EECS Special Seminar: Hoda Heidari "Distributive Justice for Machine Learning"

Automated decision-making tools are increasingly in charge of making high-stakes decisions for people. These tools can exhibit undesirable biases and disparately harm already-disadvantaged and marginalized groups and individuals. I showcase how tools and methods from computer science, economics, and political philosophy can be brought together to define, measure, and mitigate algorithmic unfairness in a principled manner. I address two key questions:
▪ Given the appropriate notion of harm/benefit, how should we measure and bound unfairness? Existing notions of fairness focus on defining conditions of fairness, but they do not offer a proper measure of unfairness. I propose inequality indices from economics as a unifying framework for measuring unfairness--both at the individual- and group-level. I utilize cardinal social welfare as an alternative measure of fairness behind a veil of ignorance and a practical method for bounding inequality.
▪ Fairness as equality of what? I offer a framework to think about this question normatively by mapping recently-proposed group-level notions of fairness to models of equality of opportunity. This mapping unifies existing notions and allows us to spell out the moral assumptions underlying each one of them. Our framework also puts forward a path to bring human judgment in the loop and construct context-aware mathematical notions of fairness in a more democratic fashion.

Bio: Hoda Heidari is currently a Postdoctoral Associate at the Department of Computer Science at Cornell University, where she collaborates with Professors Jon Kleinberg, Karen Levy, and Solon Barocas through the AIPP (Artificial Intelligence, Policy, and Practice) initiative.
Hoda’s research is broadly concerned with the societal and economic aspects of Artificial Intelligence, and in particular, the issues of unfairness and discrimination through Machine Learning. She utilizes tools and methods from computer science (algorithms, AI, and ML) and social sciences (economics and political philosophy) to quantify and mitigate the inequalities that arise when socially consequential decisions are automated.