The main goal of the Fair and Explainable Decision Making Lab is to study topics involving computational fair division, computational social choice, algorithmic game theory and algorithmic transparency. We are exploring a variety of projects on fairness and explainability. These include works on the theoretical foundations of fair allocation, how to correctly explain algorithmic decisions, and the (re)design of fair allocation mechanisms with provable fairness and efficiency guarantees.
Project Description:
Our lab develops methods for distributing resources among agents who have different preferences over them. Consider the problem of distributing chores among a group of agents - for example, dividing household chores among roommates. How should the chores be divided? Our goal is identify algorithms that can distribute chores among agents in a reasonable manner, and implement them.
Learning Objectives:
Student should be able to
explore different justice criteria by which we evaluate outcomes in the chore division domain.
create algorithmic frameworks for chore division
analyze the algorithms' properties, such as their runtime, satisfaction of justice guarantees, and interpretability.
create user interfaces that collect user preferences, translate them to meaningful inputs, and output chore allocations.
Skills to learn:
A good understanding of algorithmic analysis. For example, good performance in COMPSCI 311, or other theory-oriented CS classes is a good indicator of the skills required for this project.