Social media has greatly enabled people to participate in online activities an unprecedented rate. However, this unrestricted access also exacerbates spread of misinformation and fake news online which might cause confusion chaos unless being detected early for its mitigation. Given the rapidly nature of news events and the limited amount of annotated data,-of-the-art systems on fake news detection face challenges due to the lack large numbers of annotated training instances that are hard to come by for detection. In this work, we exploit multiple weak signals from different given by user and content engagements (referred to as weak social), and their complementary utilities to detect fake news. We jointly the limited amount of clean data along with weak signals from social to train deep neural networks in a meta-learning framework to the quality of different weak instances. Experiments on realworld demonstrate that the proposed framework outperforms state-of-the-art for early detection of fake news without using any user engagements prediction time.
No takes yet. Share an insight, caveat, or question.
Shu et al. (2020) studied this question.