This script uses an auto-encoder (AE) for end-to-end learning of a non-linear memoryless fiber channel. The determines a good constellation and a good receiver. The achievable information rate is also computed.
This code was is based on the paper
Shen Li, Christian Häger, Nil Garcia, and Henk Wymeersch, "Achievable Information Rates for Nonlinear Fiber Communication via End-to-end Autoencoder Learning," in Proc. European Conference on Optical Communication (2018) arXiv:1804.07675.