Making Sense of Dependence: Efficient Black-box Explanations Using Dependence Measure

Benchmark Model Rank Results
error-understanding-on-cub-200-2011-1HSIC-Attribution#4Average highest confidence (ResNet-101): 0.2493
error-understanding-on-cub-200-2011-resnetHSIC-Attribution#3Average highest confidence: 0.2493Insertion AUC score: 0.1446
image-attribution-on-celebaHSIC-Attribution#3Insertion AUC score (ArcFace ResNet-101): 0.5692
image-attribution-on-cub-200-2011-1HSIC-Attribution#4Insertion AUC score (ResNet-101): 0.6843
image-attribution-on-vggface2HSIC-Attribution#2Insertion AUC score (ArcFace ResNet-101): 0.6694
interpretability-techniques-for-deep-learning-1HSIC-Attribution#2Insertion AUC score: 0.5692