FedSCAl: Leveraging Server and Client Alignment for Unsupervised Federated Source-Free Domain Adaptation

2025-12-07

AI Summary

In unsupervised federated learning, large inter-client domain shifts, absence of source-domain data access, and lack of local labels cause severe client drift and unreliable pseudo-labels. To address this, we propose FedSCAl, a Federated Source-free Domain Adaptation framework. Its core innovation is the Server–Client Alignment (SCAl) mechanism, which enforces prediction consistency between local client models and the global server model on unlabeled data—thereby mitigating drift and improving pseudo-label quality. FedSCAl integrates unsupervised domain adaptation, federated collaborative optimization, and cross-device prediction alignment, enabling personalized yet globally coherent learning without requiring source-domain data. Evaluated on multiple vision benchmarks, FedSCAl significantly outperforms existing state-of-the-art methods, demonstrating superior robustness and generalization across heterogeneous domains.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Computer Vision: Adversarial Attacks & Robustness
• Multiagent Systems: Adversarial Agents

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
• Systems and Infrastructure for Web, Mobile and WoT: Decentralized Web and Fediverse systems

Abstract

We address the Federated source-Free Domain Adaptation (FFreeDA) problem, with clients holding unlabeled data with significant inter-client domain gaps. The FFreeDA setup constrains the FL frameworks to employ only a pre-trained server model as the setup restricts access to the source dataset during the training rounds. Often, this source domain dataset has a distinct distribution to the clients' domains. To address the challenges posed by the FFreeDA setup, adaptation of the Source-Free Domain Adaptation (SFDA) methods to FL struggles with client-drift in real-world scenarios due to extreme data heterogeneity caused by the aforementioned domain gaps, resulting in unreliable pseudo-labels. In this paper, we introduce FedSCAl, an FL framework leveraging our proposed Server-Client Alignment (SCAl) mechanism to regularize client updates by aligning the clients' and server model's predictions. We observe an improvement in the clients' pseudo-labeling accuracy post alignment, as the SCAl mechanism helps to mitigate the client-drift. Further, we present extensive experiments on benchmark vision datasets showcasing how FedSCAl consistently outperforms state-of-the-art FL methods in the FFreeDA setup for classification tasks.

Problem

• Addresses unsupervised federated learning with unlabeled client data and domain gaps.
• Mitigates client-drift in source-free domain adaptation under data heterogeneity.
• Improves pseudo-label accuracy by aligning server and client model predictions.

Innovation

• Server-Client Alignment regularizes client updates
• Aligns client and server model predictions
• Improves pseudo-labeling accuracy by mitigating client-drift