Detecting Online Hate Speech Using Both Supervised and Weakly-Supervised Approaches

Abstract

In the wake of a polarizing election, social media is laden with hateful content. Context accompanying a hate speech text is useful for identifying hate speech, which however has been largely overlooked in existing datasets and hate speech detection models. We provide an annotated corpus of hate speech with context information well kept. Then we propose two types of supervised hate speech detection models that incorporate context information, a logistic regression model with context features and a neural network model with learning components for context. Further, to address various limitations of supervised hate speech classification methods including corpus bias and huge cost of annotation, we propose a weakly supervised two-path bootstrapping approach for online hate speech detection by leveraging large-scale unlabeled data. This system significantly outperforms hate speech detection systems that are trained in a supervised manner using manually annotated data. Applying this model on a large quantity of tweets collected before, after, and on election day reveals motivations and patterns of inflammatory language.