What is the accuracy, precision report in paper?

xiangrongzeng / copy_re

Release for acl18 paper "Extracting Relational Facts by an End-to-End Neural Model with Copy Mechanism"

197 stars 36 forks source link

Closed edfall closed 4 years ago

edfall commented 4 years ago

if the triplet entities contain multiple words, your sequence decoding pipeline will never get the right triplet out, so whether the multiple word triplets are all count wrong?
or you just evaluate it based on the inaccruate tagging target?

yuwl798180 commented 4 years ago

agree with you, just look at your code, have the same question.

xiangrongzeng commented 4 years ago

For the triplet entities contain multiple words, this model cannot extract the complete entity. We only consider the last word of the entity (with multiple words).
We mentioned this in our paper section 4.3
How to extract the complete entity during the overlapping relation extraction remains an open chanllenge and there are several papers focuing on this problem.