-
Kernel Debiased Plug-in Estimation: Simultaneous, Automated Debiasing without Influence Functions for Many Target Parameters
Authors:
Brian Cho,
Yaroslav Mukhin,
Kyra Gan,
Ivana Malenica
Abstract:
When estimating target parameters in nonparametric models with nuisance parameters, substituting the unknown nuisances with nonparametric estimators can introduce ``plug-in bias.'' Traditional methods addressing this suboptimal bias-variance trade-off rely on the \emph{influence function} (IF) of the target parameter. When estimating multiple target parameters, these methods require debiasing the…
▽ More
When estimating target parameters in nonparametric models with nuisance parameters, substituting the unknown nuisances with nonparametric estimators can introduce ``plug-in bias.'' Traditional methods addressing this suboptimal bias-variance trade-off rely on the \emph{influence function} (IF) of the target parameter. When estimating multiple target parameters, these methods require debiasing the nuisance parameter multiple times using the corresponding IFs, which poses analytical and computational challenges. In this work, we leverage the \emph{targeted maximum likelihood estimation} (TMLE) framework to propose a novel method named \emph{kernel debiased plug-in estimation} (KDPE). KDPE refines an initial estimate through regularized likelihood maximization steps, employing a nonparametric model based on \emph{reproducing kernel Hilbert spaces}. We show that KDPE: (i) simultaneously debiases \emph{all} pathwise differentiable target parameters that satisfy our regularity conditions, (ii) does not require the IF for implementation, and (iii) remains computationally tractable. We numerically illustrate the use of KDPE and validate our theoretical results.
△ Less
Submitted 2 June, 2024; v1 submitted 14 June, 2023;
originally announced June 2023.
-
Cauchy-like Criterion for Differentiability of Functions of Several Variables
Authors:
Yurii V. Mukhin,
Nataliya D. Kundikova
Abstract:
In this paper, several differentiability criteria for real functions of multiple variables in n-dimensional Euclidean space are considered. Simple and easy-to-use Cauchy-like criterion is formulated and proven. Relaxed sufficient conditions for differentiability that do not require continuity of all partial derivatives are suggested. Generalization of the Cauchy-like criterion for functions on cro…
▽ More
In this paper, several differentiability criteria for real functions of multiple variables in n-dimensional Euclidean space are considered. Simple and easy-to-use Cauchy-like criterion is formulated and proven. Relaxed sufficient conditions for differentiability that do not require continuity of all partial derivatives are suggested. Generalization of the Cauchy-like criterion for functions on cross products of normed vector spaces (not necessarily Banach spaces) is discussed. The results of this study can be used in systems analysis, linear programming, optimization methods, functional analysis, topology and convex analysis.
△ Less
Submitted 10 July, 2021;
originally announced July 2021.
-
Sensitivity of Regular Estimators
Authors:
Yaroslav Mukhin
Abstract:
This paper studies local asymptotic relationship between two scalar estimates. We define sensitivity of a target estimate to a control estimate to be the directional derivative of the target functional with respect to the gradient direction of the control functional. Sensitivity according to the information metric on the model manifold is the asymptotic covariance of regular efficient estimators.…
▽ More
This paper studies local asymptotic relationship between two scalar estimates. We define sensitivity of a target estimate to a control estimate to be the directional derivative of the target functional with respect to the gradient direction of the control functional. Sensitivity according to the information metric on the model manifold is the asymptotic covariance of regular efficient estimators. Sensitivity according to a general policy metric on the model manifold can be obtained from influence functions of regular efficient estimators. Policy sensitivity has a local counterfactual interpretation, where the ceteris paribus change to a counterfactual distribution is specified by the combination of a control parameter and a Riemannian metric on the model manifold.
△ Less
Submitted 22 May, 2018;
originally announced May 2018.
-
Monte Carlo Studies of a Novel LiF Radiator for RICH Detectors
Authors:
A. Efimov,
M. Artuso,
Min Gao,
R. Mountain,
F. Muheim,
Y. Mukhin,
S. Playfer
Abstract:
We show that a multifaceted LiF radiator produces more Cherenkov light and has better resolution per photon than a flat radiator slab when used in a ring imaging Cherenkov counter. Such a system is being considered for the CLEO III upgrade.
We show that a multifaceted LiF radiator produces more Cherenkov light and has better resolution per photon than a flat radiator slab when used in a ring imaging Cherenkov counter. Such a system is being considered for the CLEO III upgrade.
△ Less
Submitted 14 February, 1995;
originally announced February 1995.
