Addressing Antisocial Behavior in Multi-Party Dialogs Through Multimodal Representation Learning
This addresses platform safety and societal well-being by improving detection in underexplored multi-party conversational settings, though it is incremental as it builds on existing methods with new data.
The paper tackled antisocial behavior detection in multi-party dialogs by using a French dataset and multimodal representation learning, achieving top performance of 0.718 in abuse detection with a late fusion model.
Antisocial behavior (ASB) on social media -- including hate speech, harassment, and cyberbullying -- poses growing risks to platform safety and societal well-being. Prior research has focused largely on networks such as X and Reddit, while \textit{multi-party conversational settings} remain underexplored due to limited data. To address this gap, we use \textit{CyberAgressionAdo-Large}, a French open-access dataset simulating ASB in multi-party conversations, and evaluate three tasks: \textit{abuse detection}, \textit{bullying behavior analysis}, and \textit{bullying peer-group identification}. We benchmark six text-based and eight graph-based \textit{representation-learning methods}, analyzing lexical cues, interactional dynamics, and their multimodal fusion. Results show that multimodal models outperform unimodal baselines. The late fusion model \texttt{mBERT + WD-SGCN} achieves the best overall results, with top performance on abuse detection (0.718) and competitive scores on peer-group identification (0.286) and bullying analysis (0.606). Error analysis highlights its effectiveness in handling nuanced ASB phenomena such as implicit aggression, role transitions, and context-dependent hostility.