bknyaz / graph_nn Goto Github PK

Hi, many thanks for your code! May I ask the following question regarding laplacian batch?
In github.com/tkipf/gcn, laplacian is normalized by divided by the largest eigen value:

def chebyshev_polynomials(adj, k):
"""Calculate Chebyshev polynomials up to order k. Return a list of sparse matrices (tuple representation)."""
print("Calculating Chebyshev polynomials up to order {}...".format(k))

adj_normalized = normalize_adj(adj)
laplacian = sp.eye(adj.shape[0]) - adj_normalized
largest_eigval, _ = eigsh(laplacian, 1, which='LM')
scaled_laplacian = (2. / largest_eigval[0]) * laplacian - sp.eye(adj.shape[0])

while in the code below:
L = D_hat.view(batch, N, 1) * A_hat * D_hat.view(batch, 1, N)
L is not normalized. May I ask what is the reason for not divided by the eigen value?

Many thanks in advance!

Hi Dear Author,

This looks a very good implementation in using GCN for graph-level classification (may also called inductive). Besides the sample running you shown in jupyter notebook, are the other datasets are also supported in running? If yes, could I ask which part should modify to change the calling of datasets? Thanks!

hi! sir. I am very interested in your code and would like to ask how to load other data,such as Cora data sets. The parameter W in GCN and GCNUNET is not trained in your code

Great work in reproducing Graph U-Net. I have one question about backprop in gpool. Why adding a gate operation enable gpool to backprop gradient? Here's a quoting in Graph U-Net.

Notably, the gate operation makes the projection vector p trainable.

I tried to realize the gpool in tensorflow without the gate operation, but it seems that the gradient could not be backproped.

In file graph_unet.ipynb line 82, function list_to_torch is not defined