Does One-shot Give the Best Shot? Mitigating Model Inconsistency in One-shot Federated Learning

Hui Zeng, Wenke Huang, Tongqing Zhou, Xinyi Wu, Guancheng Wan, Yingwen Chen, Zhiping Cai
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:74080-74097, 2025.

Abstract

Turning the multi-round vanilla Federated Learning into one-shot FL (OFL) significantly reduces the communication burden and makes a big leap toward practical deployment. However, this work empirically and theoretically unravels that existing OFL falls into a garbage (inconsistent one-shot local models) in and garbage (degraded global model) out pitfall. The inconsistency manifests as divergent feature representations and sample predictions. This work presents a novel OFL framework FAFI that enhances the one-shot training on the client side to essentially overcome inferior local uploading. Specifically, unsupervised feature alignment and category-wise prototype learning are adopted for clients’ local training to be consistent in representing local samples. On this basis, FAFI uses informativeness-aware feature fusion and prototype aggregation for global inference. Extensive experiments on three datasets demonstrate the effectiveness of FAFI, which facilitates superior performance compared with 11 OFL baselines (+10.86% accuracy). Code available at https://github.com/zenghui9977/FAFI_ICML25

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-zeng25c, title = {Does One-shot Give the Best Shot? {M}itigating Model Inconsistency in One-shot Federated Learning}, author = {Zeng, Hui and Huang, Wenke and Zhou, Tongqing and Wu, Xinyi and Wan, Guancheng and Chen, Yingwen and Cai, Zhiping}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {74080--74097}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/zeng25c/zeng25c.pdf}, url = {https://proceedings.mlr.press/v267/zeng25c.html}, abstract = {Turning the multi-round vanilla Federated Learning into one-shot FL (OFL) significantly reduces the communication burden and makes a big leap toward practical deployment. However, this work empirically and theoretically unravels that existing OFL falls into a garbage (inconsistent one-shot local models) in and garbage (degraded global model) out pitfall. The inconsistency manifests as divergent feature representations and sample predictions. This work presents a novel OFL framework FAFI that enhances the one-shot training on the client side to essentially overcome inferior local uploading. Specifically, unsupervised feature alignment and category-wise prototype learning are adopted for clients’ local training to be consistent in representing local samples. On this basis, FAFI uses informativeness-aware feature fusion and prototype aggregation for global inference. Extensive experiments on three datasets demonstrate the effectiveness of FAFI, which facilitates superior performance compared with 11 OFL baselines (+10.86% accuracy). Code available at https://github.com/zenghui9977/FAFI_ICML25} }
Endnote
%0 Conference Paper %T Does One-shot Give the Best Shot? Mitigating Model Inconsistency in One-shot Federated Learning %A Hui Zeng %A Wenke Huang %A Tongqing Zhou %A Xinyi Wu %A Guancheng Wan %A Yingwen Chen %A Zhiping Cai %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-zeng25c %I PMLR %P 74080--74097 %U https://proceedings.mlr.press/v267/zeng25c.html %V 267 %X Turning the multi-round vanilla Federated Learning into one-shot FL (OFL) significantly reduces the communication burden and makes a big leap toward practical deployment. However, this work empirically and theoretically unravels that existing OFL falls into a garbage (inconsistent one-shot local models) in and garbage (degraded global model) out pitfall. The inconsistency manifests as divergent feature representations and sample predictions. This work presents a novel OFL framework FAFI that enhances the one-shot training on the client side to essentially overcome inferior local uploading. Specifically, unsupervised feature alignment and category-wise prototype learning are adopted for clients’ local training to be consistent in representing local samples. On this basis, FAFI uses informativeness-aware feature fusion and prototype aggregation for global inference. Extensive experiments on three datasets demonstrate the effectiveness of FAFI, which facilitates superior performance compared with 11 OFL baselines (+10.86% accuracy). Code available at https://github.com/zenghui9977/FAFI_ICML25
APA
Zeng, H., Huang, W., Zhou, T., Wu, X., Wan, G., Chen, Y. & Cai, Z.. (2025). Does One-shot Give the Best Shot? Mitigating Model Inconsistency in One-shot Federated Learning. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:74080-74097 Available from https://proceedings.mlr.press/v267/zeng25c.html.

Related Material