28 February 2024 in IEEE Internet of Things Journal
Many wearable Internet of Medical Things (IoMT) devices have limited computing power and small storage space. Additionally, the healthcare data sensed by a single IoMT device is not enough to train a sophisticated deep learning model. To address these challenges, we propose a federated split learning (FedSL) framework that allows for collaborative healthcare analytics on multiple IoMT devices with limited resources. Compared to centralized learning, FedSL can protect user privacy by not sending raw data over wireless networks. Furthermore, FedSL offers more flexibility than other federated learning methods. It enables even low-end IoMT devices to participate in model training and result inference. Experimental results show that our FedSL performs well on medical imaging tasks with different data distributions.
20 September 2024 in IEEE Transactions on Consumer Electronics
The rapid advancement of the Internet of Things (IoT) and artificial intelligence has significantly increased the number of consumer electronics devices. Traditional federated learning methods typically require all these devices to collaboratively train deep neural networks (DNNs). However, this approach poses a significant challenge due to the substantial communication overhead incurred by uploading and downloading the large DNN models between the base station and resource-constrained IoT devices. To overcome this challenge, we propose a federated split learning (FedSL) framework, along with two variants (FedSL-I and FedSL-II). Furthermore, considering the limitations posed by battery capacity, heterogeneous channels and computing capabilities across all consumer devices, we jointly optimize power allocation, device scheduling and split layer selection to minimize the per-round weighted sum of training delay and energy consumption. We design alternating optimization algorithms to attain a sub-optimal solution with reduced computational complexity. Simulation results demonstrate that our proposed FedSL enables efficient collaborative model training among IoT devices, leading to a significant reduction in communication overhead, and ultimately enhancing the overall performance.
6 June 2025 in IEEE Internet of Things Magazine
Large artificial intelligence (AI) models exhibit remarkable capabilities in various application scenarios, but deploying them at the network edge poses significant challenges due to issues such as data privacy, computational resources, and latency. In this paper, we explore federated fine-tuning and collaborative reasoning techniques to facilitate the implementation of large AI models in resource-constrained wireless networks. Firstly, promising applications of large AI models within specific domains are discussed. Subsequently, federated fine-tuning methods are proposed to adapt large AI models to specific tasks or environments at the network edge, effectively addressing the challenges associated with communication overhead and enhancing communication efficiency. These methodologies follow clustered, hierarchical, and asynchronous paradigms to effectively tackle privacy issues and eliminate data silos. Furthermore, to enhance operational efficiency and reduce latency, efficient frameworks for model collaborative reasoning are developed, which include decentralized horizontal collaboration, cloud-edge-end vertical collaboration, and multi-access collaboration. Next, simulation results demonstrate the effectiveness of our proposed methods in reducing the fine-tuning loss of large AI models across various downstream tasks. Finally, several open challenges and research opportunities are outlined.
6 March 2025 in IEEE Transactions on Wireless Communications
The rapid evolution of edge computing and artificial intelligence (AI) paves the way for pervasive intelligence in the next-generation network. As a hybrid training paradigm, federated split learning (FSL) leverages data and model parallelism to enhance training efficiency. However, existing FSL encounters unacceptable waiting latency due to device heterogeneity and synchronous model aggregation. To address this issue, we propose a semi-asynchronous FSL (SAFSL) framework that enables personalized model splitting and aperiodic model aggregation. We derive the convergence upper bound by considering factors such as the number of devices, training iterations, and data heterogeneity. To minimize the long-term average training latency while maintaining high energy efficiency in resource-constrained wireless networks, we formulate a stochastic mixed-integer nonlinear programming problem. By decomposing it into multiple sub-problems in each round, we propose a Lyapunov-based alternating optimization algorithm to solve it in an online manner. Numerical results demonstrate that our SAFSL achieves faster convergence with reduced communication overhead while maintaining high prediction performance under non-independent and identically distributed data, outperforming state-of-the-art benchmarks. Moreover, our algorithm achieves a low training latency, highlighting its superior performance and effectiveness.
23 March 2026 in IEEE Transactions on Cognitive Communications and Networking
The growing architectural complexity of artificial intelligence models poses significant obstacles to the realization of pervasive intelligence on resource-constrained Internet of Things devices. Federated split learning (FSL) has emerged as a promising solution that enhances training efficiency by combining the benefits of model splitting and parallel computing. However, existing FSL encounters degraded model performance and training inefficiency due to data and device heterogeneity in wireless networks. To address these issues, we propose a gradient-compensation FSL (GCFSL) scheme that leverages complementary gradient information across devices. We derive the convergence upper bound by considering the impact of data heterogeneity and split layer selection of deep neural networks. To minimize the long-term training latency while guaranteeing learning performance in resource-constrained wireless networks, we formulate a two-timescale optimization problem. By decomposing the problem into short-term and long-term subproblems, we design iterative optimization algorithms to solve them with low computational complexities. Numerical results demonstrate that our GCFSL scheme achieves superior model accuracy with low communication overhead while attaining fast convergence under both independent and identically distributed (IID) and non-IID settings, outperforming state-of-the-art benchmarks. Moreover, the proposed algorithm can achieve a low training latency compared with its non-optimized counterpart.
29 January 2026 in IEEE Transactions on Wireless Communications
Federated split learning (FSL) has rapidly emerged as a promising paradigm for enabling ubiquitous intelligence in next-generation networks. However, current FSL approaches incur significant communication overhead and diminished training efficiency due to the frequent transmission of high-dimensional smashed data and gradients between devices and the base station. To address these issues, we propose a low-rank approximation (LoRA)-based FSL scheme, referred to as low-rank FSL. We analyze the convergence performance of low-rank FSL by considering the influence of LoRA rank on non-convex loss functions. To minimize a weighted sum of overall training latency and energy consumption in resource-constrained wireless networks, we formulate a long-term optimization problem by jointly optimizing computing frequency, power allocation, decoding order, LoRA rank, and split layer selection. An iterative optimization algorithm is then developed to solve this problem with low computational complexity. Numerical results demonstrate that our low-rank FSL reduces communication overhead by at least 300% while maintaining high learning performance. Moreover, our optimization algorithm achieves a low weighted cost in terms of training latency and energy consumption.