Detecting White Supremacist Hate Speech using Domain Specific Word Embedding with Deep Learning and BERT

White supremacists embrace a radical ideology that considers white people\nsuperior to people of other races. The critical influence of these groups is no\nlonger limited to social media; they also have a significant effect on society\nin many ways by promoting racial hatred and violence. White supremacist hate\nspeech is one of the most recently observed harmful content on social\nmedia.Traditional channels of reporting hate speech have proved inadequate due\nto the tremendous explosion of information, and therefore, it is necessary to\nfind an automatic way to detect such speech in a timely manner. This research\ninvestigates the viability of automatically detecting white supremacist hate\nspeech on Twitter by using deep learning and natural language processing\ntechniques. Through our experiments, we used two approaches, the first approach\nis by using domain-specific embeddings which are extracted from white\nsupremacist corpus in order to catch the meaning of this white supremacist\nslang with bidirectional Long Short-Term Memory (LSTM) deep learning model,\nthis approach reached a 0.74890 F1-score. The second approach is by using the\none of the most recent language model which is BERT, BERT model provides the\nstate of the art of most NLP tasks. It reached to a 0.79605 F1-score. Both\napproaches are tested on a balanced dataset given that our experiments were\nbased on textual data only. The dataset was combined from dataset created from\nTwitter and a Stormfront dataset compiled from that white supremacist forum.\n