Low-complexity nonlinear equalizer based on artificial neural network for 112 Gbit/s PAM-4 transmission using DML

•Adopting appropriate strategies can reduce the computational complexity of artificial neural network nonlinear equalizer.•A novel weight pruning method named dynamic pruning is proposed.•The dynamic pruning method can improve the performance of equalizer by recovering the mistakenly pruned weight connections.•The bit error rate obtained by dynamic pruning method is lower than weight pruning method.