Fake News Detection on Internal Merged Multimodal Misinformation Dataset
99.6AccuracyCaption-Aware
Evaluation results
| Method | Links | |
|---|---|---|
| Caption-AwareLanguage=Indic, Modalities=Text + Visual, Limitations=Low-resource2026.07 | 99.6 | |
| Contrastive LearningLanguage=Hindi, Modalities=Text + Visual, Limitations=Language-specific2026.07 | 98.6 | |
| Fusion ModelLanguage=Bangla+English+Meta, Modalities=Text+Image+Sarcasm+Geo, Limitations=Violence-aware2026.07 | 98 | |
| Entity FusionLanguage=Chinese+English, Modalities=Text + Visual, Limitations=Weak sarcasm cues2026.07 | 97.5 | |
| Robustness EvalLanguage=English, Modalities=Text + Visual, Limitations=Adversarial fragility2026.07 | 92.5 | |
| MDAM3Language=English, Modalities=Text + Image + Video, Limitations=High compute cost2026.07 | 85.3 | |
| Multimodal AnalysisLanguage=English, Modalities=Text + Visual, Limitations=English-only2026.07 | 85.25 |