Self-Normalizing Neural Network, Enabling One Shot Transfer Learning for Modeling EDFA Wavelength Dependent Gain

TL;DR

This paper introduces a semi-supervised, self-normalizing neural network framework that accurately models EDFA wavelength-dependent gain and enables effective one-shot transfer learning across different amplifier types.

Contribution

The novel framework allows for high-accuracy transfer learning of EDFA gain models with minimal data, improving flexibility and efficiency in optical network modeling.

Findings

High-accuracy transfer learning across different EDFAs

Effective modeling with minimal labeled data

Demonstrated on 22 EDFAs in real testbeds

Abstract

We present a novel ML framework for modeling the wavelength-dependent gain of multiple EDFAs, based on semi-supervised, self-normalizing neural networks, enabling one-shot transfer learning. Our experiments on 22 EDFAs in Open Ireland and COSMOS testbeds show high-accuracy transfer-learning even when operated across different amplifier types.