Image Classification on ImageNet (V2/S/R/A Accuracies)
65.6Accuracy (V2)InfluenceFunction
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| InfluenceFunctionData Removal=50%2025.10 | 65.6 | 28.36 | 29.54 | 8.72 | |
| RandomData Removal=50%2025.10 | 65.5 | 28.26 | 29.87 | 8.97 | |
| Data-OOBData Removal=50%2025.10 | 59.22 | 23.89 | 25.08 | 6.54 | |
| EV + Data-OOBData Removal=50%2025.10 | 54.76 | 21.82 | 22.75 | 5.37 | |
| KNN ShapleyData Removal=50%2025.10 | 40.39 | 18.03 | 16.95 | 7.82 | |
| EV + KNN ShapleyData Removal=50%2025.10 | 40.34 | 17.97 | 16.91 | 7.81 |