Combining techniques from causal inference, machine learning, and semiparametric statistics to develop efficient, robust estimators of treatment effects in experimental and observational study settings.
Discovering population segments (subgroups) and evaluating the population-level effects of learned dynamic treatment policies using causal machine learning.
Developing nonparametric estimators with novel properties using undersmoothing (sieve estimation) principles with semiparametric efficiency theory.
Efficient estimation of the causal impacts of vaccine-induced immune responses in complex trials.
Efficient estimation of functional target parameters based on the highly adaptive lasso minimum loss estimator (HAL-MLE).
Defining novel, more flexible causal effects for mediation analysis, primarily using the formalism of stochastic interventions.
Estimating the causal effects of stochastic treatment regimes, including conditional density estimation and two-phase sampling corrections.