/checkwhy

CHECKWHY

Glad to be awarded as an outstanding paper award and area chair award. The datasets for our ACL 2024 paper: "CHECKWHY: Causal Fact Verification via Argument Structure." is included in: https://drive.google.com/drive/folders/1jHk1k4UpoyuuvEChRbcrVSZj-QPXM0N-?usp=drive_link

Note: we provide the train.json and dev.json in this repo. If you need the test.json, please email me: jiashengsi@qlu.edu.cn

Datasets

Datasets:

  • 2 Classes:

    • train.json
    • dev.json
    • test.json

  • 3 Classes (Including 'Not Enough Info'):

    • train_nei.json
    • dev_nei.json
    • test_nei.json

Note:

  • checkwhy-id: The ID of the dataset.
  • evidences:
    • evidence_id: ID of the evidence.
    • useful: Whether this evidence is useful for verification.
    • leaf: Whether this evidence is a leaf node in the argument structure.
  • argument_structure: The output for Task 3.
  • code_out: The output for Task 4.

Citation

Please cite our paper if you use CHECKWHY or the datasets we provided in your work:

@inproceedings{si-etal-2024-checkwhy,
    title = "{CHECKWHY}: Causal Fact Verification via Argument Structure",
    author = "Si, Jiasheng  and
      Zhao, Yibo  and
      Zhu, Yingjie  and
      Zhu, Haiyang  and
      Lu, Wenpeng  and
      Zhou, Deyu",
    booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics",
    month = aug,
    year = "2024",
    address = "Bangkok, Thailand",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2024.acl-long.835",
    doi = "10.18653/v1/2024.acl-long.835",
    pages = "15636--15659",
}