Timezone: »

 
Spotlight
Coordinated Double Machine Learning
Nitai Fingerhut · Matteo Sesia · Yaniv Romano

Tue Jul 19 01:15 PM -- 01:20 PM (PDT) @ Ballroom 3 & 4
in MISC: Causality »

Double machine learning is a statistical method for leveraging complex black-box models to construct approximately unbiased treatment effect estimates given observational data with high-dimensional covariates, under the assumption of a partially linear model. The idea is to first fit on a subset of the samples two non-linear predictive models, one for the continuous outcome of interest and one for the observed treatment, and then to estimate a linear coefficient for the treatment using the remaining samples through a simple orthogonalized regression. While this methodology is flexible and can accommodate arbitrary predictive models, typically trained independently of one another, this paper argues that a carefully coordinated learning algorithm for deep neural networks may reduce the estimation bias. The improved empirical performance of the proposed method is demonstrated through numerical experiments on both simulated and real data.

Keywords: DL: Algorithms · Miscellaneous Aspects of Machine Learning · MISC: Causality

Author Information

Nitai Fingerhut (Technion)
Matteo Sesia (University of Southern California)
Yaniv Romano (Technion---Israel Institute of Technology)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors