solutions for Stanford University course CS224W: Machine Learning with Graphs Fall 2021 colabs
Course materials can be found here CS224W: Machine Learning with Graphs Materials
Lectures can be found here CS224W: Machine Learning with Graphs Video Lectures
我的笔记 ：斯坦福cs224w图机器学习课程lec1-8笔记 https://zhuanlan.zhihu.com/p/479505721
Visualiztion: https://distill.pub/2021/gnn-intro/ .
colabhttps://github.com/hdvvip/CS224W_Winter2021
https://sands.kaust.edu.sa/classes/CS294E/F21/schedule.html
PyG: https://pytorch-geometric.readthedocs.io/en/latest/notes/introduction.html
GNN Sampling Papers:
- Inductive Representation Learning on Large Graphs.
- Cluster-GCN: An Efficient Algorithm for Training Deep and Large Graph Convolutional Networks.
- GraphSAINT: Graph Sampling Based Inductive Learning Method.
There are three general types of prediction tasks on graphs: graph-level, node-level, and edge-level. 预测点, 边的性质. 比如一个分子表示成图, 就可以知道她的味道, 有几个环可能和性质有关. 边的预测, 可以判断图像物体之间的关系.
CRS compress row storage
对row_ptr的补充： [0, 3, 4, 6]中， 0：矩阵第一行数11在values的索引是0. 3: 矩阵第二行的第一个数值19 在values中的索引是3. ４：矩阵第三行的第一个数值23 在values中的索引是4. 6：显然矩阵没有第四行。6表示当矩阵有虚拟的第四行时，6就是这虚拟行的第一个值在values里的索引。因此，6也是values的元素数量。
行数 = length(row_ptr) -1 列数 = max(col_index)