参考资料
斯坦福CS224W课程: http://cs224w.stanford.edu
https://snap-stanford.github.io/cs224w-notes/
https://www.bilibili.com/video/av84385561
https://github.com/snap-stanford/cs224w-notes
• 图学习库 PGL:https://github.com/PaddlePaddle/PGL
资料推荐
• 理论:
• 综述
Graph Neural Networks: A Review of Methods and Applications https://arxiv.org/pdf/1812.08434.pdf
A Comprehensive Survey on Graph Neural Networks https://arxiv.org/abs/1901.00596
• 视频 •
理论课: 斯坦福 CS224W
• 动手实践:DeepWalk, node2vec, GCN, GAT, GraphSAGE
• 进阶:
经典论文 • DeepWalk. " DeepWalk: Online Learning of Social Representations " https://arxiv.org/pdf/1403.6652.pdf
• GCN. “ Semi-supervised Classification with Graph Convolutional Networks" https://arxiv.org/pdf/1609.02907.pdf
• GAT. ”Graph Attention Networks" https://arxiv.org/pdf/1710.10903.pdf
• GraphSAGE. " Inductive Representation Learning on Large Graphs" https://arxiv.org/pdf/1706.02216.pdf
• https://github.com/thunlp/GNNPapers
• 前沿研究方向:
• 针对 GNN 而言:图采样技术、邻居聚合、深度图卷积、图预训练
图学习
一般深度学习:难以处理不规则数据。 地图 商品广告
图学习:可以方便地处理不规则数据(图),充分利用图结构信息。
参考内容:https://baidu-pgl.gz.bcebos.com/pgl-course/lesson_1.pdf