Class-Incremental Learning on EMBER Class-IL
88.5Average AccuracyJoint
Evaluation Results
| Method | Links | |
|---|---|---|
| JointDescription=offline training on cumulative data; upper bound2026.05 | 88.5 | |
| FreeMOCALambda setting=Adap. Layerwise λ, Interpolation type=Linear2026.05 | 66.3 | |
| FreeMOCALambda setting=Fixed λ = 0.6, Interpolation type=Linear2026.05 | 65.2 | |
| FreeMOCALambda setting=Fixed λ = 0.6, Interpolation type=Spline2026.05 | 65.2 | |
| FreeMOCALambda setting=Fixed λ = 0.6, Interpolation type=Polynomial2026.05 | 65.2 | |
| WSC2026.05 | 64.1 | |
| FreeMOCALambda setting=Adap. Layerwise λ, Interpolation type=Polynomial2026.05 | 63.8 | |
| FreeMOCALambda setting=Adap. Layerwise λ, Interpolation type=Spline2026.05 | 63.5 | |
| CLeWI2026.05 | 60.1 | |
| iCaRL2026.05 | 55.6 | |
| MalCL2026.05 | 54.5 | |
| TAMiL2026.05 | 33.2 | |
| ER2026.05 | 26.9 | |
| BI-R2026.05 | 26.8 | |
| GR2026.05 | 26.6 | |
| EWC2026.05 | 25.6 | |
| NoneDescription=fine-tuning only; lower bound2026.05 | 22.9 | |
| LwF2026.05 | 10.7 | |
| SI2026.05 | 8.7 |