域自适应:相干光学系统中神经网络滤波器的关键技术
Domain Adaptation: the Key Enabler of Neural Network Equalizers in Coherent Optical Systems
摘要:
我们介绍了域自适应和随机化的方法来校准基于神经网络的均衡器的实际传输,使用合成数据。该方法呈现高达99%的训练过程减少,我们在三个实验设置证明。
英文摘要:
We introduce the domain adaptation and randomization approach for calibrating neural network-based equalizers for real transmissions, using synthetic data. The approach renders up to 99\% training process reduction, which we demonstrate in three experimental setups.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。