Controversy Detection

Controversy detection aims to automatically identify content eliciting strongly opposing viewpoints, a crucial task given its prevalence online. Current research focuses on improving the accuracy and controllability of models, exploring approaches that leverage user feedback, sentiment analysis (including dynamic sentiment tracking), and graph neural networks to analyze both textual content and the structure of online discussions. These advancements are significant for mitigating the spread of misinformation, improving online moderation, and enabling more nuanced understanding of public opinion across diverse platforms.