Non-parametric Clustering on MNIST
14Inferred K (K_hat)moVB
Evaluation results
| Method | Links | |
|---|---|---|
| moVBFeature source=pre-trained autoencoder features, Backbone=autoencoder, Prior knowledge of class count=none2026.07 | 14 | |
| DPM SamplerFeature source=pre-trained autoencoder features, Backbone=autoencoder, Prior knowledge of class count=none2026.07 | 11.3 | |
| DeepDPMFeature source=pre-trained autoencoder features, Backbone=autoencoder, Prior knowledge of class count=none2026.07 | 10 | |
| converge-to-surpriseFeature source=from scratch, Backbone=ResNet-9, Prior knowledge of class count=none2026.07 | 10 | |
| DBSCANFeature source=pre-trained autoencoder features, Backbone=autoencoder, Prior knowledge of class count=none2026.07 | 9 |