10-737 Creative AI
6–12 units, Carnegie Mellon University
Artificial intelligence AI systems now generate authentic paintings, compose music pieces, and find outofbox solutions to reallife problems in our world. Creativity, which was considered to be a moon shot for AI, does not seem to be too far any more. Is that true Are we close to see creative AI The answer is yes and no. We are moving closer with meaningful developments in Machine Learning, however there are several questions to be explored further to achieve the creative AI. What kind of creativity we want to represent How do we translate creativity into what machines can understand How do we design ML algorithms to be more creative This course is where we explore these questions through seminars and projects. Our goal is to design computational models that present the very possibility of the creative AI. The instructors who are specialized in Machine Learning Art and Robotics lead this course together. We introduce related examples and possible methods including multimodal datadriven learning, learning from demonstration, and combined learning from data and human demonstrations. Students are welcome to bring in their expertise and passion from diverse backgrounds to explore this topic together.
Prerequisites: none listed
Courseweave needs JavaScript to show sections, workloads and planning tools.