HASFL: Heterogeneity-aware split federated learning over edge computing systems

Author Identifier (ORCID)

Wei Ni’s ORCID record ORCID Logo

Abstract

Split federated learning (SFL) has emerged as a promising paradigm to democratize machine learning (ML) on edge devices by enabling layer-wise model partitioning.However, existing SFL approaches suffer significantly from the straggler effect due to the heterogeneous capabilities of edge devices. To address the fundamental challenge, we propose adaptively controlling batch sizes (BSs) and model splitting (MS) for edge devices to overcome resource heterogeneity. We first derive a tight convergence bound of SFL that quantifies the impact of varied BSs and MS on learning performance. Based on the convergence bound, we propose HASFL, a heterogeneity-aware SFL framework capable of adaptively controlling BS and MS to balance communication-computing latency and training convergence in heterogeneous edge networks. Extensive experiments with various datasets validate the effectiveness of HASFL and demonstrate its superiority over state-of-the-art benchmarks.

Keywords

batch size, federated learning, mobile edge computing, model splitting, split federated learning

Document Type

Journal Article

Date of Publication

8-1-2026

ISSN

15361233

Volume

25

Issue

8

Publication Title

IEEE Transactions on Mobile Computing

Publisher

IEEE

School

School of Engineering

Funding Information

Research Grants Council of Hong Kong (Grant Number: 27213824 and CRS HKU702/24)

Copyright

subscription content

Content Type

Metadata only

First Page

12455

Last Page

12471

Recommended Citation

Lin, Z., Chen, Z., Chen, X., Ni, W., & Gao, Y. (2026). HASFL: Heterogeneity-aware split federated learning over edge computing systems. IEEE Transactions on Mobile Computing, 25(8), 12455–12471. https://doi.org/10.1109/TMC.2026.3673358

Share

 
COinS
 

Link to publisher version (DOI)

10.1109/TMC.2026.3673358