Class-Incremental Learning on NSynth 100 (test)
100Accuracy (Session 0)DSN
Evaluation results
| Method | Links | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| DSNshot=5-shot, way=5-way2026.07 | 100 | 98.05 | 97.32 | 95.27 | 93.56 | 92.78 | 91.16 | 90.53 | 90.05 | 88.89 | 93.76 | 11.11 | |
| iCaRLshot=5-shot, way=5-way2026.07 | 99.98 | 93.3 | 88.88 | 84.82 | 79.57 | 67.65 | 65.92 | 59.65 | 54.62 | 51.01 | 74.54 | 48.97 | |
| PANshot=5-shot, way=5-way2026.07 | 99.98 | 98.32 | 97.69 | 95.18 | 94 | 92.73 | 90.59 | 89.24 | 88.28 | 87.08 | 93.31 | 12.9 | |
| Oursshot=5-shot, way=5-way2026.07 | 99.98 | 99 | 97.77 | 96.58 | 95.43 | 94.93 | 93.39 | 92.42 | 92.13 | 91.77 | 95.34 | 8.21 | |
| Finetune-Cshot=5-shot, way=5-way2026.07 | 99.96 | 86.12 | 79.43 | 74.61 | 66.23 | 43.21 | 50.31 | 44.21 | 42.23 | 40.37 | 62.67 | 59.59 | |
| Finetune-Mshot=5-shot, way=5-way2026.07 | 99.96 | 84.73 | 76.92 | 71.06 | 63.18 | 40.15 | 47.99 | 38.33 | 38.01 | 37.86 | 59.82 | 62.1 | |
| CECshot=5-shot, way=5-way2026.07 | 99.96 | 97.47 | 95.56 | 93.52 | 91.22 | 89.9 | 87.85 | 85.84 | 84.92 | 83.19 | 90.94 | 16.77 | |
| ARPshot=5-shot, way=5-way2026.07 | 99.96 | 95.95 | 93.6 | 92.06 | 90.32 | 89.14 | 86.16 | 83.47 | 82.28 | 79.69 | 89.26 | 20.27 | |
| AMFOshot=5-shot, way=5-way2026.07 | 99.95 | 98.92 | 97.54 | 96.54 | 95.25 | 94.31 | 92.52 | 92.26 | 92.03 | 91 | 95.03 | 8.95 | |
| DFSLshot=5-shot, way=5-way2026.07 | 99.93 | 96 | 92.95 | 89.26 | 86.47 | 83.66 | 80.28 | 77.68 | 76.12 | 75.01 | 85.74 | 24.92 | |
| CLOMshot=5-shot, way=5-way2026.07 | 99.91 | 96.55 | 94.42 | 92.26 | 90.51 | 89.8 | 87.75 | 86.41 | 85.4 | 83.74 | 90.67 | 16.17 | |
| FACTshot=5-shot, way=5-way2026.07 | 99.73 | 96.52 | 94.52 | 92.96 | 90.37 | 89.37 | 86.48 | 85.01 | 84.41 | 82.98 | 90.24 | 16.75 | |
| LDCshot=5-shot, way=5-way2026.07 | 99.71 | 97.32 | 94.09 | 92.57 | 90.16 | 89.49 | 87.85 | 86.71 | 86.25 | 84.87 | 90.9 | 14.84 |