Multimodal Aspect Sentiment Classification on TWITTER 2015
78.1F1 ScoreWisDoM
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| WisDoMCategory=Non-prompt-based2026.01 | 78.1 | 81.5 | |
| LRSACategory=Non-prompt-based2026.01 | 75.9 | 79.5 | |
| CoolNetCategory=Non-prompt-based2026.01 | 75.3 | 79.9 | |
| GLFFCACategory=Non-prompt-based2026.01 | 74.2 | 77.7 | |
| VLP-MABSA2022.04 | 73.8 | 78.6 | |
| VLP-MABSACategory=Non-prompt-based2026.01 | 73.8 | 78.6 | |
| CapTrBERT2022.04 | 73.2 | 78 | |
| JML-MASCevaluation_protocol=evaluates only on aspects correctly predicted by JML-MATE2022.04 | 72.7 | 78.7 | |
| MPFITCategory=Prompt-based2026.01 | 72.6 | 76.4 | |
| OursCategory=Prompt-based, Approach=Syntactic-aware generative framework2026.01 | 72.6 | 77.2 | |
| SaliencyBERTCategory=Non-prompt-based2026.01 | 72.4 | 77 | |
| TomBERT2022.04 | 71.8 | 77.2 | |
| BIPFCategory=Prompt-based2026.01 | 71.2 | 77.3 | |
| Flan-T5-VACategory=Prompt-based2026.01 | 69.7 | — | |
| MultiPointCategory=Prompt-based, Few-shot designed=true2026.01 | 66.6 | 67.3 |