Optimization of zero-delay mappings for distributed coding by deterministic annealing

TL;DR

This paper introduces a deterministic annealing-based optimization algorithm for zero-delay analog mappings in distributed coding networks, overcoming local optima issues of greedy methods and improving solution quality.

Contribution

It applies deterministic annealing to optimize zero-delay mappings in distributed coding, providing a global optimization approach that outperforms traditional greedy methods.

Findings

Deterministic annealing yields better mappings than greedy algorithms.

The method demonstrates strict superiority over descent-based approaches.

Insights into the properties of optimal mappings and their relation to digital coding.

Abstract

This paper studies the optimization of zero-delay analog mappings in a network setting that involves distributed coding. The cost surface is known to be non-convex, and known greedy methods tend to get trapped in poor locally optimal solutions that depend heavily on initialization. We derive an optimization algorithm based on the principles of "deterministic annealing", a powerful global optimization framework that has been successfully employed in several disciplines, including, in our recent work, to a simple zero-delay analog communications problem. We demonstrate strict superiority over the descent based methods, as well as present example mappings whose properties lend insights on the workings of the solution and relations with digital distributed coding.