Hater-O-Genius Aggression Classification using Capsule Networks
This addresses hate speech detection in social media, an important social problem, but is incremental with a small performance gain.
The paper tackles automatic classification of aggressive tweets into three categories (Covertly Aggressive, Overtly Aggressive, and Non-Aggressive) using an ensemble architecture, achieving a 65.2% F1 score on a Facebook test set with a 0.95% improvement over prior state-of-the-art.
Contending hate speech in social media is one of the most challenging social problems of our time. There are various types of anti-social behavior in social media. Foremost of them is aggressive behavior, which is causing many social issues such as affecting the social lives and mental health of social media users. In this paper, we propose an end-to-end ensemble-based architecture to automatically identify and classify aggressive tweets. Tweets are classified into three categories - Covertly Aggressive, Overtly Aggressive, and Non-Aggressive. The proposed architecture is an ensemble of smaller subnetworks that are able to characterize the feature embeddings effectively. We demonstrate qualitatively that each of the smaller subnetworks is able to learn unique features. Our best model is an ensemble of Capsule Networks and results in a 65.2% F1 score on the Facebook test set, which results in a performance gain of 0.95% over the TRAC-2018 winners. The code and the model weights are publicly available at https://github.com/parthpatwa/Hater-O-Genius-Aggression-Classification-using-Capsule-Networks.