{"type":"ScholarlyArticle","title":"DASFL: Dynamic Adaptive Split Federated Learning for Heterogeneous Clients","authors":[{"name":"Aijing Li","slug":"aijing-li","given":"Aijing","family":"Li","affiliation":""},{"name":"Yawen Li","slug":"yawen-li","given":"Yawen","family":"Li","affiliation":""},{"name":"Guanhua Ye","slug":"guanhua-ye","given":"Guanhua","family":"Ye","affiliation":""},{"name":"Dandan Liu","slug":"dandan-liu","given":"Dandan","family":"Liu","affiliation":""},{"name":"Tong Zhao","slug":"tong-zhao","given":"Tong","family":"Zhao","affiliation":""},{"name":"Zeli Guan","slug":"zeli-guan","given":"Zeli","family":"Guan","affiliation":""}],"abstract":"Split Federated Learning (SFL) has emerged as a pivotal paradigm for privacy-preserving distributed training on resource-constrained edge devices by partitioning neural networks between clients and a server. A critical design choice in SFL is the split layer, which determines the computation distribution and the semantic level of smashed data, directly impacting communication overhead, training latency, and model accuracy.\r\nExisting SFL methods largely ignore the time-varying nature of edge resources, relying on static resource profiles that lead to suboptimal efficiency and compromise model convergence.\r\nIn this paper, we propose Dynamic Adaptive Split Federated Learning (DASFL), a unified framework that jointly addresses dynamic resource heterogeneity and non-IID data distributions.\r\nWe employ a resource-aware dynamic split layer selection strategy that enables each client to minimize per-round latency by adapting to instantaneous local conditions. We also design an adaptive masked aggregation that robustly synchronizes client updates under split-induced structural heterogeneity.\r\nExtensive experiments on multiple benchmark models and datasets, conducted under heterogeneous, time-varying resource conditions and non-IID data distributions, demonstrate that DASFL achieves a superior accuracy-efficiency trade-off and faster convergence compared to state-of-the-art SFL baselines.","doi":"10.24963/ijcai.2026/499","url":"https://ijcai.org/proceedings/2026/499","pdf":"https://ijcai.org/proceedings/2026/0499.pdf","markdown":"https://ijcai.org/proceedings/2026/499.md","bibtex":"https://ijcai.org/proceedings/2026/bibtex/499","volume":{"slug":"2026","title":"Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence","short_title":"IJCAI 2026","url":"https://ijcai.org/proceedings/2026/","year":2026,"location":"Bremen, Germany","dates":"15-21 August 2026","editors":"Diego Calvanese","isbn":"978-1-956792-09-6","publisher":"International Joint Conferences on Artificial Intelligence Organization","doi_prefix":"10.24963/ijcai.2026"},"section":"Machine Learning","track":"Main Track","number":499,"pages":{"first":4483,"last":4491},"keywords":["Machine Learning: Federated learning"],"videos":[],"references":[],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}