Unsupervised Machine Learning for Designers

This course provides an introduction to the rapidly advancing area of research in unsupervised machine learning with a focus on generative models.

Recent advances such as DALL-E and ChatGPT have captured the public imagination with radical new possibilities for the design of products, interfaces, graphics, texts, buildings and cities. For designers, the potential of these disruptive new workflows cannot be overstated.

This course takes an under-the-hood look at the anatomy of generative models. Students will come away with a computational toolkit enabling them to leverage state of the art techniques  for visual representation, generative text, text to image synthesis and more.

The course will cover the foundational algorithms of deep learning and the techniques to implement them in machine learning frameworks such as Pytorch and Tensorflow, culminating  in a hands-on final design project.

Over the course of the semester we will trace the evolution of the field of machine learning through its milestone papers and their implementations, providing students with perspective to appreciate the breakthrough moment of unsupervised learning that we find ourselves in. As students grow closer to the craft, they will be able to better imagine the possibilities for its future. The course will also introduce critical perspectives from which to interrogate core issues such as biases in machine learning models and their implications.

Prerequisite is intermediate programming. No prior knowledge of machine learning is assumed. Python experience is welcome but not required. While this course is outlined for designers, the course material is a comprehensive introduction for anyone looking to dive into the world of unsupervised learning, including  engineers and scientific researchers.