Image Attribution On Cub 200 2011 1
评估指标
Deletion AUC score (ResNet-101)
Insertion AUC score (ResNet-101)
评测结果
各个模型在此基准测试上的表现结果
模型名称 | Deletion AUC score (ResNet-101) | Insertion AUC score (ResNet-101) | Paper Title | Repository |
|---|---|---|---|---|
| Grad-CAM | 0.0810 | 0.7224 | Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localization | |
| SMDL-Attribution (ICLR version) | 0.0613 | 0.7262 | Less is More: Fewer Interpretable Region via Submodular Subset Selection | |
| LIME | 0.1070 | 0.6812 | "Why Should I Trust You?": Explaining the Predictions of Any Classifier | |
| RISE | 0.0665 | 0.7193 | RISE: Randomized Input Sampling for Explanation of Black-box Models | |
| HSIC-Attribution | 0.0647 | 0.6843 | Making Sense of Dependence: Efficient Black-box Explanations Using Dependence Measure | |
| Integrated Gradients | 0.0728 | 0.0422 | Axiomatic Attribution for Deep Networks | |
| Saliency | 0.0682 | 0.6585 | Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps | |
| Kernel SHAP | 0.1016 | 0.6763 | A Unified Approach to Interpreting Model Predictions |
0 of 8 row(s) selected.