# Split Federated Learning Architectures for High-Accuracy and Low-Delay Model Training Source: [Split Federated Learning Architectures for High-Accuracy and Low-Delay Model Training]() Collection index: [llms.txt]() Data updated: 2026-03-11T05:00:19.252Z Authors: Yiannis Papageorgiou, Yannis Thomas, Ramin Khalili, Iordanis Koutsopoulos Published: 2026-03-09T17:53:20.000Z ## Abstract Can we find a network architecture for ML model training so as to optimize training loss (and thus, accuracy) in Split Federated Learning (SFL)? And can this architecture also reduce training delay and communication overhead? While accuracy is not influenced by how we split the model in ordinary, state-of-the-art SFL, in this work we answer the questions above in the affirmative. Recent Hierarchical SFL (HSFL) architectures adopt a three-tier training structure consisting of clients, (local) aggregators, and a central server. In this architecture, the model is partitioned at two partitioning layers into three sub-models, which are executed across the three tiers. Despite their merits, HSFL architectures overlook the impact of the partitioning layers and client-to-aggregator assignments on accuracy, delay, and overhead. This work explicitly captures the impact of the partitioning layers and client-to-aggregator assignments on accuracy, delay and overhead by formulating a joint optimization problem. We prove that the problem is NP-hard and propose the first accuracy-aware heuristic algorithm that explicitly accounts for model accuracy, while remaining delay-efficient. Simulation results on public datasets show that our approach can improve accuracy by 3%, while reducing delay by 20% and overhead by 50%, compared to state-of-the-art SFL and HSFL schemes. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Image Classification on CIFAR-10 (test) (Delay and Overhead Gain)](): Image Classification; CIFAR-10 (test); Delay Gain (%): 35. - [Image Classification on MNIST (test) (Efficiency/Overhead Metrics)](): Image Classification; MNIST (test); Delay Gain: 22. - [Image Classification on CINIC-10 (test)](): Image Classification; CINIC-10 (test); Accuracy: not recorded. ## Original sources and implementation [Original paper]() [PDF]()