略读文章
1: topological graph neural networks
GNN: oblivious to substructures;
TOGL, a novel layer that incorporates global topological information of a graph using homology. (同源性)
TOGL can be easily integrated into any type of GNN and is strictly more
expressive in terms of the Weisfeiler–Lehman test of isomorphism.
both our method and its static ablated variant perform similarly when classification performance does not appear to rely on topological feature.
2: 略
介绍了一种新型的图神经网络层——拓扑图神经网络(TOGL),该网络层能够利用图的全局拓扑信息进行节点分类任务。TOGL通过同源性来捕捉图的拓扑结构,并可以方便地集成到现有的图神经网络架构中。实验证明,在不依赖拓扑特征的任务上,即使简化版本的方法也表现出良好的性能。
3836

被折叠的 条评论
为什么被折叠?



