Offensive Language Detection on Social Media Using XLNet

TL;DR

This paper presents an offensive language detection system on social media using XLNet, demonstrating its superior performance over BERT in identifying offensive content and categories, with strategies to handle class imbalance.

Contribution

The study introduces an XLNet-based model for offensive language detection and compares its effectiveness with BERT, highlighting improvements in classification accuracy on social media data.

Findings

XLNet outperforms BERT in detecting offensive content and categorizing offenses.

Oversampling and undersampling improve classification performance.

XLNet-based models show robustness in handling imbalanced social media datasets.

Abstract

The widespread use of text-based communication on social media-through chats, comments, and microblogs-has improved user interaction but has also led to an increase in offensive content, including hate speech, racism, and other forms of abuse. Due to the enormous volume of user-generated content, manual moderation is impractical, which creates a need for automated systems that can detect offensive language. Deep learning models, particularly those using transfer learning, have demonstrated significant success in understanding natural language through large-scale pretraining. In this study, we propose an automatic offensive language detection model based on XLNet, a generalized autoregressive pretraining method, and compare its performance with BERT (Bidirectional Encoder Representations from Transformers), which is a widely used baseline in natural language processing (NLP). Both models are evaluated using the Offensive Language Identification Dataset (OLID), a benchmark Twitter dataset that includes hierarchical annotations. Our experimental results show that XLNet outperforms BERT in detecting offensive content and in categorizing the types of offenses, while BERT performs slightly better in identifying the targets of the offenses. Additionally, we find that oversampling and undersampling strategies are effective in addressing class imbalance and improving classification performance. These findings highlight the potential of transfer learning and XLNet-based architectures to create robust systems for detecting offensive language on social media platforms.