DMS518, Reg.#24253, Wed 09:00-12:50
The focus of the current version of the computational media seminars/workshops is machine learning (ML). The goal of the seminar is twofold: One, to understand important principles of numerical representation and ML and two, to explore machine learning as a set of cultural artifacts. This second approach is a response to recent media theory (cultural techniques) that reevaluates the role of technical processes as 'material conditions that constitute semantics' (Siegert2013).
Machine learning produces software systems that improve their own performance over time. ML is a branch of artificial intelligence, and materially responsible for many of the recent spectacular advances in automation that impact everyday life, financial markets, transportation and social media. Students will be exposed to fundamentals of supervised and unsupervised ML systems and learn how to 'train a machine' on data such that it can reproduce patterns detected in a given dataset.
This course will be paced for graduate students without prior exposure to ML, but with exposure to general purpose media arts code development. Our programming environment will be Python with Scikit and Tensorflow. Students will be asked to identify and explore a culturally relevant aspect of ML as a semester project. Teamwork encouraged.