We reported a recent work that applies modern Deep Learning (convolutional neural network) techniques in the context of two dimensional lattice complex scalar field theory, which has a non-trivial phase diagram at nonzero temperature and chemical potential. Especially we introuced the field configuration production with generative adversarial network (GAN), where the GAN is showed to be able to automatically capture the implicit local constraint for the physical configurations and also the underlying physical distribution. We further explored generalize the configuration production at different parameter space using conditional GAN.