Social Hatred: Efficient Multimodal Detection of Hatemongers (arXiv)

In order to clean up digital discourse and comprehend hatred as a social phenomena, automatic identification of hate speech online is essential. The researchers contend that user-level analysis is just as significant as previous work, which focuses on nasty statements made by individuals. To identify hate-mongers, we suggest a multimodal strategy that takes into account network data, user activity, and texts. Tested on Twitter, Gab, and Parler datasets, the suggested approach uses social context to beat earlier algorithms. It facilitates interventions on a variety of platforms and improves the identification of subtle indications like coded language.

https://www.arxiv.org/abs/2506.19603