| Paper | Code | Average highest confidence (ResNet-101) | Insertion AUC score (ResNet-101) | Average highest confidence (MobileNetV2) | Insertion AUC score (MobileNetV2) | Average highest confidence (EfficientNetV2-M) | Insertion AUC score (EfficientNetV2-M) | ModelName | ReleaseDate |
|---|---|---|---|---|---|---|---|---|---|
| Less is More: Fewer Interpretable Region via Submodular Subset Selection | ✓ Link | 0.4513 | 0.1772 | 0.5367 | 0.1922 | 0.3306 | 0.1748 | SMDL-Attribution (ICLR version) | 2024-02-14 |
| Grad-CAM++: Improved Visual Explanations for Deep Convolutional Networks | ✓ Link | 0.2647 | 0.1094 | 0.3462 | 0.1284 | 0.2659 | 0.1605 | Grad-CAM++ | 2017-10-30 |
| Score-CAM: Score-Weighted Visual Explanations for Convolutional Neural Networks | ✓ Link | 0.2510 | 0.1073 | 0.3141 | 0.1195 | 0.2403 | 0.1572 | Score-CAM | 2019-10-03 |
| Making Sense of Dependence: Efficient Black-box Explanations Using Dependence Measure | ✓ Link | 0.2493 | 0.1446 | 0.2914 | 0.1635 | 0.2679 | 0.1611 | HSIC-Attribution | 2022-06-13 |