Algorithm 1 Gradient descent optimization of GAN

for number of training iterations do for k steps do Sample batch of ${\mathrm{M}}_{num}$ samples from voltage distributions $\left\{{\mathrm{V}}^{\left(1\right)},\dots ,{\mathrm{V}}^{\left(\mathrm{M}\right)}\right\}$. Sample batch of ${\mathrm{M}}_{num}$ samples from conductivity distributions $\left\{{\mathsf{\sigma }}^{\left(1\right)},\dots ,{\mathsf{\sigma }}^{\left(\mathrm{M}\right)}\right\}$. Update the discriminator by ascending its stochastic gradient: ${\nabla }_{{\mathsf{\theta }}_{\mathrm{d}}}\frac{1}{\mathrm{m}}{\sum }_{\mathrm{i}=1}^{{\mathrm{M}}_{num}}\left[\mathrm{logD}\left({\mathsf{\sigma }}^{\left(\mathrm{i}\right)}\right)+\log \left(1-\mathrm{D}\left(\mathrm{G}\left({\mathrm{V}}^{\left(\mathrm{i}\right)}\right)\right)\right)\right]$. end for Sample batch of ${\mathrm{M}}_{num}$ samples from voltage distributions $\left\{{\mathrm{V}}^{\left(1\right)},\dots ,{\mathrm{V}}^{\left(\mathrm{M}\right)}\right\}$. Update the generator by descending its stochastic gradient: ${\nabla }_{{\mathsf{\theta }}_{\mathrm{g}}}\frac{1}{\mathrm{m}}{\sum }_{\mathrm{i}=1}^{{\mathrm{M}}_{num}}\log \left(1-\mathrm{D}\left(\mathrm{G}\left({\mathrm{V}}^{\left(\mathrm{i}\right)}\right)\right)\right)$. end for