ART543, Tues 18:00-20:30

Design means different things to different disciplines; design of interfaces for industrial equipment follows a different logic than the design of a website. Yet all design deals with reconciling oppositional needs. This is one reason why design problems are often difficult for computers to deal with. In the past, computer aided efficiency tools have altered specific components of design practices. CAD softwares and satellite image analysis have changed the way architecture operates, for example. A.I systems and in particular machine learning are now impacting along multiple vectors many more aspects of design practice previously reserved for human action.

This course caters to designers of all flavors interested in understanding the significance of new computational practices, in particular machine learning, for design at large.

The course will begin with an overview of computational thinking at large and try to understand how computer processes impact the perception of what a problem is in the first place, and how a solution is informed by problem definition. Thereafter we will survey several currently used machine learning techniques (classification and generative systems) and use case studies to understand how these algorithms are applied in specific contexts. From these case studies we will attempt to understand how designers use computational thinking to make specific computational methods effective, including understanding the many ways bias creeps into the design of machine learning systems.

Students will tinker with machine learning data preparation and algorithm training, and collaboratively create an image classification system.

course syllabus

course resources

« back