Predicting Hateful Discussions on Reddit using Graph Transformer Networks and Communal Context

Liam Hebert, Lukasz Golab, Robin Cohen

We propose a system to predict harmful discussions on social media platforms. Our solution uses contextual deep language models and proposes the novel idea of integrating state-of-…