Author:
Xu Zhaopeng,Dong Shuangyu,Ji Honglin,Manton Jonathan H.,Shieh William
Abstract
Low-complexity sparsely-connected multi-output neural networks are proposed for equalization in a 50-Gb/s 25-km PAM4 IM/DD system. Compared with traditional fully-connected single-output counterparts, a gross complexity reduction of 60.4%/56.7% can be achieved with 2-layer FNN/C-FNN architecture.
Cited by
5 articles.
订阅此论文施引文献
订阅此论文施引文献，注册后可以免费订阅5篇论文的施引文献，订阅后可以查看论文全部施引文献
1. Quantization of Recurrent Neural Network for Low-Complexity High-Speed IM/DD System Equalization Based on Neuron Clustering;2023 Asia Communications and Photonics Conference/2023 International Photonics and Optoelectronics Meetings (ACP/POEM);2023-11-04
2. Semi-Supervised Feature-Crosses Neural Network Equalizer in Fiber Optics;2023 Asia Communications and Photonics Conference/2023 International Photonics and Optoelectronics Meetings (ACP/POEM);2023-11-04
3. Hardware-Efficient Neural Network-Based Receiver for Intensity-Modulated Direct-Detection Short-Reach Optical Links;2023 IEEE 15th International Conference on Advanced Infocomm Technology (ICAIT);2023-10-13
4. On the Computational Complexity of Artificial Neural Networks for Short-Reach Optical Communication;2023 Opto-Electronics and Communications Conference (OECC);2023-07-02
5. Towards Neural Network Equalizer Implementations for IM/DD Transceivers;2023 Opto-Electronics and Communications Conference (OECC);2023-07-02