Showing papers by "Bernhard Spinnler published in 2020"

A Reinforcement Learning Framework for Parameter Optimization in Elastic Optical Networks

Abstract: We present a reinforcement learning (RL) framework for maximizing the total capacity of a 51-channel transmission system, which runs magnitudes faster than a genetic algorithm (GA) based optimization. The generalization capabilities and performance of the RL framework are compared to results obtained with a GA.

Simplifying the Supervised Learning of Kerr Nonlinearity Compensation Algorithms by Data Augmentation

Abstract: We propose a data augmentation technique to improve performance and decrease complexity of the supervised learning of nonlinearity compensation algorithms. We demonstrate both numerically and experimentally that the augmentation allows reducing the training dataset size up to 6 times while keeping the same post-compensation bit-error rate.

Experimental Verification of Complex-Valued Artificial Neural Network for Nonlinear Equalization in Coherent Optical Communication Systems

Abstract: We propose a novel design of neural network for mitigating the fiber nonlinearity, employing a structure based on physical modelling. The neural network achieved nearly 5 times BER reduction in a field trial when transmitting WDM 200G DP-16QAM over a 620 km legacy link.