应用于光信道均衡的知识蒸馏：解决循环连接的并行化问题
信号处理
2022-12-12 v1 机器学习
摘要
为规避基于循环神经网络的均衡器的不可并行性，我们提出知识蒸馏将RNN重构为可并行的前馈结构。后者显示出38%的延迟降低，而仅使Q因子受到0.5dB的影响。
引用
@article{arxiv.2212.04569,
title = {Knowledge Distillation Applied to Optical Channel Equalization: Solving the Parallelization Problem of Recurrent Connection},
author = {Sasipim Srivallapanondh and Pedro J. Freire and Bernhard Spinnler and Nelson Costa and Antonio Napoli and Sergei K. Turitsyn and Jaroslaw E. Prilepsky},
journal= {arXiv preprint arXiv:2212.04569},
year = {2022}
}
备注
Paper Accepted for Oral presentation - OFC 2023 (Optical Fiber Communication Conference)