By optimizing divergence choices during separate synchronization phases, the method achieves significant efficiency gains, using up to eight times fewer phases than full retraining while managing trade-offs between forgetting and utility in decentralized settings.
$f$-FUM: Federated Unlearning via min--max and $f$-divergence
arXiv:2602.06187v2 Announce Type: replace Abstract: Federated learning (FL) enables collaborative training while keeping raw data on clients. Deletion requests and the discovery of poisoned data create a need to remove selected contributions from an already trained model. This is challenging in FL because client data are decentralized and their influence is entangled through repeated aggregation. We present f-FUM, an active federated unlearning method that builds on teacher-student forget/retain optimization. Starting from a pretrained global model, the method keeps that model fixed as a teacher. Clients maximize an $f$-divergence between student and teacher predictive distributions on forget examples, then minimize KL-based teacher-student disagreement and supervised loss on retained examples. The updates are aggregated in separate, sample-weighted synchronization phases without moving raw data to the server or changing the model architecture. We evaluate forget-side divergence choices in client-level and data-level deletion settings involving backdoors, label confusion, and clean-data deletion. Under equal synchronization-phase budgets, divergence choice changes the forgetting-utility trade-off, yielding improvements in some settings and mixed results in others. In the evaluated configurations, f-FUM uses up to 8 times fewer synchronization phases than full federated retraining.
Abstract:Federated learning (FL) enables collaborative training while keeping raw data on clients. Deletion requests and the discovery of poisoned data create a need to remove selected contributions from an already trained model. This is challenging in FL because client data are decentralized and their influence is entangled through repeated aggregation. We present f-FUM, an active federated unlearning method that builds on teacher-student forget/retain optimization. Starting from a pretrained global model, the method keeps that model fixed as a teacher. Clients maximize an $f$-divergence between student and teacher predictive distributions on forget examples, then minimize KL-based teacher-student disagreement and supervised loss on retained examples. The updates are aggregated in separate, sample-weighted synchronization phases without moving raw data to the server or changing the model architecture. We evaluate forget-side divergence choices in client-level and data-level deletion settings involving backdoors, label confusion, and clean-data deletion. Under equal synchronization-phase budgets, divergence choice changes the forgetting-utility trade-off, yielding improvements in some settings and mixed results in others. In the evaluated configurations, f-FUM uses up to 8 times fewer synchronization phases than full federated retraining.
Machine Learning (cs.LG)
arXiv:2602.06187 [cs.LG]
(or arXiv:2602.06187v2 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2602.06187 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 271 cuvinte
Subiecte conexe
Exact Federated Continual Unlearning for Ridge Heads on Frozen Foundation Models
arXiv · 16 iun. 2026
New framework for auditing machine unlearning
Google Research · 10 iun. 2026
pFedUL: Layer-Aware Federated Unlearning for Personalized Federated Learning
arXiv · 16 iun. 2026
Lethe: Principled Dual-Stream Update for Persistent Knowledge Erasure in Federated Unlearning
arXiv · 6 oct. 2026