Task-incremental Learning on CIFAR-100 (20-split)
79.3AccuracyMulti-task
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Multi-taskBackbone=AlexNet, Task Setting=Disjoint, Number of runs=32026.04 | 79.3 | — | |
| BANBackbone=AlexNet, Task Setting=Disjoint, Number of runs=32026.04 | 78.9 | — | |
| DRCLBackbone=AlexNet, Task Setting=Disjoint, Number of runs=32026.04 | 78.5 | — | |
| WSNBackbone=AlexNet, Task Setting=Disjoint, Number of runs=32026.04 | 78.2 | — | |
| SB-MCLBackbone=AlexNet, Task Setting=Disjoint, Number of runs=32026.04 | 78.1 | — | |
| UPGDBackbone=AlexNet, Task Setting=Disjoint, Number of runs=32026.04 | 78 | — | |
| POCLBackbone=AlexNet, Task Setting=Disjoint, Number of runs=32026.04 | 77.9 | — | |
| SGPBackbone=AlexNet, Task Setting=Disjoint, Number of runs=32026.04 | 77.8 | — | |
| IBPCLBackbone=AlexNet, Task Setting=Disjoint, Number of runs=32026.04 | 76.3 | — | |
| UCLBackbone=AlexNet, Task Setting=Disjoint, Number of runs=32026.04 | 74.9 | — | |
| HATBackbone=AlexNet, Task Setting=Disjoint, Number of runs=32026.04 | 73.1 | — | |
| TAGBackbone=AlexNet, Task Setting=Disjoint, Number of runs=32026.04 | 69.3 | — | |
| EWCBackbone=AlexNet, Task Setting=Disjoint, Number of runs=32026.04 | 62.5 | — | |
| InfluenceCL2023.04 | 61.2 | -0.04 | |
| GCR2023.04 | 60.9 | — | |
| OCS2023.04 | 60.5 | -0.04 | |
| iCaRL2023.04 | 60.3 | -0.04 | |
| ER-MIR2023.04 | 60.2 | -0.04 | |
| Bilevel2023.04 | 60.1 | -0.04 | |
| Vanilla IF2023.04 | 60 | -0.05 | |
| GSS2023.04 | 59.7 | -0.04 | |
| A-GEMBackbone=AlexNet, Task Setting=Disjoint, Number of runs=32026.04 | 57.8 | — | |
| Stable SGD2023.04 | 57.4 | -0.07 | |
| BaselineBackbone=AlexNet, Task Setting=Disjoint, Number of runs=32026.04 | 56.2 | — | |
| A-GEM2023.04 | 50.7 | -0.19 | |
| EWC2023.04 | 49.5 | -0.48 | |
| ER2023.04 | 46.9 | -0.21 |