Sara Tonelli will give a talk at the Dept. of Computer Science, University of Turin, entitled: “What is hate speech? Insights from cyberbullying detection and Islamophobic discourse monitoring“
Social media messages are often written to attack specific groups of users based on their religion, ethnicity or social status, and they can be particularly threatening to vulnerable users such as teenagers. Several key issues are still open and need further investigation in order to better understand and deal with the massive rise of hateful, abusive and offensive messages. These concern not only the automatic identification of such messages, which would allow social media platforms to automatically hide them, but also the definition of what is offensive taking into account users characteristics and social network-specific features. In this talk, I will first introduce the work carried out within two projects on hate speech detection, one focused on cyberbullying and the other on islamophobia. I will detail how we leveraged guidelines and data to analyse abusive language on WhatsApp and Twitter, and I will describe the hate speech classifier we implemented for Germeval and Evalita shared tasks.
University of Turin, Dept. of Computer Science