Attacking Fake News Detectors via Manipulating News Social Engagement

Paper

Implementation for our WWW'23 paper below:

@inproceedings{wang2023attacking,
  title={Attacking Fake News Detectors via Manipulating News Social Engagement},
  author={Wang, Haoran and Dou, Yingtong and Chen, Canyu and Sun, Lichao and Yu, Philip S and Shu, Kai},
  booktitle={Proceedings of the ACM Web Conference 2023},
  year={2023}
}

Setup

To run the code, you need FakeNewsNet dataset. Due to Twitter privacy concerns, we cannot release user Tweet data. Please send email with the title MARL Dataset Request to [email protected] to download the file with user raw Tweet data. You can unzip the dataset file under the root directory of the project.

Run the code

cd MARL

To run the two random baselines:

python attack_base.py

To run MARL:

python attack.py