SmoothGrad - potential typo in implementation of noise level

[chr5tphr]

SmoothGrad - potential typo in implementation of noise level

[chr5tphr]

Hey Anna,
thanks for the issue. The paper states in Figure 3:

Each sensitivity map is
obtained by applying Gaussian noise $\mathcal{N}(0, \sigma^2)$ to the input pixels for 50 samples, and averaging them. The noise level corresponds to $\sigma / (x_\text{max} − x_\text{min})$.

Since we need to compute the standard deviation of the Gaussian $\sigma$ from a known noise level $\varepsilon$, we have thus
$$\varepsilon = \frac{\sigma}{(x_\text{max} - x_\text{min})} \iff \sigma = \varepsilon \cdot (x_\text{max} - x_\text{min})$$

[annahedstroem]

Hey @chr5tphr thanks so much for clarifying! and great that it wasn't a bug (but a misinterpretation from me!) I'll go ahead and close the issue. :D