Fine-grained Entity Typing on OntoNotes augmented (test)
87.1Macro F1DFET
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DFET2022.12 | 87.1 | 81.5 | |
| MLC-RoBERTa w/ NPCRFBackbone=RoBERTa2022.12 | 86.9 | 81.9 | |
| LITE2022.12 | 86.6 | 81.4 | |
| LabelGCN-RoBERTa w/ NPCRFBackbone=RoBERTa2022.12 | 86.2 | 81.3 | |
| LabelGCN-RoBERTaBackbone=RoBERTa, label correlation modeling=true2022.12 | 85.5 | 80.2 | |
| MLMETdistantly labeled training data or additional pretraining tasks=true2022.12 | 85.4 | 80.4 | |
| MLC-RoBERTaBackbone=RoBERTa2022.12 | 85.4 | 80.2 | |
| PL-BERT-BASE w/ NPCRFBackbone=BERT-Base2022.12 | 85.2 | 80 | |
| LRNlabel correlation modeling=true2022.12 | 84.8 | 80.1 | |
| PL-BERT-BASEBackbone=BERT-Base2022.12 | 84.8 | 79.6 | |
| LDETdistantly labeled training data or additional pretraining tasks=true2022.12 | 84.5 | 79.2 | |
| BOX4TYPESdistantly labeled training data or additional pretraining tasks=true2022.12 | 77.3 | 70.9 |