A problem about the the orthogonal constraint of TransH

[SunYatong]

Hi, I have read your paper and I am pretty interested in your work.

However, I have a problem with the loss function of TransH (eqn.4), which seems to ignore the orthogonal constraint of r and w_{r} in the original paper, which guarantees r is in its corresponding hyperplane.

I have cloned your code before, which seems to omit the orthogonal constraint, too.

Is this constraint not necessary or I missed its implementation?

Looking forward to your reply. Thanks in advance for considering my issue.