CSE Colloquia Series-Kira Goldner

Nov 15
11:00 a.m.
Lopata Hall, Room 101

Mechanism Design for Social Good

Abstract

While algorithms serve as the backbone of many systems today, these systems often interact with real human beings. As such, we need these algorithms to be robust to potential manipulation by individuals with strategic interests. In this talk, we'll discuss how to use ideas from algorithmic mechanism design to do this: specifically, understanding incentives and improving social good.

First, I will talk about auctions used in carbon cap-and-trade, where each manufacturer must hold a licenses for every x units of carb they emit, in order to limit the carbon emissions per year. We'll understand why the current auctions used in carbon emissions schemes are vulnerable to strategic behavior. Then, within the same class of auctions that are used in practice, we will determine how to set the parameters so as to guarantee that societal welfare is close to the intended level, even under worst-case strategic behavior.

Second, we'll discuss a scenario in which an employer contracts with health insurance plans and aims to optimize its employees' utility. When can the employer use its regulatory power to benefit the employees (patients), as opposed to just running a free market being best? We characterize market equilibrium conditions, reducing the problem to one in revenue maximization.

Biography

Kira Goldner is a postdoctoral researcher at Columbia University hosted by Tim Roughgarden in the Computer Science department. Specifically, she is an NSF Mathematical Sciences Postdoctoral Research Fellow and a Data Science Institute Postdoctoral Fellow. She received her PhD in computer science and engineering from the University of Washington under the advisement of Anna Karlin, during which she was been supported by a 2017-19 Microsoft Research PhD Fellowship and a 2016-17 Google Anita Borg Scholarship. Her research is in algorithmic mechanism design, with work ranging from relaxing traditional behavioral and informational assumptions, maximizing revenue in settings motivated by practice, and applying mechanism design to social good. Kira is also a co-founder of the Mechanism Design for Social Good (MD4SG) initiative.