ASIC Implementation of Time-Domain Digital Backpropagation with Deep-Learned Chromatic Dispersion Filters
arXiv:1806.07223v215 citations
AI Analysis
This work addresses power efficiency in optical communication systems, representing an incremental improvement with specific hardware gains.
The paper tackled the problem of power dissipation in time-domain digital backpropagation for optical communication by jointly optimizing and quantizing chromatic dispersion filters using machine learning, resulting in improved BER performance and over 40% power reduction in 28-nm CMOS.
We consider time-domain digital backpropagation with chromatic dispersion filters jointly optimized and quantized using machine-learning techniques. Compared to the baseline implementations, we show improved BER performance and >40% power dissipation reductions in 28-nm CMOS.