Cross-document Event Identity via Dense Annotation

Dataset

This dataset (CDEC-WN) is compiled from English Wikinews articles relating to "Disaster and accidents" category. CDEC-WN can be downloaded here and is released under CC-BY-4.0 license. For details on the dataset collection process, refer to our CoNLL 2021 paper.

Cross-document Annotation Tool

The annotation toolkit used for collecting the CDEC-WN dataset is available at github.com/adithya7/cdec-ann-tool.

Baselines

See baselines for details on the two baselines described in our paper.

Citation

If you find this dataset helpful in your research, consider citing our work,

@inproceedings{pratapa-etal-2021-cross,
    title = "Cross-document Event Identity via Dense Annotation",
    author = "Pratapa, Adithya  and
      Liu, Zhengzhong  and
      Hasegawa, Kimihiro  and
      Li, Linwei  and
      Yamakawa, Yukari  and
      Zhang, Shikun  and
      Mitamura, Teruko",
    booktitle = "Proceedings of the 25th Conference on Computational Natural Language Learning",
    month = nov,
    year = "2021",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2021.conll-1.39",
    pages = "496--517",
}

Issues

For any issues, questions or requests, please create a Github Issue.