Cross-Modal Knowledge Distillation in Multi-Modal Fake News Detection

This is a PyTorch implementation of our paper "Cross-Modal Knowledge Distillation in Multi-Modal Fake News Detection" (ICASSP 2022).

Datasets:

Weibo

Due to copyright issues, the dataset has to be taken from the creators of the dataset. Please contact Peng QI at [email protected].

Twitter

https://github.com/MKLab-ITI/image-verification-corpus/tree/master

Politi & Gossip

https://drive.google.com/file/d/1Cdil6K5MSU4ebOVrU-4_4DteW8ANNoa2/view?usp=sharing

Citation:

@inproceedings{wei2022cross,
  title={Cross-modal knowledge distillation in multi-modal fake news detection},
  author={Wei, Zimian and Pan, Hengyue and Qiao, Linbo and Niu, Xin and Dong, Peijie and Li, Dongsheng},
  booktitle={ICASSP 2022-2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
  pages={4733--4737},
  year={2022},
  organization={IEEE}
}

Acknowledgment

Our implementation is mainly based on follows. Thanks for their authors.

• https://github.com/HobbitLong/RepDistiller
• https://github.com/YingZhangDUT/Deep-Mutual-Learning
• https://github.com/Cadene/block.bootstrap.pytorch