Relation Extraction On Ace 2004
Métriques
Cross Sentence
NER Micro F1
RE+ Micro F1
Résultats
Résultats de performance de divers modèles sur ce benchmark
Tableau comparatif
| Nom du modèle | Cross Sentence | NER Micro F1 | RE+ Micro F1 |
|---|---|---|---|
| entity-relation-extraction-as-multi-turn | No | 83.6 | 49.4 |
| a-partition-filter-network-for-joint-entity | No | 89.3 | 62.5 |
| a-frustratingly-easy-approach-for-joint | Yes | 90.3 | 62.2 |
| pack-together-entity-and-relation-extraction | Yes | 90.4 | 66.5 |
| incremental-joint-extraction-of-entity | No | 79.7 | 45.3 |
| end-to-end-relation-extraction-using-lstms-on | No | 81.8 | 48.4 |
| going-out-on-a-limb-joint-extraction-of | No | 79.6 | 45.7 |
| two-are-better-than-one-joint-entity-and | No | 88.6 | 59.6 |
| adversarial-training-for-multi-context-joint | No | 81.64 | 47.45 |
| joint-entity-recognition-and-relation | No | 81.16 | 47.14 |
| a-general-framework-for-information | Yes | 87.4 | - |