Navigation auf


Department of Computational Linguistics

Categorizing Religious Online Hate Speech


Hate speech, often in relation to religion, is an omnipresent problem on the internet. This project addresses the following Research question:
– How can we create categories of religious hate in a data-driven manner?
– These categories could be semantic (who is hated how?) or syntactic/stylistic (how is the hate expressed?).
You will use the following methods and skills:
– training and using transformer-based supervised text classifiers for detecting hate speech
– using unsupervised methods, such as clustering, for recognizing and classifying types of hate speech


  • Python
  • Transformers