论文标题
域适应性：连贯的光学系统中神经网络均衡器的关键推动器
Domain Adaptation: the Key Enabler of Neural Network Equalizers in Coherent Optical Systems
论文作者
论文摘要
我们介绍了使用合成数据介绍用于校准基于神经网络的均衡器进行实际传输的域的适应和随机方法。该方法最多可减少99 \％的训练过程，我们在三个实验设置中证明了这一点。
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.
