Process / pipeline
图神经网络 — GCN / GAT / GraphSAGE
图神经网络（GNN）是一种直接在图结构数据上运行的深度学习架构，它通过迭代式的邻域消息传递，将节点特征与结构信息相结合。三个经典变体——Kipf和Welling于2017年提出的图卷积网络（GCN）、Veličković等人于2018年提出的图注意力网络（GAT）以及GraphSAGE——在聚合邻域信息的方式上有所不同：GCN对整个邻接矩阵应用谱卷积，GAT通过学习到的注意力分数对邻居进行加权，而GraphSAGE则通过采样并聚合局部邻域来归纳学习，从而能够泛化到未见过的节点。
阅读完整方法
仅限会员
登录使用免费账户登录即可阅读本节。
Method map
The neighbourhood of related methods — select a node to explore.
+7 more
来源
- Kipf, T.N. & Welling, M. (2017). Semi-Supervised Classification with Graph Convolutional Networks. International Conference on Learning Representations (ICLR). DOI: 10.48550/arXiv.1609.02907 ↗
- Veličković, P., Cucurull, G., Casanova, A., Romero, A., Liò, P., & Bengio, Y. (2018). Graph Attention Networks. International Conference on Learning Representations (ICLR). DOI: 10.48550/arXiv.1710.10903 ↗
- Hamilton, W.L. (2020). Graph Representation Learning. Morgan & Claypool. DOI: 10.1007/978-3-031-01588-5 ↗
如何引用本页
ScholarGate. (2026, June 1). Graph Neural Network (GCN / GAT / GraphSAGE). ScholarGate. https://scholargate.app/zh/network-analysis/graph-neural-network
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
Compare side by side →