Abstract

Modern statistical learning technologies such as deep learning have a great potential to deal with linear/nonlinear fiber impairments for future coherent optical communications. We introduce various learning techniques suited for nonlinear equalizations.

© 2014 OSA

PDF Article