1: \begin{abstract}

2: We propose an approach to better inform treatment decisions at an

3: individual level by adapting recent advances in average treatment

4: effect estimation to \emph{conditional} average treatment effect estimation.

5: Our work is based on doubly robust estimation methods, which combine

6: flexible machine learning tools to produce efficient effect estimates

7: while relaxing parametric assumptions about the data generating process.

8: Refinements to doubly robust methods have achieved faster convergence

9: by incorporating \emph{3-way cross-fitting}, which entails dividing

10: the sample into three partitions, using the first to estimate the

11: conditional probability of treatment, the second to estimate the conditional

12: expectation of the outcome, and the third to perform a first order

13: bias correction step.

14: 

15: Here, we combine the approaches of 3-way cross-fitting and pseudo-outcome

16: regression to produce personalized effect estimates. We show that

17: this approach yields fast convergence rates under a smoothness condition

18: on the conditional expectation of the outcome.

19: \end{abstract}

20: