This 5-day workshop is designed to introduce students to the challenges of fairness when applying machine learning and some of the research aimed at addressing these challenges. We will examine case studies of machine learning systems that have been demonstrated to be unfair, analyze and critique methods researchers have proposed to measure and mitigate ML unfairness, and explore the role of ML in society as a whole—including the context of systemic racism and other forms of discrimination historically and currently prominent.

With the widespread societal adoption of machine learning and related technologies, concerns about the ethics of the decisions made by these systems are becoming prevalent. Machine learning systems have been measured to repeat or amplify historical biases in data and even create new patterns of unjust discrimination. Recent research has accelerated to meet the critical need to understand these dangers and whether they can be mitigated.

Prerequisites: Students should have previously studied machine learning and have a good grasp of concepts such as classification, regression, training data, generalization error, and cross-validation.

Instructor: Dr. Bert Huang

Zoom link:

https://tufts.zoom.us/j/99884239223?pwd=MkdOMEN3cnFRa24zbThTVWVlVVkwQT09