Conference paper
Deep Learning of Geometric Constellation Shaping including Fiber Nonlinearities
Department of Photonics Engineering, Technical University of Denmark1
Ultra-fast Optical Communication, Department of Photonics Engineering, Technical University of Denmark2
Japan National Institute of Information and Communications Technology3
Centre of Excellence for Silicon Photonics for Optical Communications, Centers, Technical University of Denmark4
A new geometric shaping method is proposed, leveraging unsupervised machinelearning to optimize the constellation design. The learned constellationmitigates nonlinear effects with gains up to 0.13 bit/4D when trained with asimplified fiber channel model.
| Language: | English |
|---|---|
| Publisher: | IEEE |
| Year: | 2018 |
| Pages: | 1-3 |
| Proceedings: | 44th European Conference on Optical Communication |
| ISBN: | 153864861X , 1538648628 , 1538648636 , 9781538648612 , 9781538648629 and 9781538648636 |
| Types: | Conference paper |
| DOI: | 10.1109/ECOC.2018.8535453 |
| ORCIDs: | Yankov, Metodi Plamenov and Zibar, Darko |
Artificial neural networks Channel models Decoding Nonlinear optics Quadrature amplitude modulation deep learning fiber nonlinearities geometric constellation shaping geometrical optics optical fibre networks simplified fiber channel model unsupervised learning unsupervised machine learning
