[1-5] Graph neural networks(GNNs)
[6]  Graph convolution
[7]  graph neural message passing
[8-10]  pooling strategies  global
[11-13]  hiraichical pooling
[14] Wasserstein embedding
[15] linear optimal transport framework
[16] manifold learning
[17] preliminary idea of a workshop paper
[18-29]  Spatial Graph Convolution,propagation rule
	[18-20] weighted average
	[21-22] concatenation
	[23-25] learnable attention
	[26-27] other adaptive choices
	[28-29] the first-order vs. the second-order
[30] TOPKPOOL
[31] multilayer perceptron(MLP),SAGPool
[32-33] attention
[35-36] recommender systems
[37-38] computer vision
[39-40] pattern recognition
[41-44] nonlinear dimensionality reduction of Grassmann manifold
[45-46] optimization objectives of Grassmann manifold
[47] Deep Grassmann Networks
[48-49] Grassmann clustering
[50-51] The Lie group theory
[52] projection matrices
[53] the chordal distance and the Procrustes distance
[54] kernelized Grassmann learning
[55] bilinear mapping or the second-order statistics or regional features
[56] more stable definition
[57] Variational Graph Auto-Encoder(VGAE)
[58] k-means
[59] PyTorch Geometric(PYG)
[60] Molhiv (open graph benchmark)
[61] TUDataset benchmark
[62] virtual nodes
[63-64] EDGEPool
[65] PANPool
[66] Cora,Citeseer,Pubmed (a network)
[67] Wiki-CS(a network)
[68] Coauthor-CS(a network)
[69] ADAM