Abstract

Optical neural networks take optical neurons as the cornerstone to achieve complex functions. The coherent optical neuron has become one of the mainstream implementations because it can effectively perform natural and even complex number calculations. However, its state variability and requirement for reliability and effectiveness render traditional control methods no longer applicable. In this Letter, deep reinforcement coherent optical neuron control (DRCON) is proposed, and its effectiveness is experimentally demonstrated. Compared with the standard stochastic gradient descent, the average convergence rate of DRCON is 33% faster, while the effective number of bits increases from less than 2 bits to 5.5 bits. DRCON is a promising first step for large-scale optical neural network control.

© 2023 Optica Publishing Group

Deep reinforcement learning control of white-light continuum generation

Carlo M. Valensise, Alessandro Giuseppi, Giulio Cerullo, and Dario Polli
Optica 8(2) 239-242 (2021)

Reinforcement learning based robust control algorithms for coherent pulse stacking

Abulikemu Abuduweili, Jie Wang, Bowei Yang, Aimin Wang, and Zhigang Zhang
Opt. Express 29(16) 26068-26081 (2021)

Stochastic photonic spiking neuron for Bayesian inference with unsupervised learning

Bowen Ma, Junfeng Zhang, Xing Li, and Weiwen Zou
Opt. Lett. 48(6) 1411-1414 (2023)

Reinforcement-learning-based path planning in multilayer elastic optical networks [Invited]

Takafumi Tanaka
J. Opt. Commun. Netw. 16(1) A68-A77 (2024)

Adaptive optics control using model-based reinforcement learning

Jalo Nousiainen, Chang Rajani, Markus Kasper, and Tapio Helin
Opt. Express 29(10) 15327-15344 (2021)