Assistant Professor, Stanford University
3 papers at NeurIPS 2025
We show that the statistical optimality of estimation methods for causal inference depend in a surprising way on the distribution of the treatment noise.
We estimate average treatment effects, when the effect is moderated by extreme events and we are interested in the average effect in the extreme, which is rarely occurring in the observed data.
We offer statistically robust methods for preference learning that leverage response time in the estimation of rewards to yield large improvements in statistical efficiency.
