Art and Machine Learning
60411 (Art) 10-615 (Computer Science)
T Th 1:30-4:20 PM
CFA 111 (Studio for Creative Inquiry)
Dr. Eunsu Kang (kangeunsu at cmu.edu)
Dr. Barnabas Poczos (bapoczos at cs.cmu.edu)
TA: Jonathan Dinu (jdinu at cs.cmu.edu)
Ars, the Latin origin of the word art, means Art and Science. These two fields, which have been separated for a long time, are joining back together in many areas. One of those junctions is where Art and Machine Learning meet. Art in recent years has been moving forward along with the rise of new technologies and scientific discoveries. Machine Learning (ML) is one of the most cutting edge advancements in Computer Science. The popularity and accessibility of frameworks such as Google’s Deep Dream system, Pikazo the neural style transfer, Kulitta AI Music Generation Framework, Deep Mind’s WaveNet, Sony’s Flow Machines, and recurrent neural network based language models brought great attention to the marriage of Art and ML methods. The number of ML applications that mimic famous artworks, e.g. The Next Rembrandt project, or even create original artworks such as the robot artist TAIDA’s paintings, is rapidly growing. Increasing number of artists are also attempting to use ML methods in their artworks.
This course is project-based and aims to introduce the crossroad of Art and Machine Learning to the broad range of students including both Art and Computer Science majors. We will offer the knowledge of examples, technologies, and issues that connect Art and Machine Learning to the students. Students will study example codes and produce creative applications/artworks using ML methods. Students do not need to have pre-existing knowledge of Machine Learning or experience of Art practice. Students are required to have basic understanding of Python and be open-minded, for example, open to learn the necessary mathematical background and open to discussions on conceptual development and artistic value of their projects.