Hierarchical CVAE for Fine-Grained Hate Speech Classification

TL;DR

This paper introduces a hierarchical CVAE model for fine-grained hate speech classification, differentiating among 40 hate groups across 13 categories, and demonstrates improved accuracy over traditional models.

Contribution

The paper proposes a novel hierarchical CVAE architecture that leverages hate category information for more precise hate speech classification.

Findings

Hierarchical CVAE outperforms standard discriminative models.

Incorporating hate category information improves classification accuracy.

Model effectively differentiates among 40 hate groups across 13 categories.

Abstract

Existing work on automated hate speech detection typically focuses on binary classification or on differentiating among a small set of categories. In this paper, we propose a novel method on a fine-grained hate speech classification task, which focuses on differentiating among 40 hate groups of 13 different hate group categories. We first explore the Conditional Variational Autoencoder (CVAE) as a discriminative model and then extend it to a hierarchical architecture to utilize the additional hate category information for more accurate prediction. Experimentally, we show that incorporating the hate category information for training can significantly improve the classification performance and our proposed model outperforms commonly-used discriminative models.