Abstract
We propose a novel hardware-aware magnitude pruning technique for coherent photonic neural networks. The proposed technique can prune 99.45% of network parameters and reduce the static power consumption by 98.23% with a negligible accuracy loss.
© 2022 The Author(s)
PDF Article | Presentation VideoMore Like This
Sanmitra Banerjee, Mahdi Nikdast, and Krishnendu Chakrabarty
Th1A.22 Optical Fiber Communication Conference (OFC) 2021
Sri Krishna Vadlamani, Ryan Hamerly, and Dirk Englund
FTh1B.3 Frontiers in Optics (FiO) 2022
Bhavin J. Shastri, Matthew J. Filipovich, Zhimu Guo, Paul R. Prucnal, Sudip Shekhar, and Volker J. Sorger
CThA13B_02 Conference on Lasers and Electro-Optics/Pacific Rim (CLEO/PR) 2022