Optimizing Split Federated Learning with Unstable Client Participation
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866918433793245184 |
|---|---|
| author | Wei, Wei Lin, Zheng Liu, Xihui Du, Hongyang Niyato, Dusit Chen, Xianhao |
| author_facet | Wei, Wei Lin, Zheng Liu, Xihui Du, Hongyang Niyato, Dusit Chen, Xianhao |
| contents | To enable training of large artificial intelligence (AI) models at the network edge, split federated learning (SFL) has emerged as a promising approach by distributing computation between edge devices and a server. However, while unstable network environments pose significant challenges to SFL, prior schemes often overlook such an effect by assuming perfect client participation, rendering them impractical for real-world scenarios. In this work, we develop an optimization framework for SFL with unstable client participation. We theoretically derive the first convergence upper bound for SFL with unstable client participation by considering activation uploading failures, gradient downloading failures, and model aggregation failures. Based on the theoretical results, we formulate a joint optimization problem for client sampling and model splitting to minimize the upper bound. We then develop an efficient solution approach to solve the problem optimally. Extensive simulations on EMNIST and CIFAR-10 demonstrate the superiority of our proposed framework compared to existing benchmarks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_17398 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Optimizing Split Federated Learning with Unstable Client Participation Wei, Wei Lin, Zheng Liu, Xihui Du, Hongyang Niyato, Dusit Chen, Xianhao Networking and Internet Architecture To enable training of large artificial intelligence (AI) models at the network edge, split federated learning (SFL) has emerged as a promising approach by distributing computation between edge devices and a server. However, while unstable network environments pose significant challenges to SFL, prior schemes often overlook such an effect by assuming perfect client participation, rendering them impractical for real-world scenarios. In this work, we develop an optimization framework for SFL with unstable client participation. We theoretically derive the first convergence upper bound for SFL with unstable client participation by considering activation uploading failures, gradient downloading failures, and model aggregation failures. Based on the theoretical results, we formulate a joint optimization problem for client sampling and model splitting to minimize the upper bound. We then develop an efficient solution approach to solve the problem optimally. Extensive simulations on EMNIST and CIFAR-10 demonstrate the superiority of our proposed framework compared to existing benchmarks. |
| title | Optimizing Split Federated Learning with Unstable Client Participation |
| topic | Networking and Internet Architecture |
| url | https://arxiv.org/abs/2509.17398 |