ASFL: An Adaptive Model Splitting and Resource Allocation Framework for Split Federated Learning

Chuiyang Meng; Ming Tang; Vincent W. S. Wong

cs.LGFeb 19, 2026

ASFL: An Adaptive Model Splitting and Resource Allocation Framework for Split Federated Learning

Authors: Chuiyang Meng, Ming Tang, Vincent W. S. Wong

Organizations: Department of Electrical and Computer Engineering, The University of British Columbia, Vancouver, Canada · Department of Computer Science and Engineering, Southern University of Science and Technology, China

Abstract

Federated learning (FL) enables multiple clients to collaboratively train a machine learning model without sharing their raw data. However, the limited computation resources of the clients may result in a high delay and energy consumption on training. In this paper, we propose an adaptive split federated learning (ASFL) framework over wireless networks. ASFL exploits the computation resources of the central server to train part of the model and enables adaptive model splitting as well as resource allocation during training. To optimize the learning performance (i.e., convergence rate) and efficiency (i.e., delay and energy consumption) of ASFL, we theoretically analyze the convergence rate and formulate a joint learning performance and resource allocation optimization problem. Solving this problem is challenging due to the long-term delay and energy consumption constraints as well as the coupling of the model splitting and resource allocation decisions. We propose an online optimization enhanced block coordinate descent (OOE-BCD) algorithm to solve the problem iteratively. Experimental results show that when compared with five baseline schemes, our proposed ASFL framework converges faster and reduces the total delay and energy consumption by up to 75% and 80%, respectively.

Figures & tables

Notation	Definition	Notation	Definition
$B$	Bandwidth of a resource block	$q_{m}$	Output size of the $m$ -th layer (in bits)
$B^{\mathrm{DN}}$	Downlink bandwidth	$\mathcal{R}$	Set of all training rounds
$c_{n}^{\mathrm{DN},r}$	Downlink transmission rate for client $n$ in the $r$ -th training round	$R$	Number of training rounds
$c_{n}^{\mathrm{UP},r}$	Uplink transmission rate for client $n$ in the $r$ -th training round	$s_{n}^{r}$	Packet error rate of the intermediate output transmission of client $n$ in the $r$ -th training round
$D_{n}$	Number of training samples of client $n$	$u_{n}^{r}$	Allocation decision of the $k$ -th RB for client $n$ in the $r$ -th training round
$f_{n}$	CPU frequency of client $n$	$v_{m}^{\mathrm{FP}}$	Computation workload for FP of the $m$ -th layer (in FLOPs)

Table I : List of key notations

Figure 1: Illustration of our proposed ASFL framework.

Figure 2: An example of the adaptive model splitting process in stage 1 for a model with 7 layers. A client sends its fourth layer to the central server at the beginning of the $r$ -th training round.

Figure 3: An illustration of the delay in one ASFL training round.

Para.	Value	Para.	Value	Para.	Value
$B$	1 MHz	$P^{\max}$	1.5 W	$K$	8
$B^{\mathrm{DN}}$	8 MHz	$\kappa^{\mathrm{s}}$	$\frac{1}{32}$ cycles/FLOP	$p^{\mathrm{s}}$	5 W
$N_{0}$	$-$ 173 dBm/Hz	$\kappa^{\mathrm{c}}$	$\frac{1}{16}$ cycles/FLOP	$\iota$	0.05
$\alpha$	1	$f^{\mathrm{s}}$	$10^{10}$ cycles/s	$V$	10
$\gamma$	20 s	$\delta$	$0.5$ J	$\phi$	$10^{-28}$

Table II : List of key simulation parameters

Base Model	Baseline	CIFAR-10 (%)	CIFAR-100 (%)
VGG-19	ASFL	89.20	62.10
	FedAvg [ 1 ]	89.69	61.35
	SL [ 3 ]	83.41	58.30
	SFL [ 9 ]	88.77	62.43
	ACC-SFL [ 19 ]	88.02	61.96
	EPSL [ 20 ]	87.59	61.58

Table III : Comparison of the average testing accuracy between different baseline schemes.

Panel of Figure 4 — Figure 4: Comparison of the average testing accuracy versus (a) total delay and (b) total energy consumption on training.

Panel of Figure 10 — Figure 10: Effect of the adaptive model splitting on the average testing accuracy versus (a) total delay and (b) total energy consumption on training.

Panel of Figure 11 — Figure 11: Effect of the RB and transmit power allocation on the average testing accuracy versus (a) total delay and (b) total energy consumption on training.

Panel of Figure 12 — Figure 12: Effect of $\mu$ on the (a) average testing accuracy, (b) total delay, and (c) total energy consumption on training.

Split Federated Learning

Split federated learning divides models between clients and servers to train collaboratively while keeping raw data local.

See all 37 papers

Efficient Resource Optimization for Split Federated Learning
Aug 18, 2026Wei Wei, Xianhao ChenSplit Federated Learning
Department of Electrical and Computer Engineering, the University of Hong Kong, Pok Fu Lam, Hong Kong SAR, 999077
Split Federated Learning Architectures for High-Accuracy and Low-Delay Model Training
Mar 9, 2026Yiannis Papageorgiou, Yannis Thomas, Ramin Khalili +1Split Federated Learning Federated Learning
Department of Informatics, Athens University of Economics and Business, Athens, Greece · Huawei Technologies Duesseldorf Gmbh, Munich, Germany
CoASFL: Coupled Consistency Optimization for Asynchronous Split Federated Learning
Oct 19, 2026Junhao Cheng, Yunming Liao, Hongli Xu +2Split Federated Learning Federated Learning
School of Computer Science and Technology, University of Science and Technology of China, Hefei, Anhui, China, 230027 · Suzhou Institute for Advanced Research, University of Science and Technology of China, Suzhou, Jiangsu, China, 215123

Aug 18, 2026J/K move · Enter open · S save

Efficient Resource Optimization for Split Federated Learning

Wei Wei, Xianhao Chen

Department of Electrical and Computer Engineering, the University of Hong Kong, Pok Fu Lam, Hong Kong SAR, 999077

Split Federated Learning

Open paper AskPDF arXiv

Split federated learning (SFL) has emerged as a powerful paradigm for model training at the edge. However, SFL inherently involves discrete decision variables for model splitting and resource allocation, resulting in a challenging mixed-integer problem. Consequently, prior optimization schemes for SFL are either heuristic or computationally inefficient, which cannot handle large-scale user populations. To address this limitation, this work establishes an efficient optimization framework for SFL under resource-constrained networks. Our framework jointly optimizes model splitting and resource allocation to minimize training cost, which is defined as the weighted sum of latency and energy costs. We first study the model splitting problem and develop a polynomial-time algorithm that achieves the global optimum. Then, we extend the approach to the joint model splitting and resource allocation problem. In this case, we formulate it as a two-dimensional master problem and develop an efficient approximation method with a

(1+ε)

-approximation guarantee. Extensive experiments show that the proposed approach provides efficient solutions to strike the optimal energy--latency tradeoff.

Abstract

Split Federated Learning

Explore similar work

Efficient Resource Optimization for Split Federated Learning

Split Federated Learning Architectures for High-Accuracy and Low-Delay Model Training

CoASFL: Coupled Consistency Optimization for Asynchronous Split Federated Learning