Few Shot Ner On Few Nerd Intra
評価指標
10 way 1~2 shot
10 way 5~10 shot
5 way 1~2 shot
5 way 5~10 shot
評価結果
このベンチマークにおける各モデルのパフォーマンス結果
比較表
| モデル名 | 10 way 1~2 shot | 10 way 5~10 shot | 5 way 1~2 shot | 5 way 5~10 shot |
|---|---|---|---|---|
| language-model-pre-training-with-sparse | 40.48 | 53.04 | 47.20 | 59.67 |
| container-few-shot-named-entity-recognition | 33.84 | 47.49 | 40.43 | 53.70 |
| a-multi-task-semantic-decomposition-framework | 47.13±0.69 | 64.69±0.51 | 56.35±0.28 | 66.80±0.78 |
| decomposed-meta-learning-for-few-shot-named | 42.84±0.46 | 57.31±0.25 | 49.48±0.85 | 62.92±0.57 |
| nuner-entity-recognition-encoder-pre-training | 57.63±0.38 | 62.99±0.27 | 62.48±0.28 | 69.16±0.28 |
| few-nerd-a-few-shot-named-entity-recognition | 15.05±0.44 | 35.40±0.13 | 20.76±0.84 | 42.54±0.94 |
| an-enhanced-span-based-decomposition-method | 30.00±0.70 | 42.15±2.60 | 36.08±1.60 | 52.14±1.50 |
| モデル 8 | - | - | 62.49±0.43 | 71.62±0.44 |
| few-nerd-a-few-shot-named-entity-recognition | 18.27±0.41 | 27.38±0.53 | 25.78±0.91 | 36.18±0.79 |
| decomposed-meta-learning-for-few-shot | 43.03±0.29 | 57.58±0.26 | 49.90±0.33 | 64.36±0.20 |
| few-nerd-a-few-shot-named-entity-recognition | 21.03±1.13 | 26.42±0.60 | 30.21±0.90 | 38.00±1.29 |
| heproto-a-hierarchical-enhancing-protonet | 46.45±0.21 | 58.98±0.22 | 53.03±0.30 | 65.70±0.21 |
| type-aware-decomposed-framework-for-few-shot | 55.44±0.08 | 60.87±0.22 | 60.78±0.32 | 67.94±0.17 |