search

THUDM / GATNE

Source code and dataset for KDD 2019 paper "Representation Learning for Attributed Multiplex Heterogeneous Network"
MIT License
525 stars 141 forks source link

Question on vertex embedding #46

Closed snash4 closed 4 years ago

snash4 commented 4 years ago

Hi, I find the paper very interesting and to read through. However, I have a question regarding the final learned embedding. In Algorithm 1, line 6, the model learns an embedding of a vertex specific to the relation, which is also evident through eq. 6 and 13. This means that a vertex has different embedding for each view. I could not find or understand in experimental section which embedding are you using for link prediction, OR are you combining these different embedding, if yes, then how?

Thanks.

cenyk1230 commented 4 years ago

Hi,

Thank you for your attention. For the link prediction task of each view, we use the corresponding node embedding of that view. The final experimental results are averaged among the selected views.