Easton Huch is a postdoctoral researcher at Johns Hopkins Carey Business School. Easton's primary research area is causal inference methodology, especially in dynamic treatment regimes. Example application areas include Al systems employing reinforcement learning and/or bandit algorithms, and micro-randomized trials in mobile health research. In particular, much of Easton's research in this area develops robust randomization-based methods.
Prior to his PhD, Easton worked as a data scientist at a tech company, where he developed an interest in quantitative marketing and business analytics. Easton is broadly interested in statistical methodology for the business problems he encountered as a data scientist, including online experimentation, customer lifetime value modeling, and business forecasting.
Education
- PhD, Statistics, University of Michigan
- MS, Statistics, Brigham Young University
- BS, Economics & Statistics, Brigham Young University
Research
Selected publications
- Huch, Easton, Jieru Shi, Madeline R. Abbott, Jessica R. Golbus, Alexander Moreno, and Walter H. Dempsey (2024). "ROME: A Robust Mixed-effects Bandit Algorithm for Optimizing Mobile Health Interventions" In: Advances in Neural Information Processing Systems. Vol. 37, pp.128280-128329.
Working Papers
- "Overall Evaluation Criteria for A/B Tests: A Nonparametric Meta-analysis Approach" (with Stephanie A. Nail, Terran Melconian, Xan Vongsathorn, and Fred Feinberg)
- "Robust Bayesian Inference of Causal Effects via Randomization Distributions" (with Walter Dempsey and Fred Feinberg)
- "Computationally Efficient Models for Count Data with Varying Levels of Dispersion" (with Mason Ferlic, Kimberly Sellers, and Candace Berrett)