When Computing Power Network Meets Distributed Machine Learning: An Efficient Federated Split Learning Framework
When Computing Power Network Meets Distributed Machine Learning: An Efficient Federated Split Learning Framework
复制标题
DOI：
10.1109/iwqos57198.2023.10188789
复制
发表时间：
2023-05
期刊：
影响因子：
--
通讯作者：
Xinjing Yuan;Lingjun Pu;Lei Jiao;Xiaofei Wang;Mei Yang;Jingdong Xu
中科院分区：
文献类型：
--
作者：
Xinjing Yuan;Lingjun Pu;Lei Jiao;Xiaofei Wang;Mei Yang;Jingdong Xu
在本文中，我们提倡 CPN-FedSL，这是一种基于计算能力网络 (CPN) 的新颖且灵活的联邦分割学习 (FedSL) 框架。我们构建了一个专用模型来捕获基本设置和学习特征（例如训练流程、延迟和收敛）。基于该模型，我们引入了资源使用效率（RUE），这是一种将训练效用与系统成本相结合的新型性能指标，并通过综合考虑客户端准入、模型划分、服务器选择、路由和带宽分配（即混合整数分数规划），制定了最大化 RUE 的多元调度问题。我们设计了 Refinery，这是一种有效的方法，它首先对分数目标和非凸约束进行线性化，然后通过基于贪婪的舍入算法在多次迭代中解决转换后的问题。广泛的评估证实 CPN-FedSL 优于标准和最先进的学习框架（例如 FedAvg 和 SplitFed），此外 Refinery 是轻量级的，并且在各种设置下显着优于其变体和事实上的启发式方法。
In this paper, we advocate CPN-FedSL, a novel and flexible Federated Split Learning (FedSL) framework over Computing Power Network (CPN). We build a dedicated model to capture the basic settings and learning characteristics (e.g., training flow, latency and convergence). Based on this model, we introduce Resource Usage Effectiveness (RUE), a novel performance metric integrating training utility with system cost, and formulate a multivariate scheduling problem that maximizes RUE by comprehensively taking client admission, model partition, server selection, routing and bandwidth allocation into account (i.e., mixed-integer fractional programming). We design Refinery, an efficient approach that first linearizes the fractional objective and non-convex constraints, and then solves the transformed problem via a greedy based rounding algorithm in multiple iterations. Extensive evaluations corroborate that CPN-FedSL is superior to the standard and state-of-the-art learning frameworks (e.g., FedAvg and SplitFed), and besides Refinery is lightweight and significantly outperforms its variants and de facto heuristic methods under a variety of settings.