Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Embedding of head entity, relation and tail entity #9

Open
dr-GitHub-account opened this issue May 21, 2023 · 0 comments
Open

Embedding of head entity, relation and tail entity #9

dr-GitHub-account opened this issue May 21, 2023 · 0 comments

Comments

@dr-GitHub-account
Copy link

The true value function is designed to minimize || ei + rij - ej ||p for the embedding of positive triplet samples, and to maximize || ei + rij - ej ||p for the embedding of negative triplet samples. Thus, true value in loss enforces that || ei + rij - ej ||p of the embeddings of triplets in KGs are as small as possible. True value in loss can be considered as a restriction, making sure that positive triplets samples forms a "triangle" in vector space.

When not considering the restriction, I am wondering is it a nature of MuGNN encoder to calculate ej to be close to ei + rij? In other word, is this "triangle" relationship between ei, rij, ej only introduced by the restriction of loss, or not only introduced by the restriction of loss, but also by the nature of MuGNN encoder as well?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant