Corrective Unlearning (Interclass Confusion) on CIFAR100
0.06Unlearning CostSSD
Evaluation results
| Method | Links | |
|---|---|---|
| SSDNum. Corrupted Samples=1002026.04 | 0.06 | |
| SSDNum. Corrupted Samples=2502026.04 | 0.06 | |
| SSDNum. Corrupted Samples=5002026.04 | 0.06 | |
| RELOADNum. Corrupted Samples=1002026.04 | 0.14 | |
| RELOADNum. Corrupted Samples=5002026.04 | 0.14 | |
| RELOADNum. Corrupted Samples=2502026.04 | 0.15 | |
| CFNum. Corrupted Samples=2502026.04 | 0.25 | |
| CFNum. Corrupted Samples=5002026.04 | 0.25 | |
| CFNum. Corrupted Samples=1002026.04 | 0.26 | |
| ScrubNum. Corrupted Samples=1002026.04 | 0.31 | |
| ScrubNum. Corrupted Samples=2502026.04 | 0.31 | |
| ScrubNum. Corrupted Samples=5002026.04 | 0.31 | |
| BadTNum. Corrupted Samples=1002026.04 | 0.68 | |
| BadTNum. Corrupted Samples=2502026.04 | 0.68 | |
| BadTNum. Corrupted Samples=5002026.04 | 0.68 | |
| RewoDNum. Corrupted Samples=1002026.04 | 1 | |
| RewoDNum. Corrupted Samples=2502026.04 | 1 | |
| RewoDNum. Corrupted Samples=5002026.04 | 1 |