Our project, in collaboration with Dr. Bhimrao Ambedkar University and UnReaL-TecE LLP, aims to understand the construction and evaluation of communal and sexually threatening misogynistic content on social media in India. We have developed a multilingual, multimodal dataset containing 57,363 annotated comments, 1,142 annotated memes, and around 70 hours of annotated audio, which has been collected from various social media platforms. Additionally, we have developed an automatic system that identifies communal and misogynistic aggression in four languages.
Menu
Menu