Title: New Approaches to Multi-Objective Optimization with Applications to Fairness and Online Learning
Speaker: Jai Moondra
Venue: IEOR Seminar Room
Date and time: 1 January 2025 (Wednesday), 3:30-4:30 p.m.
Abstract: Real-world optimization problems often involve balancing competing objectives, such as fairness objectives in resource allocation problems or the trade-off between regret and runtime in online learning problems. Traditional approaches rely on predefined composite objectives, which (i) require fixing a composite objective a priori (ii) can pose challenges to policymakers in selecting appropriate trade-offs and (iii) lead to unintended biases in outcomes. In this talk, I will present new approaches for addressing these challenges. First, motivated by organ transplantation policies, we introduce "fairness portfolios" for optimization problems, which are small sets of solutions such that any fairness objective is approximately satisfied by some solution in the portfolio. We study the trade-off between portfolio size and solution quality, giving new approximation algorithms and a new primal-dual counting technique. I will also discuss their practical applications in the context of facility location and other policy problems.
Next, I will discuss the trade-off between regret and runtime in online learning problems. Drawing from applications in recommendation systems, we demonstrate how combining discrete and continuous techniques over submodular base polytopes can significantly reduce runtimes of optimal-regret algorithms such as Mirror Descent.
Bio: Jai Moondra is a PhD student in the School of Computer Science at Georgia Tech, advised by Dr. Swati Gupta (MIT) and Dr. Mohit Singh (Georgia Tech). He earned his B.Tech. in Computer Science from IIT Delhi in 2019. His research focuses on discrete optimization, with applications to algorithmic fairness, combinatorial learning, and quantum computing. His thesis explores algorithm design for balancing competing objectives, particularly in fairness and online learning contexts.