•   When: Friday, February 12, 2021 from 11:00 AM to 12:00 PM
•   Speakers: Shahin Jabbari, CRCS Postdoctoral Fellow, School of Engineering and Applied Sciences, Harvard
•   Location: ZOOM
•   Export to iCal

 Abstract

Human-designed algorithms have been deployed in many consequential settings such as hiring and lending. There has been mounting empirical evidence demonstrating the discriminatory behaviors of algorithms in such settings. On the flip side, algorithmic systems are often used by humans who can act strategically or launch adversarial attacks to benefit from the vulnerabilities of such systems. 

In the first part of my talk, I discuss my work in the field of algorithmic fairness which is centered around designing provably fair algorithmic frameworks. I show how such frameworks can be used by policy/lawmakers to systematically study the possible trade-offs between fairness violation and performance as well as highlight the limitations and challenges of fair algorithmic decision making in practice. In the second part of the talk, I discuss my work in the field of game theory to design provably unexploitable algorithms in settings motivated by applications such as finance and security. 

Bio

Shahin Jabbari is a CRCS postdoctoral fellow in the school of Engineering and Applied Sciences at Harvard. Shahin completed his Ph.D. at the University of Pennsylvania in 2019 where he was advised by Michael Kearns. Shahin's research interests span across many areas in artificial intelligence including algorithmic fairness, game theory, and multi-agent systems. Shahin has published in top-tier conferences in artificial intelligence (ICML/NeurIPS/AAAI/IJCAI) and has received Best Paper Award from the Conference on Decision and Game Theory for Security, and the German Conference in Artificial Intelligence.