I have a classification problem where the inputs are graphs, with no special features in the nodes of the graph.

I tried to use message passing layers like GCN and GIN but they were not able to capture any meaningful patterns.

Is there another type of NN that can capture just the structure of the graph?

  • $\begingroup$ Yes, GCN and GIN can both do that, however, GNNs are usually a bit more picky with regards to setup than CNNs. Without further info on the data that you have it's hard to suggest anything. $\endgroup$
    – Chillston
    Feb 8 at 10:11


You must log in to answer this question.

Browse other questions tagged .