AC$^2$P$^2$SL: Adaptive Communication-Computation Pipeline Parallel Split Learning over Edge Networks

2026-06-30

AI Summary

This work addresses the high training latency in conventional split learning over wireless edge networks, where serial execution of communication and computation impedes efficient utilization of distributed data and computational resources. To overcome this limitation, the authors propose an adaptive communication-computation pipelined parallel split learning framework that leverages micro-batching to enable fine-grained pipeline parallelism, effectively overlapping communication and computation. The approach innovatively introduces a cross-microbatch pipelining mechanism that jointly optimizes communication, computation, and memory resources, complemented by an adaptive task reallocation strategy tailored to device heterogeneity and dynamic network conditions. Experimental results demonstrate that the proposed method significantly reduces training latency while preserving data privacy, achieving superior efficiency and robustness in dynamic edge environments.

Abstract

In wireless edge networks, split learning (SL) enables base station (BS) to utilize the distributed data and computing power across user equipments (UEs) to achieve collaborative model training while protecting local data privacy. However, the inherent sequential execution of computation and communication processes in conventional SL usually leads to long training times. To overcome this limitation, this paper proposes an adaptive communication-computation pipeline parallel split learning (AC$^2$P$^2$SL) framework. By conceptualizing the communication and computation processes of UEs and the BS as a unified pipeline, AC$^2$P$^2$SL achieves fine-grained pipeline parallelism across multiple micro-batches. Through this approach, effective overlapping of communication and computation is achieved which results in significant reduction of the overall training latency. Moreover, by considering the system constraints in the communication, computation, and storage dimensions as well as the heterogeneity of UEs, we formulate a joint optimization problem to minimize the training time and propose a corresponding split and pre-allocation algorithm to further enhance the pipeline efficiency. Additionally, accounting for the practical dynamic environments for the UEs, we design an adaptive re-allocation strategy to enhance the system resilience. Extensive experimental results demonstrate the effectiveness and robustness of AC$^2$P$^2$SL in reducing training time while ensuring data privacy preservation.

Problem

• Split Learning
• Training Latency
• Edge Networks
• Communication-Computation Overlap
• Pipeline Parallelism

Innovation

• Split Learning
• Pipeline Parallelism
• Edge Computing
• Adaptive Resource Allocation
• Communication-Computation Overlap