Over-the-Air Federated Learning via Weighted Aggregation
Uloženo v:
| Název: | Over-the-Air Federated Learning via Weighted Aggregation |
|---|---|
| Autoři: | Azimi Abarghouyi, Seyed Mohammad, Tassiulas, Leandros |
| Zdroj: | IEEE Transactions on Wireless Communications. 23(12):18240-18253 |
| Témata: | Wireless networks, Servers, Transmitters, Resource management, Receivers, Performance evaluation, Convergence, Federated learning, machine learning, fading multiple access channel, over-the-air computation, analog communications |
| Popis: | This paper introduces a new federated learning scheme that leverages over-the-air computation. A novel feature of this scheme is the proposal to employ adaptive weights during aggregation, a facet treated as predefined in other over-the-air schemes. This can mitigate the impact of wireless channel conditions on learning performance, without needing channel state information at transmitter side (CSIT). We provide a mathematical methodology to derive the convergence bound for the proposed scheme in the context of computational heterogeneity and general loss functions, supplemented with design insights. Accordingly, we propose aggregation cost metrics and efficient algorithms to find optimized weights for the aggregation. Finally, through numerical experiments, we validate the effectiveness of the proposed scheme. Even with the challenges posed by channel conditions and device heterogeneity, the proposed scheme surpasses other over-the-air strategies by an accuracy improvement of 15% over the scheme using CSIT and 30% compared to the one without CSIT. |
| Popis souboru: | |
| Přístupová URL adresa: | https://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-358712 https://doi.org/10.1109/TWC.2024.3463754 |
| Databáze: | SwePub |
Full Text Finder 
Nájsť tento článok vo Web of Science | Abstrakt: | This paper introduces a new federated learning scheme that leverages over-the-air computation. A novel feature of this scheme is the proposal to employ adaptive weights during aggregation, a facet treated as predefined in other over-the-air schemes. This can mitigate the impact of wireless channel conditions on learning performance, without needing channel state information at transmitter side (CSIT). We provide a mathematical methodology to derive the convergence bound for the proposed scheme in the context of computational heterogeneity and general loss functions, supplemented with design insights. Accordingly, we propose aggregation cost metrics and efficient algorithms to find optimized weights for the aggregation. Finally, through numerical experiments, we validate the effectiveness of the proposed scheme. Even with the challenges posed by channel conditions and device heterogeneity, the proposed scheme surpasses other over-the-air strategies by an accuracy improvement of 15% over the scheme using CSIT and 30% compared to the one without CSIT. |
|---|---|
| ISSN: | 15361276 15582248 |
| DOI: | 10.1109/TWC.2024.3463754 |