How to interpret the global max pooling operation in graph neural networks?


I’m trying to use pytorch geometric for building graph convolutional networks. And I’m trying to interpret the result of the max pooling operation, which is described in this link:

If I understood it correctly, the result is a feature vector r_i (which the dimensionality of the features of the nodes in the graph), which represents the maximum, considering the feature vectors of all nodes in the graph.

But for me, it is obscure how the feature vectors are compared in order to find the vector that represents the maximum.

Can you help me?

Best regards.

Sounds to me you are getting a vector of size N (where N = number of features) and this vector contains the maximum value of the features among all the samples (i.e images).