Blind Decoding-Metric Estimation for Probabilistic Shaping via Expectation Maximization

TL;DR

This paper introduces an unsupervised expectation maximization method to estimate decoding metrics for probabilistic shaping, relying solely on channel observations without needing transmitted data.

Contribution

It presents a novel unsupervised approach for decoding metric estimation in probabilistic shaping using expectation maximization, eliminating the need for transmitted data.

Findings

Effective estimation of decoding metrics from channel observations.

Improved decoding performance in probabilistic shaping scenarios.

No requirement for transmitted data in the estimation process.

Abstract

An unsupervised learning approach based on expectation maximization is proposed to obtain the parameters of a soft decision forward error correction decoding metric for probabilistic shaping. The algorithm depends only on the channel observations and does not require transmitted data.