Few-shot classification on Bongard-LOGO (test)
92.1Free-Form ScoreHuman Expert
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Human Expert2023.09 | 92.1 | 99.3 | 90.7 | — | 94 | |
| Human Amateur2023.09 | 88 | 90 | — | 71 | 83 | |
| PredRNetEncoder=Other2023.09 | 74.6 | 75.2 | 71.1 | 68.4 | 72.3 | |
| SVM-MimicEncoder=RN122023.09 | 73.3 | 84.3 | 69.4 | 74.2 | 75.3 | |
| Prototype-MimicEncoder=RN122023.09 | 73.1 | 80.9 | 68.4 | 72 | 73.6 | |
| k-NN + standardizeEncoder=RN12, support-set standardization=true2023.09 | 73 | 84.9 | 67.8 | 71 | 74.2 | |
| Prototype + standardizeEncoder=RN12, support-set standardization=true2023.09 | 72.9 | 80.8 | 66.8 | 71.3 | 73 | |
| SVM + standardizeEncoder=RN12, support-set standardization=true2023.09 | 72.2 | 83.8 | 68.3 | 71.8 | 74 | |
| PrototypeEncoder=RN122023.09 | 68.9 | 68.8 | 64.7 | 66.2 | 67.1 | |
| Meta-Baseline-PSEncoder=RN122023.09 | 68.2 | 75.7 | 67.4 | 71.5 | 70.7 | |
| Meta-Baseline-SCEncoder=RN122023.09 | 66.3 | 73.3 | 63.5 | 63.9 | 66.8 | |
| Meta-Baseline-MoCoEncoder=RN122023.09 | 65.9 | 72.2 | 63.9 | 64.7 | 66.7 | |
| SVMEncoder=RN122023.09 | 65.7 | 75.3 | 65.4 | 70.4 | 69.2 | |
| ProtoNetEncoder=RN122023.09 | 64.8 | 72.4 | 62.4 | 65.4 | 66.3 | |
| k-NNEncoder=RN122023.09 | 62.1 | 69.3 | 62.8 | 63.3 | 64.4 | |
| MetaOptNetEncoder=RN122023.09 | 60.3 | 71.7 | 61.7 | 63.3 | 64.3 | |
| ANILEncoder=RN122023.09 | 56.6 | 59 | 59.6 | 61 | 59.1 | |
| SNAILEncoder=RN122023.09 | 56.3 | 60.2 | 60.1 | 61.3 | 59.5 | |
| TPTEncoder=CLIP2023.09 | 52.5 | 65.9 | 58.6 | 56.3 | 58.3 | |
| CNN-BaselineEncoder=RN122023.09 | 51.9 | 56.6 | 53.6 | 57.6 | 54.9 | |
| WReN-BongardEncoder=RN122023.09 | 50.1 | 50.9 | 53.8 | 54.3 | 52.3 |