Feature-Aware One-Shot Federated Learning via Hierarchical Token Sequences
Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (28) · 23819-23827
DOI: 10.1609/aaai.v40i28.39557 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Shudong Liu und 4 weiteren Autor:innen.
Abstract (Originalsprache)
One-shot federated learning (OSFL) reduces the communication cost and privacy risks of iterative federated learning by constructing a global model with a single round of communication. However, most existing methods struggle to achieve robust performance on real-world domains such as medical imaging, or are inefficient when handling non-IID (Independent and Identically Distributed) data. To address these limitations, we introduce FALCON, a novel framework that enhances the effectiveness of OSFL over non-IID image data. The core idea of FALCON is to leverage the feature-aware hierarchical token sequences generation and knowledge distillation into OSFL. First, each client leverages a pretrained visual encoder with hierarchical scale encoding to compress images into hierarchical token sequences, which capture multi-scale semantics. Second, a multi-scale autoregressive transformer generator is used to model the distribution of these token sequences and generate the synthetic sequences. Third, clients upload the synthetic sequences along with the local classifier trained on the real token sequences to the server. Finally, the server incorporates knowledge distillation into global training to reduce reliance on precise distribution modeling. Experiments on medical and natural image datasets validate the effectiveness of FALCON in diverse non-IID scenarios, outperforming the best OSFL baselines by 9.58\% in average accuracy.
Zitieren
APA
Liu, S., Zhang, H., Wang, X., Zhu, Y., & Luo, G. (2026). Feature-Aware One-Shot Federated Learning via Hierarchical Token Sequences. Proceedings of the AAAI Conference on Artificial Intelligence, 40(28), 23819-23827. https://doi.org/10.1609/aaai.v40i28.39557
Vancouver
Liu S, Zhang H, Wang X, Zhu Y, Luo G. Feature-Aware One-Shot Federated Learning via Hierarchical Token Sequences. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(28):23819-23827. doi:10.1609/aaai.v40i28.39557
BibTeX
@article{Liu2026,
author = {Liu, Shudong and Zhang, Hanwen and Wang, Xiuling and Zhu, Yuesheng and Luo, Guibo},
title = {Feature-Aware One-Shot Federated Learning via Hierarchical Token Sequences},
journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
year = {2026},
volume = {40},
number = {28},
pages = {23819--23827},
doi = {10.1609/aaai.v40i28.39557}
}
EUVIMED