Multimodal Classification on N24News
80.3Headline AccuracyUniS-MMC
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| UniS-MMCFusion AGG=false, Fusion ALI=true, Backbone Image=ViT, Backbone Text=RoBERTa2023.05 | 80.3 | 78.1 | 84.2 | |
| UniS-MMCFusion AGG=false, Fusion ALI=true, Backbone Image=ViT, Backbone Text=BERT2023.05 | 80.2 | 77.5 | 83.2 | |
| UnSupMMCFusion AGG=false, Fusion ALI=true, Backbone Image=ViT, Backbone Text=RoBERTa2023.05 | 79.9 | 78 | 83.7 | |
| SupMMCFusion AGG=false, Fusion ALI=true, Backbone Image=ViT, Backbone Text=RoBERTa2023.05 | 79.9 | 77.9 | 84 | |
| SupMMCFusion AGG=false, Fusion ALI=true, Backbone Image=ViT, Backbone Text=BERT2023.05 | 79.6 | 77.3 | 81.7 | |
| N24NewsFusion AGG=true, Fusion ALI=false, Backbone Image=ViT, Backbone Text=RoBERTa2023.05 | 79.41 | 77.45 | 83.33 | |
| UnSupMMCFusion AGG=false, Fusion ALI=true, Backbone Image=ViT, Backbone Text=BERT2023.05 | 79.3 | 76.9 | 81.9 | |
| AggMMFusion AGG=true, Fusion ALI=false, Backbone Image=ViT, Backbone Text=RoBERTa2023.05 | 78.9 | 77.9 | 83.5 | |
| AggMMFusion AGG=true, Fusion ALI=false, Backbone Image=ViT, Backbone Text=BERT2023.05 | 78.6 | 76.8 | 80.8 |