119

A General Derivative Identity for the Conditional Mean Estimator in Gaussian Noise and Some Applications

International Symposium on Information Theory (ISIT), 2020
Abstract

Consider a channel Y=X+N{\bf Y}={\bf X}+ {\bf N} where X{\bf X} is an nn-dimensional random vector, and N{\bf N} is a Gaussian vector with a covariance matrix KN{\bf \mathsf{K}}_{\bf N}. The object under consideration in this paper is the conditional mean of X{\bf X} given Y=y{\bf Y}={\bf y}, that is yE[XY=y]{\bf y} \to E[{\bf X}|{\bf Y}={\bf y}]. Several identities in the literature connect E[XY=y]E[{\bf X}|{\bf Y}={\bf y}] to other quantities such as the conditional variance, score functions, and higher-order conditional moments. The objective of this paper is to provide a unifying view of these identities. In the first part of the paper, a general derivative identity for the conditional mean is derived. Specifically, for the Markov chain UXY{\bf U} \leftrightarrow {\bf X} \leftrightarrow {\bf Y}, it is shown that the Jacobian of E[UY=y]E[{\bf U}|{\bf Y}={\bf y}] is given by KN1Cov(X,UY=y){\bf \mathsf{K}}_{{\bf N}}^{-1} {\bf Cov} ( {\bf X}, {\bf U} | {\bf Y}={\bf y}). In the second part of the paper, via various choices of U{\bf U}, the new identity is used to generalize many of the known identities and derive some new ones. First, a simple proof of the Hatsel and Nolte identity for the conditional variance is shown. Second, a simple proof of the recursive identity due to Jaffer is provided. Third, a new connection between the conditional cumulants and the conditional expectation is shown. In particular, it is shown that the kk-th derivative of E[XY=y]E[X|Y=y] is the (k+1)(k+1)-th conditional cumulant. The third part of the paper considers some applications. In a first application, the power series and the compositional inverse of E[XY=y]E[X|Y=y] are derived. In a second application, the distribution of the estimator error (XE[XY])(X-E[X|Y]) is derived. In a third application, we construct consistent estimators (empirical Bayes estimators) of the conditional cumulants from an i.i.d. sequence Y1,...,YnY_1,...,Y_n.

View on arXiv
Comments on this paper