Experimental Evaluation of Computational Complexity for Different Neural Network Equalizers in Optical Communications

TL;DR

This paper compares various neural network architectures for optical channel equalization, analyzing their performance versus complexity trade-offs in different fiber optic setups.

Contribution

It provides a comparative analysis of neural network architectures for optical equalizers, highlighting their performance and complexity trade-offs in TWC and SSMF scenarios.

Findings

Neural network architectures vary significantly in complexity and performance.

Trade-offs between equalizer accuracy and computational cost are quantified.

Results guide the selection of neural networks for optical communication systems.

Abstract

Addressing the neural network-based optical channel equalizers, we quantify the trade-off between their performance and complexity by carrying out the comparative analysis of several neural network architectures, presenting the results for TWC and SSMF set-ups.