Deep Learning of Geometric Constellation Shaping Including Fiber Nonlinearities

A new geometric shaping method is proposed, leveraging unsupervised machine learning to optimize the constellation design. The learned constellation mitigates nonlinear effects with gains up to 0.13 bit/4D when trained with a simplified fiber channel model.

[1]  Guigang Zhang,et al.  Deep Learning , 2016, Int. J. Semantic Comput..

[2]  Polina Bayvel,et al.  End-to-End Deep Learning of Optical Fiber Communications , 2018, Journal of Lightwave Technology.

[3]  Ting Wang,et al.  Coded polarization-multiplexed iterative polar modulation (PM-IPM) for beyond 400 Gb/s serial optical transmission , 2010, 2010 Conference on Optical Fiber Communication (OFC/NFOEC), collocated National Fiber Optic Engineers Conference.

[4]  Henk Wymeersch,et al.  Machine learning under the spotlight , 2017, Nature Photonics.

[5]  P. Winzer,et al.  Capacity Limits of Optical Fiber Networks , 2010, Journal of Lightwave Technology.

[6]  Georg Böcherer,et al.  On Probabilistic Shaping of Quadrature Amplitude Modulation for the Nonlinear Fiber Channel , 2016, Journal of Lightwave Technology.

[7]  Jakob Hoydis,et al.  An Introduction to Deep Learning for the Physical Layer , 2017, IEEE Transactions on Cognitive Communications and Networking.

[8]  Leif K. Oxenløwe,et al.  Constellation Shaping for WDM Systems Using 256QAM/1024QAM With Probabilistic Optimization , 2016, Journal of Lightwave Technology.

[9]  P. Poggiolini.  The GN Model of Non-Linear Propagation in Uncompensated Coherent Optical Systems , 2012, Journal of Lightwave Technology.

[10]  Ronen Dar,et al.  Properties of nonlinear noise in long, dispersion-uncompensated fiber links , 2013, Optics express.