Some questions about the paper

[Yfhu1103]

Hello, thanks for releasing your code. I think you adopt the same pooling method as SAGPool to select nodes to be perserved in a graph, but in figure3, The nodes your method chooses seem to be different from SAGPool. I mean, their distributions are different, your nodes scatter all over the graph while in SAGPool, nodes seem to gather together somewhere in graph. I think your method is better but I can not figure out why.

[matenure]

Thank you for your attention. The pooling method is different from SAGPool, as we expalined in our paper.