V
Victor Chernozhukov
Researcher at Massachusetts Institute of Technology
Publications - 374
Citations - 25283
Victor Chernozhukov is an academic researcher from Massachusetts Institute of Technology. The author has contributed to research in topics: Estimator & Quantile. The author has an hindex of 73, co-authored 370 publications receiving 20588 citations. Previous affiliations of Victor Chernozhukov include Amazon.com & New Economic School.
Papers
More filters
Posted Content
RieszNet and ForestRiesz: Automatic Debiased Machine Learning with Neural Nets and Random Forests.
TL;DR: This article proposed a multi-tasking neural net debiasing method with stochastic gradient descent minimization of a combined Riesz representer and regression loss, while sharing representation layers for the two functions.
Posted Content
SortedEffects: Sorted Causal Effects in R
TL;DR: SortedEffects as mentioned in this paper is a R package that implements the estimation and inference methods therein and provides tools to visualize the results, including tools for visualizing the results of sorted effects.
Journal ArticleDOI
Rearranging Edgeworth-Cornish-Fisher Expansions
TL;DR: In this paper, the authors apply a regularization procedure called increasing rearrangement to monotonize Edgeworth and Cornish-Fisher expansions and any other related approximations of distribution and quantile functions of sample statistics.
Posted Content
Learning Financial Network with Focally Sparse Structure
TL;DR: In this article, a double regularized high-dimensional generalized method of moments (GMM) framework is proposed to obtain high-quality estimator for parameters of interest, and this framework also facilitates the inference.
Journal ArticleDOI
Posterior Inference in Curved Exponential Families Under Increasing Dimensions
TL;DR: In this paper, the authors study the large sample properties of posterior-based inference in the curved exponential family under increasing dimensions and establish conditions under which the posterior distribution is approximately normal, which implies various good properties of estimation and inference procedures based on the posterior.