Federated Variational Inference: Towards Improved Personalization and Generalization
This work addresses the challenge of training effective models in heterogeneous federated learning environments, which is crucial for applications like cross-device AI, but it is incremental as it builds on existing variational inference and federated learning techniques.
The paper tackles the problem of data and model heterogeneity in federated learning by proposing Federated Variational Inference (FedVI), a method that uses a hierarchical generative model and variational inference to improve personalization and generalization, achieving state-of-the-art results on FEMNIST and CIFAR-100 image classification tasks.
Conventional federated learning algorithms train a single global model by leveraging all participating clients' data. However, due to heterogeneity in client generative distributions and predictive models, these approaches may not appropriately approximate the predictive process, converge to an optimal state, or generalize to new clients. We study personalization and generalization in stateless cross-device federated learning setups assuming heterogeneity in client data distributions and predictive models. We first propose a hierarchical generative model and formalize it using Bayesian Inference. We then approximate this process using Variational Inference to train our model efficiently. We call this algorithm Federated Variational Inference (FedVI). We use PAC-Bayes analysis to provide generalization bounds for FedVI. We evaluate our model on FEMNIST and CIFAR-100 image classification and show that FedVI beats the state-of-the-art on both tasks.