5 way 1~2 shot: 30.21±0.905 way 5~10 shot: 38.00±1.2910 way 1~2 shot: 21.03±1.1310 way 5~10 shot: 26.42±0.60
Papers with Code · CC BY-SA 4.0
Few-shot NER
Few-Shot Named Entity Recognition (NER) is the task of recognising a 'named entity' like a person, organization, time and so on in a piece of text e.g. "Alan Mathison [person] visited the Turing Institute [organization] in June [time].
논문
5 way 1~2 shot: 56.35±0.285 way 5~10 shot: 66.80±0.7810 way 1~2 shot: 47.13±0.6910 way 5~10 shot: 64.69±0.51
5 way 1~2 shot: 36.08±1.605 way 5~10 shot: 52.14±1.5010 way 1~2 shot: 30.00±0.7010 way 5~10 shot: 42.15±2.60
5 way 1~2 shot: 40.435 way 5~10 shot: 53.7010 way 1~2 shot: 33.8410 way 5~10 shot: 47.49
5 way 1~2 shot: 49.48±0.855 way 5~10 shot: 62.92±0.5710 way 1~2 shot: 42.84±0.4610 way 5~10 shot: 57.31±0.25
5 way 1~2 shot: 49.90±0.335 way 5~10 shot: 64.36±0.2010 way 1~2 shot: 43.03±0.2910 way 5~10 shot: 57.58±0.26
5 way 1~2 shot: 53.03±0.305 way 5~10 shot: 65.70±0.2110 way 1~2 shot: 46.45±0.2110 way 5~10 shot: 58.98±0.22
5 way 1~2 shot: 47.205 way 5~10 shot: 59.6710 way 1~2 shot: 40.4810 way 5~10 shot: 53.04
5 way 1~2 shot: 62.48±0.285 way 5~10 shot: 69.16±0.2810 way 1~2 shot: 57.63±0.3810 way 5~10 shot: 62.99±0.27
5 way 1~2 shot: 60.78±0.325 way 5~10 shot: 67.94±0.1710 way 1~2 shot: 55.44±0.0810 way 5~10 shot: 60.87±0.22
1:1 Accuracy: 92
Avg F1: 0.85