Domain Adaptation: the Key Enabler of Neural Network Equalizers in Coherent Optical Systems

02/25/2022
by   Pedro J. Freire, et al.
0

We introduce the domain adaptation and randomization approach for calibrating neural network-based equalizers for real transmissions, using synthetic data. The approach renders up to 99% training process reduction, which we demonstrate in three experimental setups.

READ FULL TEXT

Please sign up or login with your details

Forgot password? Click here to reset