AI art as a large model

is AI art supervised learning or unsupervised ?


Well that depends on the type of AI art application. On one hand, you have the style transfer technique where the model is trained on existing images so as to generate a new image with a similar style. In this case it is supervised learning because the labels exist. Then you have the image generation itself which is done either through Transformers, encoder-decoder kind of models, or Generative Adversarial Networks (GANs) which use a generator and a discriminator network, etc… Those model use unsupervised learning techniques because there are no labels and the model learns on itself to generate the image out of noise as an example. I hope this gives you a better idea about the subject!