Preventing User Radicalization by Automatically Detecting Misinformation Videos

Date
2022-12
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract

Online misinformation has become a very large problem with the rise in information sharing capabilities of social media. The rise in such misinformation has in turn influenced and lead to many cases of radicalization which has resulted in many unfortunate situations. Detecting misinformation is computationally expensive and usually by the time it is found and removed, the content has spread to many places. Additionally, online echo-chambers often exacerbate the spread of misinformation and without any oversight can lead to users getting radicalized. In this thesis, I will look at leading research papers on how to detect misinformation using multiple methods and how to detect potential echo-chambers. Using that knowledge, I will propose a video ranking algorithm that will attempt to prevent user radicalization by automatically detecting potential misinformation content.


Embargo status: Restricted until 01/2174. To request the author grant access, click on the PDF link to the left.

Description
Keywords
Machine Learning, Misinformation, Deep-Learning, Bidirectional Encoder Representations from Transformers (BERT), Radicalization
Citation