I am following your Hypobolic series work.
In this paper, I got lost when I met this concept,
My question is,
1) What does (D^n)^p mean?
2) How do we get that embedding? for example, the 50x2D, do we have to train 50 independent 2D Poincare embedding and then combine them?
Dear Alex,
I am following your Hypobolic series work. In this paper, I got lost when I met this concept,
My question is, 1) What does (D^n)^p mean? 2) How do we get that embedding? for example, the 50x2D, do we have to train 50 independent 2D Poincare embedding and then combine them?