Model-Based Machine Learning for Joint Digital Backpropagation and PMD Compensation

This work addresses signal degradation in optical fiber communications, offering a hardware-friendly solution for joint digital backpropagation and PMD compensation, though it appears incremental as it builds on existing split-step methods with machine learning enhancements.

The authors tackled the problem of compensating for polarization mode dispersion (PMD) in polarization-multiplexed optical communication systems by proposing a model-based machine learning approach that parameterizes the split-step method for the Manakov-PMD equation, achieving performance close to the PMD-free case.