English

Orthogonal Machine Learning: Power and Limitations

Machine Learning 2018-08-03 v6 Econometrics Statistics Theory Machine Learning Statistics Theory

Abstract

Double machine learning provides n\sqrt{n}-consistent estimates of parameters of interest even when high-dimensional or nonparametric nuisance parameters are estimated at an n1/4n^{-1/4} rate. The key is to employ Neyman-orthogonal moment equations which are first-order insensitive to perturbations in the nuisance parameters. We show that the n1/4n^{-1/4} requirement can be improved to n1/(2k+2)n^{-1/(2k+2)} by employing a kk-th order notion of orthogonality that grants robustness to more complex or higher-dimensional nuisance parameters. In the partially linear regression setting popular in causal inference, we show that we can construct second-order orthogonal moments if and only if the treatment residual is not normally distributed. Our proof relies on Stein's lemma and may be of independent interest. We conclude by demonstrating the robustness benefits of an explicit doubly-orthogonal estimation procedure for treatment effect.

Keywords

Cite

@article{arxiv.1711.00342,
  title  = {Orthogonal Machine Learning: Power and Limitations},
  author = {Lester Mackey and Vasilis Syrgkanis and Ilias Zadik},
  journal= {arXiv preprint arXiv:1711.00342},
  year   = {2018}
}
R2 v1 2026-06-22T22:32:58.499Z