Continual Learning on Washington RGB-D (WRGBD) (test)
96.7Accuracy (Other)PNN
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| PNNWeight Decay=10^-52023.07 | 96.7 | 99 | 100 | 90.7 | 99.7 | 94.1 | 96.7 | |
| MC DropoutMC Dropout=true2023.07 | 96.3 | 99.3 | 100 | 92.7 | 99.8 | 90.4 | 96.4 | |
| Zero Mean & IsotropicPrior Distribution Type=Zero Mean & Isotropic2023.07 | 96.3 | 95.5 | 100 | 91.9 | 100 | 87.7 | 95.2 | |
| BPNN (Learned Prior)Prior Distribution Type=Learned2023.07 | 96.2 | 98.4 | 100 | 93.9 | 100 | 95.1 | 97.3 | |
| IsotropicPrior Distribution Type=Isotropic2023.07 | 96.1 | 98.7 | 100 | 93.1 | 100 | 87.8 | 96 | |
| PNNWeight Decay=10^-32023.07 | 95.7 | 98.5 | 100 | 91.7 | 93.8 | 93.3 | 95.5 |