Open World Semi Supervised Learning On Cifar
Metrics
All accuracy (10% Labeled)
All accuracy (50% Labeled)
Novel accuracy (10% Labeled)
Novel accuracy (50% Labeled)
Seen accuracy (10% Labeled)
Seen accuracy (50% Labeled)
Results
Performance results of various models on this benchmark
Comparison Table
| Model Name | All accuracy (10% Labeled) | All accuracy (50% Labeled) | Novel accuracy (10% Labeled) | Novel accuracy (50% Labeled) | Seen accuracy (10% Labeled) | Seen accuracy (50% Labeled) |
|---|---|---|---|---|---|---|
| openldn-learning-to-discover-novel-classes | 92.8 | 95.4 | 93.2 | 95.1 | 92.4 | 95.7 |
| generalized-category-discovery | - | 91.5 | - | 88.2 | - | 97.9 |
| open-world-semi-supervised-learning-1 | 84.1 | 89.7 | 85.5 | 90.4 | 82.8 | 88.2 |
| towards-realistic-semi-supervised-learning | 92.2 | - | 89.6 | - | 94.9 | - |
| a-simple-parametric-classification-baseline | - | 97.0 | - | 98.5 | - | 93.9 |