Multi-Task Learning to Enhance Generalizability of Neural Network Equalizers in Coherent Optical Systems
Srivallapanondh, Sasipim, Freire, Pedro J., Alam, Ashraful, Costa, Nelson, Spinnler, Bernhard, Napoli, Antonio, Sedov, Egor, Turitsyn, Sergei K., Prilepsky, Jaroslaw E.
–arXiv.org Artificial Intelligence
For the first time, multi-task learning is proposed to improve the flexibility of NN-based equalizers in coherent systems. A "single" NN-based equalizer improves Q-factor by up to 4 dB compared to CDC, without re-training, even with variations in launch power, symbol rate, or transmission distance.
arXiv.org Artificial Intelligence
Nov-3-2023
- Country:
- Europe
- United Kingdom (0.04)
- Portugal (0.04)
- Germany > Bavaria
- Upper Bavaria > Munich (0.04)
- Europe
- Genre:
- Research Report > New Finding (0.47)
- Technology: