Straggler-Aware Resource Allocation in Semi-Decentralized Federated Learning for Large-Scale Models Over OTNs
We propose a straggler-aware resource allocation in semi-decentralized federated learning for large-scale machine-learning models over OTNs. Simulations show it achieves only 2.55% gap to MILP while reducing runtime by 99.78%, and outperforms the SoTA heuristic with 7% higher task success rate and 41% less runtime. ©2025 The Author(s)
Authors
- Qiaolun Zhang (ORCID: https://orcid.org/0000-0002-7012-950X)
- Yongli Zhao (ORCID: https://orcid.org/0000-0003-3716-8248)
- Massimo Tornatore (ORCID: https://orcid.org/0000-0003-0740-1061)
- Francesco Musumeci (ORCID: https://orcid.org/0000-0002-5554-3366)
- Jie Zhang (ORCID: https://orcid.org/0000-0001-7750-2197)
- Meng Lian (ORCID: https://orcid.org/0000-0002-2202-0454)
Institutions
- Beijing University of Posts and Telecommunications (CN)
- State Key Laboratory of Information Photonics and Optical Communications
- Politecnico di Milano (IT)
Publication Details
- Published
- 2025-09-28
- DOI
- https://doi.org/10.1109/ecoc66593.2025.11263140
- Primary Topic
- Privacy-Preserving Technologies in Data
- Type
- article
- Field-Weighted Citation Impact
- 0.00