Efficiently Assemble Normalization Layers and Regularization for Federated Domain Generalization

TL;DR

This work tackles domain shift in Federated Domain Generalization (FedDG) under privacy constraints. It introduces gPerXAN, which combines a personalized normalization scheme (PerXAN) that globally aggregates Instance Normalization and locally personalizes Batch Normalization, with a simple guiding regularizer to encourage domain-invariant representations for the global classifier. The key contributions are the Explicitly Assembled Normalization (XAN) and PerXAN, plus Regularization as Guidance, validated on PACS, Office-Home, and Camelyon17 where it outperforms state-of-the-art FedDG baselines. The approach preserves privacy by sharing only model parameters and offers efficiency advantages over data-sharing-based methods, providing a practical, scalable solution for DG in federated settings.

Abstract

Domain shift is a formidable issue in Machine Learning that causes a model to suffer from performance degradation when tested on unseen domains. Federated Domain Generalization (FedDG) attempts to train a global model using collaborative clients in a privacy-preserving manner that can generalize well to unseen clients possibly with domain shift. However, most existing FedDG methods either cause additional privacy risks of data leakage or induce significant costs in client communication and computation, which are major concerns in the Federated Learning paradigm. To circumvent these challenges, here we introduce a novel architectural method for FedDG, namely gPerXAN, which relies on a normalization scheme working with a guiding regularizer. In particular, we carefully design Personalized eXplicitly Assembled Normalization to enforce client models selectively filtering domain-specific features that are biased towards local data while retaining discrimination of those features. Then, we incorporate a simple yet effective regularizer to guide these models in directly capturing domain-invariant representations that the global model's classifier can leverage. Extensive experimental results on two benchmark datasets, i.e., PACS and Office-Home, and a real-world medical dataset, Camelyon17, indicate that our proposed method outperforms other existing methods in addressing this particular problem.

Figures (3)

• Figure 1: An overview of our proposed gPerXAN method for solving the FedDG problem.
• Figure 2: Contribution of the regularizer on different methods.
• Figure 3: Visualization of representation extracted from the global model on the PACS dataset.