Enhanced Recurrent Neural Network Equalization based on Hidden Feature Extraction Learning for Optical Interconnect.
Optical Fiber Communications Conference and Exhibition（2024）
Abstract
We propose a hidden feature extraction learning method for RNN equalization to improve training efficiency without increasing computational burden. Superior BER is demonstrated in 288 Gb/s 100 m VCSEL-MMF interconnect compared with black-box training strategy.
MoreTranslated text
Key words
Neural Network,Recurrent Neural Network,Optical Interconnects,Hidden Feature Extraction,Bit Error Rate,Activation Function,Deep Neural Network,Dimensionality Reduction,Signal Detection,Hidden Layer,Direct Detection,Optical Signal,Digital Signal Processing,Forward Error Correction,Digital Oscilloscope,Arbitrary Waveform Generator,Nonlinear Distortion,Inter-symbol Interference
AI Read Science
Must-Reading Tree
Example
Generate MRT to find the research sequence of this paper
Chat Paper
Summary is being generated by the instructions you defined