Counter speech ranking

Abstract

This thesis presents the development of a novel ranking method for intent-conditioned counter speeches. The primary aim of this thesis is to analyze hate speech on online social platforms and identify the most effective strategies to counter it. The research undertakes an exhaustive analysis of the IntentCONANv2 dataset containing 3488 hate speeches and 13952 counterspeeches across four intents per hate speech. The investigation identifies four parameters mainly- Intensity, Figurative speech, Information presence , and Target group for analysis of the hate speeches. These parameters have further fine-grained definitions to comprehensively analyse hate speech. This thesis significantly contributes to the field of hate speech and technology by developing a model that effectively analyses hate speech and further assists in the generation of counterspeech.