Sentiment Quadruplet Extraction on Rest SemEval 2015 (test)
47.78PrecisionILO
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| ILOBackbone=T5-base, Decoding=Beam Search (k=5)2022.10 | 47.78 | 50.38 | 49.05 | |
| DLOBackbone=T5-base, Decoding=Beam Search (k=5)2022.10 | 47.08 | 49.33 | 48.18 | |
| ParaphraseModel Type=Generative2022.10 | 46.16 | 47.72 | 46.93 | |
| GASModel Type=Generative2022.10 | 45.31 | 46.7 | 45.98 | |
| TASO-BERT-CRFModel Type=Discriminative/Tagging2022.10 | 44.24 | 28.66 | 34.78 | |
| TASO-BERT-LinearModel Type=Discriminative/Tagging2022.10 | 41.86 | 26.5 | 32.46 | |
| Extract-Classify-ACOSModel Type=Two-stage2022.10 | 35.64 | 37.25 | 36.42 | |
| HGCN-BERT+BERT-TFMModel Type=Discriminative/Tagging2022.10 | 25.55 | 22.01 | 23.65 | |
| HGCN-BERT+BERT-LinearModel Type=Discriminative/Tagging2022.10 | 24.43 | 20.25 | 22.15 |