FedBE: Making Bayesian Model Ensemble Applicable to Federated Learning
This addresses the problem of model aggregation in federated learning for users with non-i.i.d. data, offering an incremental improvement as a compatible module.
The paper tackles the challenge of aggregating local models into a global model in federated learning, especially with non-i.i.d. data, by proposing FedBE, a Bayesian model ensemble method that improves robustness and performance, as validated empirically.
Federated learning aims to collaboratively train a strong global model by accessing users' locally trained models but not their own data. A crucial step is therefore to aggregate local models into a global model, which has been shown challenging when users have non-i.i.d. data. In this paper, we propose a novel aggregation algorithm named FedBE, which takes a Bayesian inference perspective by sampling higher-quality global models and combining them via Bayesian model Ensemble, leading to much robust aggregation. We show that an effective model distribution can be constructed by simply fitting a Gaussian or Dirichlet distribution to the local models. Our empirical studies validate FedBE's superior performance, especially when users' data are not i.i.d. and when the neural networks go deeper. Moreover, FedBE is compatible with recent efforts in regularizing users' model training, making it an easily applicable module: you only need to replace the aggregation method but leave other parts of your federated learning algorithm intact. Our code is publicly available at https://github.com/hongyouc/FedBE.