Hadi Elzayn
| HOME | RESEARCH | CV |
I am a Research Scientist at Meta. My research interests generally lie at the intersection of computer science with economics, policy, and societal concerns. In particular, I have focused on both theoretical and applied questions around algorithmic fairness, games, and auctions.
Prior to starting at Meta, I was a Postdoctoral Fellow at Stanford University’s RegLab,
where I was advised by Dan Ho and Jacob Goldin; there, I worked on algorithmic and equity concerns in applying machine learning to closing the tax gap.
Prior to the RegLab, I did my PhD at the University of Pennsylvania, where I was advised by Michael Kearns; my dissertation was called Essays in Algorithms, Markets, and Society. During my PhD, I interned at Facebook, Microsoft Research, and the Federal Reserve Bank of Philadelphia, and before graduate school I worked for several years at TGG Group, a Chicago-based consulting firm applying econometrics and behavioral economics to solve problems for business, non-profits, and government. I received my B.A. in Mathematics and Economics from Columbia University.
News
- (September 2024) Our paper, Measuring and Mitigating Racial Disparities in Tax Audits is forthcoming in the February issue of the Quarterly Journal of Economics. The paper can be found online here.
- (May 2024) Our paper, Measuring Fairness in the U.S. Mortgage Market is now available as a working draft. An interactive online dashboard is also available and allows the viewer to explore mortgage market fairness trends for themselves along various measures and axes.
- (February 2024) Our paper, Estimating and Implementing Conventional Fairness Metrics With Probabilistic Protected Features, was accepted to IEEE's Secure and Trustworthy Machine Learning.
- (January 2023) We released our working paper Measuring and Mitigating Racial Disparities in Tax Audits to media, legislator, and policymaker engagement. The paper develops a methodology to bound disparity using probabilistic protected features and applies it to the tax audit setting, finding that Black taxpayers are audited at 2.9 to 4.7 times the rate of non-Black taxpayers.
- (March 2022) I started as a Research Scientist at Meta.
- (Jan 2021) I started my Postdoc at Stanford University in the RegLab.
hads@fb.com
hselzayn@law.stanford.edu
13) Measuring and Mitigating Racial Disparities in Tax Audits
Quarterly Journal of Economics. H. Elzayn, E. Smith, T. Hertz, A. Ramesh, R. Fisher, D. Ho, J. Goldin. (Publication Version) .
Twitter Explainer: Overview | Methodology
Engagement: NYT | NPR Morning Edition | House Ways & Means | NPR 1A
| USA Today | WWJ News Radio | Tax Notes | Axios | Slate Money
| Mother Jones | The Hill | NPR Cincinnati Edition | Tax Chats podcast | ABC11 Raleigh-Durham | Senate Finance Committee |
Warren Letter | Werfel Letter | Supreme Court
12) Measuring Fairness in the U.S. Mortgage Market.
Work-in-progress. H. Elzayn, R. Kobler, S. Freyaldenhoven, M. Shin. (Slides.)
11) Estimating and Implementing Conventional Fairness Metrics with Probabilistic Protected Features
IEEE SaTML 2024. H. Elzayn, E. Black, P. Vossler, N. Jo, J. Goldin, D. Ho. (Arxiv Version.).
10) Optimal Data Acquisition with Privacy-Aware Agents
IEEE SaTML 2023 (Best Paper Award). R.Cummings, H. Elzayn, E. Pountourakis, V. Gkatzelis, J. Ziani. (ArXiv Version.) (Publication Version.)
9) Algorithmic Fairness and Vertical Equity: Income Fairness with Tax Audit Models
ACM FAccT 2022. E. Black,H. Elzayn, A. Chouldechova, J. Goldin, D Ho. (Publication Version.)
8) Equilibria in Auctions with Ad Types
TheWebConf 2022. H. Elzayn, R. Colini-Baldeschi, B. Lan, O. Schrijvers. (Arxiv Version.) (Publication Version.)
7) Algorithms and Learning for Fair Portfolio Design
EC 2021. E. Diana, T. Dick, H. Elzayn , M. Kearns, A. Roth, Z. Schutzman, S. Sharifi-Malvajerdi, and J. Ziani.
(Publication Version).
(Arxiv Version).
6) Differentially Private Call Auctions and Market Impact.
EC 2020. E. Diana, H. Elzayn, M. Kearns, A. Roth, S. Sharifi-Malvajerdi, and J. Ziani.
(Publication Version) (Arxiv Version)
5) The Effects of Competition and Regulation on Error Inequality in Data-Driven Markets.
ACM FAT* 2020. H. Elzayn, B. Fish.
(Publication Version)
Also presented as an oral/poster presentation at the NeuRIPS 2019 AI for Social Good Workshop. AISG Best poster winner.
4) Equilibrium Characterization for Data-Acquisition Games
IJCAI 2019.
J. Dong, H. Elzayn, S. Jabbari, M. Kearns, Z. Schutzman.
(Publication Version) (Arxiv Version)
3) Price of Privacy in the Keynesian Beauty Contest.
EC 2019. H. Elzayn, Z. Schutzman.
(Publication Version) (Arxiv Version)
2) Hidden Information, Teamwork, and Prediction in Trick-Taking Card Games.
RLDM 2019 (Extended Abstract). H. Elzayn, M. Fereydounian, M. Hayhoe, H. Kumar.
(RLDM Version) (Working Paper Version)
1) Fair Algorithms for Learning in Allocation Problems.
ACM FAT* 2019. H. Elzayn, S. Jabbari, C. Jung, M. Kearns, S. Neel, A. Roth, Z. Schutzman
(Publication Version) (Arxiv Version)