Time: Tuesday/Thursday 11:15-12:30pm in person

Machine learning has stirred tremendous excitement in research and popular science, as it is being applied to problems of all shapes and sizes. This involves learning and predicting things that humans do naturally (see, hear, understand language), signals we give off unconsciously as well as decisions and ideas we have (make decisions, evaluate options, think creatively). These are captured under the broad banner of human-centered machine learning (HCML), an emergent research and practical area combining human interaction with machine learning and studying the impacts of systems in the world.

In this class, we will study what human-centered machine learning is, how this idea is built and applied in technical systems, and critiques of machine learning from other domains that question if ML can be human-centered. There will be a strong emphasis in this seminar on reading and discussing current research on HCML, from technical, critical, and social perspectives. Later, we will put these principles into practice in applying what we have learned in a group project.

This class is intended to be interdisciplinary, with students drawn from CS and STEM as well as social sciences who use ML for their research and are thinking through essential questions on the impacts of ML in society.

Students who complete this course will be able to:

  • Identify and categorize the sociotechnical, human-centered approach to machine learning a
  • Describe the state-of-the-art in research to make machine learning more human-centered, triangulated along values embedded in algorithms
  • Critique machine learning and its ability to be applied to problems with social and ethical concerns, and propose solutions to mitigate these problems
  • Design and execute a project focused on topics of human-centered machine learning

Prereqs and Necessary Background

There are some base prerequisites that are expected for students because of the technical content of the readings. CSCI 5521 (Introduction to Machine Learning), or any other course that cover an introduction to machine learning. This class will not cover fundamentals or engineering.

Main Course Components

  • Class Participation: 15%
  • Reading Responses (called RRs): 25%
  • Leading Discussion: 10%
  • Project: 50%