Linguistic Methods for the Detection of Implicit Abuse

Recent years have seen a massive rise in abusive content on the web. Automatic classification methods are sought to assist operators of online platforms in finding such content. Since much abusive content is expressed in the form of written comments, natural language processing is a key technology in tackling this issue. The effectiveness of state-of-the-art methods for abusive language detection is limited. While explicit abuse, that is, abuse conveyed by unambiguously abusive words, such as swearwords, can now be fairly reliably detected, we currently have no indication that classifiers can also detect implicit forms of abuse. In this project, we want to address the classification of a set of subtypes of implicit abuse to fill this important gap in current research. In order to do so, we will create datasets that suitably represent these forms of abuse and develop classification methods that can also be evaluated on those datasets. Linguistic features will play a key role for classification. They are more important for detecting implicitly abusive language than for detecting explicitly abusive language.