DocRed
DocRed copied to clipboard
hongwang600
Question about SentModel
Thank you for your work! I have questions about the SentModel mentioned in your paper.
- Is the SentModel based on BiLstm or Bert?
- I can't find the code about SentModel. Can you point it up?
Thank you for the interest!
The SentModel is based on BiLSTM, and the you can find the code in the branch of sentence-level encoder.
Thank you for the interest!
The SentModel is based on BiLSTM, and the you can find the code in the branch of sentence-level encoder.
Thank you for your reply! I am confused by "Then we locate each entity within a specific sentence and compute its embedding by averaging the word embedding of the entity name". If an entity has more than one mention, how to get its embedding? Does "specific sentence" mean a random sampled sentence?
If there are multiple mentions, you can average these entity embedding of all mentions. The specific sentence means the sentence that the entity located.
Best, Hong
On Sat, Apr 11, 2020 at 7:41 AM zhengnt [email protected] wrote:
Thank you for the interest!
The SentModel is based on BiLSTM, and the you can find the code in the branch of sentence-level encoder.
Thank you for your reply! I am confused by "Then we locate each entity within a specific sentence and compute its embedding by averaging the word embedding of the entity name". If an entity has more than one mention, how to get its embedding? Does "specific sentence" mean a random sampled sentence?
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/hongwang600/DocRed/issues/2#issuecomment-612436310, or unsubscribe https://github.com/notifications/unsubscribe-auth/AKM3IKCAGIUPBIPVGGLZWKLRMB6QRANCNFSM4KZ3ASBA .