Edge-level representation learning

Hello, I am working on a project that tries to obtain edge features. Unfortunately, I find it hard to get references since the majority of works focus on node-level tasks. I’m wondering where can I find good edge-level representation model to start.

1 Like

I have a Node-Edge convolution layer that has worked very well in our context that I’m happy to share with you.

Plus, DGL provide many examples here, you may search by the keyword “link prediction” for those that may be helpful.