The Price of Distributed: Rate Loss in the CEO Problem

TL;DR

This paper investigates the rate loss in distributed source coding, especially in Gaussian noise scenarios, highlighting the cost of decentralization compared to centralized encoding.

Contribution

It surveys and extends existing results on rate loss in distributed CEO problems, focusing on Gaussian noise with general sources.

Findings

Quantifies rate loss in various distributed settings

Extends known results to broader source classes

Highlights the impact of Gaussian noise on rate loss

Abstract

In the distributed remote (CEO) source coding problem, many separate encoders observe independently noisy copies of an underlying source. The rate loss is the difference between the rate required in this distributed setting and the rate that would be required in a setting where the encoders can fully cooperate. In this sense, the rate loss characterizes the price of distributed processing. We survey and extend the known results on the rate loss in various settings, with a particular emphasis on the case where the noise in the observations is Gaussian, but the underlying source is general.