End-to-End Deep Learning of Long-Haul Coherent Optical Fiber Communications via Regular Perturbation Model

This work addresses performance enhancement for optical communication systems, but it is incremental as it builds on existing autoencoder and perturbative models.

The paper tackled the problem of improving data transmission in long-haul coherent optical fiber communications by jointly optimizing constellation shaping and nonlinear pre-emphasis, achieving a mutual information gain of 0.18 bits/sym./pol. in simulations over a 30x80 km link.