Hi, this is the code of our paper "SANTA: Separate Strategies for Inaccurate and Incomplete Annotation Noise in Distantly-Supervised Named Entity Recognition" accepted by ACL 2023 Findings. Our paper is available here.
Accepted by ACL 2023 Findings. 2023.05
Code released at Github. 2023.08
Download different pretrained LMs into resource/
Use environments.yaml to get the right environments
For EC:
sh train_EC.sh
For Webpage:
sh train_NEWS.sh
For OntoNotes5.0:
sh train_onto.sh
For BC5CDR:
sh train_bc5cdr.sh
We got our results in single A40.