Algorithm 1: Original GAN algorithm

for$\mathbf{do}\#$ of training iterations:     for k do steps         Sample minibatch $\left\{{\mathit{z}}^{\left(1\right)},\dots ,{\mathit{z}}^{\left(m\right)}\right\}$∼$p_g\left(\mathbf{z}\right)$         Sample minibatch $\left\{{\mathit{x}}^{\left(1\right)},\dots ,{\mathit{x}}^{\left(m\right)}\right\}$∼$p_{\mathrm{d}}\left(\mathbf{x}\right)$         Update the weights of the discriminator by ascending its stochastic gradient: $${\nabla }_{{\mathit{\theta }}_D}\frac{1}{m}\sum _{i=1}^m\left[log\left(Dis\left({\mathbf{x}}^{\left(i\right)}\right)\right)+log\left(1-Dis\left(Gen\left({\mathbf{z}}^{\left(i\right)}\right)\right)\right)\right]$$     end for     Sample minibatch $\left\{{\mathit{z}}^{\left(1\right)},\dots ,{\mathit{z}}^{\left(m\right)}\right\}$∼$p_g\left(\mathbf{z}\right)$     Update the weights of the generator by descending its stochastic gradient: $${\nabla }_{{\mathit{\theta }}_G}\frac{1}{m}\sum _{i=1}^{m}log\left(1-Dis\left(Gen\left({\mathbf{z}}^{\left(i\right)}\right)\right)\right).$$ end for