Federated Out-of-Distribution Generalization: A Causal Augmentation View

2025-04-28

AI Summary

In federated learning, client data distribution shifts impair model generalization to out-of-distribution (OOD) samples, while existing data augmentation methods—limited by generation fidelity and semantic plausibility—fail to mitigate spurious correlations. To address this, we propose FedCAug, the first data-free, privacy-preserving causal augmentation framework for federated OOD generalization. FedCAug jointly localizes causal regions across clients and generates semantically consistent counterfactual images via federated collaboration, effectively disentangling background from object. It incorporates intra-client context-aware augmentation and a shared-data-free training architecture, eliminating raw data and label exchange. Evaluated on three benchmark datasets, FedCAug significantly reduces background dependency and achieves an average OOD accuracy gain of 3.2–5.7% over state-of-the-art methods, simultaneously ensuring strong privacy guarantees and robust generalization.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Computer Vision: Bias, Fairness & Privacy
• Reasoning under Uncertainty: Causality

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Federated learning aims to collaboratively model by integrating multi-source information to obtain a model that can generalize across all client data. Existing methods often leverage knowledge distillation or data augmentation to mitigate the negative impact of data bias across clients. However, the limited performance of teacher models on out-of-distribution samples and the inherent quality gap between augmented and original data hinder their effectiveness and they typically fail to leverage the advantages of incorporating rich contextual information. To address these limitations, this paper proposes a Federated Causal Augmentation method, termed FedCAug, which employs causality-inspired data augmentation to break the spurious correlation between attributes and categories. Specifically, it designs a causal region localization module to accurately identify and decouple the background and objects in the image, providing rich contextual information for causal data augmentation. Additionally, it designs a causality-inspired data augmentation module that integrates causal features and within-client context to generate counterfactual samples. This significantly enhances data diversity, and the entire process does not require any information sharing between clients, thereby contributing to the protection of data privacy. Extensive experiments conducted on three datasets reveal that FedCAug markedly reduces the model's reliance on background to predict sample labels, achieving superior performance compared to state-of-the-art methods.

Problem

• Mitigates data bias impact in federated learning
• Breaks spurious attribute-category correlations causally
• Enhances data diversity without compromising privacy

Innovation

• Causality-inspired data augmentation breaks spurious correlations
• Causal region localization decouples background and objects
• Generates counterfactual samples without client information sharing