Hu Bo, Mao Zhendong, Zhang Yongdong
School of Information Science and Technology, University of Science and Technology of China, Hefei 230022, China.
Fundam Res. 2024 Feb 22;5(1):332-346. doi: 10.1016/j.fmre.2024.01.017. eCollection 2025 Jan.
With the rapid development and popularization of Internet technology, the propagation and diffusion of information become much easier and faster. While making life more convenient, the Internet also promotes the wide spread of fake news, which will have a great negative impact on countries, societies, and individuals. Therefore, a lot of research efforts have been made to combat fake news. Fake news detection is typically a classification problem aiming at verifying the veracity of news contents, which may include texts, images and videos. This article provides a comprehensive survey of fake news detection. We first summarize three intrinsic characteristics of fake news by analyzing its entire diffusion process, namely intentional creation, heteromorphic transmission, and controversial reception. The first refers to why users publish fake news, the second denotes how fake news propagates and distributes, and the last means what viewpoints different users may hold for fake news. We then discuss existing fake news detection approaches according to these characteristics. Thus, this review will enable readers to better understand this field from a new perspective. We finally discuss the trends of technological advances in this field and also outline some potential directions for future research.
随着互联网技术的迅速发展和普及，信息的传播与扩散变得更加容易和快捷。互联网在使生活更便捷的同时，也助长了虚假新闻的广泛传播，这将对国家、社会和个人产生极大的负面影响。因此，人们为打击虚假新闻付出了诸多研究努力。虚假新闻检测通常是一个分类问题，旨在核实新闻内容的真实性，这些内容可能包括文本、图像和视频。本文对虚假新闻检测进行了全面综述。我们首先通过分析虚假新闻的整个传播过程，总结出其三个内在特征，即故意编造、异形传播和争议性接受。第一个特征指用户发布虚假新闻的原因，第二个特征表示虚假新闻的传播和分布方式，最后一个特征意味着不同用户对虚假新闻可能持有的观点。然后，我们根据这些特征讨论现有的虚假新闻检测方法。因此，这篇综述将使读者能够从一个新的视角更好地理解这个领域。我们最后讨论了该领域技术进步的趋势，并概述了一些未来研究的潜在方向。
