search

dmlc / dgl

Python package built to ease deep learning on graph, on top of existing DL frameworks.
http://dgl.ai
Apache License 2.0
13.44k stars 3.01k forks source link

how to construct a negative heterogeneous graph example? #7172

Open wang1028-yu opened 7 months ago

wang1028-yu commented 7 months ago

Hello everyone, I am trying to using rgcn form dgl to get an embedding of each node of a heterogeneous graph with features, and I want to construct a negative heterogeneous graph to compute the loss by using positive graph embedding subtract negative graph embedding. I am wondering how to construct a negative graph. Considering each edge in heterogeneous graph as a triple (src, relation, dst), here are three solutions I have thought.

  1. In each relation, find a src and change the dst.
  2. find an src and an dst, change the relation
  3. both 1 and 2,
github-actions[bot] commented 6 months ago

This issue has been automatically marked as stale due to lack of activity. It will be closed if no further activity occurs. Thank you