Cyberbullying detection across social media platforms via platform-aware adversarial encoding

TL;DR

This paper introduces XP-CB, a cross-platform cyberbullying detection framework using Transformers and adversarial learning, which improves generalisability across social media platforms by leveraging unlabelled data.

Contribution

It presents a novel platform-aware adversarial encoding method that enhances Transformer models for cross-platform cyberbullying detection, addressing a key gap in existing research.

Findings

Effective across multiple social media platforms

Improves generalisation with unlabelled data

Works with BERT and RoBERTa models

Abstract

Despite the increasing interest in cyberbullying detection, existing efforts have largely been limited to experiments on a single platform and their generalisability across different social media platforms have received less attention. We propose XP-CB, a novel cross-platform framework based on Transformers and adversarial learning. XP-CB can enhance a Transformer leveraging unlabelled data from the source and target platforms to come up with a common representation while preventing platform-specific training. To validate our proposed framework, we experiment on cyberbullying datasets from three different platforms through six cross-platform configurations, showing its effectiveness with both BERT and RoBERTa as the underlying Transformer models.