Compare to The Knowledge: Graph Neural Fake News Detection with External Knowledge

Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing, pages 754–763 August 1–6, 2021. ©2021 Association for Computational Linguistics 754 Compare to The Knowledge: Graph Neural Fake News Detection with External Knowledge Linmei Hu1, Tianchi Yang1, Luhao Zhang2, Wanjun Zhong3, Duyu Tang4, Chuan Shi1∗, Nan Duan4, Ming Zhou4 1School of Computer Science, Beijing University of Posts and Telecommunications 2Meituan 3Sun Yat-sen University 4Microsoft Research Asia {hulinmei,yangtianchi,shichuan}@bupt.edu.cn zhangluhao@meituan.com zhongwj25@mail2.sysu.edu.cn {dutang,nanduan,mingzhou}@microsoft.com Abstract Nowadays, fake news detection, which aims to verify whether a news document is trusted or fake, has become urgent and important. Most existing methods rely heavily on linguistic and semantic features from the news content, and fail to effectively exploit external knowledge which could help determine whether the news document is trusted. In this paper, we pro- pose a novel end-to-end graph neural model called CompareNet, which compares the news to the knowledge base (KB) through entities for fake news detection. Considering that fake news detection is correlated with topics, we also incorporate topics to enrich the news rep- resentation. Specifically, we first construct a directed heterogeneous document graph for each news incorporating topics and entities. Based on the graph, we develop a heteroge- neous graph attention network for learning the topic-enriched news representation as well as the contextual entity representations that en- code the semantics of the news content. The contextual entity representations are then com- pared to the corresponding KB-based entity representations through a carefully designed entity comparison network, to capture the con- sistency between the news content and KB. Fi- nally, the topic-enriched news representation combining the entity comparison features are fed into a fake news classifier. Experimen- tal results on two benchmark datasets demon- strate that CompareNet significantly outper- forms state-of-the-art methods. 1 Introduction With the rapid development of the Internet, there are increasingly huge opportunities for fake news ∗The work was done while visiting Micorosft Research Asia. production, dissemination and consumption. Fake news are news documents that are intentionally and verifiably false, and could mislead readers (Allcott and Gentzkow, 2017). Fake news can easily mis- guide public opinion, cause the crisis of confidence, and disturb the social order (Vosoughi et al., 2018). It is well known that fake news exerted an influence in the past 2016 US presidential elections (Allcott and Gentzkow, 2017). Thus, it is very important to develop effective methods for early fake news detection based on the textual content of the news document. Some existing fake news detection methods rely heavily on various hand-crafted linguistic and se- mantic features for differentiating between news documents (Conroy et al., 2015; Rubin et al., 2016; Rashkin et al., 2017; Khurana and Intelligentie, 2017; Shu et al., 2020). To avoid feature engi- neering, deep neural models such as Bi-LSTM and convolutional neural networks (CNN) have been employed (Oshikawa et al., 2020; Wang, 2017; Rodr´ıguez and Iglesias, 2019). However, they fail to consider the sentence interactions in the docu- ment. Vaibhav et al. showed …

Content truncated to 500 words