Rumor Detection on Twitter with Claim-Guided Hierarchical Graph Attention Networks

Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 10035–10047 November 7–11, 2021. c⃝2021 Association for Computational Linguistics 10035 Rumor Detection on Twitter with Claim-Guided Hierarchical Graph Attention Networks Hongzhan Lin1,2, Jing Ma2,∗, Mingfei Cheng1, Zhiwei Yang3, Liangliang Chen1, Guang Chen1,∗ 1Beijing University of Posts and Telecommunications, Beijing, China 2Hong Kong Baptist University, Hong Kong SAR, China 3Jilin University, Changchun, China {linhongzhan,mingfeicheng,outside,chenguang}@bupt.edu.cn majing@comp.hkbu.edu.cn, yangzw18@mails.jlu.edu.cn Abstract Rumors are rampant in the era of social me- dia. Conversation structures provide valuable clues to differentiate between real and fake claims. However, existing rumor detection methods are either limited to the strict relation of user responses or oversimplify the conversa- tion structure. In this study, to substantially re- inforces the interaction of user opinions while alleviating the negative impact imposed by ir- relevant posts, we first represent the conversa- tion thread as an undirected interaction graph. We then present a Claim-guided Hierarchical Graph Attention Network for rumor classifica- tion, which enhances the representation learn- ing for responsive posts considering the entire social contexts and attends over the posts that can semantically infer the target claim. Ex- tensive experiments on three Twitter datasets demonstrate that our rumor detection method achieves much better performance than state- of-the-art methods and exhibits a superior ca- pacity for detecting rumors at early stages. 1 Introduction Rumor is one type of social diseases in the era of social media. The spread of false rumors has a far-reaching destructive impact on both society and individuals (Ma et al., 2019b). For instance, the global COVID-19 pandemic has created fertile soil for the widespread of various rumors, conspiracy theories, hoaxes, and fake news, heavily disrupt- ing people’s peaceful lives and leading to unprece- dented information chaos. A strange, new rumor claiming that “wearing a mask to prevent the spread of COVID-19 is unnecessary because the disease can also be spread via farts" 1 may mislead masses to belittle the importance of those potentially life- saving masks in epidemic prevention. Therefore, it is necessary to develop automatic approaches to facilitate rumor detection, especially amid crises. ∗Corresponding authors. 1https://www.snopes.com/fact-check/ farting-negate-covid19-masks/ Social psychology literature defines a rumor as a story or a statement whose truth value is unverified or intentionally false (DiFonzo and Bordia, 2007). Rumor detection aims to determine the veracity of a given story or statement. For automating rumor detection, previous studies focus on text mining from sequential microblog streams with supervised classifiers based on feature engineering (Castillo et al., 2011; Yang et al., 2012; Kwon et al., 2013; Liu et al., 2015; Ma et al., 2015) and feature learn- ing (Ma et al., 2016; Yu et al., 2017). The inter- actions among users generally show conductive to provide useful clues for debunking rumors. Struc- tured information is generally observed on social media platforms such as Twitter. Structure-based methods (Ma et al., 2017, 2018) are thus proposed to capture the interactive characteristics of rumor diffusion. We discuss briefly two types of state- of-the-art approaches: Transformer-based (Khoo et al., 2020; Ma …

Content truncated to 500 words