Multi-Task Learning: Towards Low-Complexity Neural Network Nonlinearity Compensation in Coherent Optical Systems
We propose multi-task learning based NN equalization for nonlinearity compensation in coherent optical system. Compared with conventional NN equalization, 30% complexity reduction is achieved in an 800-Gb/s PDM-16QAM system with 5 output symbols.
Authors
- Weisheng Hu (ORCID: https://orcid.org/0000-0002-6168-2688)
- Tonghui Ji (ORCID: https://orcid.org/0000-0003-3453-0248)
- Zhaopeng Xu (ORCID: https://orcid.org/0000-0001-6476-9001)
- Jinlong Wei (ORCID: https://orcid.org/0000-0001-7714-5003)
- Honglin Ji (ORCID: https://orcid.org/0000-0003-3437-8573)
- Linsheng Fan (ORCID: https://orcid.org/0009-0002-7259-084X)
- Qi Wu (ORCID: https://orcid.org/0000-0002-1196-226X)
- Jianwei Tang (ORCID: https://orcid.org/0009-0007-6506-0046)
- Lulu Liu (ORCID: https://orcid.org/0000-0002-9601-6909)
- Zhongliang Sun (ORCID: https://orcid.org/0000-0002-5559-4892)
- Yang Yu (ORCID: https://orcid.org/0000-0002-6145-8811)
- Junpeng Liang (ORCID: https://orcid.org/0000-0002-0594-9579)
- Shangcheng Wang (ORCID: https://orcid.org/0009-0000-6791-8519)
- Juhao Li (ORCID: https://orcid.org/0000-0002-4913-6844)
Institutions
- Shanghai Jiao Tong University (CN)
- Peking University (CN)
- Peng Cheng Laboratory (CN)
Publication Details
- Published
- 2024-11-02
- DOI
- https://doi.org/10.1109/acp/ipoc63121.2024.10809684
- Citations
- 2
- Primary Topic
- Spectroscopy Techniques in Biomedical and Chemical Research
- Type
- article
- Field-Weighted Citation Impact
- 1.32