A review of web infodemic analysis and detection trends across multi-modalities using deep neural network

A review of web infodemic analysis and detection trends across multi-modalities using deep neural network

연구 분야: Cryptography

학회: International Journal of Data Science and Analytics

The proliferation of disinformation and misinformation across diverse digital platforms poses a significant societal challenge. Previous work in this area adequately addresses the false news detection on the online text shared. The technological platforms that have enough study and research carried out on them include the web news, as well as Twitter. This emerging field is gaining grounds for Facebook, Reddit, WhatsApp, YouTube, among other social media apps. Online data are analyzed using different modalities of text, images, videos, and speech with other sources of influence. It ended with the effectiveness of multi-modal integration in the identification of misinformation. To combat this issue, the development of robust and accurate detection techniques is imperative. This review delves into the multifaceted nature of this problem, exploring the intricacies of multi-modal fake news detection. It examines how the integration of text, image, video, and audio modalities can enhance detection accuracy. At present, there is an abundance of surveys consolidating textual fake news detection algorithms. This review primarily deals with multi-modal fake news detection techniques that include images, videos, and their combinations with text. We provide a comprehensive literature survey of eighty articles presenting state-of-the-art detection techniques, thereby identifying research gaps and building a pathway for researchers to further advance this domain.