A Few-Shot Learning Equalizer Applied to Optical Communication System

We propose a fully connected neural network equalizer based on few-shot learning in coherent optical fiber communication systems. The results show that the equalizer we constructed can achieve significant equalization effect with fewer eigenvalues.