Skip to content
/ CofCED Public

COLING 2022: A Coarse-to-fine Cascaded Evidence-Distillation Neural Network for Explainable Fake News Detection.

License

Notifications You must be signed in to change notification settings

Nicozwy/CofCED

Repository files navigation

CofCED

Wisdom of crowds: CofCED

🚩 The codes and datasets have been uploaded!

A Coarse-to-fine Cascaded Evidence-Distillation Neural Network for Explainable Fake News Detection is accepted by COLING 2022. CofCED is an explainable method proposed by this paper. We present the first study on explainable fake news detection directly utilizing the wisdom of crowds (raw reports), alleviating the dependency on fact-checked reports.

🚩 If possible, could you please star this project. ⭐ ↗️

Codes

Installing requirement packages

conda create -n fact22 python=3.8
source activate fact22
conda install pytorch==1.12.0 torchvision==0.13.0 torchaudio==0.12.0 cudatoolkit=11.3 -c pytorch
pip install transformers pandas==1.1.2 tqdm==4.50.0 nltk==3.5 rouge-score==0.0.4 sklearn
pip install sentence_transformers   # for evaluation
pip install torch>=1.8

Datasets

We constructed two realistic datasets, i.e., RAWFC and LIAR-RAW, consisting of raw reports for each claim.

Please cite this paper as follows (BibTeX):

@inproceedings{yang2022cofced,
  title={A Coarse-to-fine Cascaded Evidence-Distillation Neural Network for Explainable Fake News Detection},
  author={Yang, Zhiwei and Ma, Jing and Chen, Hechang and Lin, Hongzhan and Luo, Ziyang and Chang Yi},
  booktitle={Proceedings of the 29th International Conference on Computational Linguistics (COLING)},
  pages={2608--2621},
  month={oct},
  year={2022},
  url={https://aclanthology.org/2022.coling-1.230},
}

PDF: https://aclanthology.org/2022.coling-1.230.pdf

About

COLING 2022: A Coarse-to-fine Cascaded Evidence-Distillation Neural Network for Explainable Fake News Detection.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages