Open-world Semi-Supervised Learning on CIFAR-100 (50% known, 50% novel split)
74.1Known AccuracyOpenLDN
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| OpenLDNSSL Framework=UDA2022.07 | 74.1 | 44.5 | 59.3 | |
| OpenLDNSSL Framework=MixMatch2022.07 | 73.5 | 46.8 | 60.1 | |
| TRSSL2022.07 | 68.5 | 52.1 | 60.3 | |
| UNO2022.07 | 68.3 | 36.5 | 51.5 | |
| ORCA2022.07 | 66.9 | 43 | 48.1 | |
| DS3L2022.07 | 55.1 | 23.7 | 24 | |
| UNO2022.07 | 53.7 | 33.6 | 42.7 | |
| ORCA2022.07 | 52.5 | 31.8 | 38.6 | |
| CGDL2022.07 | 49.3 | 22.5 | 23.5 | |
| FixMatch2022.07 | 39.6 | 23.5 | 20.3 | |
| RankStats2022.07 | 36.4 | 28.4 | 23.1 | |
| DS3L2022.07 | 33.7 | 15.8 | 15.1 | |
| DTC2022.07 | 31.3 | 22.9 | 18.3 | |
| FixMatch2022.07 | 30.9 | 18.5 | 15.3 | |
| DTC2022.07 | 22.1 | 10.5 | 13.7 | |
| RankStats2022.07 | 20.4 | 16.7 | 17.8 |