Explores the interactions between machine learning and people: the ones that create predictive models, leverage them to build data-driven systems, and consume these systems. How can we predict unintended consequences that rise from biases in design choices and data? How can we design systems backed by smart agents when the technological capabilities are not well-understood? How can we offer personalization while protecting privacy?

To begin answering these questions, students will read and discuss recent papers from human computer-interaction, product design, cognitive science, machine learning, graphics, vision, and natural language processing. Students will work in teams on a multi-week project to build data-driven tools to solve real-world design problems. Practical data mining and machine learning knowledge is emphasized: crowdsourcing and web scraping, model and feature selection, parameter tuning. The course has no formal prerequisites, but students should be algorithmically and programmatically mature.