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

Ahmadreza Eslaminia; Klara Nahrstedt; Chenhui Shao

cs.LGOct 4, 2026

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

Authors: Ahmadreza Eslaminia, Klara Nahrstedt, Chenhui Shao

Organizations: Coordinated Science Laboratory University of Illinois Urbana-Champaign · Dept. of Mechanical Engineering University of Michigan

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.

Figures & tables

Figure 1: Onboarding a late-joining site. (A) The base federated session leaves behind a shared classifier, one reconstruction expert per known class, and representation-level prototypes. (B) The joining site screens its own unlabeled pool against those experts and clusters what it flags into candidate conditions. (C) The classifier expands by one output per recovered group, with the layers producing $\phi_{\theta}$ held fixed. Prototype replay and local rehearsal are alternative sources of old-class evidence.

Name	Old outputs	$\phi_{\theta}$	Prototypes	Records	Applicable when
	Held fixed		Old-class evidence
Finetune	no	no	no	no	no old-class evidence is available
FrozenHead	yes	no	no	no	only the model may be reused
Global	yes	yes	yes	no	class means may be released, nothing retained
Local	yes	yes	no	$b$ /class	nothing released, records retained locally
Dual	yes	yes	yes	$b$ /class	both disclosures are permitted

Table 1: Adaptation configurations, what each holds fixed, and the disclosure each requires.

b

is the number of records the joining site retains per old class. Freezing

\phi_{\theta}

keeps prototypes and retained records in a coordinate system that does not shift while the classifier is expanded.

Appendix figures & tables5 assets

Supplementary material from the paper’s appendix.

Appendix

Score	ROC-AUC	Average precision	Recall
Maximum softmax probability	$0.884\!\pm\!0.033$	$0.923\!\pm\!0.032$	$0.614\!\pm\!0.085$
Energy score	$0.875\!\pm\!0.040$	$0.932\!\pm\!0.020$	$0.636\!\pm\!0.095$
Class-conditional reconstruction (ours)	$0.941\!\pm\!0.018$	$0.950\!\pm\!0.034$	$0.872\!\pm\!0.057$

Appendix

Table A1: Post-hoc screening scores on the joining site’s pool, over 5 seeds. Recall is measured with each score thresholded at the 0.9 quantile of its own familiar-record scores, so all three operate at the same acceptance rate, 0.89. Mean

\pm

sample standard deviation.

Figure A1: Regularization sweeps for the reference methods against the replay-based configurations, 5 seeds. Each curve traces one method as its regularization weight increases. Old-class accuracy is bought with new-class accuracy along each sweep; the replay configurations lie on neither.

Figure A2: Joining-site data budget under Local , over 5 splits with all three conditions arriving at once. $n$ is the number of labeled records per arriving condition and $b$ the number retained per old class. (a) old-class accuracy over all participating sites and (b) new-class accuracy at the joining site are the two quantities Table 2 reports; (c) is their harmonic mean, which is high only where both are. The outlined block marks the region $n\geq 3{}$ , $b\geq 1{}$ discussed in the text.

Seed	Groups recovered	ARI	Unassigned
7	3	0.764	12
11	3	0.507	9
42	3	0.734	11
123	3	0.679	13
314	3	0.667	9

Appendix

Table A2: Grouping outcomes for the complete runs. ARI is the adjusted Rand index between the recovered partition and the reference conditions, computed on the assigned records.

Symbol	Meaning
$M$ , $K$	number of established sites; number of old classes
$R$	number of new conditions recovered and added as outputs
$\mathcal{U}$	unlabeled pool held by the joining site
$\phi_{\theta}(x)$	early-layer activation, shared by all stages
$g_{c}$ , $e_{c}(x)$	reconstruction expert for class $c$ ; its squared error
$s_{c}$	scale of expert $c$ , its median error on its own class

Appendix

Table 7

Explore similar work

Jun 24, 2026stat.ML

FedReLa: Imbalanced Federated Learning via Re-Labeling

Federated learning has emerged as the foremost approach for decentralized model training with privacy preservation. The global class imbalance and cross-client data heterogeneity naturally coexist, and the mismatch between local and global imbalances exacerbates the performance degradation of the aggregated model. The agnosticism of global class distribution poses significant challenges for data-level methods, especially under extreme conditions with severe class absence across clients. In this paper, we propose FedReLa, a novel data-level approach that tackles the coexistence of data heterogeneity and class imbalance in federated learning. By re-labeling samples with a feature-dependent label re-allocator, FedReLa corrects biased global decision boundaries without requiring knowledge of the global class distribution. This modular, model-agnostic approach can be integrated with algorithmic methods to deliver consistent improvements without additional communication overhead. Through extensive experiments, our method significantly improves the accuracy of minority classes and the overall accuracy on stepwise-imbalanced and long-tailed datasets, outperforming the previous state of the art.

Guangzheng Hu, Patricia Menéndez, Feng Liu +3

School of Mathematics and Statistics, University of Melbourne, Victoria, Australia · School of Computing and Information Systems, University of Melbourne, Victoria, Australia · School of Statistics and Data Science, LPMC, KLMDASR, and LEBPS, Nankai University

Federated Learning Imbalanced Classification Fedavg Heterogeneity Label Propagation

Ask PDF

Date pendingcs.LG

Class-wise Contribution Estimation via Logit Maximization for Federated Learning

Federated learning (FL) enables collaborative learning of computer vision models, where privacy and regulatory constraints prevent centralizing data across devices or organizations. However, practical FL deployments often exhibit severe class imbalance and label skew, causing standard aggregation protocols to overfit dominant clients and degrade minority-class performance. We propose a data-free, class-wise contribution estimation and aggregation framework based on logit maximization (CELM) that does not require sharing raw data, client metadata, or auxiliary public datasets. The FL server probes client updates to obtain class-wise evidence scores and assembles a cross-client evidence matrix, which quantifies both per-class competence and class coverage. Using this matrix, we compute contribution weights that upweight clients providing strong, discriminative evidence for underrepresented classes. The resulting aggregation is stable due to simplex constraints and momentum smoothing, and it remains compatible with standard FL training pipelines. We evaluate the approach on representative vision benchmarks under controlled non-IID and pathological label splits, demonstrating that CELM-based aggregation improves robustness to imbalance and statistical heterogeneity, while yielding better performance without requiring any additional data exchange.

Asim Ukaye, Nurbek Tastan, Mubarak Abdu-Aguye +1

Mohamed bin Zayed University of Artificial Intelligence, UAE · Michigan State University, USA

Federated Learning Contributions Logit Lens

Ask PDF

Aug 2, 2026cs.CV

Lethe: How Hard Is It to Forget? A Benchmark for Federated Unlearning in Medical Imaging

Federated learning enables medical-imaging models to be trained across hospitals, and privacy law, most explicitly the GDPR ``right to be forgotten'', turns removing a hospital's, a class's, or a patient's influence from such a model into a federated unlearning problem. This need is most acute in medicine, where patients withdraw consent and hospitals leave collaborations. Yet nearly all unlearning evidence comes from natural images, whose heterogeneity and task structure differ sharply from clinical data, so it is unclear whether existing methods transfer, and no shared protocol covers clinical data. We present Lethe, a benchmark for federated unlearning in medical imaging. It evaluates twelve methods across eight task families, from classification and segmentation to denoising, cross-modality synthesis, and vision-language question answering, at three forgetting granularities and against a retrained gold standard on utility, privacy, and cost. The central result is that what separates methods is the difficulty of the forgetting request, not the method itself. The easy removals that dominate the literature leave the methods that preserve utility indistinguishable, while only hard ones separate them. More striking, on the many medical tasks that generalize across sites, forgetting a client barely changes task performance, leaving residual membership as the signal that must be erased.

Shengchao Chen, Ting Shu

University of Technology Sydney · Shenzhen University

Medical Imaging Forgetting Privacy Hard

Ask PDF