Code for the paper “Multi-Modal Rumor Detection with Scene Graph Attention Networks Integrating External Knowledge and Evidence”.

The framework of the proposed model:

Datasets

• For the original MR2 data, please download it from Google Drive link or Baidu Cloud link.
• For the Pheme and Weibo-2 datasets, please download it from Baidu Cloud link.
• We have also uploaded the preprocessed ELA images and scene graphs to Baidu Cloud. The download link is: https://pan.baidu.com/s/1I3WCdVh67Ci1D-OIMjx0tA?pwd=s9fn
• please refer to this repo for the evidence collection process.

Dependencies

• tqdm==4.63.1
• torch==1.12.1+cu113
• seaborn==0.11.2
• matplotlib==3.3.4
• Pillow==8.4.0
• pytorch_lightning==1.9.0
• transformers==4.18.0
• pycocotools
• torchmetrics==1.4.0.post0
• nltk==3.7
• scikit_learn==0.24.2
• wordcloud==1.8.1
• torchvision==0.13.1+cu113
• requests==2.27.1
• tabulate==0.8.9
• packaging==21.3
• pandas==1.1.5
• dgl_cu111==0.6.1
• Cython==0.29.28
• numpy==1.21.0
• FactualSceneGraph
• sentence_transformers
• SPARQLWrapper==2.0.0
• jieba==0.42.1
• spacy==3.8.2

Run

python main.py --dataset twitter --model SGKE --batch 32

If you find this project helps your research, please kindly consider citing our paper in your publications.

@article{ JSJX20250520001,
author = { 黄学坚 and  马廷淮 and  荣欢 and  王根生 and  廖国琼 and  刘德喜 },
title = {融合外部知识与证据的场景图注意力网络多模态谣言检测},
journal = {计算机学报},
year = {2025},
url={http://kns.cnki.net/kcms/detail/11.1826.tp.20250520.1501.002.html},
pages = {1-23},
issn = {0254-4164},
}

Acknowledgements

Thank you to Xuming Hu (Tsinghua University, Beijing, China), Zhijiang Guo (University of Cambridge, Cambridge, United Kingdom), Junzhe Chen (Tsinghua University, Beijing, China), Lijie Wen (Tsinghua University, Beijing, China), and Philip S Yu (University of Illinois at Chicago, Chicago, IL, USA) for providing the dataset.