When the Cross-Silo Federation Goes Offline: Continual Learning for Site Onboarding with Limited Unlabeled Data

2026-10-04

AI Summary

This study addresses the challenges of model expansion and catastrophic forgetting when new sites join offline without labels following federated training. To this end, we propose a three-stage autonomous adaptation framework. The method achieves dynamic model expansion by integrating reconstruction experts, cluster analysis, and class mean estimation within a shared representation space, while effectively mitigating forgetting without access to raw data through prototype-based or locally recorded information, thereby balancing performance across old and new classes. Experimental results on industrial datasets demonstrate that the proposed approach maintains an accuracy of ≥0.868 on old classes with a forgetting rate of ≤0.063, significantly outperforming baseline methods such as knowledge distillation.

Abstract

An organization often holds too little labeled data to train a model that generalizes, and the records that would supply the rest sit with organizations that cannot release them. Cross-silo federated learning offers a way through, since participants exchange model parameters rather than records, but it ordinarily settles two aspects of the arrangement in advance, the participating sites and the classes the model can predict, and deployment can breach both. A new site joins after training, once the established sites have finished their engagement and gone offline, and its records arrive unlabeled, mixing conditions the model already recognizes with conditions no participant has observed. We present an autonomous three-stage procedure that expands the model entirely at the joining site: reconstruction experts screen for novelty, clustering separates the flagged records into candidate conditions, and class means describe the old classes, all inside one shared representation. Those classes were learned from records that never leave their owners, so the usual defenses against forgetting are unavailable, and the procedure supplies the evidence they would have carried from either of two dissimilar sources, prototypes held by the federation or records held by the joining site. On a real industrial condition-monitoring dataset, run end to end with no label consulted, either source holds old-class accuracy at 0.868 or above with forgetting at most 0.063, and the two differ by 0.021, so a configuration can be chosen by the disclosure it permits rather than the accuracy it delivers. Both keep old- and new-class accuracy in balance where every alternative we measure gives up one for the other, and both retain more of the old classes than distillation- and regularization-based baselines. The balance still holds with only 6 labeled records per arriving condition and 3 retained per old class.

Problem

• Cross-silo federated learning
• Continual learning
• Catastrophic forgetting
• Unlabeled data
• Site onboarding

Innovation

• Cross-silo federated learning
• Continual learning
• Novelty detection
• Catastrophic forgetting
• Unlabeled data