A GAN is modeled as a dynamic equilibrium because: The generator maximizes its own loss function while the discriminator minimzes its own loss function. The equilibrium point changes from batch to batch. The discriminator maximizes while the generator minimizes an onverall loss function. None of the above