search

JD-AI-Research-Silicon-Valley / SACN

End-to-end Structure-Aware Convolutional Networks for Knowledge Base Completion
MIT License
112 stars 30 forks source link

questions about paper please #7

Open ConanCui opened 5 years ago

ConanCui commented 5 years ago
  1. The degree of nodes are different, so we usually need normalization in GCN (by using degree matrix). In equation 3, 4, 5, I can't find the normalization.
  2. In Table 3, why DistMult performs better than R-GCN on FB15K-237? This is a contradiction with the result in R-GCN paper.
  3. There is an attribute triple (entity, relation, attribute) example (s = Tom, r = people.person.gender, a = male. The paper said that ”Note that each type of attribute corresponds to a node.". I understand this by taking the attributes "male" and "female " as two nodes. However, the paper also said "For instance, in our example, gender is represented by a single node rather than two nodes for “male” and “female”.“ I understand this by taking the relation "people.person.gender" as nodes. I feel confused.
chaoshangcs commented 5 years ago

Hi, thank you so much for your questions.

  1. We didn't use the normalization here. You can try to use it and see if it can improve the performance.
  2. Please check the ConvE paper. We are consistent with this paper. https://arxiv.org/pdf/1707.01476.pdf
  3. You are right. The "male" and "female" are two nodes in the dataset. You can also make these two nodes as one node "gender" if you wanna to decrease the size of attribute nodes. Thanks!