Double/debiased machine learning for treatment and structural parameters

By Victor Chernozhukov, D. Chetverikov, Mert Demirer, Esther Duflo, Christian Hansen, Whitney K. Newey, James M. Robins. 2017

3,836 citations recorded by the source corpus.

Subfields: Statistics and Probability, Statistical Learning Theory

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Statistics and Probability

In Statistical Learning Theory

Citations in the corpus

Causal Regularization (2019, background): “, [3], addressing different challenges, for instance, high dimensional confounders [4] or the case where some variables other than the common causes are observed [5], just to mention a few of them.”
The hardness of conditional independence testing and the generalised covariance measure (2018, methodology): “[14] uses related ideas to ours here to obtain 1/ √ n convergent estimates and confidence intervals for parameters such as average treatment effects in causal inference settings.”
Deep Neural Networks for Estimation and Inference: Application to Causal Effects and Other Semiparametric Estimands (2018, methodology): “s the recent literature in econometrics showing that the double robustness implies valid inference under weaker conditions 17 on the rst step nonparametric estimates (Belloni et al.,2014;Farrell,2015;Chernozhukov et al., 2018a). Indeed, it is these conditions that we verify using the results of Theo”
Double/De-Biased Machine Learning of Global and Local Parameters Using Regularized Riesz Representers (2018, methodology): “Using closed-form solutions for Riesz representers in several examples, Chernozhukov et al. (2016, 2018a) defined DML estimators in high dimensional settings and established their good properties.”
Approximate residual balancing: debiased inference of average treatment effects in high dimensions (2016, methodology): “…will help to highlight the statistical phenomena that make approximate residual balancing work, and to explain why—unlike the methods of Belloni et al. (2017), Chernozhukov et al. (2017) or Farrell (2015)—our method does not require consistent estimability of the treatment propensity function e.x/.”
Efficiently Breaking the Curse of Horizon in Off-Policy Evaluation with Double Reinforcement Learning (2019, background): “For example, metric entropy or Donsker assumptions can be avoided by using a cross-ﬁtting strategy (Chernozhukov et al., 2018).”
Learning When-to-Treat Policies (2019, positive): “8The idea of cross-fitting has gained growing popularity recently to reduce the effect of own-observation bias and to enable results on semiparametric rates of convergence using generic nuisance component estimates (Schick 1986; Chernozhukov et al. 2018; Athey and Wager 2020).”
Double Reinforcement Learning for Efficient Off-Policy Evaluation in Markov Decision Processes (2019, methodology): “This is achieved by following the cross-fold sample-splitting strategy developed by Chernozhukov et al. (2018).”
Efficient Policy Learning (2017, methodology): “As preliminaries, we note that the results of Chernozhukov et al. (2018b) immediately imply that, given Assumption 2, Â(1) is a semiparametrically efficient estimate of A(1), where we use “1” as shorthand for the “always treat” policy.”
Quasi-oracle estimation of heterogeneous treatment effects (2017, methodology): “as well as feasible analogues obtained by cross-fitting (Chernozhukov et al., 2017; Schick, 1986):”
Two-Step Estimation and Inference with Possibly Many Included Covariates (2018, background): “See also Chernozhukov et al. (2018) and Newey and Robins (2018) for related approaches.”
Machine Learning Advances for Time Series Forecasting (2020, neutral): “Arrange the set of tuples (𝑦 𝑡+ℎ ,𝒙 ′𝑡 ) , 𝑡 = ℎ + 1,…,𝑇 , in the form of a matrix 𝑽 of dimension (𝑇 − ℎ) × 𝑛 .”
Bayesian Regression Tree Models for Causal Inference: Regularization, Confounding, and Heterogeneous Effects (with Discussion) (2020, positive): “Chernozhukov et al. (2018) further pointed out that it is necessary to combine high-dimensional (e.g. machine learning) models for estimating the propensity score and outcome models in order to achieve √ N consistency in estimating the average treatment effect.”
On cross-validated Lasso in high dimensions (2016, neutral): “5) [0, 5] [6, 10] [11, 15] [16, 20] [21, 25] [26, p] [31, p] [36, p] (n, p)=(100, 40) 0.”
MALTS: Matching After Learning to Stretch (2018, background): “Machine learning approaches generalize regression approaches and can create models that are extremely flexible and predict outcomes accurately for both treatment and control groups (Hill 2011, Chernozhukov et al. 2018, Hahn et al. 2017). However, complicated regression methods lose the interpretabil”
Local Inference in Additive Models with Decorrelated Local Linear Estimator (2019, methodology): “Beyond the high-dimensional linear regression, [10] and [46] studied the inference procedure to the partial linear model.”