TR2020-155

End-to-End Deep Learning for Phase Noise-Robust Multi-Dimensional Geometric Shaping


    •  Talreja, V., Koike-Akino, T., Wang, Y., Millar, D.S., Kojima, K., Parsons, K., "End-to-End Deep Learning for Phase Noise-Robust Multi-Dimensional Geometric Shaping", European Conference on Optical Communication (ECOC), November 2020.
      BibTeX TR2020-155 PDF Video
      • @inproceedings{Talreja2020nov,
      • author = {Talreja, Veeru and Koike-Akino, Toshiaki and Wang, Ye and Millar, David S. and Kojima, Keisuke and Parsons, Kieran},
      • title = {End-to-End Deep Learning for Phase Noise-Robust Multi-Dimensional Geometric Shaping},
      • booktitle = {European Conference on Optical Communication (ECOC)},
      • year = 2020,
      • month = nov,
      • url = {https://www.merl.com/publications/TR2020-155}
      • }
  • MERL Contacts:
  • Research Areas:

    Artificial Intelligence, Communications, Multi-Physical Modeling, Optimization, Signal Processing

Abstract:

We propose an end-to-end deep learning model for phase noise-robust optical communications. A convolutional embedding layer is integrated with a deep autoencoder for multi-dimensional constellation design to achieve shaping gain. The proposed model offers a significant gain up to 2 dB.

 

  • Related Videos