ToxBuster: In-game Chat Toxicity Buster with BERT

TL;DR

ToxBuster is a scalable BERT-based model designed to detect toxicity in online game chat, leveraging chat history and metadata to improve accuracy, with promising results for real-time moderation and transferability across games.

Contribution

Introduces ToxBuster, a novel scalable toxicity detection model trained on a large annotated dataset, improving precision and recall over existing methods by using chat history and metadata.

Findings

Achieves 82.95% precision and 83.56% recall.

Effective for real-time and post-game moderation.

Demonstrates transferability between different games.

Abstract

Detecting toxicity in online spaces is challenging and an ever more pressing problem given the increase in social media and gaming consumption. We introduce ToxBuster, a simple and scalable model trained on a relatively large dataset of 194k lines of game chat from Rainbow Six Siege and For Honor, carefully annotated for different kinds of toxicity. Compared to the existing state-of-the-art, ToxBuster achieves 82.95% (+7) in precision and 83.56% (+57) in recall. This improvement is obtained by leveraging past chat history and metadata. We also study the implication towards real-time and post-game moderation as well as the model transferability from one game to another.