search

snap-stanford / neural-subgraph-learning-GNN

329 stars 64 forks source link

Why only use 'add edge' in negative sampling? #24

Open XieWenJhin opened 2 years ago

XieWenJhin commented 2 years ago

https://github.com/snap-stanford/neural-subgraph-learning-GNN/blob/4d0 74cbc0fa9d81defef746302e62b1b9a97791d/common/data.py#L153 image

I guess the else condition won't be satisfied all the time, so could I konw the reason？ Thanks a lot.