New Media Installation: Art that Learns

Doherty Hall C316 Tue, Thu 01:30-4:20 PM

Instructor: Osman Khan & Carlos Guestrin
Email: okhan@andrew.cmu.edu, guestrin@cs.cmu.edu

Office Hours:
Osman Khan – Doherty Hall D311 Thursday 4:30-6 PM
Carlos Guestrin –

Description
A project course where we design and build art installations using machine learning.

It is hard to imagine anything more fascinating than automated systems that improve their own performance. In the recent years, machine learning techniques have allowed us to make such systems a reality.  The impact of these techniques can be seen everywhere: from movie suggestions in Netflix, through computer vision systems that can recognize objects from images, to computational biology algorithms to understand our genome.

In this course, we will explore a new dimension to machine learning: can we create aesthetically appealing interactive art installations, and even question our understanding of world, by using machine learning techniques?  Our “media” will encompass data sources from images, video, text documents, graphs, sounds and music, and even sensors Machine learning can then yield new thought-provoking interpretations from these data.

The course will culminate with an exposition of the students’ project at the Children’s Museum of Pittsburgh

This is an art studio before it is a computer art studio. We want to make art using computers and technologies, not simply learn how to use technology. This means that learning the skills necessary to become an artist (in any medium): vision, invention/imagination, motivation, self-direction and self reliance, self criticism/self assessment are central to this class. Mastery of tools is good, but it is not an end in itself. We expect proficiency with the tools and inventive application of them to your creative goal.
A course material fee is required. In addition, students can expect to purchase individual items that are unique to their projects.

Skills
Exposure to a range of technologies and software with emphasize on Machine Learning technologies and methodologies. Enhancing ability to present work. Increasing vocabulary appropriate for discussing visual work. Introduction to relevant historical precedents.

Learning Objectives
Upon satisfactory completion, student will be able to:
•    Engage new media and machine learning technologies and methodologies for art production and deployment.
•    Employ historical and contemporary contexts around machine learning methodologies and contemporary art to inform their own art practice.
•    Develop projects involving new media technologies and machine learning methodologies to help inform their own art practice.