Federated Graph Neural Networks: Overview, Techniques and Challenges
来源:arXiv 2022
1. Abstract
In this paper, we bridge this gap by offering a comprehensive survey of this emerging field. We propose a unique 3-tiered taxonomy of the FedGNNs literature to provide a clear view into how GNNs work in the context of Federated Learning (FL).
本文提出了一种独特的 FedGNNs 文献的三层分类法,提供了一个关于GNN 在联邦学习(FL)环境中如何工作的清晰视角。
2.1. Terminology (专业术语)
- GNN
- adjacency matrix: \(A \in \mathbb{R}^{N \times N}\)
- 节点特征 node features: \(\mathbf{X} \in \mathbb{R}^{N \times f}\)
- FL
- clients: data owners with sensitive local data
- server: 协调 clients
值得注意的是 GNN 和 FL 中都有 Aggregation 这个概念
GNN Aggregation
- 给定一个节点,通过聚合其邻居节点的信息来更新它的嵌入, Aggregation 操作可以是 mean, weighted average, or max/min pooling methods
FL Aggregation
- 服务器(用某种算法,eg. FedAvg)根据数据拥有方的上传的本地模型参数去聚合更新全局模型的参数
2.2 The Proposed 3-Tiered FedGNN Taxonomy
上图就是此篇论文提出的三层分类法
标签:FL,Aggregation,第一次,Federated,汇报,times,GNN,节点 From: https://www.cnblogs.com/Summerworm/p/16830492.html