From: Feature partitioning for robust tree ensembles and their certification in adversarial scenarios
Model | Parameters | Accuracy | |||||||
---|---|---|---|---|---|---|---|---|---|
b | r | p | ml | \(|{\mathcal {T}}|\) | \(\phantom {\dot {i}\!}ACC_{A_{0}}\) | k | \(\phantom {\dot {i}\!}ACC_{A_{k}}\) | Δ f-FPF | |
f-FPF | 2 | 60 | 16 | 300 | 0.918 | 1 | 0.837 | 0.000 | |
4 | 33 | 16 | 297 | 0.866 | 2 | 0.782 | 0.000 | ||
h-FPF | 2 | 60 | 16 | 300 | 0.913 | 1 | 0.834 | −0.003 | |
4 | 33 | 8 | 297 | 0.865 | 2 | 0.775 | −0.007 | ||
RSM | 0.2 | 16 | 300 | 0.909 | 1 | 0.837 | 0.000 | ||
0.2 | 16 | 300 | 0.909 | 2 | 0.736 | −0.046 | |||
RF | 8 | 300 | 0.923 | 1 | 0.766 | −0.071 | |||
4 | 300 | 0.902 | 2 | 0.449 | −0.333 | ||||
RT | 3 | 8 | 300 | 0.715 | 1 | 0.433 | −0.422 | ||
3 | 8 | 300 | 0.715 | 2 | 0.201 | −0.655 |