Skip to content

Latest commit

 

History

History
executable file
·
80 lines (64 loc) · 2.63 KB

README.md

File metadata and controls

executable file
·
80 lines (64 loc) · 2.63 KB
Raw

PRiSM

Source code for our Findings of IJCNLP-AACL 2023 paper PRiSM: Enhancing Low-Resource Document-Level Relation Extraction with Relation-Aware Score Calibration.

Requirements

  • Python (tested on 3.8.16)
  • CUDA (tested on 11.7)
  • PyTorch (tested on 1.13.1)
  • Transformers (tested on 4.30.0)
  • numpy (tested on 1.22.4)
  • wandb
  • tqdm

Datasets

Datasets can be downloaded here: DocRED, Re-DocRED, DWIE. The expected structure of files is:

[working directory]
 |-- data
 |    |-- DocRED
 |    |    |-- train_distant.json        
 |    |    |-- train.json
 |    |    |-- dev.json
 |    |    |-- test.json
 |    |    |-- label_map.json
 |    |    |-- rel_info.json
 |    |    |-- rel_desc.json
 |    |-- Re-DocRED
 |    |    |-- train_distant.json        
 |    |    |-- train.json
 |    |    |-- dev.json
 |    |    |-- test.json
 |    |    |-- label_map.json
 |    |    |-- rel_info.json
 |    |    |-- rel_desc.json
 |    |-- DWIE
 |    |    |-- train/
 |    |    |-- dev/
 |    |    |-- test/
 |    |    |-- label_map.json
 |    |    |-- rel_desc.json

Training and Evaluation

Train the model with the following command:

>> bash scripts/train.sh

Evaluate the model with the following command:

>> bash scripts/evaluate.sh

Citation

If you make use of this code in your work, please kindly cite our paper:

@inproceedings{choi2023prism,
               author={Choi, Minseok and Lim, Hyesu and Choo, Jaegul},
               title={P{R}i{S}{M}: Enhancing Low-Resource Document-Level Relation Extraction with Relation-Aware Score Calibration},
               booktitle={Proceedings of the 13th International Joint Conference on Natural Language Processing and the 3rd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics},
               month={November},
               year={2023},
               address={Nusa Dua, Bali},
               publisher={Association for Computational Linguistics},
               pages={39--47},
               url={https://aclanthology.org/2023.findings-ijcnlp.4}
}

Acknowledgements

This work was supported by Institute of Information & communications Technology Planning & Evaluation (IITP) grant funded by the Korea government (MSIT) (No.2019-0-00075, Artificial Intelligence Graduate School Program (KAIST)), the National Supercomputing Center with supercomputing resources including technical support (KSC-2022-CRE-0312), and Samsung Electronics Co., Ltd.