Efficient Cluster Selection for Personalized Federated Learning: A Multi-Armed Bandit Approach
This work addresses the problem of efficient user clustering for personalized federated learning in dynamic networks, which is incremental as it builds on existing PFL and MAB methods.
The paper tackles the challenge of clustering users in personalized federated learning (PFL) within dynamic networks by introducing a dynamic Upper Confidence Bound (dUCB) algorithm based on multi-armed bandit principles, which helps new users find optimal clusters by balancing exploration and exploitation, with evaluations showing its effectiveness in dynamic scenarios.
Federated learning (FL) offers a decentralized training approach for machine learning models, prioritizing data privacy. However, the inherent heterogeneity in FL networks, arising from variations in data distribution, size, and device capabilities, poses challenges in user federation. Recognizing this, Personalized Federated Learning (PFL) emphasizes tailoring learning processes to individual data profiles. In this paper, we address the complexity of clustering users in PFL, especially in dynamic networks, by introducing a dynamic Upper Confidence Bound (dUCB) algorithm inspired by the multi-armed bandit (MAB) approach. The dUCB algorithm ensures that new users can effectively find the best cluster for their data distribution by balancing exploration and exploitation. The performance of our algorithm is evaluated in various cases, showing its effectiveness in handling dynamic federated learning scenarios.