Innovations in machine learning have led to many engineering breakthroughs, from real time voice recognition to automatic categorization (and in some cases production) of news stories. Since these techniques are at their essence novel ways to work with data, they also have the potential to transform how the empirical sciences produce knowledge. This course focuses on applications of machine learning. To facilitate such applications, it addresses a few questions about these new techniques: (i) How do they work and what kinds of statistical guarantees can be made about their performance? (ii) How can they be used to test theories or create social policies? (iii) What is the relationship to causal inference?

The goal is to create a working understanding of where and how these tools can be profitably applied, and so the focus is on conceptual understanding; we will not cover mechanics either implementation or mathematical – students will be expected to know that material or learn it on their own.