通过优化原型分布提升联邦学习分类性能
Heterogeneous Federated Learning with Prototype Alignment and Upscaling
- 基于球面优化和放大原型，增强类别间分离度
- 在多个基准数据集上优于现有异构联邦学习方法
- 通信效率高，适合资源受限设备使用
数据分布与模型架构的异质性仍是联邦学习（FL）中的主要挑战。近期提出的异构联邦学习（HtFL）方法中，基于原型的联邦学习（PBFL）因其仅共享最后一层前的每类均值激活而具有实用性。然而，此类方法常因原型分离不足而限制判别能力。本文提出原型归一化（ProtoNorm），包含两个关键组件：原型对齐（PA）与原型扩增（PU）。PA借鉴经典物理学中的汤姆孙问题，在单位球面上优化全局原型配置以最大化角度分离；随后PU通过增大原型幅值，增强欧氏空间中的分离效果。在多个基准数据集上的广泛评估表明，该方法显著改善了原型分离，持续优于现有HtFL方法。由于ProtoNorm继承了PBFL的通信效率，且PA操作在服务器端完成，因此特别适用于资源受限环境。
原文摘要 · Abstract (English)
Heterogeneity in data distributions and model architectures remains a significant challenge in federated learning (FL). Various heterogeneous FL (HtFL) approaches have recently been proposed to address this challenge. Among them, prototype-based FL (PBFL) has emerged as a practical framework that only shares per-class mean activations from the penultimate layer. However, PBFL approaches often suffer from suboptimal prototype separation, limiting their discriminative power. We propose Prototype Normalization (ProtoNorm), a novel PBFL framework that addresses this limitation through two key components: Prototype Alignment (PA) and Prototype Upscaling (PU). The PA method draws inspiration from the Thomson problem in classical physics, optimizing global prototype configurations on a unit sphere to maximize angular separation; subsequently, the PU method increases prototype magnitudes to enhance separation in Euclidean space. Extensive evaluations on benchmark datasets show that our approach better separates prototypes and thus consistently outperforms existing HtFL approaches. Notably, since ProtoNorm inherits the communication efficiency of PBFL and the PA is performed server-side, it is particularly suitable for resource-constrained environments.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。