MEMEWEAVER: Inter-Meme Graph Reasoning for Sexism and Misogyny Detection
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866914251956813824 |
|---|---|
| author | Italiani, Paolo Gimeno-Gomez, David Ragazzi, Luca Moro, Gianluca Rosso, Paolo |
| author_facet | Italiani, Paolo Gimeno-Gomez, David Ragazzi, Luca Moro, Gianluca Rosso, Paolo |
| contents | Women are twice as likely as men to face online harassment due to their gender. Despite recent advances in multimodal content moderation, most approaches still overlook the social dynamics behind this phenomenon, where perpetrators reinforce prejudices and group identity within like-minded communities. Graph-based methods offer a promising way to capture such interactions, yet existing solutions remain limited by heuristic graph construction, shallow modality fusion, and instance-level reasoning. In this work, we present MemeWeaver, an end-to-end trainable multimodal framework for detecting sexism and misogyny through a novel inter-meme graph reasoning mechanism. We systematically evaluate multiple visual--textual fusion strategies and show that our approach consistently outperforms state-of-the-art baselines on the MAMI and EXIST benchmarks, while achieving faster training convergence. Further analyses reveal that the learned graph structure captures semantically meaningful patterns, offering valuable insights into the relational nature of online hate. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_08684 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | MEMEWEAVER: Inter-Meme Graph Reasoning for Sexism and Misogyny Detection Italiani, Paolo Gimeno-Gomez, David Ragazzi, Luca Moro, Gianluca Rosso, Paolo Artificial Intelligence Computer Vision and Pattern Recognition Women are twice as likely as men to face online harassment due to their gender. Despite recent advances in multimodal content moderation, most approaches still overlook the social dynamics behind this phenomenon, where perpetrators reinforce prejudices and group identity within like-minded communities. Graph-based methods offer a promising way to capture such interactions, yet existing solutions remain limited by heuristic graph construction, shallow modality fusion, and instance-level reasoning. In this work, we present MemeWeaver, an end-to-end trainable multimodal framework for detecting sexism and misogyny through a novel inter-meme graph reasoning mechanism. We systematically evaluate multiple visual--textual fusion strategies and show that our approach consistently outperforms state-of-the-art baselines on the MAMI and EXIST benchmarks, while achieving faster training convergence. Further analyses reveal that the learned graph structure captures semantically meaningful patterns, offering valuable insights into the relational nature of online hate. |
| title | MEMEWEAVER: Inter-Meme Graph Reasoning for Sexism and Misogyny Detection |
| topic | Artificial Intelligence Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2601.08684 |