社交媒体平台谣言的早期自动检测

Early Detection of Rumors in Social Media

在社交媒体服务迅速发展与普及的今天,谣言传播以前所未有的迅猛之势对人类社会产生着巨大的影响。同时,人工智能技术的异军突起,也为社交媒体平台的谣言自动检测提供了可能。谣言检测现有方法通常是,通过学习某条社交媒体信息的所有转发或评论的语义表示,来预测该条社交媒体信息是否为谣言。然而,是否能在谣言引起严重的社会影响之前尽早有效做出判断(谣言早期检测)至关重要,这一问题在以往的研究中尚未得到很好的解决。本文总结了现有社交媒体平台谣言自动检测的主要技术路线,并探讨了进行谣言早期检测的可能性。

With the rapid development and popularization of social media services,rumors are spreading with unprecedented rapidity and have a tremendous impact on human society.Meanwhile,the development of artificial intelligence technologies provides a promising approach for social media platform to automatically detect rumors.Most existing methods of rumor detection typically predict based on all comments when users forward the message.However,it is very important to make an early and effective judgment(i.e.early detection of rumors)before rumors cause serious social impact.This article summarizes existing methods for rumor detection,and explores the possibility of early rumor detection.