Why using transpose of adjacent matrix in your GCN module?

[tdingquan]

MTGNN/net.py

Line 121 in 12869c7

x = self.gconv1[i](x, adp)+self.gconv2[i](x, adp.transpose(1,0))

Dear author,

As provided in the above link, I notice that you use the transpose of the adjacent matrix to calculate another GCN result.

Why is this step necessary and what is the difference between using an undirected acyclic graph in a single GCN and using two GCN with a DAG and its transpose?