TY - GEN

T1 - Optimality of the Plug-in Estimator for Differential Entropy Estimation under Gaussian Convolutions

AU - Goldfeld, Ziv

AU - Greenewald, Kristjan

AU - Weed, Jonathan

AU - Polyanskiy, Yury

N1 - Publisher Copyright:
© 2019 IEEE.

PY - 2019/7/1

Y1 - 2019/7/1

N2 - This paper establishes the optimality of the plugin estimator for the problem of differential entropy estimation under Gaussian convolutions. Specifically, we consider the estimation of the differential entropy h(X + Z), where X and Z are independent d-dimensional random variables with Z{\sim}\mathcal{N}( {0,{σ ^2}{{\text{I}}-d}} ). The distribution of X is unknown and belongs to some nonparametric class, but n independently and identically distributed samples from it are available. We first show that despite the regularizing effect of noise, any good estimator (within an additive gap) for this problem must have an exponential in d sample complexity. We then analyze the absolute-error risk of the plug-in estimator and show that it converges as frac{{{c^d}}}{{n }}, thus attaining the parametric estimation rate. This implies the optimality of the plug-in estimator for the considered problem. We provide numerical results comparing the performance of the plug-in estimator to general-purpose (unstructured) differential entropy estimators (based on kernel density estimation (KDE) or k nearest neighbors (kNN) techniques) applied to samples of X + Z. These results reveal a significant empirical superiority of the plug-in to state-of-the-art KDE- and kNN-based methods.

AB - This paper establishes the optimality of the plugin estimator for the problem of differential entropy estimation under Gaussian convolutions. Specifically, we consider the estimation of the differential entropy h(X + Z), where X and Z are independent d-dimensional random variables with Z{\sim}\mathcal{N}( {0,{σ ^2}{{\text{I}}-d}} ). The distribution of X is unknown and belongs to some nonparametric class, but n independently and identically distributed samples from it are available. We first show that despite the regularizing effect of noise, any good estimator (within an additive gap) for this problem must have an exponential in d sample complexity. We then analyze the absolute-error risk of the plug-in estimator and show that it converges as frac{{{c^d}}}{{n }}, thus attaining the parametric estimation rate. This implies the optimality of the plug-in estimator for the considered problem. We provide numerical results comparing the performance of the plug-in estimator to general-purpose (unstructured) differential entropy estimators (based on kernel density estimation (KDE) or k nearest neighbors (kNN) techniques) applied to samples of X + Z. These results reveal a significant empirical superiority of the plug-in to state-of-the-art KDE- and kNN-based methods.

UR - http://www.scopus.com/inward/record.url?scp=85073146341&partnerID=8YFLogxK

U2 - 10.1109/ISIT.2019.8849414

DO - 10.1109/ISIT.2019.8849414

M3 - Conference contribution

AN - SCOPUS:85073146341

T3 - IEEE International Symposium on Information Theory - Proceedings

SP - 892

EP - 896

BT - 2019 IEEE International Symposium on Information Theory, ISIT 2019 - Proceedings

PB - Institute of Electrical and Electronics Engineers

T2 - 2019 IEEE International Symposium on Information Theory, ISIT 2019

Y2 - 7 July 2019 through 12 July 2019

ER -