Hateful Meme Detection through Context-Sensitive Prompting and Fine-Grained Labeling
This addresses automated moderation for social media content, but it appears incremental as it builds on existing multi-modal classification methods.
The paper tackled hateful meme detection by proposing an end-to-end optimization framework that integrates context-sensitive prompting and fine-grained labeling, achieving the highest accuracy and AUROC in experiments.
The prevalence of multi-modal content on social media complicates automated moderation strategies. This calls for an enhancement in multi-modal classification and a deeper understanding of understated meanings in images and memes. Although previous efforts have aimed at improving model performance through fine-tuning, few have explored an end-to-end optimization pipeline that accounts for modalities, prompting, labeling, and fine-tuning. In this study, we propose an end-to-end conceptual framework for model optimization in complex tasks. Experiments support the efficacy of this traditional yet novel framework, achieving the highest accuracy and AUROC. Ablation experiments demonstrate that isolated optimizations are not ineffective on their own.