Robust One Round Federated Learning with Predictive Space Bayesian Inference
This work addresses communication bottlenecks and robustness issues in federated learning, particularly for applications with heterogeneous data, though it is incremental as it builds on existing Bayesian approaches.
The paper tackles the dual challenges of robustness and communication efficiency in federated learning by proposing a Bayesian method that aggregates client predictive posteriors in one round, showing competitive performance with existing techniques and outperforming them in heterogeneous data settings.
Making predictions robust is an important challenge. A separate challenge in federated learning (FL) is to reduce the number of communication rounds, particularly since doing so reduces performance in heterogeneous data settings. To tackle both issues, we take a Bayesian perspective on the problem of learning a global model. We show how the global predictive posterior can be approximated using client predictive posteriors. This is unlike other works which aggregate the local model space posteriors into the global model space posterior, and are susceptible to high approximation errors due to the posterior's high dimensional multimodal nature. In contrast, our method performs the aggregation on the predictive posteriors, which are typically easier to approximate owing to the low-dimensionality of the output space. We present an algorithm based on this idea, which performs MCMC sampling at each client to obtain an estimate of the local posterior, and then aggregates these in one round to obtain a global ensemble model. Through empirical evaluation on several classification and regression tasks, we show that despite using one round of communication, the method is competitive with other FL techniques, and outperforms them on heterogeneous settings. The code is publicly available at https://github.com/hasanmohsin/FedPredSpace_1Round.