Multi-label Classification on NUS-WIDE (P/R/F1@K3, P/R/F1@K5, mAP)
30.9Precision @ K=3Our Approach
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Our ApproachTask=GZSL2021.01 | 30.9 | 13.6 | 18.9 | 26 | 19.1 | 22 | 8.9 | |
| Our ApproachTask=ZSL2021.01 | 26.6 | 42.8 | 32.8 | 20.1 | 53.6 | 29.3 | 25.7 | |
| LESA (M=10)Task=ZSL2021.01 | 25.7 | 41.1 | 31.6 | 19.7 | 52.5 | 28.7 | 19.4 | |
| LESA (M=10)Task=GZSL2021.01 | 23.6 | 10.4 | 14.4 | 19.8 | 14.6 | 16.8 | 5.6 | |
| Fast0TagTask=ZSL2021.01 | 22.6 | 36.2 | 27.8 | 18.2 | 48.4 | 26.4 | 15.1 | |
| One Attention per LabelTask=ZSL2021.01 | 20.9 | 33.5 | 25.8 | 16.2 | 43.2 | 23.6 | 10.4 | |
| One Attention per Cluster (M=10)Task=ZSL2021.01 | 20 | 31.9 | 24.6 | 15.7 | 41.9 | 22.9 | 12.9 | |
| Fast0TagTask=GZSL2021.01 | 18.8 | 8.3 | 11.5 | 15.9 | 11.7 | 13.5 | 3.7 | |
| One Attention per LabelTask=GZSL2021.01 | 17.9 | 7.9 | 10.9 | 15.6 | 11.5 | 13.2 | 3.7 | |
| CONSETask=ZSL2021.01 | 17.5 | 28 | 21.6 | 13.9 | 37 | 20.2 | 9.4 | |
| LabelEMTask=ZSL2021.01 | 15.6 | 25 | 19.2 | 13.4 | 35.7 | 19.5 | 7.1 | |
| LabelEMTask=GZSL2021.01 | 15.5 | 6.8 | 9.5 | 13.4 | 9.8 | 11.3 | 2.2 | |
| CONSETask=GZSL2021.01 | 11.5 | 5.1 | 7 | 9.6 | 7.1 | 8.1 | 2.1 | |
| One Attention per Cluster (M=10)Task=GZSL2021.01 | 10.4 | 4.6 | 6.4 | 9.1 | 6.7 | 7.7 | 2.6 |