Our primary motivation is to apply a GAN-derived model, to the generation of sequential comics art. Generative Adversarial Networks are one of the most successful image synthesis programming architectures in the past few years. This model pits two or more neural networks against each other in adversarial training to produce generative models. The first network, called the generator (G), generates samples that are intended to come from the same probability distribution as the training data. The other network, denoted as the discriminator (D), examines the samples to determine whether they are coming from the dataset (real) or not (fake). The two networks compete in what is known in AI as unsupervised learning, unfolding through a zero-sum game, until the generated samples are indistinguishable from those that are in the dataset.
Graphic narratives are not only important in comics or in general domains of artistic expression. They are tools whose multimodal expressive communication has become our primary modalitiy in sharing and shaping representation of our worlds. From data infographics and communication strategies to community building and graphic journalism there is a story to be told. Our skills and acquired sets of knowledges can find multiple applications in a graphic narrative-rich Internet environment.