Paper information:

Research on multi-modal fake news detection technology based on deep learning

The VAEMTL folder represents the models mentioned in the paper where process_data_twitter.py, process_data_weibo.py represents the preprocessing file of the data MBPAM+Decoder_twitter.py, MBPAM+Decoder_weibo.py represent the content of the fifth chapter in the paper, including three innovative model improvements:

1. Multi-modal fake news detection algorithm based on dual branch adversarial network
2. Multimodal False News Detection Algorithm Based on Combined Fusion Mechanism
3. Multimodal Fake News Detection Algorithm Based on Variational Autoencoder for Multi-task Learning pytorch_compact_bilinear_pooling.py represents the implementation process of bilinear pooling The crawled_data folder represents some news data crawled templates+app.py+mbpam_decoder_predict.py represents the content of Chapter 6 of the paper, and can build a simple interface.

requirements In the requirements.txt file