Neural Character-based Composition Models for Abuse Detection

TL;DR

This paper introduces a neural character-based model that composes embeddings for unseen words, improving abuse detection accuracy by addressing obfuscation tactics in social media content.

Contribution

It presents a novel character-based composition model for embeddings that enhances abuse detection, especially against obfuscated words, outperforming existing RNN-based methods.

Findings

Significantly improves abuse detection accuracy on Twitter and Wikipedia datasets.

Effectively distinguishes obfuscated words from non-obfuscated or rare words.

Outperforms current state-of-the-art methods in abuse detection.

Abstract

The advent of social media in recent years has fed into some highly undesirable phenomena such as proliferation of offensive language, hate speech, sexist remarks, etc. on the Internet. In light of this, there have been several efforts to automate the detection and moderation of such abusive content. However, deliberate obfuscation of words by users to evade detection poses a serious challenge to the effectiveness of these efforts. The current state of the art approaches to abusive language detection, based on recurrent neural networks, do not explicitly address this problem and resort to a generic OOV (out of vocabulary) embedding for unseen words. However, in using a single embedding for all unseen words we lose the ability to distinguish between obfuscated and non-obfuscated or rare words. In this paper, we address this problem by designing a model that can compose embeddings for unseen words. We experimentally demonstrate that our approach significantly advances the current state of the art in abuse detection on datasets from two different domains, namely Twitter and Wikipedia talk page.