Hate speech detection using static BERT embeddings

TL;DR

This paper investigates hate speech detection on social media using static BERT embeddings, demonstrating improved specificity over other embedding methods and fine-tuned BERT, with a focus on reducing false positives.

Contribution

It introduces the use of static BERT embeddings for hate speech detection and compares its performance with other embedding techniques and fine-tuned BERT.

Findings

Static BERT embeddings outperform other embeddings in hate speech detection.

Neural networks with static BERT embeddings show higher specificity.

Using static BERT embeddings reduces false positives compared to fine-tuned BERT.

Abstract

With increasing popularity of social media platforms hate speech is emerging as a major concern, where it expresses abusive speech that targets specific group characteristics, such as gender, religion or ethnicity to spread violence. Earlier people use to verbally deliver hate speeches but now with the expansion of technology, some people are deliberately using social media platforms to spread hate by posting, sharing, commenting, etc. Whether it is Christchurch mosque shootings or hate crimes against Asians in west, it has been observed that the convicts are very much influenced from hate text present online. Even though AI systems are in place to flag such text but one of the key challenges is to reduce the false positive rate (marking non hate as hate), so that these systems can detect hate speech without undermining the freedom of expression. In this paper, we use ETHOS hate speech detection dataset and analyze the performance of hate speech detection classifier by replacing or integrating the word embeddings (fastText (FT), GloVe (GV) or FT + GV) with static BERT embeddings (BE). With the extensive experimental trails it is observed that the neural network performed better with static BE compared to using FT, GV or FT + GV as word embeddings. In comparison to fine-tuned BERT, one metric that significantly improved is specificity.