Open-world Semi-Supervised Learning on Single-cell
89.9Seen AccuracyORCA
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| ORCA2021.02 | 89.9 | 65.2 | 72.9 | |
| ORCA-FNM2021.02 | 89.7 | 48.6 | 58.7 | |
| ORCA-ZM2021.02 | 89.5 | 35.1 | 47.6 | |
| ORCA2021.02 | 80.6 | 52.8 | 61.3 | |
| ORCA-ZMvariant=Zero-Margin2021.02 | 79.9 | 31.8 | 41.7 | |
| ORCA-FNMvariant=Fixed Negative Margin2021.02 | 79.4 | 40.3 | 45.3 | |
| DS3LExtended to detect novel classes=true2021.02 | 76.2 | 29.7 | 26.4 | |
| CGDLExtended to detect novel classes=true2021.02 | 74.1 | 30.4 | 25.6 | |
| DS3L2021.02 | 70.8 | 26.4 | 21.8 | |
| RankStatsExtended to recognize seen classes=true2021.02 | 42.3 | 31.9 | 38.6 | |
| RankStatsextended_for_seen_classes=true2021.02 | 40.1 | 27.6 | 35.3 | |
| DTCExtended to recognize seen classes=true2021.02 | 29.6 | 25.3 | 27.8 | |
| DTCextended_for_seen_classes=true2021.02 | 25.8 | 20.2 | 22.4 |