Clustering on SVHN
93.83AccuracyFB-NLL(-)
Evaluation results
| Method | Links | ||
|---|---|---|---|
| FB-NLL(-)Number of Tasks=Five Tasks, Noise Type=Uniform noise2026.04 | 93.83 | — | |
| FB-NLL(-)Number of Tasks=Three Tasks, Noise Type=Uniform noise2026.04 | 88.86 | — | |
| IFCA-PFLNumber of Tasks=Five Tasks, Noise Type=Uniform noise2026.04 | 88.43 | — | |
| FB-NLL(-)Number of Tasks=Two Tasks, Noise Type=Uniform noise2026.04 | 85.25 | — | |
| IFCA-PFLNumber of Tasks=Two Tasks, Noise Type=Uniform noise2026.04 | 85.17 | — | |
| IFCA-PFLNumber of Tasks=Three Tasks, Noise Type=Uniform noise2026.04 | 83.96 | — | |
| Single global modelNumber of Tasks=Two Tasks, Noise Type=Uniform noise2026.04 | 64.05 | — | |
| DTI-Sprites# runs=10, Clustering domain=pixels, Number of layers=multi-layer2026.04 | 63.1 | — | |
| Single global modelNumber of Tasks=Three Tasks, Noise Type=Uniform noise2026.04 | 59.92 | — | |
| IMSAT# runs=12 (5), Clustering domain=learned features, Data augmentation=true2026.04 | 57.3 | — | |
| SCAE# runs=5, Clustering domain=learned features, Data augmentation=true2026.04 | 55.3 | — | |
| SCAN# runs=5, Clustering domain=learned features, Data augmentation=true2026.04 | 54.2 | — | |
| Ours-C# runs=10, Clustering domain=pixels, Number of layers=2 layers2026.04 | 52.4 | — | |
| DTI-Clustering# runs=10, Clustering domain=pixels2026.04 | 44.5 | — | |
| Single global modelNumber of Tasks=Five Tasks, Noise Type=Uniform noise2026.04 | 40.74 | — | |
| ADC# runs=20, Clustering domain=learned features, Data augmentation=true2026.04 | 38.6 | — | |
| Ours-C# runs=10, Clustering domain=pixels, Number of layers=1 layer2026.04 | 37.6 | — | |
| K-means# runs=10, Clustering domain=pixels2026.04 | 12.2 | — | |
| BIRCHVSA Representation=MAP, Data Source=Main Network Outputs2024.10 | — | -0.07 | |
| BIRCHVSA Representation=HLB, Data Source=Main Network Outputs2024.10 | — | -0.07 | |
| GMMVSA Representation=VTB, Data Source=Main Network Inputs2024.10 | — | 1.37 | |
| GMMVSA Representation=HLB, Data Source=Main Network Outputs2024.10 | — | -0.04 | |
| HDBSCANVSA Representation=HRR, Data Source=Main Network Inputs2024.10 | — | -0.24 | |
| K-MEANSVSA Representation=HRR, Data Source=Main Network Inputs2024.10 | — | -0.01 | |
| K-MEANSVSA Representation=HRR, Data Source=Main Network Outputs2024.10 | — | 0.06 | |
| K-MEANSVSA Representation=VTB, Data Source=Main Network Outputs2024.10 | — | 0.13 |