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

TL;DR

This paper introduces ASFL, an adaptive split federated learning framework that optimizes model splitting and resource allocation over wireless networks, significantly improving convergence speed and reducing delay and energy use.

Contribution

The paper presents a novel adaptive framework with an optimization algorithm that jointly manages model splitting and resource allocation for efficient federated learning.

Findings

Converges faster than baseline schemes.

Reduces total delay by up to 75%.

Reduces energy consumption by up to 80%.

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.