Instrumental variable analysis is a powerful tool for estimating causal
effects when randomization or full control of confounders is not possible. The
application of standard methods such as 2SLS, GMM, and more recent variants are
significantly impeded when the causal effects are complex, the instruments are
high-dimensional, and/or the treatment is high-dimensional. In this paper, we
propose the DeepGMM algorithm to overcome this. Our algorithm is based on a new
variational reformulation of GMM with optimal inverse-covariance weighting that
allows us to efficiently control very many moment conditions. We further
develop practical techniques for optimization and model selection that make it
particularly successful in practice. Our algorithm is also computationally
tractable and can handle large-scale datasets. Numerical results show our
algorithm matches the performance of the best tuned methods in standard
settings and continues to work in high-dimensional settings where even recent
methods break.
Description
[1905.12495] Deep Generalized Method of Moments for Instrumental Variable Analysis
%0 Conference Paper
%1 bennett2019generalized
%A Bennett, Andrew
%A Kallus, Nathan
%A Schnabel, Tobias
%D 2019
%K generative-models neurips2019
%T Deep Generalized Method of Moments for Instrumental Variable Analysis
%U http://arxiv.org/abs/1905.12495
%X Instrumental variable analysis is a powerful tool for estimating causal
effects when randomization or full control of confounders is not possible. The
application of standard methods such as 2SLS, GMM, and more recent variants are
significantly impeded when the causal effects are complex, the instruments are
high-dimensional, and/or the treatment is high-dimensional. In this paper, we
propose the DeepGMM algorithm to overcome this. Our algorithm is based on a new
variational reformulation of GMM with optimal inverse-covariance weighting that
allows us to efficiently control very many moment conditions. We further
develop practical techniques for optimization and model selection that make it
particularly successful in practice. Our algorithm is also computationally
tractable and can handle large-scale datasets. Numerical results show our
algorithm matches the performance of the best tuned methods in standard
settings and continues to work in high-dimensional settings where even recent
methods break.
@inproceedings{bennett2019generalized,
abstract = {Instrumental variable analysis is a powerful tool for estimating causal
effects when randomization or full control of confounders is not possible. The
application of standard methods such as 2SLS, GMM, and more recent variants are
significantly impeded when the causal effects are complex, the instruments are
high-dimensional, and/or the treatment is high-dimensional. In this paper, we
propose the DeepGMM algorithm to overcome this. Our algorithm is based on a new
variational reformulation of GMM with optimal inverse-covariance weighting that
allows us to efficiently control very many moment conditions. We further
develop practical techniques for optimization and model selection that make it
particularly successful in practice. Our algorithm is also computationally
tractable and can handle large-scale datasets. Numerical results show our
algorithm matches the performance of the best tuned methods in standard
settings and continues to work in high-dimensional settings where even recent
methods break.},
added-at = {2019-12-11T00:34:14.000+0100},
author = {Bennett, Andrew and Kallus, Nathan and Schnabel, Tobias},
biburl = {https://www.bibsonomy.org/bibtex/2f05bbf6ad1101a3947a3b32947b2be5a/kirk86},
description = {[1905.12495] Deep Generalized Method of Moments for Instrumental Variable Analysis},
interhash = {f697c52cc669de5682095acf8c75e12c},
intrahash = {f05bbf6ad1101a3947a3b32947b2be5a},
keywords = {generative-models neurips2019},
note = {cite arxiv:1905.12495},
timestamp = {2019-12-11T00:34:14.000+0100},
title = {Deep Generalized Method of Moments for Instrumental Variable Analysis},
url = {http://arxiv.org/abs/1905.12495},
year = 2019
}