Machine learning + images: what's possible

This workshop will introduce students to the basics of manipulating and generating images with neural networks.

There will be a short presentation at the beginning showing work of artists currently working with ML, and a discussion about our ethical obligations as artists+technologists working with images. If you plan to attend, please come for this part of the class!

The second half of the class will be a hands on workshop.

We'll use style transfer as an introduction to ML models, and discuss how to use pre-trained models, tweak parameters on models, and how to train your own models.

Finally, we'll take a look at libraries such as pix2pix and openpose which offer beginner-friendly tools for working with ML and images.

I'll be using Spell (spell.run) to show the demos (but you're welcome to use another platform). I'll also give a quick intro to remote training and the difference between GPU/CPU.