| Paper | Code | Insertion AUC score (ArcFace ResNet-101) | Deletion AUC score (ArcFace ResNet-101) | ModelName | ReleaseDate |
|---|---|---|---|---|---|
| Less is More: Fewer Interpretable Region via Submodular Subset Selection | ✓ Link | 0.6705 | 0.1304 | SMDL-Attribution (ICLR version) | 2024-02-14 |
| Making Sense of Dependence: Efficient Black-box Explanations Using Dependence Measure | ✓ Link | 0.6694 | 0.1317 | HSIC-Attribution | 2022-06-13 |
| RISE: Randomized Input Sampling for Explanation of Black-box Models | ✓ Link | 0.6530 | 0.1375 | RISE | 2018-06-19 |
| "Why Should I Trust You?": Explaining the Predictions of Any Classifier | ✓ Link | 0.6185 | 0.2119 | LIME | 2016-02-16 |
| A Unified Approach to Interpreting Model Predictions | ✓ Link | 0.6132 | 0.2034 | Kernel SHAP | 2017-05-22 |
| Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps | ✓ Link | 0.5612 | 0.1907 | Saliency | 2013-12-20 |
| Axiomatic Attribution for Deep Networks | ✓ Link | 0.5399 | 0.0749 | Integrated Gradients | 2017-03-04 |
| Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localization | ✓ Link | 0.4733 | 0.3103 | Grad-CAM | 2016-10-07 |