Audio Classification Attribution on VGG-Sound (val) benchmark leaderboard · SOTA2 Research

Audio Classification Attribution on VGG-Sound (val)

14.56Deletion AUCKernel SHAP

Evaluation Results

Method	Links
Kernel SHAP2025.04		14.56	17.85
LIME2025.04		13.75	14.22
Kernel SHAP2025.04		13.16	14.36
LIME2025.04		10.92	21.64
Saliency2025.04		10.24	21.84
Saliency2025.04		9.99	18.23
Sobol-Attribution2025.04		9.56	22.35
VarGrad2025.04		9.55	26.43
Occlusion2025.04		9.11	20.52
HSIC-Attribution2025.04		9.04	22.7
HSIC-Attribution2025.04		8.91	14.18
Square-Grad2025.04		8.88	26.97
Sobol-Attribution2025.04		8.81	14.29
LIMA2025.04		8.06	53.62
RISE2025.04		8	28.79
RISE2025.04		7.82	19.52
VarGrad2025.04		7.55	21.21
Square-Grad2025.04		7.48	21.2
SmoothGrad2025.04		7.34	9.55
Occlusion2025.04		7.07	16.5
SmoothGrad2025.04		5.35	16.64
LIMA2025.04		4.58	51.72
LIMA2025.04		4.25	41.64
LIMA2025.04		3.29	38.99
Gradient-Input2025.04		3.26	12.57
Integrated Gradients2025.04		3.13	12.92
Gradient-Input2025.04		2.43	13.88
Integrated Gradients2025.04		1.65	21.53