William Herlands
I research machine learning and statistics focussing on public policy and public health. I hold a dual PhD in Machine Learning and Public Policy from Carnegie Mellon University where I was advised by Daniel Neill and Andrew Gordon Wilson.
From a methods perspective I develop novel ways to automate and generalize econometric tools. Specifically, I use anomalous pattern detection to identify natural experiments for causal inference in observational data. Additionally, using Bayesian nonparametrics and deep learning I develop spatiotemporal models to characterize and predict nuanced real-world phenomena.
Applying these methods I study dynamics of disease, crime, and transportation specifically within urban environments where strong spatiotemporal correlations require careful methodological consideration.
Ultimately, my research is oriented towards helping policy-makers create more targeted and effective interventions. Through better analytics and clearer communication I believe we can make government smarter, fairer, and more responsive.