Universal compression of Gaussian sources with unknown parameters

TL;DR

This paper investigates the universal compression of Gaussian sources with unknown mean and variance, analyzing the attenuation of the optimal universal pdf and providing exact growth rates with sample size.

Contribution

It introduces the concept of attenuation for continuous distributions and derives exact bounds for Gaussian sources with unknown parameters.

Findings

Attenuation is finite and grows linearly with sample size for joint mean and variance uncertainty.

Attenuation grows as the square root of the sample size when only one parameter varies.

Exact attenuation bounds are provided for cases with single-parameter uncertainty.

Abstract

For a collection of distributions over a countable support set, the worst case universal compression formulation by Shtarkov attempts to assign a universal distribution over the support set. The formulation aims to ensure that the universal distribution does not underestimate the probability of any element in the support set relative to distributions in the collection. When the alphabet is uncountable and we have a collection $\cal P$ of Lebesgue continuous measures instead, we ask if there is a corresponding universal probability density function (pdf) that does not underestimate the value of the density function at any point in the support relative to pdfs in $\cal P$. Analogous to the worst case redundancy of a collection of distributions over a countable alphabet, we define the \textit{attenuation} of a class to be $A$ when the worst case optimal universal pdf at any point $x$ in the support is always at least the value any pdf in the collection $\cal P$ assigns to $x$ divided by $A$. We analyze the attenuation of the worst optimal universal pdf over length-$n$ samples generated \textit{i.i.d.} from a Gaussian distribution whose mean can be anywhere between $-\alpha/2$ to $\alpha/2$ and variance between $\sigma_m^2$ and $\sigma_M^2$. We show that this attenuation is finite, grows with the number of samples as ${\cal O}(n)$, and also specify the attentuation exactly without approximations. When only one parameter is allowed to vary, we show that the attenuation grows as ${\cal O}(\sqrt{n})$, again keeping in line with results from prior literature that fix the order of magnitude as a factor of $\sqrt{n}$ per parameter. In addition, we also specify the attenuation exactly without approximation when only the mean or only the variance is allowed to vary.