Ai Art

As someone who works with numbers and codes, I’m interested in exploring how algorithmic models intersect with artistic creativity.

This series of artwork is inspired by Deep Learning algorithms that are fine-tuned to generate original contents. A few types of frameworks are applied: Style Transfer (ST), Generative Adversarial Network(GAN), text-to-image(Stable Diffusion, Midjourney). ST is a deep learning model that blends style elements from one image and content from another to produce a new image. Meanwhile, GAN trains two neural networks – the discriminator and the generator, to produce images that mimic the style/genre of the training data set.

In 2022, two major text-to-image models, Stable Diffusion and Midjourney, became the game changers in the world of AI generative art. Both systems are capable of generating highly original and nuanced images based on the user’s text prompts. My personal bias is that Midjourney is more refined in its output and more sensitive to the input text prompting.

Some of the pieces are for sale on Fine Art America(see the ‘SHOP’ tab), I am working on uploading all of them to the FAA platform.