Graph Neural Network: A First Glance

tangwing

于 2019-09-06 17:44:03 发布

阅读量227

点赞数

CC 4.0 BY-SA版权

分类专栏： # Data-Driven Decision Making 文章标签： GNN

本文链接：https://blog.youkuaiyun.com/tangwing/article/details/100582584

22 篇文章

订阅专栏

本文深入探讨了图神经网络(GNN)的基本概念及其在图数据上的应用，对比了传统图卷积方法与GNN的区别，解析了GNN的不足及改进方案GGNN，强调了边权重的学习与更新在图表示学习中的重要性。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

@[TOC]GNN

Resources

To make $f$ a Contraction Map: Penalize Jacobian Matrix of $f$ over $H$ . I.e. Bound its derivative.
GNN: stop when converged.
GNN drawbacks
- Edges serve only as connections not learned
- Not suitable for learning Graph Representation: all nodes share info with each other.
GGNN: replace convergent $f$ with a Gated Unit like in RNN. Use BPTT instead of AP and can output before convergence. Edges now have weights that can be updated.