Multi-Instance Learning TPAUC Maximization on MUSK2 (test)
0.867Mean TPAUCAUC-M (att)
Evaluation Results
| Method | Links | |
|---|---|---|
| AUC-M (att)(TPR lower bound, FPR upper bound)=(0.1, 0.9), Pooling strategy=attention-based2023.10 | 0.867 | |
| SONT (att)(TPR lower bound, FPR upper bound)=(0.1, 0.9), Pooling strategy=attention-based2023.10 | 0.867 | |
| MIDAM (att)(TPR lower bound, FPR upper bound)=(0.1, 0.9), Pooling strategy=attention-based2023.10 | 0.819 | |
| SOTAS (att)(TPR lower bound, FPR upper bound)=(0.1, 0.9), Pooling strategy=attention-based2023.10 | 0.819 | |
| MIDAM (smx)(TPR lower bound, FPR upper bound)=(0.1, 0.9), Pooling strategy=stochastic smoothed-max2023.10 | 0.8 | |
| SONT (att)(TPR lower bound, FPR upper bound)=(0.3, 0.7), Pooling strategy=attention-based2023.10 | 0.8 | |
| AUC-M (att)(TPR lower bound, FPR upper bound)=(0.3, 0.7), Pooling strategy=attention-based2023.10 | 0.783 | |
| MIDAM (att)(TPR lower bound, FPR upper bound)=(0.3, 0.7), Pooling strategy=attention-based2023.10 | 0.717 | |
| SONT (att)(TPR lower bound, FPR upper bound)=(0.5, 0.5), Pooling strategy=attention-based2023.10 | 0.7 | |
| SOTAS (att)(TPR lower bound, FPR upper bound)=(0.3, 0.7), Pooling strategy=attention-based2023.10 | 0.683 | |
| AUC-M (att)(TPR lower bound, FPR upper bound)=(0.5, 0.5), Pooling strategy=attention-based2023.10 | 0.675 | |
| MIDAM (smx)(TPR lower bound, FPR upper bound)=(0.3, 0.7), Pooling strategy=stochastic smoothed-max2023.10 | 0.667 | |
| MIDAM (att)(TPR lower bound, FPR upper bound)=(0.5, 0.5), Pooling strategy=attention-based2023.10 | 0.6 | |
| SOTAS (att)(TPR lower bound, FPR upper bound)=(0.5, 0.5), Pooling strategy=attention-based2023.10 | 0.6 | |
| MIDAM (smx)(TPR lower bound, FPR upper bound)=(0.5, 0.5), Pooling strategy=stochastic smoothed-max2023.10 | 0.525 |