Jason Poulos

Postdoctoral Fellow,
Harvard Medical School

I recently completed a Postdoctoral Fellowship in Data Science at the Department of Health Care Policy at Harvard Medical School and am now a postdoc at Brigham and Women's Hospital and Harvard Medical School. In this new role, I am working on projects involving the development of predictive models in medicine, leveraging longitudinal data to analyze electronic medical records (EMR). The focus is on developing new techniques to better predict patients at risk for diabetes treatment failure in both outpatient and inpatient settings.

I earned a Ph.D. in Political Science with a Designated Emphasis in Computational Science and Engineering from UC Berkeley, where I was also an NSF Graduate Research Fellow.

My research portfolio centers on the application of machine learning—particularly deep learning—to improve causal inference in observational studies across social sciences and health data sciences. I have published work in various domains, including missing data imputation, handwritten text recognition, and adversarial machine learning.

Preprints

Targeted Learning in Observational Studies with Multi-Level Treatments: An Evaluation of Antipsychotic Drug Treatment Safety.
(with Marcela Horvitz-Lennon, Katya Zelevinsky, Thomas Huijskens, Pooja Tyagi, Jiaju Yan, Jordi Diaz, Tudor Cristea-Platon, and Sharon-Lise Normand).
[Code]

State-Building through Public Land Disposal? An Application of Matrix Completion for Counterfactual Prediction.
[Code]

Articles

Antipsychotics and the Risk of Diabetes and Death among Adults with Serious Mental Illnesses.
(with Sharon-Lise Normand, Katya Zelevinsky, John Newcomer, Denis Agniel, Haley Abing, and Marcela Horvitz-Lennon).
Psychological Medicine, 2023.

[Code]

Adversarial Machine Learning: Bayesian Perspectives.
(with David Rios Insua, Roi Naveiro, and Víctor Gallego).
JASA, 2023.

[arXiv] [Code]

Gender Gaps in Frontier Entrepreneurship? Evidence from 1901 Oklahoma Land Lottery Winners.
Journal of Historical Political Economy, 2023.
[arXiv] [Code]

Are Deep Learning Models Superior for Missing Data Imputation in Surveys? Evidence from an Empirical Comparison.
(with Zhenhua Wang, Olanrewaju Akande, and Fan Li).
Survey Methodology, 2022.

[arXiv] [Code and Supplementary Material]

RNN-Based Counterfactual Prediction, with an Application to Homestead Policy and Public Schooling.
(with Shuxi Zeng). Journal of the Royal Statistical Society, 2021.
[arXiv] [Code]

Amnesty Policy and Elite Persistence in the Postbellum South: Evidence from a Regression Discontinuity Design. Journal of Historical Political Economy, 2021.
[arXiv] [Code]

Character-Based Handwritten Text Transcription with Attention Networks.
(with Rafael Valle). Neural Computing & Applications, 2021.
[arXiv] [Code]

Estimating Population Average Treatment Effects from Experiments with Noncompliance.
(with Kellie Ottoboni). Journal of Causal Inference, 2020.
[arXiv] [Code]

Land Lotteries, Long-Term Wealth, and Political Selection. Public Choice, 2019.
[Code]

Missing Data Imputation for Supervised Learning.
(with Rafael Valle). Applied Artificial Intelligence, 2018.
[arXiv] [Supplementary Material] [Code]