Pegasus@Dravidian-CodeMix-HASOC2021: Analyzing Social Media Content for Detection of Offensive Text

TL;DR

This paper presents two Transformer-based methods for detecting offensive social media content in Tamil and Malayalam, addressing challenges of informal, unstructured, and code-mixed text, achieving top-tier results in a shared task.

Contribution

Introduction of two novel Transformer-based approaches for offensive content detection in Dravidian languages, with code availability and competitive performance.

Findings

Achieved top 8 placement in all shared task categories.

Effectively handled informal, unstructured, and code-mixed social media text.

Demonstrated the effectiveness of Transformer models for multilingual offensive content detection.

Abstract

To tackle the conundrum of detecting offensive comments/posts which are considerably informal, unstructured, miswritten and code-mixed, we introduce two inventive methods in this research paper. Offensive comments/posts on the social media platforms, can affect an individual, a group or underage alike. In order to classify comments/posts in two popular Dravidian languages, Tamil and Malayalam, as a part of the HASOC - DravidianCodeMix FIRE 2021 shared task, we employ two Transformer-based prototypes which successfully stood in the top 8 for all the tasks. The codes for our approach can be viewed and utilized.