PARDON: Privacy-Aware and Robust Federated Domain Generalization

TL;DR

This paper introduces FISC, a federated learning method that enhances privacy and robustness in domain generalization by extracting style representations and using contrastive learning, outperforming existing approaches on multiple datasets.

Contribution

FISC is a novel federated domain generalization framework that handles complex domain distributions securely without sharing sensitive data, improving unseen domain performance.

Findings

FISC outperforms state-of-the-art methods on multiple datasets.

Achieves up to 57.22% accuracy improvement on unseen domains.

Effectively learns multi-domain representations with privacy preservation.

Abstract

Federated Learning (FL) shows promise in preserving privacy and enabling collaborative learning. However, most current solutions focus on private data collected from a single domain. A significant challenge arises when client data comes from diverse domains (i.e., domain shift), leading to poor performance on unseen domains. Existing Federated Domain Generalization approaches address this problem but assume each client holds data for an entire domain, limiting their practicality in real-world scenarios with domain-based heterogeneity and client sampling. In addition, certain methods enable information sharing among clients, raising privacy concerns as this information could be used to reconstruct sensitive private data. To overcome this, we introduce FISC, a novel FedDG paradigm designed to robustly handle more complicated domain distributions between clients while ensuring security. FISC enables learning across domains by extracting an interpolative style from local styles and employing contrastive learning. This strategy gives clients multi-domain representations and unbiased convergent targets. Empirical results on multiple datasets, including PACS, Office-Home, and IWildCam, show FISC outperforms state-of-the-art (SOTA) methods. Our method achieves accuracy on unseen domains, with improvements ranging from 3.64% to 57.22% on unseen domains. Our code is available at https://github.com/judydnguyen/PARDON-FedDG.