After reading the paper, I'm wondering whether the the similarity ranking results are computed from all data pairs in the second task of the experiment section (also the unsupervised training). Cuz I found even the approximation version of GED would take a long time to compute on so many graph pairs.
And if you did calculate the GED for all data pairs, how long has it taken to finish the computation?
Hi @yunshengb, thanks for sharing the code!
After reading the paper, I'm wondering whether the the similarity ranking results are computed from all data pairs in the second task of the experiment section (also the unsupervised training). Cuz I found even the approximation version of GED would take a long time to compute on so many graph pairs.
And if you did calculate the GED for all data pairs, how long has it taken to finish the computation?
Thanks!