WLV-RIT at GermEval 2021: Multitask Learning with Transformers to Detect Toxic, Engaging, and Fact-Claiming Comments

TL;DR

This paper presents a multitask learning approach using transformer models to simultaneously detect toxic, engaging, and fact-claiming comments on social media, demonstrating improved performance over single-task methods.

Contribution

The paper introduces a multitask learning framework with transformers for multi-label comment classification, outperforming traditional single-task models on GermEval-2021 data.

Findings

Multitask learning outperforms single-task models in all three comment classification tasks.

Transformer-based models effectively handle German social media comments.

The approach achieved top results in the GermEval-2021 shared task.

Abstract

This paper addresses the identification of toxic, engaging, and fact-claiming comments on social media. We used the dataset made available by the organizers of the GermEval-2021 shared task containing over 3,000 manually annotated Facebook comments in German. Considering the relatedness of the three tasks, we approached the problem using large pre-trained transformer models and multitask learning. Our results indicate that multitask learning achieves performance superior to the more common single task learning approach in all three tasks. We submit our best systems to GermEval-2021 under the team name WLV-RIT.