A State-independent and Time-evolving Network for Early Rumor Detection in Social Media

Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing, pages 9042–9051, November 16–20, 2020. c⃝2020 Association for Computational Linguistics 9042 A State-independent and Time-evolving Network for Early Rumor Detection in Social Media Rui Xia∗, Kaizhou Xuan∗, and Jianfei Yu† School of Computer Science and Engineering, Nanjing University of Science and Technology, China rxia@njust.edu.cn, kaizhouxuan@gmail.com, jfyu@njust.edu.cn Abstract In this paper, we study automatic rumor de- tection for in social media at the event level where an event consists of a sequence of posts organized according to the posting time. is common that the state of an event is dy-It namically evolving. However, most of the ex- isting methods to this task ignored this prob- lem, and established a global representation based on all the posts in the event’s life cy- cle. Such coarse-grained methods failed to capture the event’s unique features in differ- ent states. To address this limitation, we pro- pose a state-independent and time-evolving Network (STN) for rumor detection based on fine-grained event state detection and segmen- tation. Given an event composed of a sequence of posts, STN first predicts the correspond- ing sequence of states and segments the event into several state-independent sub-events. For each sub-event, STN independently trains an encoder to learn the feature representation for that sub-event and incrementally fuses the rep- resentation of the current sub-event with pre- vious ones for rumor prediction. This frame- work can more accurately learn the representa- tion of an event in the initial stage and enable early rumor detection. Experiments on two benchmark datasets show that STN can signif- icantly improve the rumor detection accuracy in comparison with some strong baseline sys- tems. We also design a new evaluation metric to measure the performance of early rumor de- tection, under which STN shows a higher ad- vantage in comparison. 1 Introduction Rumor is defined as an unverified statement, which may be unintentionally created or deliberately fab- ricated (DiFonzo and Bordia, 2007). False rumors are damaging as they may cause public panic and ∗Equal contribution. † Corresponding author. Figure 1: An illustration of the spread of rumors, which displays the average number of posts on Twitter and Weibo datasets (Ma et al., 2016) over the propaga- tion timeline. social unrest. Social media platforms have been ideal places for spreading rumors. It is important to automatically detect the rumors and debunk them before they are widely spread. In recent years, the rumor detection task has at- tracted continuous attention from many researchers in the NLP community. We denote a statement in social media as an event consisting of a source post and its following posts such as comments or reposts (collectively called posts). Given an event, the rumor detection task is typically defined as a text classification problem (Zubiaga et al., 2018). The former aims to detect whether an event is a rumor or not. In the literature, the typical method was to first obtain a global representation of the event based on all posts in the event’s …

Content truncated to 500 words