Abstract
The way people consuming news changed from traditional news papers to social media like Weibo and Twitter with the development of information technology, which facilitates the propagation of fake news. Several studies have made great contributions to fake news detection with multimodal contents, but the relationship between news contents and user comments is not fully exploited. In this work, we propose a fake news detection method named TICCA (Text, Images, Comments Co-Attention) using not only contents including news text and news images but also user comments, the news images features and news text features are fused by text-image co-attention, the correlation between news text and user comments is extracted by text-comment co-attention, the correlation between news images and user comments is extracted by comment-image co-attention. The experimental results show the effectiveness of our method.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
