Reproduce “Defining, Understanding, and Detecting Online Toxicity: Challenges and Machine Learning Approaches” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Gautam Kishore Shahi, Tim A. Majchrzak
This systematic literature review by Shahi and Majchrzak synthesizes 140 publications on online toxicity, analyzing diverse definitions, datasets, and detection methods. It reveals significant fragmentation in research, marked by English-centric and platform-specific biases, alongside inconsistent content definitions, hindering the development of universally applicable detection models.
Key insights
- Research on online toxicity is predominantly focused on English content (over 100 studies) and specific social media platforms, especially Twitter.
- There is significant heterogeneity in how "toxic content" is defined across studies, government regulations, and platform policies, complicating consistent data annotation.
- Developing truly generalizable models for detecting toxic content across different platforms, languages, and topics remains a substantial challenge despite diverse machine learning advancements.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.